diff options
| author | Dmitry Torokhov <dmitry.torokhov@gmail.com> | 2010-03-08 23:17:32 -0800 |
|---|---|---|
| committer | Dmitry Torokhov <dmitry.torokhov@gmail.com> | 2010-03-08 23:17:32 -0800 |
| commit | ce67eef6a112bb283c6db39f9195800f31f5599a (patch) | |
| tree | 34f2e7fbd0b9838abba482dbb1a7db09e6fbb53c /drivers/gpu | |
| parent | 776943fd6f104a6e8457dc95a17282e69e963666 (diff) | |
| parent | 57d54889cd00db2752994b389ba714138652e60c (diff) | |
Merge commit 'v2.6.34-rc1' into for-linus
Diffstat (limited to 'drivers/gpu')
169 files changed, 20097 insertions, 7535 deletions
diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 96eddd17e05..305c5900396 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -66,6 +66,8 @@ config DRM_RADEON If M is selected, the module will be called radeon. +source "drivers/gpu/drm/radeon/Kconfig" + config DRM_I810 tristate "Intel I810" depends on DRM && AGP && AGP_INTEL diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 39c5aa75b8f..abe3f446ca4 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -4,7 +4,7 @@ ccflags-y := -Iinclude/drm -drm-y := drm_auth.o drm_bufs.o drm_cache.o \ +drm-y := drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \ drm_context.o drm_dma.o drm_drawable.o \ drm_drv.o drm_fops.o drm_gem.o drm_ioctl.o drm_irq.o \ drm_lock.o drm_memory.o drm_proc.o drm_stub.o drm_vm.o \ diff --git a/drivers/gpu/drm/ati_pcigart.c b/drivers/gpu/drm/ati_pcigart.c index a1fce68e3bb..17be051b7aa 100644 --- a/drivers/gpu/drm/ati_pcigart.c +++ b/drivers/gpu/drm/ati_pcigart.c @@ -113,7 +113,7 @@ int drm_ati_pcigart_init(struct drm_device *dev, struct drm_ati_pcigart_info *ga if (pci_set_dma_mask(dev->pdev, gart_info->table_mask)) { DRM_ERROR("fail to set dma mask to 0x%Lx\n", - gart_info->table_mask); + (unsigned long long)gart_info->table_mask); ret = 1; goto done; } diff --git a/drivers/gpu/drm/drm_buffer.c b/drivers/gpu/drm/drm_buffer.c new file mode 100644 index 00000000000..55d03ed0500 --- /dev/null +++ b/drivers/gpu/drm/drm_buffer.c @@ -0,0 +1,184 @@ +/************************************************************************** + * + * Copyright 2010 Pauli Nieminen. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * + **************************************************************************/ +/* + * Multipart buffer for coping data which is larger than the page size. + * + * Authors: + * Pauli Nieminen <suokkos-at-gmail-dot-com> + */ + +#include "drm_buffer.h" + +/** + * Allocate the drm buffer object. + * + * buf: Pointer to a pointer where the object is stored. + * size: The number of bytes to allocate. + */ +int drm_buffer_alloc(struct drm_buffer **buf, int size) +{ + int nr_pages = size / PAGE_SIZE + 1; + int idx; + + /* Allocating pointer table to end of structure makes drm_buffer + * variable sized */ + *buf = kzalloc(sizeof(struct drm_buffer) + nr_pages*sizeof(char *), + GFP_KERNEL); + + if (*buf == NULL) { + DRM_ERROR("Failed to allocate drm buffer object to hold" + " %d bytes in %d pages.\n", + size, nr_pages); + return -ENOMEM; + } + + (*buf)->size = size; + + for (idx = 0; idx < nr_pages; ++idx) { + + (*buf)->data[idx] = + kmalloc(min(PAGE_SIZE, size - idx * PAGE_SIZE), + GFP_KERNEL); + + + if ((*buf)->data[idx] == NULL) { + DRM_ERROR("Failed to allocate %dth page for drm" + " buffer with %d bytes and %d pages.\n", + idx + 1, size, nr_pages); + goto error_out; + } + + } + + return 0; + +error_out: + + /* Only last element can be null pointer so check for it first. */ + if ((*buf)->data[idx]) + kfree((*buf)->data[idx]); + + for (--idx; idx >= 0; --idx) + kfree((*buf)->data[idx]); + + kfree(*buf); + return -ENOMEM; +} +EXPORT_SYMBOL(drm_buffer_alloc); + +/** + * Copy the user data to the begin of the buffer and reset the processing + * iterator. + * + * user_data: A pointer the data that is copied to the buffer. + * size: The Number of bytes to copy. + */ +extern int drm_buffer_copy_from_user(struct drm_buffer *buf, + void __user *user_data, int size) +{ + int nr_pages = size / PAGE_SIZE + 1; + int idx; + + if (size > buf->size) { + DRM_ERROR("Requesting to copy %d bytes to a drm buffer with" + " %d bytes space\n", + size, buf->size); + return -EFAULT; + } + + for (idx = 0; idx < nr_pages; ++idx) { + + if (DRM_COPY_FROM_USER(buf->data[idx], + user_data + idx * PAGE_SIZE, + min(PAGE_SIZE, size - idx * PAGE_SIZE))) { + DRM_ERROR("Failed to copy user data (%p) to drm buffer" + " (%p) %dth page.\n", + user_data, buf, idx); + return -EFAULT; + + } + } + buf->iterator = 0; + return 0; +} +EXPORT_SYMBOL(drm_buffer_copy_from_user); + +/** + * Free the drm buffer object + */ +void drm_buffer_free(struct drm_buffer *buf) +{ + + if (buf != NULL) { + + int nr_pages = buf->size / PAGE_SIZE + 1; + int idx; + for (idx = 0; idx < nr_pages; ++idx) + kfree(buf->data[idx]); + + kfree(buf); + } +} +EXPORT_SYMBOL(drm_buffer_free); + +/** + * Read an object from buffer that may be split to multiple parts. If object + * is not split function just returns the pointer to object in buffer. But in + * case of split object data is copied to given stack object that is suplied + * by caller. + * + * The processing location of the buffer is also advanced to the next byte + * after the object. + * + * objsize: The size of the objet in bytes. + * stack_obj: A pointer to a memory location where object can be copied. + */ +void *drm_buffer_read_object(struct drm_buffer *buf, + int objsize, void *stack_obj) +{ + int idx = drm_buffer_index(buf); + int page = drm_buffer_page(buf); + void *obj = 0; + + if (idx + objsize <= PAGE_SIZE) { + obj = &buf->data[page][idx]; + } else { + /* The object is split which forces copy to temporary object.*/ + int beginsz = PAGE_SIZE - idx; + memcpy(stack_obj, &buf->data[page][idx], beginsz); + + memcpy(stack_obj + beginsz, &buf->data[page + 1][0], + objsize - beginsz); + + obj = stack_obj; + } + + drm_buffer_advance(buf, objsize); + return obj; +} +EXPORT_SYMBOL(drm_buffer_read_object); diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 7d0f00a935f..f2aaf39be39 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -836,11 +836,7 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) mode_changed = true; } else if (set->fb == NULL) { mode_changed = true; - } else if ((set->fb->bits_per_pixel != - set->crtc->fb->bits_per_pixel) || - set->fb->depth != set->crtc->fb->depth) - fb_changed = true; - else + } else fb_changed = true; } diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 766c46875a2..f3c58e2bd75 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -125,28 +125,28 @@ static struct drm_ioctl_desc drm_ioctls[] = { DRM_IOCTL_DEF(DRM_IOCTL_UPDATE_DRAW, drm_update_drawable_info, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), - DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, 0), - DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH), - - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_MASTER), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_mode_attachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_mode_detachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_MASTER | DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW) + DRM_IOCTL_DEF(DRM_IOCTL_GEM_CLOSE, drm_gem_close_ioctl, DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED), + + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_MASTER|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_mode_attachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_mode_detachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_MASTER | DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED) }; #define DRM_CORE_IOCTL_COUNT ARRAY_SIZE( drm_ioctls ) diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index defcaf10846..f97e7c42ac8 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -60,8 +60,7 @@ #define EDID_QUIRK_FIRST_DETAILED_PREFERRED (1 << 5) /* use +hsync +vsync for detailed mode */ #define EDID_QUIRK_DETAILED_SYNC_PP (1 << 6) -/* define the number of Extension EDID block */ -#define MAX_EDID_EXT_NUM 4 + #define LEVEL_DMT 0 #define LEVEL_GTF 1 @@ -114,14 +113,14 @@ static const u8 edid_header[] = { }; /** - * edid_is_valid - sanity check EDID data + * drm_edid_is_valid - sanity check EDID data * @edid: EDID data * * Sanity check the EDID block by looking at the header, the version number * and the checksum. Return 0 if the EDID doesn't check out, or 1 if it's * valid. */ -static bool edid_is_valid(struct edid *edid) +bool drm_edid_is_valid(struct edid *edid) { int i, score = 0; u8 csum = 0; @@ -163,6 +162,7 @@ bad: } return 0; } +EXPORT_SYMBOL(drm_edid_is_valid); /** * edid_vendor - match a string against EDID's obfuscated vendor field @@ -598,6 +598,50 @@ struct drm_display_mode *drm_mode_std(struct drm_device *dev, return mode; } +/* + * EDID is delightfully ambiguous about how interlaced modes are to be + * encoded. Our internal representation is of frame height, but some + * HDTV detailed timings are encoded as field height. + * + * The format list here is from CEA, in frame size. Technically we + * should be checking refresh rate too. Whatever. + */ +static void +drm_mode_do_interlace_quirk(struct drm_display_mode *mode, + struct detailed_pixel_timing *pt) +{ + int i; + static const struct { + int w, h; + } cea_interlaced[] = { + { 1920, 1080 }, + { 720, 480 }, + { 1440, 480 }, + { 2880, 480 }, + { 720, 576 }, + { 1440, 576 }, + { 2880, 576 }, + }; + static const int n_sizes = + sizeof(cea_interlaced)/sizeof(cea_interlaced[0]); + + if (!(pt->misc & DRM_EDID_PT_INTERLACED)) + return; + + for (i = 0; i < n_sizes; i++) { + if ((mode->hdisplay == cea_interlaced[i].w) && + (mode->vdisplay == cea_interlaced[i].h / 2)) { + mode->vdisplay *= 2; + mode->vsync_start *= 2; + mode->vsync_end *= 2; + mode->vtotal *= 2; + mode->vtotal |= 1; + } + } + + mode->flags |= DRM_MODE_FLAG_INTERLACE; +} + /** * drm_mode_detailed - create a new mode from an EDID detailed timing section * @dev: DRM device (needed to create new mode) @@ -633,8 +677,7 @@ static struct drm_display_mode *drm_mode_detailed(struct drm_device *dev, return NULL; } if (!(pt->misc & DRM_EDID_PT_SEPARATE_SYNC)) { - printk(KERN_WARNING "integrated sync not supported\n"); - return NULL; + printk(KERN_WARNING "composite sync not supported\n"); } /* it is incorrect if hsync/vsync width is zero */ @@ -681,8 +724,7 @@ static struct drm_display_mode *drm_mode_detailed(struct drm_device *dev, drm_mode_set_name(mode); - if (pt->misc & DRM_EDID_PT_INTERLACED) - mode->flags |= DRM_MODE_FLAG_INTERLACE; + drm_mode_do_interlace_quirk(mode, pt); if (quirks & EDID_QUIRK_DETAILED_SYNC_PP) { pt->misc |= DRM_EDID_PT_HSYNC_POSITIVE | DRM_EDID_PT_VSYNC_POSITIVE; @@ -1070,8 +1112,8 @@ static int add_detailed_info_eedid(struct drm_connector *connector, } /* Chose real EDID extension number */ - edid_ext_num = edid->extensions > MAX_EDID_EXT_NUM ? - MAX_EDID_EXT_NUM : edid->extensions; + edid_ext_num = edid->extensions > DRM_MAX_EDID_EXT_NUM ? + DRM_MAX_EDID_EXT_NUM : edid->extensions; /* Find CEA extension */ for (i = 0; i < edid_ext_num; i++) { @@ -1153,7 +1195,7 @@ static int drm_ddc_read_edid(struct drm_connector *connector, for (i = 0; i < 4; i++) { if (drm_do_probe_ddc_edid(adapter, buf, len)) return -1; - if (edid_is_valid((struct edid *)buf)) + if (drm_edid_is_valid((struct edid *)buf)) return 0; } @@ -1178,7 +1220,7 @@ struct edid *drm_get_edid(struct drm_connector *connector, int ret; struct edid *edid; - edid = kmalloc(EDID_LENGTH * (MAX_EDID_EXT_NUM + 1), + edid = kmalloc(EDID_LENGTH * (DRM_MAX_EDID_EXT_NUM + 1), GFP_KERNEL); if (edid == NULL) { dev_warn(&connector->dev->pdev->dev, @@ -1196,14 +1238,14 @@ struct edid *drm_get_edid(struct drm_connector *connector, if (edid->extensions != 0) { int edid_ext_num = edid->extensions; - if (edid_ext_num > MAX_EDID_EXT_NUM) { + if (edid_ext_num > DRM_MAX_EDID_EXT_NUM) { dev_warn(&connector->dev->pdev->dev, "The number of extension(%d) is " "over max (%d), actually read number (%d)\n", - edid_ext_num, MAX_EDID_EXT_NUM, - MAX_EDID_EXT_NUM); + edid_ext_num, DRM_MAX_EDID_EXT_NUM, + DRM_MAX_EDID_EXT_NUM); /* Reset EDID extension number to be read */ - edid_ext_num = MAX_EDID_EXT_NUM; + edid_ext_num = DRM_MAX_EDID_EXT_NUM; } /* Read EDID including extensions too */ ret = drm_ddc_read_edid(connector, adapter, (char *)edid, @@ -1246,8 +1288,8 @@ bool drm_detect_hdmi_monitor(struct edid *edid) goto end; /* Chose real EDID extension number */ - edid_ext_num = edid->extensions > MAX_EDID_EXT_NUM ? - MAX_EDID_EXT_NUM : edid->extensions; + edid_ext_num = edid->extensions > DRM_MAX_EDID_EXT_NUM ? + DRM_MAX_EDID_EXT_NUM : edid->extensions; /* Find CEA extension */ for (i = 0; i < edid_ext_num; i++) { @@ -1304,7 +1346,7 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid) if (edid == NULL) { return 0; } - if (!edid_is_valid(edid)) { + if (!drm_edid_is_valid(edid)) { dev_warn(&connector->dev->pdev->dev, "%s: EDID invalid.\n", drm_get_connector_name(connector)); return 0; diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index 1c2b7d44ec0..50549703584 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -27,6 +27,7 @@ * Dave Airlie <airlied@linux.ie> * Jesse Barnes <jesse.barnes@intel.com> */ +#include <linux/kernel.h> #include <linux/sysrq.h> #include <linux/fb.h> #include "drmP.h" @@ -50,21 +51,6 @@ int drm_fb_helper_add_connector(struct drm_connector *connector) } EXPORT_SYMBOL(drm_fb_helper_add_connector); -static int my_atoi(const char *name) -{ - int val = 0; - - for (;; name++) { - switch (*name) { - case '0' ... '9': - val = 10*val+(*name-'0'); - break; - default: - return val; - } - } -} - /** * drm_fb_helper_connector_parse_command_line - parse command line for connector * @connector - connector to parse line for @@ -111,7 +97,7 @@ static bool drm_fb_helper_connector_parse_command_line(struct drm_connector *con namelen = i; if (!refresh_specified && !bpp_specified && !yres_specified) { - refresh = my_atoi(&name[i+1]); + refresh = simple_strtol(&name[i+1], NULL, 10); refresh_specified = 1; if (cvt || rb) cvt = 0; @@ -121,7 +107,7 @@ static bool drm_fb_helper_connector_parse_command_line(struct drm_connector *con case '-': namelen = i; if (!bpp_specified && !yres_specified) { - bpp = my_atoi(&name[i+1]); + bpp = simple_strtol(&name[i+1], NULL, 10); bpp_specified = 1; if (cvt || rb) cvt = 0; @@ -130,7 +116,7 @@ static bool drm_fb_helper_connector_parse_command_line(struct drm_connector *con break; case 'x': if (!yres_specified) { - yres = my_atoi(&name[i+1]); + yres = simple_strtol(&name[i+1], NULL, 10); yres_specified = 1; } else goto done; @@ -170,7 +156,7 @@ static bool drm_fb_helper_connector_parse_command_line(struct drm_connector *con } } if (i < 0 && yres_specified) { - xres = my_atoi(name); + xres = simple_strtol(name, NULL, 10); res_specified = 1; } done: @@ -389,7 +375,7 @@ int drm_fb_helper_blank(int blank, struct fb_info *info) break; /* Display: Off; HSync: On, VSync: On */ case FB_BLANK_NORMAL: - drm_fb_helper_off(info, DRM_MODE_DPMS_ON); + drm_fb_helper_off(info, DRM_MODE_DPMS_STANDBY); break; /* Display: Off; HSync: Off, VSync: On */ case FB_BLANK_HSYNC_SUSPEND: @@ -694,7 +680,7 @@ int drm_fb_helper_set_par(struct fb_info *info) int i; if (var->pixclock != 0) { - DRM_ERROR("PIXEL CLCOK SET\n"); + DRM_ERROR("PIXEL CLOCK SET\n"); return -EINVAL; } diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index e9dbb481c46..aa89d4b0b4c 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -142,19 +142,6 @@ drm_gem_object_alloc(struct drm_device *dev, size_t size) if (IS_ERR(obj->filp)) goto free; - /* Basically we want to disable the OOM killer and handle ENOMEM - * ourselves by sacrificing pages from cached buffers. - * XXX shmem_file_[gs]et_gfp_mask() - */ - mapping_set_gfp_mask(obj->filp->f_path.dentry->d_inode->i_mapping, - GFP_HIGHUSER | - __GFP_COLD | - __GFP_FS | - __GFP_RECLAIMABLE | - __GFP_NORETRY | - __GFP_NOWARN | - __GFP_NOMEMALLOC); - kref_init(&obj->refcount); kref_init(&obj->handlecount); obj->size = size; @@ -205,9 +192,7 @@ drm_gem_handle_delete(struct drm_file *filp, u32 handle) idr_remove(&filp->object_idr, handle); spin_unlock(&filp->table_lock); - mutex_lock(&dev->struct_mutex); - drm_gem_object_handle_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_handle_unreference_unlocked(obj); return 0; } @@ -338,9 +323,7 @@ again: } err: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); return ret; } @@ -371,9 +354,7 @@ drm_gem_open_ioctl(struct drm_device *dev, void *data, return -ENOENT; ret = drm_gem_handle_create(file_priv, obj, &handle); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); if (ret) return ret; @@ -403,7 +384,7 @@ drm_gem_object_release_handle(int id, void *ptr, void *data) { struct drm_gem_object *obj = ptr; - drm_gem_object_handle_unreference(obj); + drm_gem_object_handle_unreference_unlocked(obj); return 0; } @@ -416,16 +397,25 @@ drm_gem_object_release_handle(int id, void *ptr, void *data) void drm_gem_release(struct drm_device *dev, struct drm_file *file_private) { - mutex_lock(&dev->struct_mutex); idr_for_each(&file_private->object_idr, &drm_gem_object_release_handle, NULL); idr_destroy(&file_private->object_idr); - mutex_unlock(&dev->struct_mutex); +} + +static void +drm_gem_object_free_common(struct drm_gem_object *obj) +{ + struct drm_device *dev = obj->dev; + fput(obj->filp); + atomic_dec(&dev->object_count); + atomic_sub(obj->size, &dev->object_memory); + kfree(obj); } /** * Called after the last reference to the object has been lost. + * Must be called holding struct_ mutex * * Frees the object */ @@ -440,14 +430,40 @@ drm_gem_object_free(struct kref *kref) if (dev->driver->gem_free_object != NULL) dev->driver->gem_free_object(obj); - fput(obj->filp); - atomic_dec(&dev->object_count); - atomic_sub(obj->size, &dev->object_memory); - kfree(obj); + drm_gem_object_free_common(obj); } EXPORT_SYMBOL(drm_gem_object_free); /** + * Called after the last reference to the object has been lost. + * Must be called without holding struct_mutex + * + * Frees the object + */ +void +drm_gem_object_free_unlocked(struct kref *kref) +{ + struct drm_gem_object *obj = (struct drm_gem_object *) kref; + struct drm_device *dev = obj->dev; + + if (dev->driver->gem_free_object_unlocked != NULL) + dev->driver->gem_free_object_unlocked(obj); + else if (dev->driver->gem_free_object != NULL) { + mutex_lock(&dev->struct_mutex); + dev->driver->gem_free_object(obj); + mutex_unlock(&dev->struct_mutex); + } + + drm_gem_object_free_common(obj); +} +EXPORT_SYMBOL(drm_gem_object_free_unlocked); + +static void drm_gem_object_ref_bug(struct kref *list_kref) +{ + BUG(); +} + +/** * Called after the last handle to the object has been closed * * Removes any name for the object. Note that this must be @@ -471,8 +487,10 @@ drm_gem_object_handle_free(struct kref *kref) /* * The object name held a reference to this object, drop * that now. + * + * This cannot be the last reference, since the handle holds one too. */ - drm_gem_object_unreference(obj); + kref_put(&obj->refcount, drm_gem_object_ref_bug); } else spin_unlock(&dev->object_name_lock); @@ -490,11 +508,8 @@ EXPORT_SYMBOL(drm_gem_vm_open); void drm_gem_vm_close(struct vm_area_struct *vma) { struct drm_gem_object *obj = vma->vm_private_data; - struct drm_device *dev = obj->dev; - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); } EXPORT_SYMBOL(drm_gem_vm_close); diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c index cdec3297712..2ac074c8f5d 100644 --- a/drivers/gpu/drm/drm_mm.c +++ b/drivers/gpu/drm/drm_mm.c @@ -405,7 +405,8 @@ struct drm_mm_node *drm_mm_search_free_in_range(const struct drm_mm *mm, wasted += alignment - tmp; } - if (entry->size >= size + wasted) { + if (entry->size >= size + wasted && + (entry->start + wasted + size) <= end) { if (!best_match) return entry; if (entry->size < best_size) { diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c index 7e42b7e9d43..014ce24761b 100644 --- a/drivers/gpu/drm/drm_sysfs.c +++ b/drivers/gpu/drm/drm_sysfs.c @@ -70,19 +70,17 @@ static int drm_class_resume(struct device *dev) return 0; } -/* Display the version of drm_core. This doesn't work right in current design */ -static ssize_t version_show(struct class *dev, char *buf) -{ - return sprintf(buf, "%s %d.%d.%d %s\n", CORE_NAME, CORE_MAJOR, - CORE_MINOR, CORE_PATCHLEVEL, CORE_DATE); -} - static char *drm_devnode(struct device *dev, mode_t *mode) { return kasprintf(GFP_KERNEL, "dri/%s", dev_name(dev)); } -static CLASS_ATTR(version, S_IRUGO, version_show, NULL); +static CLASS_ATTR_STRING(version, S_IRUGO, + CORE_NAME " " + __stringify(CORE_MAJOR) "." + __stringify(CORE_MINOR) "." + __stringify(CORE_PATCHLEVEL) " " + CORE_DATE); /** * drm_sysfs_create - create a struct drm_sysfs_class structure @@ -109,7 +107,7 @@ struct class *drm_sysfs_create(struct module *owner, char *name) class->suspend = drm_class_suspend; class->resume = drm_class_resume; - err = class_create_file(class, &class_attr_version); + err = class_create_file(class, &class_attr_version.attr); if (err) goto err_out_class; @@ -132,7 +130,7 @@ void drm_sysfs_destroy(void) { if ((drm_class == NULL) || (IS_ERR(drm_class))) return; - class_remove_file(drm_class, &class_attr_version); + class_remove_file(drm_class, &class_attr_version.attr); class_destroy(drm_class); } diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 9c9998c4dce..1376dfe44c9 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -162,7 +162,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data) struct drm_device *dev = node->minor->dev; drm_i915_private_t *dev_priv = dev->dev_private; - if (!IS_IRONLAKE(dev)) { + if (!HAS_PCH_SPLIT(dev)) { seq_printf(m, "Interrupt enable: %08x\n", I915_READ(IER)); seq_printf(m, "Interrupt identity: %08x\n", @@ -290,7 +290,7 @@ static int i915_batchbuffer_info(struct seq_file *m, void *data) list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { obj = obj_priv->obj; if (obj->read_domains & I915_GEM_DOMAIN_COMMAND) { - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, 0); if (ret) { DRM_ERROR("Failed to get pages: %d\n", ret); spin_unlock(&dev_priv->mm.active_list_lock); @@ -350,6 +350,36 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data) return 0; } +static const char *pin_flag(int pinned) +{ + if (pinned > 0) + return " P"; + else if (pinned < 0) + return " p"; + else + return ""; +} + +static const char *tiling_flag(int tiling) +{ + switch (tiling) { + default: + case I915_TILING_NONE: return ""; + case I915_TILING_X: return " X"; + case I915_TILING_Y: return " Y"; + } +} + +static const char *dirty_flag(int dirty) +{ + return dirty ? " dirty" : ""; +} + +static const char *purgeable_flag(int purgeable) +{ + return purgeable ? " purgeable" : ""; +} + static int i915_error_state(struct seq_file *m, void *unused) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -357,6 +387,7 @@ static int i915_error_state(struct seq_file *m, void *unused) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_error_state *error; unsigned long flags; + int i, page, offset, elt; spin_lock_irqsave(&dev_priv->error_lock, flags); if (!dev_priv->first_error) { @@ -368,6 +399,7 @@ static int i915_error_state(struct seq_file *m, void *unused) seq_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec, error->time.tv_usec); + seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device); seq_printf(m, "EIR: 0x%08x\n", error->eir); seq_printf(m, " PGTBL_ER: 0x%08x\n", error->pgtbl_er); seq_printf(m, " INSTPM: 0x%08x\n", error->instpm); @@ -379,6 +411,59 @@ static int i915_error_state(struct seq_file *m, void *unused) seq_printf(m, " INSTPS: 0x%08x\n", error->instps); seq_printf(m, " INSTDONE1: 0x%08x\n", error->instdone1); } + seq_printf(m, "seqno: 0x%08x\n", error->seqno); + + if (error->active_bo_count) { + seq_printf(m, "Buffers [%d]:\n", error->active_bo_count); + + for (i = 0; i < error->active_bo_count; i++) { + seq_printf(m, " %08x %8zd %08x %08x %08x%s%s%s%s", + error->active_bo[i].gtt_offset, + error->active_bo[i].size, + error->active_bo[i].read_domains, + error->active_bo[i].write_domain, + error->active_bo[i].seqno, + pin_flag(error->active_bo[i].pinned), + tiling_flag(error->active_bo[i].tiling), + dirty_flag(error->active_bo[i].dirty), + purgeable_flag(error->active_bo[i].purgeable)); + + if (error->active_bo[i].name) + seq_printf(m, " (name: %d)", error->active_bo[i].name); + if (error->active_bo[i].fence_reg != I915_FENCE_REG_NONE) + seq_printf(m, " (fence: %d)", error->active_bo[i].fence_reg); + + seq_printf(m, "\n"); + } + } + + for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++) { + if (error->batchbuffer[i]) { + struct drm_i915_error_object *obj = error->batchbuffer[i]; + + seq_printf(m, "--- gtt_offset = 0x%08x\n", obj->gtt_offset); + offset = 0; + for (page = 0; page < obj->page_count; page++) { + for (elt = 0; elt < PAGE_SIZE/4; elt++) { + seq_printf(m, "%08x : %08x\n", offset, obj->pages[page][elt]); + offset += 4; + } + } + } + } + + if (error->ringbuffer) { + struct drm_i915_error_object *obj = error->ringbuffer; + + seq_printf(m, "--- ringbuffer = 0x%08x\n", obj->gtt_offset); + offset = 0; + for (page = 0; page < obj->page_count; page++) { + for (elt = 0; elt < PAGE_SIZE/4; elt++) { + seq_printf(m, "%08x : %08x\n", offset, obj->pages[page][elt]); + offset += 4; + } + } + } out: spin_unlock_irqrestore(&dev_priv->error_lock, flags); @@ -386,6 +471,165 @@ out: return 0; } +static int i915_rstdby_delays(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + u16 crstanddelay = I915_READ16(CRSTANDVID); + + seq_printf(m, "w/ctx: %d, w/o ctx: %d\n", (crstanddelay >> 8) & 0x3f, (crstanddelay & 0x3f)); + + return 0; +} + +static int i915_cur_delayinfo(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + u16 rgvswctl = I915_READ16(MEMSWCTL); + + seq_printf(m, "Last command: 0x%01x\n", (rgvswctl >> 13) & 0x3); + seq_printf(m, "Command status: %d\n", (rgvswctl >> 12) & 1); + seq_printf(m, "P%d DELAY 0x%02x\n", (rgvswctl >> 8) & 0xf, + rgvswctl & 0x3f); + + return 0; +} + +static int i915_delayfreq_table(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + u32 delayfreq; + int i; + + for (i = 0; i < 16; i++) { + delayfreq = I915_READ(PXVFREQ_BASE + i * 4); + seq_printf(m, "P%02dVIDFREQ: 0x%08x\n", i, delayfreq); + } + + return 0; +} + +static inline int MAP_TO_MV(int map) +{ + return 1250 - (map * 25); +} + +static int i915_inttoext_table(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + u32 inttoext; + int i; + + for (i = 1; i <= 32; i++) { + inttoext = I915_READ(INTTOEXT_BASE_ILK + i * 4); + seq_printf(m, "INTTOEXT%02d: 0x%08x\n", i, inttoext); + } + + return 0; +} + +static int i915_drpc_info(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + u32 rgvmodectl = I915_READ(MEMMODECTL); + + seq_printf(m, "HD boost: %s\n", (rgvmodectl & MEMMODE_BOOST_EN) ? + "yes" : "no"); + seq_printf(m, "Boost freq: %d\n", + (rgvmodectl & MEMMODE_BOOST_FREQ_MASK) >> + MEMMODE_BOOST_FREQ_SHIFT); + seq_printf(m, "HW control enabled: %s\n", + rgvmodectl & MEMMODE_HWIDLE_EN ? "yes" : "no"); + seq_printf(m, "SW control enabled: %s\n", + rgvmodectl & MEMMODE_SWMODE_EN ? "yes" : "no"); + seq_printf(m, "Gated voltage change: %s\n", + rgvmodectl & MEMMODE_RCLK_GATE ? "yes" : "no"); + seq_printf(m, "Starting frequency: P%d\n", + (rgvmodectl & MEMMODE_FSTART_MASK) >> MEMMODE_FSTART_SHIFT); + seq_printf(m, "Max frequency: P%d\n", + (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT); + seq_printf(m, "Min frequency: P%d\n", (rgvmodectl & MEMMODE_FMIN_MASK)); + + return 0; +} + +static int i915_fbc_status(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct drm_crtc *crtc; + drm_i915_private_t *dev_priv = dev->dev_private; + bool fbc_enabled = false; + + if (!dev_priv->display.fbc_enabled) { + seq_printf(m, "FBC unsupported on this chipset\n"); + return 0; + } + + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + if (!crtc->enabled) + continue; + if (dev_priv->display.fbc_enabled(crtc)) + fbc_enabled = true; + } + + if (fbc_enabled) { + seq_printf(m, "FBC enabled\n"); + } else { + seq_printf(m, "FBC disabled: "); + switch (dev_priv->no_fbc_reason) { + case FBC_STOLEN_TOO_SMALL: + seq_printf(m, "not enough stolen memory"); + break; + case FBC_UNSUPPORTED_MODE: + seq_printf(m, "mode not supported"); + break; + case FBC_MODE_TOO_LARGE: + seq_printf(m, "mode too large"); + break; + case FBC_BAD_PLANE: + seq_printf(m, "FBC unsupported on plane"); + break; + case FBC_NOT_TILED: + seq_printf(m, "scanout buffer not tiled"); + break; + default: + seq_printf(m, "unknown reason"); + } + seq_printf(m, "\n"); + } + return 0; +} + +static int i915_sr_status(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + bool sr_enabled = false; + + if (IS_I965G(dev) || IS_I945G(dev) || IS_I945GM(dev)) + sr_enabled = I915_READ(FW_BLC_SELF) & FW_BLC_SELF_EN; + else if (IS_I915GM(dev)) + sr_enabled = I915_READ(INSTPM) & INSTPM_SELF_EN; + else if (IS_PINEVIEW(dev)) + sr_enabled = I915_READ(DSPFW3) & PINEVIEW_SELF_REFRESH_EN; + + seq_printf(m, "self-refresh: %s\n", sr_enabled ? "enabled" : + "disabled"); + + return 0; +} + static int i915_wedged_open(struct inode *inode, struct file *filp) @@ -503,6 +747,13 @@ static struct drm_info_list i915_debugfs_list[] = { {"i915_ringbuffer_info", i915_ringbuffer_info, 0}, {"i915_batchbuffers", i915_batchbuffer_info, 0}, {"i915_error_state", i915_error_state, 0}, + {"i915_rstdby_delays", i915_rstdby_delays, 0}, + {"i915_cur_delayinfo", i915_cur_delayinfo, 0}, + {"i915_delayfreq_table", i915_delayfreq_table, 0}, + {"i915_inttoext_table", i915_inttoext_table, 0}, + {"i915_drpc_info", i915_drpc_info, 0}, + {"i915_fbc_status", i915_fbc_status, 0}, + {"i915_sr_status", i915_sr_status, 0}, }; #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list) diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index e660ac07f3b..8bfc0bbf13e 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -35,6 +35,9 @@ #include "i915_drv.h" #include "i915_trace.h" #include <linux/vgaarb.h> +#include <linux/acpi.h> +#include <linux/pnp.h> +#include <linux/vga_switcheroo.h> /* Really want an OS-independent resettable timer. Would like to have * this loop run for (eg) 3 sec, but have the timer reset every time @@ -735,8 +738,10 @@ static int i915_cmdbuffer(struct drm_device *dev, void *data, if (cmdbuf->num_cliprects) { cliprects = kcalloc(cmdbuf->num_cliprects, sizeof(struct drm_clip_rect), GFP_KERNEL); - if (cliprects == NULL) + if (cliprects == NULL) { + ret = -ENOMEM; goto fail_batch_free; + } ret = copy_from_user(cliprects, cmdbuf->cliprects, cmdbuf->num_cliprects * @@ -931,6 +936,120 @@ static int i915_get_bridge_dev(struct drm_device *dev) return 0; } +#define MCHBAR_I915 0x44 +#define MCHBAR_I965 0x48 +#define MCHBAR_SIZE (4*4096) + +#define DEVEN_REG 0x54 +#define DEVEN_MCHBAR_EN (1 << 28) + +/* Allocate space for the MCH regs if needed, return nonzero on error */ +static int +intel_alloc_mchbar_resource(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; + u32 temp_lo, temp_hi = 0; + u64 mchbar_addr; + int ret = 0; + + if (IS_I965G(dev)) + pci_read_config_dword(dev_priv->bridge_dev, reg + 4, &temp_hi); + pci_read_config_dword(dev_priv->bridge_dev, reg, &temp_lo); + mchbar_addr = ((u64)temp_hi << 32) | temp_lo; + + /* If ACPI doesn't have it, assume we need to allocate it ourselves */ +#ifdef CONFIG_PNP + if (mchbar_addr && + pnp_range_reserved(mchbar_addr, mchbar_addr + MCHBAR_SIZE)) { + ret = 0; + goto out; + } +#endif + + /* Get some space for it */ + ret = pci_bus_alloc_resource(dev_priv->bridge_dev->bus, &dev_priv->mch_res, + MCHBAR_SIZE, MCHBAR_SIZE, + PCIBIOS_MIN_MEM, + 0, pcibios_align_resource, + dev_priv->bridge_dev); + if (ret) { + DRM_DEBUG_DRIVER("failed bus alloc: %d\n", ret); + dev_priv->mch_res.start = 0; + goto out; + } + + if (IS_I965G(dev)) + pci_write_config_dword(dev_priv->bridge_dev, reg + 4, + upper_32_bits(dev_priv->mch_res.start)); + + pci_write_config_dword(dev_priv->bridge_dev, reg, + lower_32_bits(dev_priv->mch_res.start)); +out: + return ret; +} + +/* Setup MCHBAR if possible, return true if we should disable it again */ +static void +intel_setup_mchbar(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int mchbar_reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; + u32 temp; + bool enabled; + + dev_priv->mchbar_need_disable = false; + + if (IS_I915G(dev) || IS_I915GM(dev)) { + pci_read_config_dword(dev_priv->bridge_dev, DEVEN_REG, &temp); + enabled = !!(temp & DEVEN_MCHBAR_EN); + } else { + pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); + enabled = temp & 1; + } + + /* If it's already enabled, don't have to do anything */ + if (enabled) + return; + + if (intel_alloc_mchbar_resource(dev)) + return; + + dev_priv->mchbar_need_disable = true; + + /* Space is allocated or reserved, so enable it. */ + if (IS_I915G(dev) || IS_I915GM(dev)) { + pci_write_config_dword(dev_priv->bridge_dev, DEVEN_REG, + temp | DEVEN_MCHBAR_EN); + } else { + pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); + pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp | 1); + } +} + +static void +intel_teardown_mchbar(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int mchbar_reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; + u32 temp; + + if (dev_priv->mchbar_need_disable) { + if (IS_I915G(dev) || IS_I915GM(dev)) { + pci_read_config_dword(dev_priv->bridge_dev, DEVEN_REG, &temp); + temp &= ~DEVEN_MCHBAR_EN; + pci_write_config_dword(dev_priv->bridge_dev, DEVEN_REG, temp); + } else { + pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); + temp &= ~1; + pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp); + } + } + + if (dev_priv->mch_res.start) + release_resource(&dev_priv->mch_res); +} + /** * i915_probe_agp - get AGP bootup configuration * @pdev: PCI device @@ -976,59 +1095,123 @@ static int i915_probe_agp(struct drm_device *dev, uint32_t *aperture_size, * Some of the preallocated space is taken by the GTT * and popup. GTT is 1K per MB of aperture size, and popup is 4K. */ - if (IS_G4X(dev) || IS_PINEVIEW(dev) || IS_IRONLAKE(dev)) + if (IS_G4X(dev) || IS_PINEVIEW(dev) || IS_IRONLAKE(dev) || IS_GEN6(dev)) overhead = 4096; else overhead = (*aperture_size / 1024) + 4096; - switch (tmp & INTEL_GMCH_GMS_MASK) { - case INTEL_855_GMCH_GMS_DISABLED: - DRM_ERROR("video memory is disabled\n"); - return -1; - case INTEL_855_GMCH_GMS_STOLEN_1M: - stolen = 1 * 1024 * 1024; - break; - case INTEL_855_GMCH_GMS_STOLEN_4M: - stolen = 4 * 1024 * 1024; - break; - case INTEL_855_GMCH_GMS_STOLEN_8M: - stolen = 8 * 1024 * 1024; - break; - case INTEL_855_GMCH_GMS_STOLEN_16M: - stolen = 16 * 1024 * 1024; - break; - case INTEL_855_GMCH_GMS_STOLEN_32M: - stolen = 32 * 1024 * 1024; - break; - case INTEL_915G_GMCH_GMS_STOLEN_48M: - stolen = 48 * 1024 * 1024; - break; - case INTEL_915G_GMCH_GMS_STOLEN_64M: - stolen = 64 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_128M: - stolen = 128 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_256M: - stolen = 256 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_96M: - stolen = 96 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_160M: - stolen = 160 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_224M: - stolen = 224 * 1024 * 1024; - break; - case INTEL_GMCH_GMS_STOLEN_352M: - stolen = 352 * 1024 * 1024; - break; - default: - DRM_ERROR("unexpected GMCH_GMS value: 0x%02x\n", - tmp & INTEL_GMCH_GMS_MASK); - return -1; + if (IS_GEN6(dev)) { + /* SNB has memory control reg at 0x50.w */ + pci_read_config_word(dev->pdev, SNB_GMCH_CTRL, &tmp); + + switch (tmp & SNB_GMCH_GMS_STOLEN_MASK) { + case INTEL_855_GMCH_GMS_DISABLED: + DRM_ERROR("video memory is disabled\n"); + return -1; + case SNB_GMCH_GMS_STOLEN_32M: + stolen = 32 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_64M: + stolen = 64 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_96M: + stolen = 96 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_128M: + stolen = 128 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_160M: + stolen = 160 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_192M: + stolen = 192 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_224M: + stolen = 224 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_256M: + stolen = 256 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_288M: + stolen = 288 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_320M: + stolen = 320 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_352M: + stolen = 352 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_384M: + stolen = 384 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_416M: + stolen = 416 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_448M: + stolen = 448 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_480M: + stolen = 480 * 1024 * 1024; + break; + case SNB_GMCH_GMS_STOLEN_512M: + stolen = 512 * 1024 * 1024; + break; + default: + DRM_ERROR("unexpected GMCH_GMS value: 0x%02x\n", + tmp & SNB_GMCH_GMS_STOLEN_MASK); + return -1; + } + } else { + switch (tmp & INTEL_GMCH_GMS_MASK) { + case INTEL_855_GMCH_GMS_DISABLED: + DRM_ERROR("video memory is disabled\n"); + return -1; + case INTEL_855_GMCH_GMS_STOLEN_1M: + stolen = 1 * 1024 * 1024; + break; + case INTEL_855_GMCH_GMS_STOLEN_4M: + stolen = 4 * 1024 * 1024; + break; + case INTEL_855_GMCH_GMS_STOLEN_8M: + stolen = 8 * 1024 * 1024; + break; + case INTEL_855_GMCH_GMS_STOLEN_16M: + stolen = 16 * 1024 * 1024; + break; + case INTEL_855_GMCH_GMS_STOLEN_32M: + stolen = 32 * 1024 * 1024; + break; + case INTEL_915G_GMCH_GMS_STOLEN_48M: + stolen = 48 * 1024 * 1024; + break; + case INTEL_915G_GMCH_GMS_STOLEN_64M: + stolen = 64 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_128M: + stolen = 128 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_256M: + stolen = 256 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_96M: + stolen = 96 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_160M: + stolen = 160 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_224M: + stolen = 224 * 1024 * 1024; + break; + case INTEL_GMCH_GMS_STOLEN_352M: + stolen = 352 * 1024 * 1024; + break; + default: + DRM_ERROR("unexpected GMCH_GMS value: 0x%02x\n", + tmp & INTEL_GMCH_GMS_MASK); + return -1; + } } + *preallocated_size = stolen - overhead; *start = overhead; @@ -1062,7 +1245,7 @@ static unsigned long i915_gtt_to_phys(struct drm_device *dev, int gtt_offset, gtt_size; if (IS_I965G(dev)) { - if (IS_G4X(dev) || IS_IRONLAKE(dev)) { + if (IS_G4X(dev) || IS_IRONLAKE(dev) || IS_GEN6(dev)) { gtt_offset = 2*1024*1024; gtt_size = 2*1024*1024; } else { @@ -1131,6 +1314,7 @@ static void i915_setup_compression(struct drm_device *dev, int size) /* Leave 1M for line length buffer & misc. */ compressed_fb = drm_mm_search_free(&dev_priv->vram, size, 4096, 0); if (!compressed_fb) { + dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL; i915_warn_stolen(dev); return; } @@ -1138,6 +1322,7 @@ static void i915_setup_compression(struct drm_device *dev, int size) compressed_fb = drm_mm_get_block(compressed_fb, size, 4096); if (!compressed_fb) { i915_warn_stolen(dev); + dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL; return; } @@ -1197,6 +1382,32 @@ static unsigned int i915_vga_set_decode(void *cookie, bool state) return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM; } +static void i915_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; + if (state == VGA_SWITCHEROO_ON) { + printk(KERN_INFO "i915: switched off\n"); + /* i915 resume handler doesn't set to D0 */ + pci_set_power_state(dev->pdev, PCI_D0); + i915_resume(dev); + } else { + printk(KERN_ERR "i915: switched off\n"); + i915_suspend(dev, pmm); + } +} + +static bool i915_switcheroo_can_switch(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + bool can_switch; + + spin_lock(&dev->count_lock); + can_switch = (dev->open_count == 0); + spin_unlock(&dev->count_lock); + return can_switch; +} + static int i915_load_modeset_init(struct drm_device *dev, unsigned long prealloc_start, unsigned long prealloc_size, @@ -1258,6 +1469,12 @@ static int i915_load_modeset_init(struct drm_device *dev, if (ret) goto destroy_ringbuffer; + ret = vga_switcheroo_register_client(dev->pdev, + i915_switcheroo_set_state, + i915_switcheroo_can_switch); + if (ret) + goto destroy_ringbuffer; + intel_modeset_init(dev); ret = drm_irq_install(dev); @@ -1279,7 +1496,9 @@ static int i915_load_modeset_init(struct drm_device *dev, return 0; destroy_ringbuffer: + mutex_lock(&dev->struct_mutex); i915_gem_cleanup_ringbuffer(dev); + mutex_unlock(&dev->struct_mutex); out: return ret; } @@ -1443,11 +1662,14 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) dev->driver->get_vblank_counter = i915_get_vblank_counter; dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */ - if (IS_G4X(dev) || IS_IRONLAKE(dev)) { + if (IS_G4X(dev) || IS_IRONLAKE(dev) || IS_GEN6(dev)) { dev->max_vblank_count = 0xffffffff; /* full 32 bit counter */ dev->driver->get_vblank_counter = gm45_get_vblank_counter; } + /* Try to make sure MCHBAR is enabled before poking at it */ + intel_setup_mchbar(dev); + i915_gem_load(dev); /* Init HWS */ @@ -1521,6 +1743,8 @@ int i915_driver_unload(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; + i915_destroy_error_state(dev); + destroy_workqueue(dev_priv->wq); del_timer_sync(&dev_priv->hangcheck_timer); @@ -1542,6 +1766,7 @@ int i915_driver_unload(struct drm_device *dev) dev_priv->child_dev_num = 0; } drm_irq_uninstall(dev); + vga_switcheroo_unregister_client(dev->pdev); vga_client_register(dev->pdev, NULL, NULL, NULL); } @@ -1567,6 +1792,8 @@ int i915_driver_unload(struct drm_device *dev) intel_cleanup_overlay(dev); } + intel_teardown_mchbar(dev); + pci_dev_put(dev_priv->bridge_dev); kfree(dev->dev_private); @@ -1609,6 +1836,7 @@ void i915_driver_lastclose(struct drm_device * dev) if (!dev_priv || drm_core_check_feature(dev, DRIVER_MODESET)) { drm_fb_helper_restore(); + vga_switcheroo_process_delayed_switch(); return; } diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 46d88965852..1b2e95455c0 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -49,6 +49,7 @@ unsigned int i915_lvds_downclock = 0; module_param_named(lvds_downclock, i915_lvds_downclock, int, 0400); static struct drm_driver driver; +extern int intel_agp_enabled; #define INTEL_VGA_DEVICE(id, info) { \ .class = PCI_CLASS_DISPLAY_VGA << 8, \ @@ -120,7 +121,7 @@ const static struct intel_device_info intel_gm45_info = { const static struct intel_device_info intel_pineview_info = { .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .is_i9xx = 1, - .has_pipe_cxsr = 1, + .need_gfx_hws = 1, .has_hotplug = 1, }; @@ -136,6 +137,16 @@ const static struct intel_device_info intel_ironlake_m_info = { .has_hotplug = 1, }; +const static struct intel_device_info intel_sandybridge_d_info = { + .is_i965g = 1, .is_i9xx = 1, .need_gfx_hws = 1, + .has_hotplug = 1, +}; + +const static struct intel_device_info intel_sandybridge_m_info = { + .is_i965g = 1, .is_mobile = 1, .is_i9xx = 1, .need_gfx_hws = 1, + .has_hotplug = 1, +}; + const static struct pci_device_id pciidlist[] = { INTEL_VGA_DEVICE(0x3577, &intel_i830_info), INTEL_VGA_DEVICE(0x2562, &intel_845g_info), @@ -167,6 +178,8 @@ const static struct pci_device_id pciidlist[] = { INTEL_VGA_DEVICE(0xa011, &intel_pineview_info), INTEL_VGA_DEVICE(0x0042, &intel_ironlake_d_info), INTEL_VGA_DEVICE(0x0046, &intel_ironlake_m_info), + INTEL_VGA_DEVICE(0x0102, &intel_sandybridge_d_info), + INTEL_VGA_DEVICE(0x0106, &intel_sandybridge_m_info), {0, 0, 0} }; @@ -174,26 +187,20 @@ const static struct pci_device_id pciidlist[] = { MODULE_DEVICE_TABLE(pci, pciidlist); #endif -static int i915_suspend(struct drm_device *dev, pm_message_t state) +static int i915_drm_freeze(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - if (!dev || !dev_priv) { - DRM_ERROR("dev: %p, dev_priv: %p\n", dev, dev_priv); - DRM_ERROR("DRM not initialized, aborting suspend.\n"); - return -ENODEV; - } - - if (state.event == PM_EVENT_PRETHAW) - return 0; - pci_save_state(dev->pdev); /* If KMS is active, we do the leavevt stuff here */ if (drm_core_check_feature(dev, DRIVER_MODESET)) { - if (i915_gem_idle(dev)) + int error = i915_gem_idle(dev); + if (error) { dev_err(&dev->pdev->dev, - "GEM idle failed, resume may fail\n"); + "GEM idle failed, resume might fail\n"); + return error; + } drm_irq_uninstall(dev); } @@ -201,26 +208,42 @@ static int i915_suspend(struct drm_device *dev, pm_message_t state) intel_opregion_free(dev, 1); + /* Modeset on resume, not lid events */ + dev_priv->modeset_on_lid = 0; + + return 0; +} + +int i915_suspend(struct drm_device *dev, pm_message_t state) +{ + int error; + + if (!dev || !dev->dev_private) { + DRM_ERROR("dev: %p\n", dev); + DRM_ERROR("DRM not initialized, aborting suspend.\n"); + return -ENODEV; + } + + if (state.event == PM_EVENT_PRETHAW) + return 0; + + error = i915_drm_freeze(dev); + if (error) + return error; + if (state.event == PM_EVENT_SUSPEND) { /* Shut down the device */ pci_disable_device(dev->pdev); pci_set_power_state(dev->pdev, PCI_D3hot); } - /* Modeset on resume, not lid events */ - dev_priv->modeset_on_lid = 0; - return 0; } -static int i915_resume(struct drm_device *dev) +static int i915_drm_thaw(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - int ret = 0; - - if (pci_enable_device(dev->pdev)) - return -1; - pci_set_master(dev->pdev); + int error = 0; i915_restore_state(dev); @@ -231,21 +254,28 @@ static int i915_resume(struct drm_device *dev) mutex_lock(&dev->struct_mutex); dev_priv->mm.suspended = 0; - ret = i915_gem_init_ringbuffer(dev); - if (ret != 0) - ret = -1; + error = i915_gem_init_ringbuffer(dev); mutex_unlock(&dev->struct_mutex); drm_irq_install(dev); - } - if (drm_core_check_feature(dev, DRIVER_MODESET)) { + /* Resume the modeset for every activated CRTC */ drm_helper_resume_force_mode(dev); } dev_priv->modeset_on_lid = 0; - return ret; + return error; +} + +int i915_resume(struct drm_device *dev) +{ + if (pci_enable_device(dev->pdev)) + return -EIO; + + pci_set_master(dev->pdev); + + return i915_drm_thaw(dev); } /** @@ -386,57 +416,62 @@ i915_pci_remove(struct pci_dev *pdev) drm_put_dev(dev); } -static int -i915_pci_suspend(struct pci_dev *pdev, pm_message_t state) +static int i915_pm_suspend(struct device *dev) { - struct drm_device *dev = pci_get_drvdata(pdev); + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + int error; - return i915_suspend(dev, state); -} + if (!drm_dev || !drm_dev->dev_private) { + dev_err(dev, "DRM not initialized, aborting suspend.\n"); + return -ENODEV; + } -static int -i915_pci_resume(struct pci_dev *pdev) -{ - struct drm_device *dev = pci_get_drvdata(pdev); + error = i915_drm_freeze(drm_dev); + if (error) + return error; - return i915_resume(dev); -} + pci_disable_device(pdev); + pci_set_power_state(pdev, PCI_D3hot); -static int -i915_pm_suspend(struct device *dev) -{ - return i915_pci_suspend(to_pci_dev(dev), PMSG_SUSPEND); + return 0; } -static int -i915_pm_resume(struct device *dev) +static int i915_pm_resume(struct device *dev) { - return i915_pci_resume(to_pci_dev(dev)); -} + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); -static int -i915_pm_freeze(struct device *dev) -{ - return i915_pci_suspend(to_pci_dev(dev), PMSG_FREEZE); + return i915_resume(drm_dev); } -static int -i915_pm_thaw(struct device *dev) +static int i915_pm_freeze(struct device *dev) { - /* thaw during hibernate, do nothing! */ - return 0; + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + + if (!drm_dev || !drm_dev->dev_private) { + dev_err(dev, "DRM not initialized, aborting suspend.\n"); + return -ENODEV; + } + + return i915_drm_freeze(drm_dev); } -static int -i915_pm_poweroff(struct device *dev) +static int i915_pm_thaw(struct device *dev) { - return i915_pci_suspend(to_pci_dev(dev), PMSG_HIBERNATE); + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + + return i915_drm_thaw(drm_dev); } -static int -i915_pm_restore(struct device *dev) +static int i915_pm_poweroff(struct device *dev) { - return i915_pci_resume(to_pci_dev(dev)); + struct pci_dev *pdev = to_pci_dev(dev); + struct drm_device *drm_dev = pci_get_drvdata(pdev); + + return i915_drm_freeze(drm_dev); } const struct dev_pm_ops i915_pm_ops = { @@ -445,7 +480,7 @@ const struct dev_pm_ops i915_pm_ops = { .freeze = i915_pm_freeze, .thaw = i915_pm_thaw, .poweroff = i915_pm_poweroff, - .restore = i915_pm_restore, + .restore = i915_pm_resume, }; static struct vm_operations_struct i915_gem_vm_ops = { @@ -524,6 +559,11 @@ static struct drm_driver driver = { static int __init i915_init(void) { + if (!intel_agp_enabled) { + DRM_ERROR("drm/i915 can't work without intel_agp module!\n"); + return -ENODEV; + } + driver.num_ioctls = i915_max_ioctl; i915_gem_shrinker_init(); @@ -549,6 +589,11 @@ static int __init i915_init(void) driver.driver_features &= ~DRIVER_MODESET; #endif + if (!(driver.driver_features & DRIVER_MODESET)) { + driver.suspend = i915_suspend; + driver.resume = i915_resume; + } + return drm_init(&driver); } diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 2c1669488b5..979439cfb82 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -150,7 +150,27 @@ struct drm_i915_error_state { u32 instps; u32 instdone1; u32 seqno; + u64 bbaddr; struct timeval time; + struct drm_i915_error_object { + int page_count; + u32 gtt_offset; + u32 *pages[0]; + } *ringbuffer, *batchbuffer[2]; + struct drm_i915_error_buffer { + size_t size; + u32 name; + u32 seqno; + u32 gtt_offset; + u32 read_domains; + u32 write_domain; + u32 fence_reg; + s32 pinned:2; + u32 tiling:2; + u32 dirty:1; + u32 purgeable:1; + } *active_bo; + u32 active_bo_count; }; struct drm_i915_display_funcs { @@ -192,6 +212,14 @@ struct intel_device_info { u8 cursor_needs_physical : 1; }; +enum no_fbc_reason { + FBC_STOLEN_TOO_SMALL, /* not enough space to hold compressed buffers */ + FBC_UNSUPPORTED_MODE, /* interlace or doublescanned mode */ + FBC_MODE_TOO_LARGE, /* mode too large for compression */ + FBC_BAD_PLANE, /* fbc not supported on plane */ + FBC_NOT_TILED, /* buffer not tiled */ +}; + typedef struct drm_i915_private { struct drm_device *dev; @@ -452,6 +480,7 @@ typedef struct drm_i915_private { u32 savePIPEB_DATA_N1; u32 savePIPEB_LINK_M1; u32 savePIPEB_LINK_N1; + u32 saveMCHBAR_RENDER_STANDBY; struct { struct drm_mm gtt_space; @@ -493,6 +522,15 @@ typedef struct drm_i915_private { struct list_head flushing_list; /** + * List of objects currently pending a GPU write flush. + * + * All elements on this list will belong to either the + * active_list or flushing_list, last_rendering_seqno can + * be used to differentiate between the two elements. + */ + struct list_head gpu_write_list; + + /** * LRU list of objects which are not in the ringbuffer and * are ready to unbind, but are still in the GTT. * @@ -581,6 +619,14 @@ typedef struct drm_i915_private { int child_dev_num; struct child_device_config *child_dev; struct drm_connector *int_lvds_connector; + + bool mchbar_need_disable; + + u8 cur_delay; + u8 min_delay; + u8 max_delay; + + enum no_fbc_reason no_fbc_reason; } drm_i915_private_t; /** driver private structure attached to each drm_gem_object */ @@ -592,6 +638,8 @@ struct drm_i915_gem_object { /** This object's place on the active/flushing/inactive lists */ struct list_head list; + /** This object's place on GPU write list */ + struct list_head gpu_write_list; /** This object's place on the fenced object LRU */ struct list_head fence_list; @@ -725,6 +773,8 @@ extern unsigned int i915_fbpercrtc; extern unsigned int i915_powersave; extern unsigned int i915_lvds_downclock; +extern int i915_suspend(struct drm_device *dev, pm_message_t state); +extern int i915_resume(struct drm_device *dev); extern void i915_save_display(struct drm_device *dev); extern void i915_restore_display(struct drm_device *dev); extern int i915_master_create(struct drm_device *dev, struct drm_master *master); @@ -750,6 +800,7 @@ extern int i965_reset(struct drm_device *dev, u8 flags); /* i915_irq.c */ void i915_hangcheck_elapsed(unsigned long data); +void i915_destroy_error_state(struct drm_device *dev); extern int i915_irq_emit(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int i915_irq_wait(struct drm_device *dev, void *data, @@ -872,7 +923,7 @@ int i915_gem_attach_phys_object(struct drm_device *dev, void i915_gem_detach_phys_object(struct drm_device *dev, struct drm_gem_object *obj); void i915_gem_free_all_phys_object(struct drm_device *dev); -int i915_gem_object_get_pages(struct drm_gem_object *obj); +int i915_gem_object_get_pages(struct drm_gem_object *obj, gfp_t gfpmask); void i915_gem_object_put_pages(struct drm_gem_object *obj); void i915_gem_release(struct drm_device * dev, struct drm_file *file_priv); void i915_gem_object_flush_write_domain(struct drm_gem_object *obj); @@ -886,7 +937,8 @@ void i915_gem_object_do_bit_17_swizzle(struct drm_gem_object *obj); void i915_gem_object_save_bit_17_swizzle(struct drm_gem_object *obj); bool i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode); -bool i915_obj_fenceable(struct drm_device *dev, struct drm_gem_object *obj); +bool i915_gem_object_fence_offset_ok(struct drm_gem_object *obj, + int tiling_mode); /* i915_gem_debug.c */ void i915_gem_dump_object(struct drm_gem_object *obj, int len, @@ -1015,7 +1067,7 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); #define IS_845G(dev) ((dev)->pci_device == 0x2562) #define IS_I85X(dev) ((dev)->pci_device == 0x3582) #define IS_I865G(dev) ((dev)->pci_device == 0x2572) -#define IS_I8XX(dev) (INTEL_INFO(dev)->is_i8xx) +#define IS_GEN2(dev) (INTEL_INFO(dev)->is_i8xx) #define IS_I915G(dev) (INTEL_INFO(dev)->is_i915g) #define IS_I915GM(dev) ((dev)->pci_device == 0x2592) #define IS_I945G(dev) ((dev)->pci_device == 0x2772) @@ -1034,8 +1086,29 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); #define IS_I9XX(dev) (INTEL_INFO(dev)->is_i9xx) #define IS_MOBILE(dev) (INTEL_INFO(dev)->is_mobile) +#define IS_GEN3(dev) (IS_I915G(dev) || \ + IS_I915GM(dev) || \ + IS_I945G(dev) || \ + IS_I945GM(dev) || \ + IS_G33(dev) || \ + IS_PINEVIEW(dev)) +#define IS_GEN4(dev) ((dev)->pci_device == 0x2972 || \ + (dev)->pci_device == 0x2982 || \ + (dev)->pci_device == 0x2992 || \ + (dev)->pci_device == 0x29A2 || \ + (dev)->pci_device == 0x2A02 || \ + (dev)->pci_device == 0x2A12 || \ + (dev)->pci_device == 0x2E02 || \ + (dev)->pci_device == 0x2E12 || \ + (dev)->pci_device == 0x2E22 || \ + (dev)->pci_device == 0x2E32 || \ + (dev)->pci_device == 0x2A42 || \ + (dev)->pci_device == 0x2E42) + #define I915_NEED_GFX_HWS(dev) (INTEL_INFO(dev)->need_gfx_hws) +#define IS_GEN6(dev) ((dev)->pci_device == 0x0102) + /* With the 945 and later, Y tiling got adjusted so that it was 32 128-byte * rows, which changed the alignment requirements and fence programming. */ @@ -1056,6 +1129,9 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); #define I915_HAS_FBC(dev) (INTEL_INFO(dev)->has_fbc) #define I915_HAS_RC6(dev) (INTEL_INFO(dev)->has_rc6) +#define HAS_PCH_SPLIT(dev) (IS_IRONLAKE(dev) || \ + IS_GEN6(dev)) + #define PRIMARY_RINGBUFFER_SIZE (128*1024) #endif diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 0c67924ca80..fba37e9f775 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -128,9 +128,7 @@ i915_gem_create_ioctl(struct drm_device *dev, void *data, return -ENOMEM; ret = drm_gem_handle_create(file_priv, obj, &handle); - mutex_lock(&dev->struct_mutex); - drm_gem_object_handle_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_handle_unreference_unlocked(obj); if (ret) return ret; @@ -277,7 +275,7 @@ i915_gem_shmem_pread_fast(struct drm_device *dev, struct drm_gem_object *obj, mutex_lock(&dev->struct_mutex); - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, 0); if (ret != 0) goto fail_unlock; @@ -321,40 +319,24 @@ fail_unlock: return ret; } -static inline gfp_t -i915_gem_object_get_page_gfp_mask (struct drm_gem_object *obj) -{ - return mapping_gfp_mask(obj->filp->f_path.dentry->d_inode->i_mapping); -} - -static inline void -i915_gem_object_set_page_gfp_mask (struct drm_gem_object *obj, gfp_t gfp) -{ - mapping_set_gfp_mask(obj->filp->f_path.dentry->d_inode->i_mapping, gfp); -} - static int i915_gem_object_get_pages_or_evict(struct drm_gem_object *obj) { int ret; - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, __GFP_NORETRY | __GFP_NOWARN); /* If we've insufficient memory to map in the pages, attempt * to make some space by throwing out some old buffers. */ if (ret == -ENOMEM) { struct drm_device *dev = obj->dev; - gfp_t gfp; ret = i915_gem_evict_something(dev, obj->size); if (ret) return ret; - gfp = i915_gem_object_get_page_gfp_mask(obj); - i915_gem_object_set_page_gfp_mask(obj, gfp & ~__GFP_NORETRY); - ret = i915_gem_object_get_pages(obj); - i915_gem_object_set_page_gfp_mask (obj, gfp); + ret = i915_gem_object_get_pages(obj, 0); } return ret; @@ -504,7 +486,7 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data, */ if (args->offset > obj->size || args->size > obj->size || args->offset + args->size > obj->size) { - drm_gem_object_unreference(obj); + drm_gem_object_unreference_unlocked(obj); return -EINVAL; } @@ -517,7 +499,7 @@ i915_gem_pread_ioctl(struct drm_device *dev, void *data, file_priv); } - drm_gem_object_unreference(obj); + drm_gem_object_unreference_unlocked(obj); return ret; } @@ -790,7 +772,7 @@ i915_gem_shmem_pwrite_fast(struct drm_device *dev, struct drm_gem_object *obj, mutex_lock(&dev->struct_mutex); - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, 0); if (ret != 0) goto fail_unlock; @@ -977,7 +959,7 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, */ if (args->offset > obj->size || args->size > obj->size || args->offset + args->size > obj->size) { - drm_gem_object_unreference(obj); + drm_gem_object_unreference_unlocked(obj); return -EINVAL; } @@ -1011,7 +993,7 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, DRM_INFO("pwrite failed %d\n", ret); #endif - drm_gem_object_unreference(obj); + drm_gem_object_unreference_unlocked(obj); return ret; } @@ -1154,9 +1136,7 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, PROT_READ | PROT_WRITE, MAP_SHARED, args->offset); up_write(¤t->mm->mmap_sem); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); if (IS_ERR((void *)addr)) return addr; @@ -1568,6 +1548,8 @@ i915_gem_object_move_to_inactive(struct drm_gem_object *obj) else list_move_tail(&obj_priv->list, &dev_priv->mm.inactive_list); + BUG_ON(!list_empty(&obj_priv->gpu_write_list)); + obj_priv->last_rendering_seqno = 0; if (obj_priv->active) { obj_priv->active = 0; @@ -1576,6 +1558,38 @@ i915_gem_object_move_to_inactive(struct drm_gem_object *obj) i915_verify_inactive(dev, __FILE__, __LINE__); } +static void +i915_gem_process_flushing_list(struct drm_device *dev, + uint32_t flush_domains, uint32_t seqno) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_i915_gem_object *obj_priv, *next; + + list_for_each_entry_safe(obj_priv, next, + &dev_priv->mm.gpu_write_list, + gpu_write_list) { + struct drm_gem_object *obj = obj_priv->obj; + + if ((obj->write_domain & flush_domains) == + obj->write_domain) { + uint32_t old_write_domain = obj->write_domain; + + obj->write_domain = 0; + list_del_init(&obj_priv->gpu_write_list); + i915_gem_object_move_to_active(obj, seqno); + + /* update the fence lru list */ + if (obj_priv->fence_reg != I915_FENCE_REG_NONE) + list_move_tail(&obj_priv->fence_list, + &dev_priv->mm.fence_list); + + trace_i915_gem_object_change_domain(obj, + obj->read_domains, + old_write_domain); + } + } +} + /** * Creates a new sequence number, emitting a write of it to the status page * plus an interrupt, which will trigger i915_user_interrupt_handler. @@ -1634,27 +1648,8 @@ i915_add_request(struct drm_device *dev, struct drm_file *file_priv, /* Associate any objects on the flushing list matching the write * domain we're flushing with our flush. */ - if (flush_domains != 0) { - struct drm_i915_gem_object *obj_priv, *next; - - list_for_each_entry_safe(obj_priv, next, - &dev_priv->mm.flushing_list, list) { - struct drm_gem_object *obj = obj_priv->obj; - - if ((obj->write_domain & flush_domains) == - obj->write_domain) { - uint32_t old_write_domain = obj->write_domain; - - obj->write_domain = 0; - i915_gem_object_move_to_active(obj, seqno); - - trace_i915_gem_object_change_domain(obj, - obj->read_domains, - old_write_domain); - } - } - - } + if (flush_domains != 0) + i915_gem_process_flushing_list(dev, flush_domains, seqno); if (!dev_priv->mm.suspended) { mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD); @@ -1834,7 +1829,7 @@ i915_do_wait_request(struct drm_device *dev, uint32_t seqno, int interruptible) return -EIO; if (!i915_seqno_passed(i915_get_gem_seqno(dev), seqno)) { - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) ier = I915_READ(DEIER) | I915_READ(GTIER); else ier = I915_READ(IER); @@ -2003,6 +1998,7 @@ int i915_gem_object_unbind(struct drm_gem_object *obj) { struct drm_device *dev = obj->dev; + drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv = obj->driver_private; int ret = 0; @@ -2058,8 +2054,10 @@ i915_gem_object_unbind(struct drm_gem_object *obj) } /* Remove ourselves from the LRU list if present. */ + spin_lock(&dev_priv->mm.active_list_lock); if (!list_empty(&obj_priv->list)) list_del_init(&obj_priv->list); + spin_unlock(&dev_priv->mm.active_list_lock); if (i915_gem_object_is_purgeable(obj_priv)) i915_gem_object_truncate(obj); @@ -2097,10 +2095,33 @@ i915_gem_find_inactive_object(struct drm_device *dev, int min_size) } static int -i915_gem_evict_everything(struct drm_device *dev) +i915_gpu_idle(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; + bool lists_empty; uint32_t seqno; + + spin_lock(&dev_priv->mm.active_list_lock); + lists_empty = list_empty(&dev_priv->mm.flushing_list) && + list_empty(&dev_priv->mm.active_list); + spin_unlock(&dev_priv->mm.active_list_lock); + + if (lists_empty) + return 0; + + /* Flush everything onto the inactive list. */ + i915_gem_flush(dev, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS); + seqno = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS); + if (seqno == 0) + return -ENOMEM; + + return i915_wait_request(dev, seqno); +} + +static int +i915_gem_evict_everything(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; int ret; bool lists_empty; @@ -2114,15 +2135,12 @@ i915_gem_evict_everything(struct drm_device *dev) return -ENOSPC; /* Flush everything (on to the inactive lists) and evict */ - i915_gem_flush(dev, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS); - seqno = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS); - if (seqno == 0) - return -ENOMEM; - - ret = i915_wait_request(dev, seqno); + ret = i915_gpu_idle(dev); if (ret) return ret; + BUG_ON(!list_empty(&dev_priv->mm.flushing_list)); + ret = i915_gem_evict_from_inactive_list(dev); if (ret) return ret; @@ -2230,7 +2248,8 @@ i915_gem_evict_something(struct drm_device *dev, int min_size) } int -i915_gem_object_get_pages(struct drm_gem_object *obj) +i915_gem_object_get_pages(struct drm_gem_object *obj, + gfp_t gfpmask) { struct drm_i915_gem_object *obj_priv = obj->driver_private; int page_count, i; @@ -2256,7 +2275,10 @@ i915_gem_object_get_pages(struct drm_gem_object *obj) inode = obj->filp->f_path.dentry->d_inode; mapping = inode->i_mapping; for (i = 0; i < page_count; i++) { - page = read_mapping_page(mapping, i, NULL); + page = read_cache_page_gfp(mapping, i, + mapping_gfp_mask (mapping) | + __GFP_COLD | + gfpmask); if (IS_ERR(page)) { ret = PTR_ERR(page); i915_gem_object_put_pages(obj); @@ -2271,6 +2293,28 @@ i915_gem_object_get_pages(struct drm_gem_object *obj) return 0; } +static void sandybridge_write_fence_reg(struct drm_i915_fence_reg *reg) +{ + struct drm_gem_object *obj = reg->obj; + struct drm_device *dev = obj->dev; + drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_i915_gem_object *obj_priv = obj->driver_private; + int regnum = obj_priv->fence_reg; + uint64_t val; + + val = (uint64_t)((obj_priv->gtt_offset + obj->size - 4096) & + 0xfffff000) << 32; + val |= obj_priv->gtt_offset & 0xfffff000; + val |= (uint64_t)((obj_priv->stride / 128) - 1) << + SANDYBRIDGE_FENCE_PITCH_SHIFT; + + if (obj_priv->tiling_mode == I915_TILING_Y) + val |= 1 << I965_FENCE_TILING_Y_SHIFT; + val |= I965_FENCE_REG_VALID; + + I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + (regnum * 8), val); +} + static void i965_write_fence_reg(struct drm_i915_fence_reg *reg) { struct drm_gem_object *obj = reg->obj; @@ -2367,6 +2411,58 @@ static void i830_write_fence_reg(struct drm_i915_fence_reg *reg) I915_WRITE(FENCE_REG_830_0 + (regnum * 4), val); } +static int i915_find_fence_reg(struct drm_device *dev) +{ + struct drm_i915_fence_reg *reg = NULL; + struct drm_i915_gem_object *obj_priv = NULL; + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_gem_object *obj = NULL; + int i, avail, ret; + + /* First try to find a free reg */ + avail = 0; + for (i = dev_priv->fence_reg_start; i < dev_priv->num_fence_regs; i++) { + reg = &dev_priv->fence_regs[i]; + if (!reg->obj) + return i; + + obj_priv = reg->obj->driver_private; + if (!obj_priv->pin_count) + avail++; + } + + if (avail == 0) + return -ENOSPC; + + /* None available, try to steal one or wait for a user to finish */ + i = I915_FENCE_REG_NONE; + list_for_each_entry(obj_priv, &dev_priv->mm.fence_list, + fence_list) { + obj = obj_priv->obj; + + if (obj_priv->pin_count) + continue; + + /* found one! */ + i = obj_priv->fence_reg; + break; + } + + BUG_ON(i == I915_FENCE_REG_NONE); + + /* We only have a reference on obj from the active list. put_fence_reg + * might drop that one, causing a use-after-free in it. So hold a + * private reference to obj like the other callers of put_fence_reg + * (set_tiling ioctl) do. */ + drm_gem_object_reference(obj); + ret = i915_gem_object_put_fence_reg(obj); + drm_gem_object_unreference(obj); + if (ret != 0) + return ret; + + return i; +} + /** * i915_gem_object_get_fence_reg - set up a fence reg for an object * @obj: object to map through a fence reg @@ -2387,8 +2483,7 @@ i915_gem_object_get_fence_reg(struct drm_gem_object *obj) struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv = obj->driver_private; struct drm_i915_fence_reg *reg = NULL; - struct drm_i915_gem_object *old_obj_priv = NULL; - int i, ret, avail; + int ret; /* Just update our place in the LRU if our fence is getting used. */ if (obj_priv->fence_reg != I915_FENCE_REG_NONE) { @@ -2416,86 +2511,27 @@ i915_gem_object_get_fence_reg(struct drm_gem_object *obj) break; } - /* First try to find a free reg */ - avail = 0; - for (i = dev_priv->fence_reg_start; i < dev_priv->num_fence_regs; i++) { - reg = &dev_priv->fence_regs[i]; - if (!reg->obj) - break; - - old_obj_priv = reg->obj->driver_private; - if (!old_obj_priv->pin_count) - avail++; - } - - /* None available, try to steal one or wait for a user to finish */ - if (i == dev_priv->num_fence_regs) { - struct drm_gem_object *old_obj = NULL; - - if (avail == 0) - return -ENOSPC; - - list_for_each_entry(old_obj_priv, &dev_priv->mm.fence_list, - fence_list) { - old_obj = old_obj_priv->obj; - - if (old_obj_priv->pin_count) - continue; - - /* Take a reference, as otherwise the wait_rendering - * below may cause the object to get freed out from - * under us. - */ - drm_gem_object_reference(old_obj); - - /* i915 uses fences for GPU access to tiled buffers */ - if (IS_I965G(dev) || !old_obj_priv->active) - break; - - /* This brings the object to the head of the LRU if it - * had been written to. The only way this should - * result in us waiting longer than the expected - * optimal amount of time is if there was a - * fence-using buffer later that was read-only. - */ - i915_gem_object_flush_gpu_write_domain(old_obj); - ret = i915_gem_object_wait_rendering(old_obj); - if (ret != 0) { - drm_gem_object_unreference(old_obj); - return ret; - } - - break; - } - - /* - * Zap this virtual mapping so we can set up a fence again - * for this object next time we need it. - */ - i915_gem_release_mmap(old_obj); - - i = old_obj_priv->fence_reg; - reg = &dev_priv->fence_regs[i]; - - old_obj_priv->fence_reg = I915_FENCE_REG_NONE; - list_del_init(&old_obj_priv->fence_list); - - drm_gem_object_unreference(old_obj); - } + ret = i915_find_fence_reg(dev); + if (ret < 0) + return ret; - obj_priv->fence_reg = i; + obj_priv->fence_reg = ret; + reg = &dev_priv->fence_regs[obj_priv->fence_reg]; list_add_tail(&obj_priv->fence_list, &dev_priv->mm.fence_list); reg->obj = obj; - if (IS_I965G(dev)) + if (IS_GEN6(dev)) + sandybridge_write_fence_reg(reg); + else if (IS_I965G(dev)) i965_write_fence_reg(reg); else if (IS_I9XX(dev)) i915_write_fence_reg(reg); else i830_write_fence_reg(reg); - trace_i915_gem_object_get_fence(obj, i, obj_priv->tiling_mode); + trace_i915_gem_object_get_fence(obj, obj_priv->fence_reg, + obj_priv->tiling_mode); return 0; } @@ -2514,9 +2550,12 @@ i915_gem_clear_fence_reg(struct drm_gem_object *obj) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv = obj->driver_private; - if (IS_I965G(dev)) + if (IS_GEN6(dev)) { + I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + + (obj_priv->fence_reg * 8), 0); + } else if (IS_I965G(dev)) { I915_WRITE64(FENCE_REG_965_0 + (obj_priv->fence_reg * 8), 0); - else { + } else { uint32_t fence_reg; if (obj_priv->fence_reg < 8) @@ -2550,6 +2589,12 @@ i915_gem_object_put_fence_reg(struct drm_gem_object *obj) if (obj_priv->fence_reg == I915_FENCE_REG_NONE) return 0; + /* If we've changed tiling, GTT-mappings of the object + * need to re-fault to ensure that the correct fence register + * setup is in place. + */ + i915_gem_release_mmap(obj); + /* On the i915, GPU access to tiled buffers is via a fence, * therefore we must wait for any outstanding access to complete * before clearing the fence. @@ -2558,12 +2603,12 @@ i915_gem_object_put_fence_reg(struct drm_gem_object *obj) int ret; i915_gem_object_flush_gpu_write_domain(obj); - i915_gem_object_flush_gtt_write_domain(obj); ret = i915_gem_object_wait_rendering(obj); if (ret != 0) return ret; } + i915_gem_object_flush_gtt_write_domain(obj); i915_gem_clear_fence_reg (obj); return 0; @@ -2579,7 +2624,7 @@ i915_gem_object_bind_to_gtt(struct drm_gem_object *obj, unsigned alignment) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj_priv = obj->driver_private; struct drm_mm_node *free_space; - bool retry_alloc = false; + gfp_t gfpmask = __GFP_NORETRY | __GFP_NOWARN; int ret; if (obj_priv->madv != I915_MADV_WILLNEED) { @@ -2623,15 +2668,7 @@ i915_gem_object_bind_to_gtt(struct drm_gem_object *obj, unsigned alignment) DRM_INFO("Binding object of size %zd at 0x%08x\n", obj->size, obj_priv->gtt_offset); #endif - if (retry_alloc) { - i915_gem_object_set_page_gfp_mask (obj, - i915_gem_object_get_page_gfp_mask (obj) & ~__GFP_NORETRY); - } - ret = i915_gem_object_get_pages(obj); - if (retry_alloc) { - i915_gem_object_set_page_gfp_mask (obj, - i915_gem_object_get_page_gfp_mask (obj) | __GFP_NORETRY); - } + ret = i915_gem_object_get_pages(obj, gfpmask); if (ret) { drm_mm_put_block(obj_priv->gtt_space); obj_priv->gtt_space = NULL; @@ -2641,9 +2678,9 @@ i915_gem_object_bind_to_gtt(struct drm_gem_object *obj, unsigned alignment) ret = i915_gem_evict_something(dev, obj->size); if (ret) { /* now try to shrink everyone else */ - if (! retry_alloc) { - retry_alloc = true; - goto search_free; + if (gfpmask) { + gfpmask = 0; + goto search_free; } return ret; @@ -2711,7 +2748,6 @@ static void i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj) { struct drm_device *dev = obj->dev; - uint32_t seqno; uint32_t old_write_domain; if ((obj->write_domain & I915_GEM_GPU_DOMAINS) == 0) @@ -2720,9 +2756,8 @@ i915_gem_object_flush_gpu_write_domain(struct drm_gem_object *obj) /* Queue the GPU write cache flushing we need. */ old_write_domain = obj->write_domain; i915_gem_flush(dev, 0, obj->write_domain); - seqno = i915_add_request(dev, NULL, obj->write_domain); - obj->write_domain = 0; - i915_gem_object_move_to_active(obj, seqno); + (void) i915_add_request(dev, NULL, obj->write_domain); + BUG_ON(obj->write_domain); trace_i915_gem_object_change_domain(obj, obj->read_domains, @@ -3261,7 +3296,8 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj, obj_priv->tiling_mode != I915_TILING_NONE; /* Check fence reg constraints and rebind if necessary */ - if (need_fence && !i915_obj_fenceable(dev, obj)) + if (need_fence && !i915_gem_object_fence_offset_ok(obj, + obj_priv->tiling_mode)) i915_gem_object_unbind(obj); /* Choose the GTT offset for our buffer and put it there. */ @@ -3331,6 +3367,16 @@ i915_gem_object_pin_and_relocate(struct drm_gem_object *obj, } /* Validate that the target is in a valid r/w GPU domain */ + if (reloc->write_domain & (reloc->write_domain - 1)) { + DRM_ERROR("reloc with multiple write domains: " + "obj %p target %d offset %d " + "read %08x write %08x", + obj, reloc->target_handle, + (int) reloc->offset, + reloc->read_domains, + reloc->write_domain); + return -EINVAL; + } if (reloc->write_domain & I915_GEM_DOMAIN_CPU || reloc->read_domains & I915_GEM_DOMAIN_CPU) { DRM_ERROR("reloc with read/write CPU domains: " @@ -3584,6 +3630,9 @@ i915_gem_put_relocs_to_user(struct drm_i915_gem_exec_object2 *exec_list, uint32_t reloc_count = 0, i; int ret = 0; + if (relocs == NULL) + return 0; + for (i = 0; i < buffer_count; i++) { struct drm_i915_gem_relocation_entry __user *user_relocs; int unwritten; @@ -3673,7 +3722,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, struct drm_gem_object *batch_obj; struct drm_i915_gem_object *obj_priv; struct drm_clip_rect *cliprects = NULL; - struct drm_i915_gem_relocation_entry *relocs; + struct drm_i915_gem_relocation_entry *relocs = NULL; int ret = 0, ret2, i, pinned = 0; uint64_t exec_offset; uint32_t seqno, flush_domains, reloc_index; @@ -3699,8 +3748,10 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (args->num_cliprects != 0) { cliprects = kcalloc(args->num_cliprects, sizeof(*cliprects), GFP_KERNEL); - if (cliprects == NULL) + if (cliprects == NULL) { + ret = -ENOMEM; goto pre_mutex_err; + } ret = copy_from_user(cliprects, (struct drm_clip_rect __user *) @@ -3742,6 +3793,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (object_list[i] == NULL) { DRM_ERROR("Invalid object handle %d at index %d\n", exec_list[i].handle, i); + /* prevent error path from reading uninitialized data */ + args->buffer_count = i + 1; ret = -EBADF; goto err; } @@ -3750,6 +3803,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (obj_priv->in_execbuffer) { DRM_ERROR("Object %p appears more than once in object list\n", object_list[i]); + /* prevent error path from reading uninitialized data */ + args->buffer_count = i + 1; ret = -EBADF; goto err; } @@ -3863,16 +3918,23 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, i915_gem_flush(dev, dev->invalidate_domains, dev->flush_domains); - if (dev->flush_domains) + if (dev->flush_domains & I915_GEM_GPU_DOMAINS) (void)i915_add_request(dev, file_priv, dev->flush_domains); } for (i = 0; i < args->buffer_count; i++) { struct drm_gem_object *obj = object_list[i]; + struct drm_i915_gem_object *obj_priv = obj->driver_private; uint32_t old_write_domain = obj->write_domain; obj->write_domain = obj->pending_write_domain; + if (obj->write_domain) + list_move_tail(&obj_priv->gpu_write_list, + &dev_priv->mm.gpu_write_list); + else + list_del_init(&obj_priv->gpu_write_list); + trace_i915_gem_object_change_domain(obj, obj->read_domains, old_write_domain); @@ -3946,6 +4008,7 @@ err: mutex_unlock(&dev->struct_mutex); +pre_mutex_err: /* Copy the updated relocations out regardless of current error * state. Failure to update the relocs would mean that the next * time userland calls execbuf, it would do so with presumed offset @@ -3960,7 +4023,6 @@ err: ret = ret2; } -pre_mutex_err: drm_free_large(object_list); kfree(cliprects); @@ -4383,6 +4445,7 @@ int i915_gem_init_object(struct drm_gem_object *obj) obj_priv->obj = obj; obj_priv->fence_reg = I915_FENCE_REG_NONE; INIT_LIST_HEAD(&obj_priv->list); + INIT_LIST_HEAD(&obj_priv->gpu_write_list); INIT_LIST_HEAD(&obj_priv->fence_list); obj_priv->madv = I915_MADV_WILLNEED; @@ -4442,8 +4505,7 @@ int i915_gem_idle(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; - uint32_t seqno, cur_seqno, last_seqno; - int stuck, ret; + int ret; mutex_lock(&dev->struct_mutex); @@ -4452,115 +4514,36 @@ i915_gem_idle(struct drm_device *dev) return 0; } - /* Hack! Don't let anybody do execbuf while we don't control the chip. - * We need to replace this with a semaphore, or something. - */ - dev_priv->mm.suspended = 1; - del_timer(&dev_priv->hangcheck_timer); - - /* Cancel the retire work handler, wait for it to finish if running - */ - mutex_unlock(&dev->struct_mutex); - cancel_delayed_work_sync(&dev_priv->mm.retire_work); - mutex_lock(&dev->struct_mutex); - - i915_kernel_lost_context(dev); - - /* Flush the GPU along with all non-CPU write domains - */ - i915_gem_flush(dev, I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS); - seqno = i915_add_request(dev, NULL, I915_GEM_GPU_DOMAINS); - - if (seqno == 0) { + ret = i915_gpu_idle(dev); + if (ret) { mutex_unlock(&dev->struct_mutex); - return -ENOMEM; + return ret; } - dev_priv->mm.waiting_gem_seqno = seqno; - last_seqno = 0; - stuck = 0; - for (;;) { - cur_seqno = i915_get_gem_seqno(dev); - if (i915_seqno_passed(cur_seqno, seqno)) - break; - if (last_seqno == cur_seqno) { - if (stuck++ > 100) { - DRM_ERROR("hardware wedged\n"); - atomic_set(&dev_priv->mm.wedged, 1); - DRM_WAKEUP(&dev_priv->irq_queue); - break; - } + /* Under UMS, be paranoid and evict. */ + if (!drm_core_check_feature(dev, DRIVER_MODESET)) { + ret = i915_gem_evict_from_inactive_list(dev); + if (ret) { + mutex_unlock(&dev->struct_mutex); + return ret; } - msleep(10); - last_seqno = cur_seqno; - } - dev_priv->mm.waiting_gem_seqno = 0; - - i915_gem_retire_requests(dev); - - spin_lock(&dev_priv->mm.active_list_lock); - if (!atomic_read(&dev_priv->mm.wedged)) { - /* Active and flushing should now be empty as we've - * waited for a sequence higher than any pending execbuffer - */ - WARN_ON(!list_empty(&dev_priv->mm.active_list)); - WARN_ON(!list_empty(&dev_priv->mm.flushing_list)); - /* Request should now be empty as we've also waited - * for the last request in the list - */ - WARN_ON(!list_empty(&dev_priv->mm.request_list)); } - /* Empty the active and flushing lists to inactive. If there's - * anything left at this point, it means that we're wedged and - * nothing good's going to happen by leaving them there. So strip - * the GPU domains and just stuff them onto inactive. + /* Hack! Don't let anybody do execbuf while we don't control the chip. + * We need to replace this with a semaphore, or something. + * And not confound mm.suspended! */ - while (!list_empty(&dev_priv->mm.active_list)) { - struct drm_gem_object *obj; - uint32_t old_write_domain; - - obj = list_first_entry(&dev_priv->mm.active_list, - struct drm_i915_gem_object, - list)->obj; - old_write_domain = obj->write_domain; - obj->write_domain &= ~I915_GEM_GPU_DOMAINS; - i915_gem_object_move_to_inactive(obj); - - trace_i915_gem_object_change_domain(obj, - obj->read_domains, - old_write_domain); - } - spin_unlock(&dev_priv->mm.active_list_lock); - - while (!list_empty(&dev_priv->mm.flushing_list)) { - struct drm_gem_object *obj; - uint32_t old_write_domain; - - obj = list_first_entry(&dev_priv->mm.flushing_list, - struct drm_i915_gem_object, - list)->obj; - old_write_domain = obj->write_domain; - obj->write_domain &= ~I915_GEM_GPU_DOMAINS; - i915_gem_object_move_to_inactive(obj); - - trace_i915_gem_object_change_domain(obj, - obj->read_domains, - old_write_domain); - } - - - /* Move all inactive buffers out of the GTT. */ - ret = i915_gem_evict_from_inactive_list(dev); - WARN_ON(!list_empty(&dev_priv->mm.inactive_list)); - if (ret) { - mutex_unlock(&dev->struct_mutex); - return ret; - } + dev_priv->mm.suspended = 1; + del_timer(&dev_priv->hangcheck_timer); + i915_kernel_lost_context(dev); i915_gem_cleanup_ringbuffer(dev); + mutex_unlock(&dev->struct_mutex); + /* Cancel the retire work handler, which should be idle now. */ + cancel_delayed_work_sync(&dev_priv->mm.retire_work); + return 0; } @@ -4604,8 +4587,13 @@ i915_gem_init_hws(struct drm_device *dev) } dev_priv->hws_obj = obj; memset(dev_priv->hw_status_page, 0, PAGE_SIZE); - I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr); - I915_READ(HWS_PGA); /* posting read */ + if (IS_GEN6(dev)) { + I915_WRITE(HWS_PGA_GEN6, dev_priv->status_gfx_addr); + I915_READ(HWS_PGA_GEN6); /* posting read */ + } else { + I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr); + I915_READ(HWS_PGA); /* posting read */ + } DRM_DEBUG_DRIVER("hws offset: 0x%08x\n", dev_priv->status_gfx_addr); return 0; @@ -4834,6 +4822,7 @@ i915_gem_load(struct drm_device *dev) spin_lock_init(&dev_priv->mm.active_list_lock); INIT_LIST_HEAD(&dev_priv->mm.active_list); INIT_LIST_HEAD(&dev_priv->mm.flushing_list); + INIT_LIST_HEAD(&dev_priv->mm.gpu_write_list); INIT_LIST_HEAD(&dev_priv->mm.inactive_list); INIT_LIST_HEAD(&dev_priv->mm.request_list); INIT_LIST_HEAD(&dev_priv->mm.fence_list); @@ -4846,7 +4835,8 @@ i915_gem_load(struct drm_device *dev) spin_unlock(&shrink_list_lock); /* Old X drivers will take 0-2 for front, back, depth buffers */ - dev_priv->fence_reg_start = 3; + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + dev_priv->fence_reg_start = 3; if (IS_I965G(dev) || IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) dev_priv->num_fence_regs = 16; @@ -4946,7 +4936,7 @@ void i915_gem_detach_phys_object(struct drm_device *dev, if (!obj_priv->phys_obj) return; - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, 0); if (ret) goto out; @@ -5004,7 +4994,7 @@ i915_gem_attach_phys_object(struct drm_device *dev, obj_priv->phys_obj = dev_priv->mm.phys_objs[id - 1]; obj_priv->phys_obj->cur_obj = obj; - ret = i915_gem_object_get_pages(obj); + ret = i915_gem_object_get_pages(obj, 0); if (ret) { DRM_ERROR("failed to get page list\n"); goto out; diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c index df278b2685b..b5c55d88ff7 100644 --- a/drivers/gpu/drm/i915/i915_gem_tiling.c +++ b/drivers/gpu/drm/i915/i915_gem_tiling.c @@ -25,8 +25,6 @@ * */ -#include <linux/acpi.h> -#include <linux/pnp.h> #include "linux/string.h" #include "linux/bitops.h" #include "drmP.h" @@ -83,120 +81,6 @@ * to match what the GPU expects. */ -#define MCHBAR_I915 0x44 -#define MCHBAR_I965 0x48 -#define MCHBAR_SIZE (4*4096) - -#define DEVEN_REG 0x54 -#define DEVEN_MCHBAR_EN (1 << 28) - -/* Allocate space for the MCH regs if needed, return nonzero on error */ -static int -intel_alloc_mchbar_resource(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - int reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; - u32 temp_lo, temp_hi = 0; - u64 mchbar_addr; - int ret = 0; - - if (IS_I965G(dev)) - pci_read_config_dword(dev_priv->bridge_dev, reg + 4, &temp_hi); - pci_read_config_dword(dev_priv->bridge_dev, reg, &temp_lo); - mchbar_addr = ((u64)temp_hi << 32) | temp_lo; - - /* If ACPI doesn't have it, assume we need to allocate it ourselves */ -#ifdef CONFIG_PNP - if (mchbar_addr && - pnp_range_reserved(mchbar_addr, mchbar_addr + MCHBAR_SIZE)) { - ret = 0; - goto out; - } -#endif - - /* Get some space for it */ - ret = pci_bus_alloc_resource(dev_priv->bridge_dev->bus, &dev_priv->mch_res, - MCHBAR_SIZE, MCHBAR_SIZE, - PCIBIOS_MIN_MEM, - 0, pcibios_align_resource, - dev_priv->bridge_dev); - if (ret) { - DRM_DEBUG_DRIVER("failed bus alloc: %d\n", ret); - dev_priv->mch_res.start = 0; - goto out; - } - - if (IS_I965G(dev)) - pci_write_config_dword(dev_priv->bridge_dev, reg + 4, - upper_32_bits(dev_priv->mch_res.start)); - - pci_write_config_dword(dev_priv->bridge_dev, reg, - lower_32_bits(dev_priv->mch_res.start)); -out: - return ret; -} - -/* Setup MCHBAR if possible, return true if we should disable it again */ -static bool -intel_setup_mchbar(struct drm_device *dev) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - int mchbar_reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; - u32 temp; - bool need_disable = false, enabled; - - if (IS_I915G(dev) || IS_I915GM(dev)) { - pci_read_config_dword(dev_priv->bridge_dev, DEVEN_REG, &temp); - enabled = !!(temp & DEVEN_MCHBAR_EN); - } else { - pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); - enabled = temp & 1; - } - - /* If it's already enabled, don't have to do anything */ - if (enabled) - goto out; - - if (intel_alloc_mchbar_resource(dev)) - goto out; - - need_disable = true; - - /* Space is allocated or reserved, so enable it. */ - if (IS_I915G(dev) || IS_I915GM(dev)) { - pci_write_config_dword(dev_priv->bridge_dev, DEVEN_REG, - temp | DEVEN_MCHBAR_EN); - } else { - pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); - pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp | 1); - } -out: - return need_disable; -} - -static void -intel_teardown_mchbar(struct drm_device *dev, bool disable) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - int mchbar_reg = IS_I965G(dev) ? MCHBAR_I965 : MCHBAR_I915; - u32 temp; - - if (disable) { - if (IS_I915G(dev) || IS_I915GM(dev)) { - pci_read_config_dword(dev_priv->bridge_dev, DEVEN_REG, &temp); - temp &= ~DEVEN_MCHBAR_EN; - pci_write_config_dword(dev_priv->bridge_dev, DEVEN_REG, temp); - } else { - pci_read_config_dword(dev_priv->bridge_dev, mchbar_reg, &temp); - temp &= ~1; - pci_write_config_dword(dev_priv->bridge_dev, mchbar_reg, temp); - } - } - - if (dev_priv->mch_res.start) - release_resource(&dev_priv->mch_res); -} - /** * Detects bit 6 swizzling of address lookup between IGD access and CPU * access through main memory. @@ -207,9 +91,8 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev) drm_i915_private_t *dev_priv = dev->dev_private; uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN; uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN; - bool need_disable; - if (IS_IRONLAKE(dev)) { + if (IS_IRONLAKE(dev) || IS_GEN6(dev)) { /* On Ironlake whatever DRAM config, GPU always do * same swizzling setup. */ @@ -224,9 +107,6 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev) } else if (IS_MOBILE(dev)) { uint32_t dcc; - /* Try to make sure MCHBAR is enabled before poking at it */ - need_disable = intel_setup_mchbar(dev); - /* On mobile 9xx chipsets, channel interleave by the CPU is * determined by DCC. For single-channel, neither the CPU * nor the GPU do swizzling. For dual channel interleaved, @@ -266,8 +146,6 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev) swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN; swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN; } - - intel_teardown_mchbar(dev, need_disable); } else { /* The 965, G33, and newer, have a very flexible memory * configuration. It will enable dual-channel mode @@ -302,39 +180,6 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev) dev_priv->mm.bit_6_swizzle_y = swizzle_y; } - -/** - * Returns whether an object is currently fenceable. If not, it may need - * to be unbound and have its pitch adjusted. - */ -bool -i915_obj_fenceable(struct drm_device *dev, struct drm_gem_object *obj) -{ - struct drm_i915_gem_object *obj_priv = obj->driver_private; - - if (IS_I965G(dev)) { - /* The 965 can have fences at any page boundary. */ - if (obj->size & 4095) - return false; - return true; - } else if (IS_I9XX(dev)) { - if (obj_priv->gtt_offset & ~I915_FENCE_START_MASK) - return false; - } else { - if (obj_priv->gtt_offset & ~I830_FENCE_START_MASK) - return false; - } - - /* Power of two sized... */ - if (obj->size & (obj->size - 1)) - return false; - - /* Objects must be size aligned as well */ - if (obj_priv->gtt_offset & (obj->size - 1)) - return false; - return true; -} - /* Check pitch constriants for all chips & tiling formats */ bool i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode) @@ -391,7 +236,7 @@ i915_tiling_ok(struct drm_device *dev, int stride, int size, int tiling_mode) return true; } -static bool +bool i915_gem_object_fence_offset_ok(struct drm_gem_object *obj, int tiling_mode) { struct drm_device *dev = obj->dev; @@ -438,9 +283,7 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, obj_priv = obj->driver_private; if (!i915_tiling_ok(dev, args->stride, obj->size, args->tiling_mode)) { - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); return -EINVAL; } @@ -493,12 +336,6 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, goto err; } - /* If we've changed tiling, GTT-mappings of the object - * need to re-fault to ensure that the correct fence register - * setup is in place. - */ - i915_gem_release_mmap(obj); - obj_priv->tiling_mode = args->tiling_mode; obj_priv->stride = args->stride; } diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 89a071a3e6f..5388354da0d 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -166,7 +166,7 @@ void intel_enable_asle (struct drm_device *dev) { drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) ironlake_enable_display_irq(dev_priv, DE_GSE); else i915_enable_pipestat(dev_priv, 1, @@ -269,6 +269,57 @@ static void i915_hotplug_work_func(struct work_struct *work) drm_sysfs_hotplug_event(dev); } +static void i915_handle_rps_change(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + u32 busy_up, busy_down, max_avg, min_avg; + u16 rgvswctl; + u8 new_delay = dev_priv->cur_delay; + + I915_WRITE(MEMINTRSTS, I915_READ(MEMINTRSTS) & ~MEMINT_EVAL_CHG); + busy_up = I915_READ(RCPREVBSYTUPAVG); + busy_down = I915_READ(RCPREVBSYTDNAVG); + max_avg = I915_READ(RCBMAXAVG); + min_avg = I915_READ(RCBMINAVG); + + /* Handle RCS change request from hw */ + if (busy_up > max_avg) { + if (dev_priv->cur_delay != dev_priv->max_delay) + new_delay = dev_priv->cur_delay - 1; + if (new_delay < dev_priv->max_delay) + new_delay = dev_priv->max_delay; + } else if (busy_down < min_avg) { + if (dev_priv->cur_delay != dev_priv->min_delay) + new_delay = dev_priv->cur_delay + 1; + if (new_delay > dev_priv->min_delay) + new_delay = dev_priv->min_delay; + } + + DRM_DEBUG("rps change requested: %d -> %d\n", + dev_priv->cur_delay, new_delay); + + rgvswctl = I915_READ(MEMSWCTL); + if (rgvswctl & MEMCTL_CMD_STS) { + DRM_ERROR("gpu busy, RCS change rejected\n"); + return; /* still busy with another command */ + } + + /* Program the new state */ + rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | + (new_delay << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; + I915_WRITE(MEMSWCTL, rgvswctl); + POSTING_READ(MEMSWCTL); + + rgvswctl |= MEMCTL_CMD_STS; + I915_WRITE(MEMSWCTL, rgvswctl); + + dev_priv->cur_delay = new_delay; + + DRM_DEBUG("rps changed\n"); + + return; +} + irqreturn_t ironlake_irq_handler(struct drm_device *dev) { drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; @@ -309,12 +360,33 @@ irqreturn_t ironlake_irq_handler(struct drm_device *dev) if (de_iir & DE_GSE) ironlake_opregion_gse_intr(dev); + if (de_iir & DE_PLANEA_FLIP_DONE) { + intel_prepare_page_flip(dev, 0); + intel_finish_page_flip(dev, 0); + } + + if (de_iir & DE_PLANEB_FLIP_DONE) { + intel_prepare_page_flip(dev, 1); + intel_finish_page_flip(dev, 1); + } + + if (de_iir & DE_PIPEA_VBLANK) + drm_handle_vblank(dev, 0); + + if (de_iir & DE_PIPEB_VBLANK) + drm_handle_vblank(dev, 1); + /* check event from PCH */ if ((de_iir & DE_PCH_EVENT) && (pch_iir & SDE_HOTPLUG_MASK)) { queue_work(dev_priv->wq, &dev_priv->hotplug_work); } + if (de_iir & DE_PCU_EVENT) { + I915_WRITE(MEMINTRSTS, I915_READ(MEMINTRSTS)); + i915_handle_rps_change(dev); + } + /* should clear PCH hotplug event before clear CPU irq */ I915_WRITE(SDEIIR, pch_iir); I915_WRITE(GTIIR, gt_iir); @@ -360,6 +432,121 @@ static void i915_error_work_func(struct work_struct *work) } } +static struct drm_i915_error_object * +i915_error_object_create(struct drm_device *dev, + struct drm_gem_object *src) +{ + struct drm_i915_error_object *dst; + struct drm_i915_gem_object *src_priv; + int page, page_count; + + if (src == NULL) + return NULL; + + src_priv = src->driver_private; + if (src_priv->pages == NULL) + return NULL; + + page_count = src->size / PAGE_SIZE; + + dst = kmalloc(sizeof(*dst) + page_count * sizeof (u32 *), GFP_ATOMIC); + if (dst == NULL) + return NULL; + + for (page = 0; page < page_count; page++) { + void *s, *d = kmalloc(PAGE_SIZE, GFP_ATOMIC); + if (d == NULL) + goto unwind; + s = kmap_atomic(src_priv->pages[page], KM_USER0); + memcpy(d, s, PAGE_SIZE); + kunmap_atomic(s, KM_USER0); + dst->pages[page] = d; + } + dst->page_count = page_count; + dst->gtt_offset = src_priv->gtt_offset; + + return dst; + +unwind: + while (page--) + kfree(dst->pages[page]); + kfree(dst); + return NULL; +} + +static void +i915_error_object_free(struct drm_i915_error_object *obj) +{ + int page; + + if (obj == NULL) + return; + + for (page = 0; page < obj->page_count; page++) + kfree(obj->pages[page]); + + kfree(obj); +} + +static void +i915_error_state_free(struct drm_device *dev, + struct drm_i915_error_state *error) +{ + i915_error_object_free(error->batchbuffer[0]); + i915_error_object_free(error->batchbuffer[1]); + i915_error_object_free(error->ringbuffer); + kfree(error->active_bo); + kfree(error); +} + +static u32 +i915_get_bbaddr(struct drm_device *dev, u32 *ring) +{ + u32 cmd; + + if (IS_I830(dev) || IS_845G(dev)) + cmd = MI_BATCH_BUFFER; + else if (IS_I965G(dev)) + cmd = (MI_BATCH_BUFFER_START | (2 << 6) | + MI_BATCH_NON_SECURE_I965); + else + cmd = (MI_BATCH_BUFFER_START | (2 << 6)); + + return ring[0] == cmd ? ring[1] : 0; +} + +static u32 +i915_ringbuffer_last_batch(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u32 head, bbaddr; + u32 *ring; + + /* Locate the current position in the ringbuffer and walk back + * to find the most recently dispatched batch buffer. + */ + bbaddr = 0; + head = I915_READ(PRB0_HEAD) & HEAD_ADDR; + ring = (u32 *)(dev_priv->ring.virtual_start + head); + + while (--ring >= (u32 *)dev_priv->ring.virtual_start) { + bbaddr = i915_get_bbaddr(dev, ring); + if (bbaddr) + break; + } + + if (bbaddr == 0) { + ring = (u32 *)(dev_priv->ring.virtual_start + dev_priv->ring.Size); + while (--ring >= (u32 *)dev_priv->ring.virtual_start) { + bbaddr = i915_get_bbaddr(dev, ring); + if (bbaddr) + break; + } + } + + return bbaddr; +} + /** * i915_capture_error_state - capture an error record for later analysis * @dev: drm device @@ -372,19 +559,26 @@ static void i915_error_work_func(struct work_struct *work) static void i915_capture_error_state(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_gem_object *obj_priv; struct drm_i915_error_state *error; + struct drm_gem_object *batchbuffer[2]; unsigned long flags; + u32 bbaddr; + int count; spin_lock_irqsave(&dev_priv->error_lock, flags); - if (dev_priv->first_error) - goto out; + error = dev_priv->first_error; + spin_unlock_irqrestore(&dev_priv->error_lock, flags); + if (error) + return; error = kmalloc(sizeof(*error), GFP_ATOMIC); if (!error) { - DRM_DEBUG_DRIVER("out ot memory, not capturing error state\n"); - goto out; + DRM_DEBUG_DRIVER("out of memory, not capturing error state\n"); + return; } + error->seqno = i915_get_gem_seqno(dev); error->eir = I915_READ(EIR); error->pgtbl_er = I915_READ(PGTBL_ER); error->pipeastat = I915_READ(PIPEASTAT); @@ -395,6 +589,7 @@ static void i915_capture_error_state(struct drm_device *dev) error->ipehr = I915_READ(IPEHR); error->instdone = I915_READ(INSTDONE); error->acthd = I915_READ(ACTHD); + error->bbaddr = 0; } else { error->ipeir = I915_READ(IPEIR_I965); error->ipehr = I915_READ(IPEHR_I965); @@ -402,14 +597,101 @@ static void i915_capture_error_state(struct drm_device *dev) error->instps = I915_READ(INSTPS); error->instdone1 = I915_READ(INSTDONE1); error->acthd = I915_READ(ACTHD_I965); + error->bbaddr = I915_READ64(BB_ADDR); } - do_gettimeofday(&error->time); + bbaddr = i915_ringbuffer_last_batch(dev); + + /* Grab the current batchbuffer, most likely to have crashed. */ + batchbuffer[0] = NULL; + batchbuffer[1] = NULL; + count = 0; + list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { + struct drm_gem_object *obj = obj_priv->obj; + + if (batchbuffer[0] == NULL && + bbaddr >= obj_priv->gtt_offset && + bbaddr < obj_priv->gtt_offset + obj->size) + batchbuffer[0] = obj; + + if (batchbuffer[1] == NULL && + error->acthd >= obj_priv->gtt_offset && + error->acthd < obj_priv->gtt_offset + obj->size && + batchbuffer[0] != obj) + batchbuffer[1] = obj; - dev_priv->first_error = error; + count++; + } + + /* We need to copy these to an anonymous buffer as the simplest + * method to avoid being overwritten by userpace. + */ + error->batchbuffer[0] = i915_error_object_create(dev, batchbuffer[0]); + error->batchbuffer[1] = i915_error_object_create(dev, batchbuffer[1]); + + /* Record the ringbuffer */ + error->ringbuffer = i915_error_object_create(dev, dev_priv->ring.ring_obj); + + /* Record buffers on the active list. */ + error->active_bo = NULL; + error->active_bo_count = 0; + + if (count) + error->active_bo = kmalloc(sizeof(*error->active_bo)*count, + GFP_ATOMIC); + + if (error->active_bo) { + int i = 0; + list_for_each_entry(obj_priv, &dev_priv->mm.active_list, list) { + struct drm_gem_object *obj = obj_priv->obj; + + error->active_bo[i].size = obj->size; + error->active_bo[i].name = obj->name; + error->active_bo[i].seqno = obj_priv->last_rendering_seqno; + error->active_bo[i].gtt_offset = obj_priv->gtt_offset; + error->active_bo[i].read_domains = obj->read_domains; + error->active_bo[i].write_domain = obj->write_domain; + error->active_bo[i].fence_reg = obj_priv->fence_reg; + error->active_bo[i].pinned = 0; + if (obj_priv->pin_count > 0) + error->active_bo[i].pinned = 1; + if (obj_priv->user_pin_count > 0) + error->active_bo[i].pinned = -1; + error->active_bo[i].tiling = obj_priv->tiling_mode; + error->active_bo[i].dirty = obj_priv->dirty; + error->active_bo[i].purgeable = obj_priv->madv != I915_MADV_WILLNEED; + + if (++i == count) + break; + } + error->active_bo_count = i; + } -out: + do_gettimeofday(&error->time); + + spin_lock_irqsave(&dev_priv->error_lock, flags); + if (dev_priv->first_error == NULL) { + dev_priv->first_error = error; + error = NULL; + } spin_unlock_irqrestore(&dev_priv->error_lock, flags); + + if (error) + i915_error_state_free(dev, error); +} + +void i915_destroy_error_state(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_error_state *error; + + spin_lock(&dev_priv->error_lock); + error = dev_priv->first_error; + dev_priv->first_error = NULL; + spin_unlock(&dev_priv->error_lock); + + if (error) + i915_error_state_free(dev, error); } /** @@ -560,7 +842,7 @@ irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS) atomic_inc(&dev_priv->irq_received); - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return ironlake_irq_handler(dev); iir = I915_READ(IIR); @@ -721,7 +1003,7 @@ void i915_user_irq_get(struct drm_device *dev) spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); if (dev->irq_enabled && (++dev_priv->user_irq_refcount == 1)) { - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) ironlake_enable_graphics_irq(dev_priv, GT_USER_INTERRUPT); else i915_enable_irq(dev_priv, I915_USER_INTERRUPT); @@ -737,7 +1019,7 @@ void i915_user_irq_put(struct drm_device *dev) spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); BUG_ON(dev->irq_enabled && dev_priv->user_irq_refcount <= 0); if (dev->irq_enabled && (--dev_priv->user_irq_refcount == 0)) { - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) ironlake_disable_graphics_irq(dev_priv, GT_USER_INTERRUPT); else i915_disable_irq(dev_priv, I915_USER_INTERRUPT); @@ -844,11 +1126,11 @@ int i915_enable_vblank(struct drm_device *dev, int pipe) if (!(pipeconf & PIPEACONF_ENABLE)) return -EINVAL; - if (IS_IRONLAKE(dev)) - return 0; - spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); - if (IS_I965G(dev)) + if (HAS_PCH_SPLIT(dev)) + ironlake_enable_display_irq(dev_priv, (pipe == 0) ? + DE_PIPEA_VBLANK: DE_PIPEB_VBLANK); + else if (IS_I965G(dev)) i915_enable_pipestat(dev_priv, pipe, PIPE_START_VBLANK_INTERRUPT_ENABLE); else @@ -866,13 +1148,14 @@ void i915_disable_vblank(struct drm_device *dev, int pipe) drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; unsigned long irqflags; - if (IS_IRONLAKE(dev)) - return; - spin_lock_irqsave(&dev_priv->user_irq_lock, irqflags); - i915_disable_pipestat(dev_priv, pipe, - PIPE_VBLANK_INTERRUPT_ENABLE | - PIPE_START_VBLANK_INTERRUPT_ENABLE); + if (HAS_PCH_SPLIT(dev)) + ironlake_disable_display_irq(dev_priv, (pipe == 0) ? + DE_PIPEA_VBLANK: DE_PIPEB_VBLANK); + else + i915_disable_pipestat(dev_priv, pipe, + PIPE_VBLANK_INTERRUPT_ENABLE | + PIPE_START_VBLANK_INTERRUPT_ENABLE); spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags); } @@ -880,7 +1163,7 @@ void i915_enable_interrupt (struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - if (!IS_IRONLAKE(dev)) + if (!HAS_PCH_SPLIT(dev)) opregion_enable_asle(dev); dev_priv->irq_enabled = 1; } @@ -956,7 +1239,11 @@ void i915_hangcheck_elapsed(unsigned long data) struct drm_device *dev = (struct drm_device *)data; drm_i915_private_t *dev_priv = dev->dev_private; uint32_t acthd; - + + /* No reset support on this chip yet. */ + if (IS_GEN6(dev)) + return; + if (!IS_I965G(dev)) acthd = I915_READ(ACTHD); else @@ -1015,13 +1302,14 @@ static int ironlake_irq_postinstall(struct drm_device *dev) { drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private; /* enable kind of interrupts always enabled */ - u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT; + u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT | + DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE; u32 render_mask = GT_USER_INTERRUPT; u32 hotplug_mask = SDE_CRT_HOTPLUG | SDE_PORTB_HOTPLUG | SDE_PORTC_HOTPLUG | SDE_PORTD_HOTPLUG; dev_priv->irq_mask_reg = ~display_mask; - dev_priv->de_irq_enable_reg = display_mask; + dev_priv->de_irq_enable_reg = display_mask | DE_PIPEA_VBLANK | DE_PIPEB_VBLANK; /* should always can generate irq */ I915_WRITE(DEIIR, I915_READ(DEIIR)); @@ -1046,6 +1334,13 @@ static int ironlake_irq_postinstall(struct drm_device *dev) I915_WRITE(SDEIER, dev_priv->pch_irq_enable_reg); (void) I915_READ(SDEIER); + if (IS_IRONLAKE_M(dev)) { + /* Clear & enable PCU event interrupts */ + I915_WRITE(DEIIR, DE_PCU_EVENT); + I915_WRITE(DEIER, I915_READ(DEIER) | DE_PCU_EVENT); + ironlake_enable_display_irq(dev_priv, DE_PCU_EVENT); + } + return 0; } @@ -1058,7 +1353,7 @@ void i915_driver_irq_preinstall(struct drm_device * dev) INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func); INIT_WORK(&dev_priv->error_work, i915_error_work_func); - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { ironlake_irq_preinstall(dev); return; } @@ -1090,7 +1385,7 @@ int i915_driver_irq_postinstall(struct drm_device *dev) dev_priv->vblank_pipe = DRM_I915_VBLANK_PIPE_A | DRM_I915_VBLANK_PIPE_B; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return ironlake_irq_postinstall(dev); /* Unmask the interrupts that we always want on. */ @@ -1178,7 +1473,7 @@ void i915_driver_irq_uninstall(struct drm_device * dev) dev_priv->vblank_pipe = 0; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { ironlake_irq_uninstall(dev); return; } diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 847006c5218..3d59862c7cc 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -53,6 +53,25 @@ #define INTEL_GMCH_GMS_STOLEN_224M (0xc << 4) #define INTEL_GMCH_GMS_STOLEN_352M (0xd << 4) +#define SNB_GMCH_CTRL 0x50 +#define SNB_GMCH_GMS_STOLEN_MASK 0xF8 +#define SNB_GMCH_GMS_STOLEN_32M (1 << 3) +#define SNB_GMCH_GMS_STOLEN_64M (2 << 3) +#define SNB_GMCH_GMS_STOLEN_96M (3 << 3) +#define SNB_GMCH_GMS_STOLEN_128M (4 << 3) +#define SNB_GMCH_GMS_STOLEN_160M (5 << 3) +#define SNB_GMCH_GMS_STOLEN_192M (6 << 3) +#define SNB_GMCH_GMS_STOLEN_224M (7 << 3) +#define SNB_GMCH_GMS_STOLEN_256M (8 << 3) +#define SNB_GMCH_GMS_STOLEN_288M (9 << 3) +#define SNB_GMCH_GMS_STOLEN_320M (0xa << 3) +#define SNB_GMCH_GMS_STOLEN_352M (0xb << 3) +#define SNB_GMCH_GMS_STOLEN_384M (0xc << 3) +#define SNB_GMCH_GMS_STOLEN_416M (0xd << 3) +#define SNB_GMCH_GMS_STOLEN_448M (0xe << 3) +#define SNB_GMCH_GMS_STOLEN_480M (0xf << 3) +#define SNB_GMCH_GMS_STOLEN_512M (0x10 << 3) + /* PCI config space */ #define HPLLCC 0xc0 /* 855 only */ @@ -61,6 +80,7 @@ #define GC_CLOCK_100_200 (1 << 0) #define GC_CLOCK_100_133 (2 << 0) #define GC_CLOCK_166_250 (3 << 0) +#define GCFGC2 0xda #define GCFGC 0xf0 /* 915+ only */ #define GC_LOW_FREQUENCY_ENABLE (1 << 7) #define GC_DISPLAY_CLOCK_190_200_MHZ (0 << 4) @@ -234,6 +254,9 @@ #define I965_FENCE_REG_VALID (1<<0) #define I965_FENCE_MAX_PITCH_VAL 0x0400 +#define FENCE_REG_SANDYBRIDGE_0 0x100000 +#define SANDYBRIDGE_FENCE_PITCH_SHIFT 32 + /* * Instruction and interrupt control regs */ @@ -265,6 +288,7 @@ #define INSTDONE1 0x0207c /* 965+ only */ #define ACTHD_I965 0x02074 #define HWS_PGA 0x02080 +#define HWS_PGA_GEN6 0x04080 #define HWS_ADDRESS_MASK 0xfffff000 #define HWS_START_ADDRESS_SHIFT 4 #define PWRCTXA 0x2088 /* 965GM+ only */ @@ -282,7 +306,7 @@ #define I915_PIPE_CONTROL_NOTIFY_INTERRUPT (1<<18) #define I915_DISPLAY_PORT_INTERRUPT (1<<17) #define I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT (1<<15) -#define I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT (1<<14) +#define I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT (1<<14) /* p-state */ #define I915_HWB_OOM_INTERRUPT (1<<13) #define I915_SYNC_STATUS_INTERRUPT (1<<12) #define I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT (1<<11) @@ -306,11 +330,14 @@ #define I915_ERROR_MEMORY_REFRESH (1<<1) #define I915_ERROR_INSTRUCTION (1<<0) #define INSTPM 0x020c0 +#define INSTPM_SELF_EN (1<<12) /* 915GM only */ #define ACTHD 0x020c8 #define FW_BLC 0x020d8 #define FW_BLC2 0x020dc #define FW_BLC_SELF 0x020e0 /* 915+ only */ -#define FW_BLC_SELF_EN (1<<15) +#define FW_BLC_SELF_EN_MASK (1<<31) +#define FW_BLC_SELF_FIFO_MASK (1<<16) /* 945 only */ +#define FW_BLC_SELF_EN (1<<15) /* 945 only */ #define MM_BURST_LENGTH 0x00700000 #define MM_FIFO_WATERMARK 0x0001F000 #define LM_BURST_LENGTH 0x00000700 @@ -324,6 +351,7 @@ #define CM0_COLOR_EVICT_DISABLE (1<<3) #define CM0_DEPTH_WRITE_DISABLE (1<<1) #define CM0_RC_OP_FLUSH_DISABLE (1<<0) +#define BB_ADDR 0x02140 /* 8 bytes */ #define GFX_FLSH_CNTL 0x02170 /* 915+ only */ @@ -338,6 +366,7 @@ #define FBC_CTL_PERIODIC (1<<30) #define FBC_CTL_INTERVAL_SHIFT (16) #define FBC_CTL_UNCOMPRESSIBLE (1<<14) +#define FBC_C3_IDLE (1<<13) #define FBC_CTL_STRIDE_SHIFT (5) #define FBC_CTL_FENCENO (1<<0) #define FBC_COMMAND 0x0320c @@ -783,10 +812,144 @@ #define CLKCFG_MEM_800 (3 << 4) #define CLKCFG_MEM_MASK (7 << 4) -/** GM965 GM45 render standby register */ -#define MCHBAR_RENDER_STANDBY 0x111B8 +#define CRSTANDVID 0x11100 +#define PXVFREQ_BASE 0x11110 /* P[0-15]VIDFREQ (0x1114c) (Ironlake) */ +#define PXVFREQ_PX_MASK 0x7f000000 +#define PXVFREQ_PX_SHIFT 24 +#define VIDFREQ_BASE 0x11110 +#define VIDFREQ1 0x11110 /* VIDFREQ1-4 (0x1111c) (Cantiga) */ +#define VIDFREQ2 0x11114 +#define VIDFREQ3 0x11118 +#define VIDFREQ4 0x1111c +#define VIDFREQ_P0_MASK 0x1f000000 +#define VIDFREQ_P0_SHIFT 24 +#define VIDFREQ_P0_CSCLK_MASK 0x00f00000 +#define VIDFREQ_P0_CSCLK_SHIFT 20 +#define VIDFREQ_P0_CRCLK_MASK 0x000f0000 +#define VIDFREQ_P0_CRCLK_SHIFT 16 +#define VIDFREQ_P1_MASK 0x00001f00 +#define VIDFREQ_P1_SHIFT 8 +#define VIDFREQ_P1_CSCLK_MASK 0x000000f0 +#define VIDFREQ_P1_CSCLK_SHIFT 4 +#define VIDFREQ_P1_CRCLK_MASK 0x0000000f +#define INTTOEXT_BASE_ILK 0x11300 +#define INTTOEXT_BASE 0x11120 /* INTTOEXT1-8 (0x1113c) */ +#define INTTOEXT_MAP3_SHIFT 24 +#define INTTOEXT_MAP3_MASK (0x1f << INTTOEXT_MAP3_SHIFT) +#define INTTOEXT_MAP2_SHIFT 16 +#define INTTOEXT_MAP2_MASK (0x1f << INTTOEXT_MAP2_SHIFT) +#define INTTOEXT_MAP1_SHIFT 8 +#define INTTOEXT_MAP1_MASK (0x1f << INTTOEXT_MAP1_SHIFT) +#define INTTOEXT_MAP0_SHIFT 0 +#define INTTOEXT_MAP0_MASK (0x1f << INTTOEXT_MAP0_SHIFT) +#define MEMSWCTL 0x11170 /* Ironlake only */ +#define MEMCTL_CMD_MASK 0xe000 +#define MEMCTL_CMD_SHIFT 13 +#define MEMCTL_CMD_RCLK_OFF 0 +#define MEMCTL_CMD_RCLK_ON 1 +#define MEMCTL_CMD_CHFREQ 2 +#define MEMCTL_CMD_CHVID 3 +#define MEMCTL_CMD_VMMOFF 4 +#define MEMCTL_CMD_VMMON 5 +#define MEMCTL_CMD_STS (1<<12) /* write 1 triggers command, clears + when command complete */ +#define MEMCTL_FREQ_MASK 0x0f00 /* jitter, from 0-15 */ +#define MEMCTL_FREQ_SHIFT 8 +#define MEMCTL_SFCAVM (1<<7) +#define MEMCTL_TGT_VID_MASK 0x007f +#define MEMIHYST 0x1117c +#define MEMINTREN 0x11180 /* 16 bits */ +#define MEMINT_RSEXIT_EN (1<<8) +#define MEMINT_CX_SUPR_EN (1<<7) +#define MEMINT_CONT_BUSY_EN (1<<6) +#define MEMINT_AVG_BUSY_EN (1<<5) +#define MEMINT_EVAL_CHG_EN (1<<4) +#define MEMINT_MON_IDLE_EN (1<<3) +#define MEMINT_UP_EVAL_EN (1<<2) +#define MEMINT_DOWN_EVAL_EN (1<<1) +#define MEMINT_SW_CMD_EN (1<<0) +#define MEMINTRSTR 0x11182 /* 16 bits */ +#define MEM_RSEXIT_MASK 0xc000 +#define MEM_RSEXIT_SHIFT 14 +#define MEM_CONT_BUSY_MASK 0x3000 +#define MEM_CONT_BUSY_SHIFT 12 +#define MEM_AVG_BUSY_MASK 0x0c00 +#define MEM_AVG_BUSY_SHIFT 10 +#define MEM_EVAL_CHG_MASK 0x0300 +#define MEM_EVAL_BUSY_SHIFT 8 +#define MEM_MON_IDLE_MASK 0x00c0 +#define MEM_MON_IDLE_SHIFT 6 +#define MEM_UP_EVAL_MASK 0x0030 +#define MEM_UP_EVAL_SHIFT 4 +#define MEM_DOWN_EVAL_MASK 0x000c +#define MEM_DOWN_EVAL_SHIFT 2 +#define MEM_SW_CMD_MASK 0x0003 +#define MEM_INT_STEER_GFX 0 +#define MEM_INT_STEER_CMR 1 +#define MEM_INT_STEER_SMI 2 +#define MEM_INT_STEER_SCI 3 +#define MEMINTRSTS 0x11184 +#define MEMINT_RSEXIT (1<<7) +#define MEMINT_CONT_BUSY (1<<6) +#define MEMINT_AVG_BUSY (1<<5) +#define MEMINT_EVAL_CHG (1<<4) +#define MEMINT_MON_IDLE (1<<3) +#define MEMINT_UP_EVAL (1<<2) +#define MEMINT_DOWN_EVAL (1<<1) +#define MEMINT_SW_CMD (1<<0) +#define MEMMODECTL 0x11190 +#define MEMMODE_BOOST_EN (1<<31) +#define MEMMODE_BOOST_FREQ_MASK 0x0f000000 /* jitter for boost, 0-15 */ +#define MEMMODE_BOOST_FREQ_SHIFT 24 +#define MEMMODE_IDLE_MODE_MASK 0x00030000 +#define MEMMODE_IDLE_MODE_SHIFT 16 +#define MEMMODE_IDLE_MODE_EVAL 0 +#define MEMMODE_IDLE_MODE_CONT 1 +#define MEMMODE_HWIDLE_EN (1<<15) +#define MEMMODE_SWMODE_EN (1<<14) +#define MEMMODE_RCLK_GATE (1<<13) +#define MEMMODE_HW_UPDATE (1<<12) +#define MEMMODE_FSTART_MASK 0x00000f00 /* starting jitter, 0-15 */ +#define MEMMODE_FSTART_SHIFT 8 +#define MEMMODE_FMAX_MASK 0x000000f0 /* max jitter, 0-15 */ +#define MEMMODE_FMAX_SHIFT 4 +#define MEMMODE_FMIN_MASK 0x0000000f /* min jitter, 0-15 */ +#define RCBMAXAVG 0x1119c +#define MEMSWCTL2 0x1119e /* Cantiga only */ +#define SWMEMCMD_RENDER_OFF (0 << 13) +#define SWMEMCMD_RENDER_ON (1 << 13) +#define SWMEMCMD_SWFREQ (2 << 13) +#define SWMEMCMD_TARVID (3 << 13) +#define SWMEMCMD_VRM_OFF (4 << 13) +#define SWMEMCMD_VRM_ON (5 << 13) +#define CMDSTS (1<<12) +#define SFCAVM (1<<11) +#define SWFREQ_MASK 0x0380 /* P0-7 */ +#define SWFREQ_SHIFT 7 +#define TARVID_MASK 0x001f +#define MEMSTAT_CTG 0x111a0 +#define RCBMINAVG 0x111a0 +#define RCUPEI 0x111b0 +#define RCDNEI 0x111b4 +#define MCHBAR_RENDER_STANDBY 0x111b8 #define RCX_SW_EXIT (1<<23) #define RSX_STATUS_MASK 0x00700000 +#define VIDCTL 0x111c0 +#define VIDSTS 0x111c8 +#define VIDSTART 0x111cc /* 8 bits */ +#define MEMSTAT_ILK 0x111f8 +#define MEMSTAT_VID_MASK 0x7f00 +#define MEMSTAT_VID_SHIFT 8 +#define MEMSTAT_PSTATE_MASK 0x00f8 +#define MEMSTAT_PSTATE_SHIFT 3 +#define MEMSTAT_MON_ACTV (1<<2) +#define MEMSTAT_SRC_CTL_MASK 0x0003 +#define MEMSTAT_SRC_CTL_CORE 0 +#define MEMSTAT_SRC_CTL_TRB 1 +#define MEMSTAT_SRC_CTL_THM 2 +#define MEMSTAT_SRC_CTL_STDBY 3 +#define RCPREVBSYTUPAVG 0x113b8 +#define RCPREVBSYTDNAVG 0x113bc #define PEG_BAND_GAP_DATA 0x14d68 /* diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c index a3b90c9561d..ac0d1a73ac2 100644 --- a/drivers/gpu/drm/i915/i915_suspend.c +++ b/drivers/gpu/drm/i915/i915_suspend.c @@ -682,6 +682,8 @@ void i915_restore_display(struct drm_device *dev) I915_WRITE(PCH_PP_OFF_DELAYS, dev_priv->savePP_OFF_DELAYS); I915_WRITE(PCH_PP_DIVISOR, dev_priv->savePP_DIVISOR); I915_WRITE(PCH_PP_CONTROL, dev_priv->savePP_CONTROL); + I915_WRITE(MCHBAR_RENDER_STANDBY, + dev_priv->saveMCHBAR_RENDER_STANDBY); } else { I915_WRITE(PFIT_PGM_RATIOS, dev_priv->savePFIT_PGM_RATIOS); I915_WRITE(BLC_PWM_CTL, dev_priv->saveBLC_PWM_CTL); @@ -745,11 +747,16 @@ int i915_save_state(struct drm_device *dev) dev_priv->saveGTIMR = I915_READ(GTIMR); dev_priv->saveFDI_RXA_IMR = I915_READ(FDI_RXA_IMR); dev_priv->saveFDI_RXB_IMR = I915_READ(FDI_RXB_IMR); + dev_priv->saveMCHBAR_RENDER_STANDBY = + I915_READ(MCHBAR_RENDER_STANDBY); } else { dev_priv->saveIER = I915_READ(IER); dev_priv->saveIMR = I915_READ(IMR); } + if (IS_IRONLAKE_M(dev)) + ironlake_disable_drps(dev); + /* Cache mode state */ dev_priv->saveCACHE_MODE_0 = I915_READ(CACHE_MODE_0); @@ -820,6 +827,9 @@ int i915_restore_state(struct drm_device *dev) /* Clock gating state */ intel_init_clock_gating(dev); + if (IS_IRONLAKE_M(dev)) + ironlake_enable_drps(dev); + /* Cache mode state */ I915_WRITE (CACHE_MODE_0, dev_priv->saveCACHE_MODE_0 | 0xffff0000); diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 15fbc1b5a83..70c9d4ba704 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -247,6 +247,7 @@ static void parse_general_features(struct drm_i915_private *dev_priv, struct bdb_header *bdb) { + struct drm_device *dev = dev_priv->dev; struct bdb_general_features *general; /* Set sensible defaults in case we can't find the general block */ @@ -263,7 +264,7 @@ parse_general_features(struct drm_i915_private *dev_priv, if (IS_I85X(dev_priv->dev)) dev_priv->lvds_ssc_freq = general->ssc_freq ? 66 : 48; - else if (IS_IRONLAKE(dev_priv->dev)) + else if (IS_IRONLAKE(dev_priv->dev) || IS_GEN6(dev)) dev_priv->lvds_ssc_freq = general->ssc_freq ? 100 : 120; else diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index ddefc871edf..fccf07470c8 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -39,7 +39,7 @@ static void intel_crt_dpms(struct drm_encoder *encoder, int mode) struct drm_i915_private *dev_priv = dev->dev_private; u32 temp, reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) reg = PCH_ADPA; else reg = ADPA; @@ -113,7 +113,7 @@ static void intel_crt_mode_set(struct drm_encoder *encoder, else dpll_md_reg = DPLL_B_MD; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) adpa_reg = PCH_ADPA; else adpa_reg = ADPA; @@ -122,7 +122,7 @@ static void intel_crt_mode_set(struct drm_encoder *encoder, * Disable separate mode multiplier used when cloning SDVO to CRT * XXX this needs to be adjusted when we really are cloning */ - if (IS_I965G(dev) && !IS_IRONLAKE(dev)) { + if (IS_I965G(dev) && !HAS_PCH_SPLIT(dev)) { dpll_md = I915_READ(dpll_md_reg); I915_WRITE(dpll_md_reg, dpll_md & ~DPLL_MD_UDI_MULTIPLIER_MASK); @@ -136,11 +136,11 @@ static void intel_crt_mode_set(struct drm_encoder *encoder, if (intel_crtc->pipe == 0) { adpa |= ADPA_PIPE_A_SELECT; - if (!IS_IRONLAKE(dev)) + if (!HAS_PCH_SPLIT(dev)) I915_WRITE(BCLRPAT_A, 0); } else { adpa |= ADPA_PIPE_B_SELECT; - if (!IS_IRONLAKE(dev)) + if (!HAS_PCH_SPLIT(dev)) I915_WRITE(BCLRPAT_B, 0); } @@ -157,6 +157,9 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector) adpa = I915_READ(PCH_ADPA); adpa &= ~ADPA_CRT_HOTPLUG_MASK; + /* disable HPD first */ + I915_WRITE(PCH_ADPA, adpa); + (void)I915_READ(PCH_ADPA); adpa |= (ADPA_CRT_HOTPLUG_PERIOD_128 | ADPA_CRT_HOTPLUG_WARMUP_10MS | @@ -199,7 +202,7 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) u32 hotplug_en; int i, tries = 0; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return intel_ironlake_crt_detect_hotplug(connector); /* @@ -521,7 +524,7 @@ void intel_crt_init(struct drm_device *dev) &intel_output->enc); /* Set up the DDC bus. */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) i2c_reg = PCH_GPIOA; else { i2c_reg = GPIOA; diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 45da78ef4a9..9cd6de5f990 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -232,7 +232,7 @@ struct intel_limit { #define G4X_P2_DISPLAY_PORT_FAST 10 #define G4X_P2_DISPLAY_PORT_LIMIT 0 -/* Ironlake */ +/* Ironlake / Sandybridge */ /* as we calculate clock using (register_value + 2) for N/M1/M2, so here the range value for them is (actual_value-2). */ @@ -240,33 +240,86 @@ struct intel_limit { #define IRONLAKE_DOT_MAX 350000 #define IRONLAKE_VCO_MIN 1760000 #define IRONLAKE_VCO_MAX 3510000 -#define IRONLAKE_N_MIN 1 -#define IRONLAKE_N_MAX 6 -#define IRONLAKE_M_MIN 79 -#define IRONLAKE_M_MAX 127 #define IRONLAKE_M1_MIN 12 #define IRONLAKE_M1_MAX 22 #define IRONLAKE_M2_MIN 5 #define IRONLAKE_M2_MAX 9 -#define IRONLAKE_P_SDVO_DAC_MIN 5 -#define IRONLAKE_P_SDVO_DAC_MAX 80 -#define IRONLAKE_P_LVDS_MIN 28 -#define IRONLAKE_P_LVDS_MAX 112 -#define IRONLAKE_P1_MIN 1 -#define IRONLAKE_P1_MAX 8 -#define IRONLAKE_P2_SDVO_DAC_SLOW 10 -#define IRONLAKE_P2_SDVO_DAC_FAST 5 -#define IRONLAKE_P2_LVDS_SLOW 14 /* single channel */ -#define IRONLAKE_P2_LVDS_FAST 7 /* double channel */ #define IRONLAKE_P2_DOT_LIMIT 225000 /* 225Mhz */ -#define IRONLAKE_P_DISPLAY_PORT_MIN 10 -#define IRONLAKE_P_DISPLAY_PORT_MAX 20 -#define IRONLAKE_P2_DISPLAY_PORT_FAST 10 -#define IRONLAKE_P2_DISPLAY_PORT_SLOW 10 -#define IRONLAKE_P2_DISPLAY_PORT_LIMIT 0 -#define IRONLAKE_P1_DISPLAY_PORT_MIN 1 -#define IRONLAKE_P1_DISPLAY_PORT_MAX 2 +/* We have parameter ranges for different type of outputs. */ + +/* DAC & HDMI Refclk 120Mhz */ +#define IRONLAKE_DAC_N_MIN 1 +#define IRONLAKE_DAC_N_MAX 5 +#define IRONLAKE_DAC_M_MIN 79 +#define IRONLAKE_DAC_M_MAX 127 +#define IRONLAKE_DAC_P_MIN 5 +#define IRONLAKE_DAC_P_MAX 80 +#define IRONLAKE_DAC_P1_MIN 1 +#define IRONLAKE_DAC_P1_MAX 8 +#define IRONLAKE_DAC_P2_SLOW 10 +#define IRONLAKE_DAC_P2_FAST 5 + +/* LVDS single-channel 120Mhz refclk */ +#define IRONLAKE_LVDS_S_N_MIN 1 +#define IRONLAKE_LVDS_S_N_MAX 3 +#define IRONLAKE_LVDS_S_M_MIN 79 +#define IRONLAKE_LVDS_S_M_MAX 118 +#define IRONLAKE_LVDS_S_P_MIN 28 +#define IRONLAKE_LVDS_S_P_MAX 112 +#define IRONLAKE_LVDS_S_P1_MIN 2 +#define IRONLAKE_LVDS_S_P1_MAX 8 +#define IRONLAKE_LVDS_S_P2_SLOW 14 +#define IRONLAKE_LVDS_S_P2_FAST 14 + +/* LVDS dual-channel 120Mhz refclk */ +#define IRONLAKE_LVDS_D_N_MIN 1 +#define IRONLAKE_LVDS_D_N_MAX 3 +#define IRONLAKE_LVDS_D_M_MIN 79 +#define IRONLAKE_LVDS_D_M_MAX 127 +#define IRONLAKE_LVDS_D_P_MIN 14 +#define IRONLAKE_LVDS_D_P_MAX 56 +#define IRONLAKE_LVDS_D_P1_MIN 2 +#define IRONLAKE_LVDS_D_P1_MAX 8 +#define IRONLAKE_LVDS_D_P2_SLOW 7 +#define IRONLAKE_LVDS_D_P2_FAST 7 + +/* LVDS single-channel 100Mhz refclk */ +#define IRONLAKE_LVDS_S_SSC_N_MIN 1 +#define IRONLAKE_LVDS_S_SSC_N_MAX 2 +#define IRONLAKE_LVDS_S_SSC_M_MIN 79 +#define IRONLAKE_LVDS_S_SSC_M_MAX 126 +#define IRONLAKE_LVDS_S_SSC_P_MIN 28 +#define IRONLAKE_LVDS_S_SSC_P_MAX 112 +#define IRONLAKE_LVDS_S_SSC_P1_MIN 2 +#define IRONLAKE_LVDS_S_SSC_P1_MAX 8 +#define IRONLAKE_LVDS_S_SSC_P2_SLOW 14 +#define IRONLAKE_LVDS_S_SSC_P2_FAST 14 + +/* LVDS dual-channel 100Mhz refclk */ +#define IRONLAKE_LVDS_D_SSC_N_MIN 1 +#define IRONLAKE_LVDS_D_SSC_N_MAX 3 +#define IRONLAKE_LVDS_D_SSC_M_MIN 79 +#define IRONLAKE_LVDS_D_SSC_M_MAX 126 +#define IRONLAKE_LVDS_D_SSC_P_MIN 14 +#define IRONLAKE_LVDS_D_SSC_P_MAX 42 +#define IRONLAKE_LVDS_D_SSC_P1_MIN 2 +#define IRONLAKE_LVDS_D_SSC_P1_MAX 6 +#define IRONLAKE_LVDS_D_SSC_P2_SLOW 7 +#define IRONLAKE_LVDS_D_SSC_P2_FAST 7 + +/* DisplayPort */ +#define IRONLAKE_DP_N_MIN 1 +#define IRONLAKE_DP_N_MAX 2 +#define IRONLAKE_DP_M_MIN 81 +#define IRONLAKE_DP_M_MAX 90 +#define IRONLAKE_DP_P_MIN 10 +#define IRONLAKE_DP_P_MAX 20 +#define IRONLAKE_DP_P2_FAST 10 +#define IRONLAKE_DP_P2_SLOW 10 +#define IRONLAKE_DP_P2_LIMIT 0 +#define IRONLAKE_DP_P1_MIN 1 +#define IRONLAKE_DP_P1_MAX 2 static bool intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, @@ -474,33 +527,78 @@ static const intel_limit_t intel_limits_pineview_lvds = { .find_pll = intel_find_best_PLL, }; -static const intel_limit_t intel_limits_ironlake_sdvo = { +static const intel_limit_t intel_limits_ironlake_dac = { .dot = { .min = IRONLAKE_DOT_MIN, .max = IRONLAKE_DOT_MAX }, .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX }, - .n = { .min = IRONLAKE_N_MIN, .max = IRONLAKE_N_MAX }, - .m = { .min = IRONLAKE_M_MIN, .max = IRONLAKE_M_MAX }, + .n = { .min = IRONLAKE_DAC_N_MIN, .max = IRONLAKE_DAC_N_MAX }, + .m = { .min = IRONLAKE_DAC_M_MIN, .max = IRONLAKE_DAC_M_MAX }, .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, - .p = { .min = IRONLAKE_P_SDVO_DAC_MIN, .max = IRONLAKE_P_SDVO_DAC_MAX }, - .p1 = { .min = IRONLAKE_P1_MIN, .max = IRONLAKE_P1_MAX }, + .p = { .min = IRONLAKE_DAC_P_MIN, .max = IRONLAKE_DAC_P_MAX }, + .p1 = { .min = IRONLAKE_DAC_P1_MIN, .max = IRONLAKE_DAC_P1_MAX }, .p2 = { .dot_limit = IRONLAKE_P2_DOT_LIMIT, - .p2_slow = IRONLAKE_P2_SDVO_DAC_SLOW, - .p2_fast = IRONLAKE_P2_SDVO_DAC_FAST }, + .p2_slow = IRONLAKE_DAC_P2_SLOW, + .p2_fast = IRONLAKE_DAC_P2_FAST }, .find_pll = intel_g4x_find_best_PLL, }; -static const intel_limit_t intel_limits_ironlake_lvds = { +static const intel_limit_t intel_limits_ironlake_single_lvds = { .dot = { .min = IRONLAKE_DOT_MIN, .max = IRONLAKE_DOT_MAX }, .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX }, - .n = { .min = IRONLAKE_N_MIN, .max = IRONLAKE_N_MAX }, - .m = { .min = IRONLAKE_M_MIN, .max = IRONLAKE_M_MAX }, + .n = { .min = IRONLAKE_LVDS_S_N_MIN, .max = IRONLAKE_LVDS_S_N_MAX }, + .m = { .min = IRONLAKE_LVDS_S_M_MIN, .max = IRONLAKE_LVDS_S_M_MAX }, .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, - .p = { .min = IRONLAKE_P_LVDS_MIN, .max = IRONLAKE_P_LVDS_MAX }, - .p1 = { .min = IRONLAKE_P1_MIN, .max = IRONLAKE_P1_MAX }, + .p = { .min = IRONLAKE_LVDS_S_P_MIN, .max = IRONLAKE_LVDS_S_P_MAX }, + .p1 = { .min = IRONLAKE_LVDS_S_P1_MIN, .max = IRONLAKE_LVDS_S_P1_MAX }, .p2 = { .dot_limit = IRONLAKE_P2_DOT_LIMIT, - .p2_slow = IRONLAKE_P2_LVDS_SLOW, - .p2_fast = IRONLAKE_P2_LVDS_FAST }, + .p2_slow = IRONLAKE_LVDS_S_P2_SLOW, + .p2_fast = IRONLAKE_LVDS_S_P2_FAST }, + .find_pll = intel_g4x_find_best_PLL, +}; + +static const intel_limit_t intel_limits_ironlake_dual_lvds = { + .dot = { .min = IRONLAKE_DOT_MIN, .max = IRONLAKE_DOT_MAX }, + .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX }, + .n = { .min = IRONLAKE_LVDS_D_N_MIN, .max = IRONLAKE_LVDS_D_N_MAX }, + .m = { .min = IRONLAKE_LVDS_D_M_MIN, .max = IRONLAKE_LVDS_D_M_MAX }, + .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, + .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, + .p = { .min = IRONLAKE_LVDS_D_P_MIN, .max = IRONLAKE_LVDS_D_P_MAX }, + .p1 = { .min = IRONLAKE_LVDS_D_P1_MIN, .max = IRONLAKE_LVDS_D_P1_MAX }, + .p2 = { .dot_limit = IRONLAKE_P2_DOT_LIMIT, + .p2_slow = IRONLAKE_LVDS_D_P2_SLOW, + .p2_fast = IRONLAKE_LVDS_D_P2_FAST }, + .find_pll = intel_g4x_find_best_PLL, +}; + +static const intel_limit_t intel_limits_ironlake_single_lvds_100m = { + .dot = { .min = IRONLAKE_DOT_MIN, .max = IRONLAKE_DOT_MAX }, + .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX }, + .n = { .min = IRONLAKE_LVDS_S_SSC_N_MIN, .max = IRONLAKE_LVDS_S_SSC_N_MAX }, + .m = { .min = IRONLAKE_LVDS_S_SSC_M_MIN, .max = IRONLAKE_LVDS_S_SSC_M_MAX }, + .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, + .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, + .p = { .min = IRONLAKE_LVDS_S_SSC_P_MIN, .max = IRONLAKE_LVDS_S_SSC_P_MAX }, + .p1 = { .min = IRONLAKE_LVDS_S_SSC_P1_MIN,.max = IRONLAKE_LVDS_S_SSC_P1_MAX }, + .p2 = { .dot_limit = IRONLAKE_P2_DOT_LIMIT, + .p2_slow = IRONLAKE_LVDS_S_SSC_P2_SLOW, + .p2_fast = IRONLAKE_LVDS_S_SSC_P2_FAST }, + .find_pll = intel_g4x_find_best_PLL, +}; + +static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = { + .dot = { .min = IRONLAKE_DOT_MIN, .max = IRONLAKE_DOT_MAX }, + .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX }, + .n = { .min = IRONLAKE_LVDS_D_SSC_N_MIN, .max = IRONLAKE_LVDS_D_SSC_N_MAX }, + .m = { .min = IRONLAKE_LVDS_D_SSC_M_MIN, .max = IRONLAKE_LVDS_D_SSC_M_MAX }, + .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, + .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, + .p = { .min = IRONLAKE_LVDS_D_SSC_P_MIN, .max = IRONLAKE_LVDS_D_SSC_P_MAX }, + .p1 = { .min = IRONLAKE_LVDS_D_SSC_P1_MIN,.max = IRONLAKE_LVDS_D_SSC_P1_MAX }, + .p2 = { .dot_limit = IRONLAKE_P2_DOT_LIMIT, + .p2_slow = IRONLAKE_LVDS_D_SSC_P2_SLOW, + .p2_fast = IRONLAKE_LVDS_D_SSC_P2_FAST }, .find_pll = intel_g4x_find_best_PLL, }; @@ -509,34 +607,53 @@ static const intel_limit_t intel_limits_ironlake_display_port = { .max = IRONLAKE_DOT_MAX }, .vco = { .min = IRONLAKE_VCO_MIN, .max = IRONLAKE_VCO_MAX}, - .n = { .min = IRONLAKE_N_MIN, - .max = IRONLAKE_N_MAX }, - .m = { .min = IRONLAKE_M_MIN, - .max = IRONLAKE_M_MAX }, + .n = { .min = IRONLAKE_DP_N_MIN, + .max = IRONLAKE_DP_N_MAX }, + .m = { .min = IRONLAKE_DP_M_MIN, + .max = IRONLAKE_DP_M_MAX }, .m1 = { .min = IRONLAKE_M1_MIN, .max = IRONLAKE_M1_MAX }, .m2 = { .min = IRONLAKE_M2_MIN, .max = IRONLAKE_M2_MAX }, - .p = { .min = IRONLAKE_P_DISPLAY_PORT_MIN, - .max = IRONLAKE_P_DISPLAY_PORT_MAX }, - .p1 = { .min = IRONLAKE_P1_DISPLAY_PORT_MIN, - .max = IRONLAKE_P1_DISPLAY_PORT_MAX}, - .p2 = { .dot_limit = IRONLAKE_P2_DISPLAY_PORT_LIMIT, - .p2_slow = IRONLAKE_P2_DISPLAY_PORT_SLOW, - .p2_fast = IRONLAKE_P2_DISPLAY_PORT_FAST }, + .p = { .min = IRONLAKE_DP_P_MIN, + .max = IRONLAKE_DP_P_MAX }, + .p1 = { .min = IRONLAKE_DP_P1_MIN, + .max = IRONLAKE_DP_P1_MAX}, + .p2 = { .dot_limit = IRONLAKE_DP_P2_LIMIT, + .p2_slow = IRONLAKE_DP_P2_SLOW, + .p2_fast = IRONLAKE_DP_P2_FAST }, .find_pll = intel_find_pll_ironlake_dp, }; static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc) { + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = dev->dev_private; const intel_limit_t *limit; - if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) - limit = &intel_limits_ironlake_lvds; - else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) || + int refclk = 120; + + if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) { + if (dev_priv->lvds_use_ssc && dev_priv->lvds_ssc_freq == 100) + refclk = 100; + + if ((I915_READ(PCH_LVDS) & LVDS_CLKB_POWER_MASK) == + LVDS_CLKB_POWER_UP) { + /* LVDS dual channel */ + if (refclk == 100) + limit = &intel_limits_ironlake_dual_lvds_100m; + else + limit = &intel_limits_ironlake_dual_lvds; + } else { + if (refclk == 100) + limit = &intel_limits_ironlake_single_lvds_100m; + else + limit = &intel_limits_ironlake_single_lvds; + } + } else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) || HAS_eDP) limit = &intel_limits_ironlake_display_port; else - limit = &intel_limits_ironlake_sdvo; + limit = &intel_limits_ironlake_dac; return limit; } @@ -573,7 +690,7 @@ static const intel_limit_t *intel_limit(struct drm_crtc *crtc) struct drm_device *dev = crtc->dev; const intel_limit_t *limit; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) limit = intel_ironlake_limit(crtc); else if (IS_G4X(dev)) { limit = intel_g4x_limit(crtc); @@ -769,7 +886,7 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) { int lvds_reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) lvds_reg = PCH_LVDS; else lvds_reg = LVDS; @@ -914,6 +1031,8 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval) /* enable it... */ fbc_ctl = FBC_CTL_EN | FBC_CTL_PERIODIC; + if (IS_I945GM(dev)) + fbc_ctl |= FBC_C3_IDLE; /* 945 needs special SR handling */ fbc_ctl |= (dev_priv->cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT; fbc_ctl |= (interval & 0x2fff) << FBC_CTL_INTERVAL_SHIFT; if (obj_priv->tiling_mode != I915_TILING_NONE) @@ -1069,25 +1188,30 @@ static void intel_update_fbc(struct drm_crtc *crtc, if (intel_fb->obj->size > dev_priv->cfb_size) { DRM_DEBUG_KMS("framebuffer too large, disabling " "compression\n"); + dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL; goto out_disable; } if ((mode->flags & DRM_MODE_FLAG_INTERLACE) || (mode->flags & DRM_MODE_FLAG_DBLSCAN)) { DRM_DEBUG_KMS("mode incompatible with compression, " "disabling\n"); + dev_priv->no_fbc_reason = FBC_UNSUPPORTED_MODE; goto out_disable; } if ((mode->hdisplay > 2048) || (mode->vdisplay > 1536)) { DRM_DEBUG_KMS("mode too large for compression, disabling\n"); + dev_priv->no_fbc_reason = FBC_MODE_TOO_LARGE; goto out_disable; } if ((IS_I915GM(dev) || IS_I945GM(dev)) && plane != 0) { DRM_DEBUG_KMS("plane not 0, disabling compression\n"); + dev_priv->no_fbc_reason = FBC_BAD_PLANE; goto out_disable; } if (obj_priv->tiling_mode != I915_TILING_X) { DRM_DEBUG_KMS("framebuffer not tiled, disabling compression\n"); + dev_priv->no_fbc_reason = FBC_NOT_TILED; goto out_disable; } @@ -1247,7 +1371,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y, dspcntr &= ~DISPPLANE_TILED; } - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) /* must disable */ dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE; @@ -1308,7 +1432,7 @@ static void i915_disable_vga (struct drm_device *dev) u8 sr1; u32 vga_reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) vga_reg = CPU_VGACNTRL; else vga_reg = VGACNTRL; @@ -1638,6 +1762,7 @@ static void ironlake_crtc_dpms(struct drm_crtc *crtc, int mode) case DRM_MODE_DPMS_OFF: DRM_DEBUG_KMS("crtc %d dpms off\n", pipe); + drm_vblank_off(dev, pipe); /* Disable display plane */ temp = I915_READ(dspcntr_reg); if ((temp & DISPLAY_PLANE_ENABLE) != 0) { @@ -1991,7 +2116,7 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc, struct drm_display_mode *adjusted_mode) { struct drm_device *dev = crtc->dev; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { /* FDI link clock is fixed at 2.7G */ if (mode->clock * 3 > 27000 * 4) return MODE_CLOCK_HIGH; @@ -2519,6 +2644,10 @@ static void g4x_update_wm(struct drm_device *dev, int planea_clock, sr_entries = roundup(sr_entries / cacheline_size, 1); DRM_DEBUG("self-refresh entries: %d\n", sr_entries); I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN); + } else { + /* Turn off self refresh if both pipes are enabled */ + I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF) + & ~FW_BLC_SELF_EN); } DRM_DEBUG("Setting FIFO watermarks - A: %d, B: %d, SR %d\n", @@ -2562,6 +2691,10 @@ static void i965_update_wm(struct drm_device *dev, int planea_clock, srwm = 1; srwm &= 0x3f; I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN); + } else { + /* Turn off self refresh if both pipes are enabled */ + I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF) + & ~FW_BLC_SELF_EN); } DRM_DEBUG_KMS("Setting FIFO watermarks - A: 8, B: 8, C: 8, SR %d\n", @@ -2629,7 +2762,22 @@ static void i9xx_update_wm(struct drm_device *dev, int planea_clock, srwm = total_size - sr_entries; if (srwm < 0) srwm = 1; - I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN | (srwm & 0x3f)); + + if (IS_I945G(dev) || IS_I945GM(dev)) + I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_FIFO_MASK | (srwm & 0xff)); + else if (IS_I915GM(dev)) { + /* 915M has a smaller SRWM field */ + I915_WRITE(FW_BLC_SELF, srwm & 0x3f); + I915_WRITE(INSTPM, I915_READ(INSTPM) | INSTPM_SELF_EN); + } + } else { + /* Turn off self refresh if both pipes are enabled */ + if (IS_I945G(dev) || IS_I945GM(dev)) { + I915_WRITE(FW_BLC_SELF, I915_READ(FW_BLC_SELF) + & ~FW_BLC_SELF_EN); + } else if (IS_I915GM(dev)) { + I915_WRITE(INSTPM, I915_READ(INSTPM) & ~INSTPM_SELF_EN); + } } DRM_DEBUG_KMS("Setting FIFO watermarks - A: %d, B: %d, C: %d, SR %d\n", @@ -2835,7 +2983,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, refclk / 1000); } else if (IS_I9XX(dev)) { refclk = 96000; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) refclk = 120000; /* 120Mhz refclk */ } else { refclk = 48000; @@ -2893,7 +3041,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, } /* FDI link */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { int lane, link_bw, bpp; /* eDP doesn't require FDI link, so just set DP M/N according to current link config */ @@ -2970,7 +3118,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, * PCH B stepping, previous chipset stepping should be * ignoring this setting. */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { temp = I915_READ(PCH_DREF_CONTROL); /* Always enable nonspread source */ temp &= ~DREF_NONSPREAD_SOURCE_MASK; @@ -3017,7 +3165,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, reduced_clock.m2; } - if (!IS_IRONLAKE(dev)) + if (!HAS_PCH_SPLIT(dev)) dpll = DPLL_VGA_MODE_DIS; if (IS_I9XX(dev)) { @@ -3030,7 +3178,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, sdvo_pixel_multiply = adjusted_mode->clock / mode->clock; if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) dpll |= (sdvo_pixel_multiply - 1) << SDVO_MULTIPLIER_SHIFT_HIRES; - else if (IS_IRONLAKE(dev)) + else if (HAS_PCH_SPLIT(dev)) dpll |= (sdvo_pixel_multiply - 1) << PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT; } if (is_dp) @@ -3042,7 +3190,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, else { dpll |= (1 << (clock.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT; /* also FPA1 */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) dpll |= (1 << (clock.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT; if (IS_G4X(dev) && has_reduced_clock) dpll |= (1 << (reduced_clock.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT; @@ -3061,7 +3209,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14; break; } - if (IS_I965G(dev) && !IS_IRONLAKE(dev)) + if (IS_I965G(dev) && !HAS_PCH_SPLIT(dev)) dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT); } else { if (is_lvds) { @@ -3095,7 +3243,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, /* Ironlake's plane is forced to pipe, bit 24 is to enable color space conversion */ - if (!IS_IRONLAKE(dev)) { + if (!HAS_PCH_SPLIT(dev)) { if (pipe == 0) dspcntr &= ~DISPPLANE_SEL_PIPE_MASK; else @@ -3122,14 +3270,14 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, /* Disable the panel fitter if it was on our pipe */ - if (!IS_IRONLAKE(dev) && intel_panel_fitter_pipe(dev) == pipe) + if (!HAS_PCH_SPLIT(dev) && intel_panel_fitter_pipe(dev) == pipe) I915_WRITE(PFIT_CONTROL, 0); DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B'); drm_mode_debug_printmodeline(mode); /* assign to Ironlake registers */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { fp_reg = pch_fp_reg; dpll_reg = pch_dpll_reg; } @@ -3150,7 +3298,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, if (is_lvds) { u32 lvds; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) lvds_reg = PCH_LVDS; lvds = I915_READ(lvds_reg); @@ -3172,12 +3320,12 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, /* set the dithering flag */ if (IS_I965G(dev)) { if (dev_priv->lvds_dither) { - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) pipeconf |= PIPE_ENABLE_DITHER; else lvds |= LVDS_ENABLE_DITHER; } else { - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) pipeconf &= ~PIPE_ENABLE_DITHER; else lvds &= ~LVDS_ENABLE_DITHER; @@ -3196,7 +3344,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, /* Wait for the clocks to stabilize. */ udelay(150); - if (IS_I965G(dev) && !IS_IRONLAKE(dev)) { + if (IS_I965G(dev) && !HAS_PCH_SPLIT(dev)) { if (is_sdvo) { sdvo_pixel_multiply = adjusted_mode->clock / mode->clock; I915_WRITE(dpll_md_reg, (0 << DPLL_MD_UDI_DIVIDER_SHIFT) | @@ -3243,14 +3391,14 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, /* pipesrc and dspsize control the size that is scaled from, which should * always be the user's requested size. */ - if (!IS_IRONLAKE(dev)) { + if (!HAS_PCH_SPLIT(dev)) { I915_WRITE(dspsize_reg, ((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1)); I915_WRITE(dsppos_reg, 0); } I915_WRITE(pipesrc_reg, ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1)); - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { I915_WRITE(data_m1_reg, TU_SIZE(m_n.tu) | m_n.gmch_m); I915_WRITE(data_n1_reg, TU_SIZE(m_n.tu) | m_n.gmch_n); I915_WRITE(link_m1_reg, m_n.link_m); @@ -3306,7 +3454,7 @@ void intel_crtc_load_lut(struct drm_crtc *crtc) return; /* use legacy palette for Ironlake */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) palreg = (intel_crtc->pipe == 0) ? LGC_PALETTE_A : LGC_PALETTE_B; @@ -3421,11 +3569,10 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc, intel_crtc->cursor_bo = bo; return 0; -fail: - mutex_lock(&dev->struct_mutex); fail_locked: - drm_gem_object_unreference(bo); mutex_unlock(&dev->struct_mutex); +fail: + drm_gem_object_unreference_unlocked(bo); return ret; } @@ -3790,7 +3937,7 @@ static void intel_increase_pllclock(struct drm_crtc *crtc, bool schedule) int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; int dpll = I915_READ(dpll_reg); - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return; if (!dev_priv->lvds_downclock_avail) @@ -3829,7 +3976,7 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc) int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; int dpll = I915_READ(dpll_reg); - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return; if (!dev_priv->lvds_downclock_avail) @@ -3879,6 +4026,11 @@ static void intel_idle_update(struct work_struct *work) mutex_lock(&dev->struct_mutex); + if (IS_I945G(dev) || IS_I945GM(dev)) { + DRM_DEBUG_DRIVER("enable memory self refresh on 945\n"); + I915_WRITE(FW_BLC_SELF, FW_BLC_SELF_EN_MASK | FW_BLC_SELF_EN); + } + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { /* Skip inactive CRTCs */ if (!crtc->fb) @@ -3912,9 +4064,17 @@ void intel_mark_busy(struct drm_device *dev, struct drm_gem_object *obj) if (!drm_core_check_feature(dev, DRIVER_MODESET)) return; - if (!dev_priv->busy) + if (!dev_priv->busy) { + if (IS_I945G(dev) || IS_I945GM(dev)) { + u32 fw_blc_self; + + DRM_DEBUG_DRIVER("disable memory self refresh on 945\n"); + fw_blc_self = I915_READ(FW_BLC_SELF); + fw_blc_self &= ~FW_BLC_SELF_EN; + I915_WRITE(FW_BLC_SELF, fw_blc_self | FW_BLC_SELF_EN_MASK); + } dev_priv->busy = true; - else + } else mod_timer(&dev_priv->idle_timer, jiffies + msecs_to_jiffies(GPU_IDLE_TIMEOUT)); @@ -3926,6 +4086,14 @@ void intel_mark_busy(struct drm_device *dev, struct drm_gem_object *obj) intel_fb = to_intel_framebuffer(crtc->fb); if (intel_fb->obj == obj) { if (!intel_crtc->busy) { + if (IS_I945G(dev) || IS_I945GM(dev)) { + u32 fw_blc_self; + + DRM_DEBUG_DRIVER("disable memory self refresh on 945\n"); + fw_blc_self = I915_READ(FW_BLC_SELF); + fw_blc_self &= ~FW_BLC_SELF_EN; + I915_WRITE(FW_BLC_SELF, fw_blc_self | FW_BLC_SELF_EN_MASK); + } /* Non-busy -> busy, upclock */ intel_increase_pllclock(crtc, true); intel_crtc->busy = true; @@ -3949,7 +4117,8 @@ static void intel_crtc_destroy(struct drm_crtc *crtc) struct intel_unpin_work { struct work_struct work; struct drm_device *dev; - struct drm_gem_object *obj; + struct drm_gem_object *old_fb_obj; + struct drm_gem_object *pending_flip_obj; struct drm_pending_vblank_event *event; int pending; }; @@ -3960,8 +4129,9 @@ static void intel_unpin_work_fn(struct work_struct *__work) container_of(__work, struct intel_unpin_work, work); mutex_lock(&work->dev->struct_mutex); - i915_gem_object_unpin(work->obj); - drm_gem_object_unreference(work->obj); + i915_gem_object_unpin(work->old_fb_obj); + drm_gem_object_unreference(work->pending_flip_obj); + drm_gem_object_unreference(work->old_fb_obj); mutex_unlock(&work->dev->struct_mutex); kfree(work); } @@ -3984,6 +4154,12 @@ void intel_finish_page_flip(struct drm_device *dev, int pipe) spin_lock_irqsave(&dev->event_lock, flags); work = intel_crtc->unpin_work; if (work == NULL || !work->pending) { + if (work && !work->pending) { + obj_priv = work->pending_flip_obj->driver_private; + DRM_DEBUG_DRIVER("flip finish: %p (%d) not pending?\n", + obj_priv, + atomic_read(&obj_priv->pending_flip)); + } spin_unlock_irqrestore(&dev->event_lock, flags); return; } @@ -4004,8 +4180,11 @@ void intel_finish_page_flip(struct drm_device *dev, int pipe) spin_unlock_irqrestore(&dev->event_lock, flags); - obj_priv = work->obj->driver_private; - if (atomic_dec_and_test(&obj_priv->pending_flip)) + obj_priv = work->pending_flip_obj->driver_private; + + /* Initial scanout buffer will have a 0 pending flip count */ + if ((atomic_read(&obj_priv->pending_flip) == 0) || + atomic_dec_and_test(&obj_priv->pending_flip)) DRM_WAKEUP(&dev_priv->pending_flip_queue); schedule_work(&work->work); } @@ -4018,8 +4197,11 @@ void intel_prepare_page_flip(struct drm_device *dev, int plane) unsigned long flags; spin_lock_irqsave(&dev->event_lock, flags); - if (intel_crtc->unpin_work) + if (intel_crtc->unpin_work) { intel_crtc->unpin_work->pending = 1; + } else { + DRM_DEBUG_DRIVER("preparing flip with no unpin work?\n"); + } spin_unlock_irqrestore(&dev->event_lock, flags); } @@ -4035,7 +4217,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_unpin_work *work; unsigned long flags; - int ret; + int pipesrc_reg = (intel_crtc->pipe == 0) ? PIPEASRC : PIPEBSRC; + int ret, pipesrc; RING_LOCALS; work = kzalloc(sizeof *work, GFP_KERNEL); @@ -4047,12 +4230,13 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, work->event = event; work->dev = crtc->dev; intel_fb = to_intel_framebuffer(crtc->fb); - work->obj = intel_fb->obj; + work->old_fb_obj = intel_fb->obj; INIT_WORK(&work->work, intel_unpin_work_fn); /* We borrow the event spin lock for protecting unpin_work */ spin_lock_irqsave(&dev->event_lock, flags); if (intel_crtc->unpin_work) { + DRM_DEBUG_DRIVER("flip queue: crtc already busy\n"); spin_unlock_irqrestore(&dev->event_lock, flags); kfree(work); mutex_unlock(&dev->struct_mutex); @@ -4066,19 +4250,24 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, ret = intel_pin_and_fence_fb_obj(dev, obj); if (ret != 0) { + DRM_DEBUG_DRIVER("flip queue: %p pin & fence failed\n", + obj->driver_private); kfree(work); + intel_crtc->unpin_work = NULL; mutex_unlock(&dev->struct_mutex); return ret; } - /* Reference the old fb object for the scheduled work. */ - drm_gem_object_reference(work->obj); + /* Reference the objects for the scheduled work. */ + drm_gem_object_reference(work->old_fb_obj); + drm_gem_object_reference(obj); crtc->fb = fb; i915_gem_object_flush_write_domain(obj); drm_vblank_get(dev, intel_crtc->pipe); obj_priv = obj->driver_private; atomic_inc(&obj_priv->pending_flip); + work->pending_flip_obj = obj; BEGIN_LP_RING(4); OUT_RING(MI_DISPLAY_FLIP | @@ -4086,7 +4275,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, OUT_RING(fb->pitch); if (IS_I965G(dev)) { OUT_RING(obj_priv->gtt_offset | obj_priv->tiling_mode); - OUT_RING((fb->width << 16) | fb->height); + pipesrc = I915_READ(pipesrc_reg); + OUT_RING(pipesrc & 0x0fff0fff); } else { OUT_RING(obj_priv->gtt_offset); OUT_RING(MI_NOOP); @@ -4228,7 +4418,7 @@ static void intel_setup_outputs(struct drm_device *dev) if (IS_MOBILE(dev) && !IS_I830(dev)) intel_lvds_init(dev); - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { int found; if (IS_MOBILE(dev) && (I915_READ(DP_A) & DP_DETECTED)) @@ -4297,7 +4487,7 @@ static void intel_setup_outputs(struct drm_device *dev) DRM_DEBUG_KMS("probing DP_D\n"); intel_dp_init(dev, DP_D); } - } else if (IS_I8XX(dev)) + } else if (IS_GEN2(dev)) intel_dvo_init(dev); if (SUPPORTS_TV(dev)) @@ -4322,9 +4512,7 @@ static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb) intelfb_remove(dev, fb); drm_framebuffer_cleanup(fb); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(intel_fb->obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(intel_fb->obj); kfree(intel_fb); } @@ -4387,9 +4575,7 @@ intel_user_framebuffer_create(struct drm_device *dev, ret = intel_framebuffer_create(dev, mode_cmd, &fb, obj); if (ret) { - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); return NULL; } @@ -4437,6 +4623,91 @@ err_unref: return NULL; } +void ironlake_enable_drps(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u32 rgvmodectl = I915_READ(MEMMODECTL), rgvswctl; + u8 fmax, fmin, fstart, vstart; + int i = 0; + + /* 100ms RC evaluation intervals */ + I915_WRITE(RCUPEI, 100000); + I915_WRITE(RCDNEI, 100000); + + /* Set max/min thresholds to 90ms and 80ms respectively */ + I915_WRITE(RCBMAXAVG, 90000); + I915_WRITE(RCBMINAVG, 80000); + + I915_WRITE(MEMIHYST, 1); + + /* Set up min, max, and cur for interrupt handling */ + fmax = (rgvmodectl & MEMMODE_FMAX_MASK) >> MEMMODE_FMAX_SHIFT; + fmin = (rgvmodectl & MEMMODE_FMIN_MASK); + fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >> + MEMMODE_FSTART_SHIFT; + vstart = (I915_READ(PXVFREQ_BASE + (fstart * 4)) & PXVFREQ_PX_MASK) >> + PXVFREQ_PX_SHIFT; + + dev_priv->max_delay = fstart; /* can't go to fmax w/o IPS */ + dev_priv->min_delay = fmin; + dev_priv->cur_delay = fstart; + + I915_WRITE(MEMINTREN, MEMINT_CX_SUPR_EN | MEMINT_EVAL_CHG_EN); + + /* + * Interrupts will be enabled in ironlake_irq_postinstall + */ + + I915_WRITE(VIDSTART, vstart); + POSTING_READ(VIDSTART); + + rgvmodectl |= MEMMODE_SWMODE_EN; + I915_WRITE(MEMMODECTL, rgvmodectl); + + while (I915_READ(MEMSWCTL) & MEMCTL_CMD_STS) { + if (i++ > 100) { + DRM_ERROR("stuck trying to change perf mode\n"); + break; + } + msleep(1); + } + msleep(1); + + rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | + (fstart << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; + I915_WRITE(MEMSWCTL, rgvswctl); + POSTING_READ(MEMSWCTL); + + rgvswctl |= MEMCTL_CMD_STS; + I915_WRITE(MEMSWCTL, rgvswctl); +} + +void ironlake_disable_drps(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + u32 rgvswctl; + u8 fstart; + + /* Ack interrupts, disable EFC interrupt */ + I915_WRITE(MEMINTREN, I915_READ(MEMINTREN) & ~MEMINT_EVAL_CHG_EN); + I915_WRITE(MEMINTRSTS, MEMINT_EVAL_CHG); + I915_WRITE(DEIER, I915_READ(DEIER) & ~DE_PCU_EVENT); + I915_WRITE(DEIIR, DE_PCU_EVENT); + I915_WRITE(DEIMR, I915_READ(DEIMR) | DE_PCU_EVENT); + + /* Go back to the starting frequency */ + fstart = (I915_READ(MEMMODECTL) & MEMMODE_FSTART_MASK) >> + MEMMODE_FSTART_SHIFT; + rgvswctl = (MEMCTL_CMD_CHFREQ << MEMCTL_CMD_SHIFT) | + (fstart << MEMCTL_FREQ_SHIFT) | MEMCTL_SFCAVM; + I915_WRITE(MEMSWCTL, rgvswctl); + msleep(1); + rgvswctl |= MEMCTL_CMD_STS; + I915_WRITE(MEMSWCTL, rgvswctl); + msleep(1); + +} + void intel_init_clock_gating(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -4445,7 +4716,7 @@ void intel_init_clock_gating(struct drm_device *dev) * Disable clock gating reported to work incorrectly according to the * specs, but enable as much else as we can. */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { return; } else if (IS_G4X(dev)) { uint32_t dspclk_gate; @@ -4518,7 +4789,7 @@ static void intel_init_display(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; /* We always want a DPMS function */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) dev_priv->display.dpms = ironlake_crtc_dpms; else dev_priv->display.dpms = i9xx_crtc_dpms; @@ -4561,7 +4832,7 @@ static void intel_init_display(struct drm_device *dev) i830_get_display_clock_speed; /* For FIFO watermark updates */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) dev_priv->display.update_wm = NULL; else if (IS_G4X(dev)) dev_priv->display.update_wm = g4x_update_wm; @@ -4620,11 +4891,6 @@ void intel_modeset_init(struct drm_device *dev) DRM_DEBUG_KMS("%d display pipe%s available.\n", num_pipe, num_pipe > 1 ? "s" : ""); - if (IS_I85X(dev)) - pci_read_config_word(dev->pdev, HPLLCC, &dev_priv->orig_clock); - else if (IS_I9XX(dev) || IS_G4X(dev)) - pci_read_config_word(dev->pdev, GCFGC, &dev_priv->orig_clock); - for (i = 0; i < num_pipe; i++) { intel_crtc_init(dev, i); } @@ -4633,6 +4899,9 @@ void intel_modeset_init(struct drm_device *dev) intel_init_clock_gating(dev); + if (IS_IRONLAKE_M(dev)) + ironlake_enable_drps(dev); + INIT_WORK(&dev_priv->idle_work, intel_idle_update); setup_timer(&dev_priv->idle_timer, intel_gpu_idle_timer, (unsigned long)dev); @@ -4680,6 +4949,9 @@ void intel_modeset_cleanup(struct drm_device *dev) drm_gem_object_unreference(dev_priv->pwrctx); } + if (IS_IRONLAKE_M(dev)) + ironlake_disable_drps(dev); + mutex_unlock(&dev->struct_mutex); drm_mode_config_cleanup(dev); diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 439506cefc1..3ef3a0d0edd 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -231,7 +231,7 @@ intel_dp_aux_ch(struct intel_output *intel_output, */ if (IS_eDP(intel_output)) aux_clock_divider = 225; /* eDP input clock at 450Mhz */ - else if (IS_IRONLAKE(dev)) + else if (HAS_PCH_SPLIT(dev)) aux_clock_divider = 62; /* IRL input clock fixed at 125Mhz */ else aux_clock_divider = intel_hrawclk(dev) / 2; @@ -584,7 +584,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode, intel_dp_compute_m_n(3, lane_count, mode->clock, adjusted_mode->clock, &m_n); - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { if (intel_crtc->pipe == 0) { I915_WRITE(TRANSA_DATA_M1, ((m_n.tu - 1) << PIPE_GMCH_DATA_M_TU_SIZE_SHIFT) | @@ -1176,7 +1176,7 @@ intel_dp_detect(struct drm_connector *connector) dp_priv->has_audio = false; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return ironlake_dp_detect(connector); temp = I915_READ(PORT_HOTPLUG_EN); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index a51573da1ff..3a467ca5785 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -209,6 +209,8 @@ extern void intel_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, extern void intel_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, u16 *blue, int regno); extern void intel_init_clock_gating(struct drm_device *dev); +extern void ironlake_enable_drps(struct drm_device *dev); +extern void ironlake_disable_drps(struct drm_device *dev); extern int intel_framebuffer_create(struct drm_device *dev, struct drm_mode_fb_cmd *mode_cmd, diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c index 371d753e362..8cd791dc5b2 100644 --- a/drivers/gpu/drm/i915/intel_fb.c +++ b/drivers/gpu/drm/i915/intel_fb.c @@ -35,6 +35,7 @@ #include <linux/delay.h> #include <linux/fb.h> #include <linux/init.h> +#include <linux/vga_switcheroo.h> #include "drmP.h" #include "drm.h" @@ -148,7 +149,7 @@ static int intelfb_create(struct drm_device *dev, uint32_t fb_width, mutex_lock(&dev->struct_mutex); - ret = i915_gem_object_pin(fbo, PAGE_SIZE); + ret = i915_gem_object_pin(fbo, 64*1024); if (ret) { DRM_ERROR("failed to pin fb: %d\n", ret); goto out_unref; @@ -235,6 +236,7 @@ static int intelfb_create(struct drm_device *dev, uint32_t fb_width, obj_priv->gtt_offset, fbo); mutex_unlock(&dev->struct_mutex); + vga_switcheroo_client_fb_set(dev->pdev, info); return 0; out_unpin: diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 0e268deed76..a30f8bfc198 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -82,7 +82,7 @@ static void intel_hdmi_dpms(struct drm_encoder *encoder, int mode) /* HW workaround, need to toggle enable bit off and on for 12bpc, but * we do this anyway which shows more stable in testing. */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { I915_WRITE(hdmi_priv->sdvox_reg, temp & ~SDVO_ENABLE); POSTING_READ(hdmi_priv->sdvox_reg); } @@ -99,7 +99,7 @@ static void intel_hdmi_dpms(struct drm_encoder *encoder, int mode) /* HW workaround, need to write this twice for issue that may result * in first write getting masked. */ - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { I915_WRITE(hdmi_priv->sdvox_reg, temp); POSTING_READ(hdmi_priv->sdvox_reg); } diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c index 8673c735b8a..fcc753ca5d9 100644 --- a/drivers/gpu/drm/i915/intel_i2c.c +++ b/drivers/gpu/drm/i915/intel_i2c.c @@ -128,7 +128,7 @@ intel_i2c_reset_gmbus(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { I915_WRITE(PCH_GMBUS0, 0); } else { I915_WRITE(GMBUS0, 0); diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index aa74e59bec6..14e516fdc2d 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -56,7 +56,7 @@ static void intel_lvds_set_backlight(struct drm_device *dev, int level) struct drm_i915_private *dev_priv = dev->dev_private; u32 blc_pwm_ctl, reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) reg = BLC_PWM_CPU_CTL; else reg = BLC_PWM_CTL; @@ -74,7 +74,7 @@ static u32 intel_lvds_get_max_backlight(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; u32 reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) reg = BLC_PWM_PCH_CTL2; else reg = BLC_PWM_CTL; @@ -89,17 +89,22 @@ static u32 intel_lvds_get_max_backlight(struct drm_device *dev) static void intel_lvds_set_power(struct drm_device *dev, bool on) { struct drm_i915_private *dev_priv = dev->dev_private; - u32 pp_status, ctl_reg, status_reg; + u32 pp_status, ctl_reg, status_reg, lvds_reg; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { ctl_reg = PCH_PP_CONTROL; status_reg = PCH_PP_STATUS; + lvds_reg = PCH_LVDS; } else { ctl_reg = PP_CONTROL; status_reg = PP_STATUS; + lvds_reg = LVDS; } if (on) { + I915_WRITE(lvds_reg, I915_READ(lvds_reg) | LVDS_PORT_EN); + POSTING_READ(lvds_reg); + I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON); do { @@ -115,6 +120,9 @@ static void intel_lvds_set_power(struct drm_device *dev, bool on) do { pp_status = I915_READ(status_reg); } while (pp_status & PP_ON); + + I915_WRITE(lvds_reg, I915_READ(lvds_reg) & ~LVDS_PORT_EN); + POSTING_READ(lvds_reg); } } @@ -137,7 +145,7 @@ static void intel_lvds_save(struct drm_connector *connector) u32 pp_on_reg, pp_off_reg, pp_ctl_reg, pp_div_reg; u32 pwm_ctl_reg; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { pp_on_reg = PCH_PP_ON_DELAYS; pp_off_reg = PCH_PP_OFF_DELAYS; pp_ctl_reg = PCH_PP_CONTROL; @@ -174,7 +182,7 @@ static void intel_lvds_restore(struct drm_connector *connector) u32 pp_on_reg, pp_off_reg, pp_ctl_reg, pp_div_reg; u32 pwm_ctl_reg; - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { pp_on_reg = PCH_PP_ON_DELAYS; pp_off_reg = PCH_PP_OFF_DELAYS; pp_ctl_reg = PCH_PP_CONTROL; @@ -297,7 +305,7 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder, } /* full screen scale for now */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) goto out; /* 965+ wants fuzzy fitting */ @@ -327,7 +335,7 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder, * to register description and PRM. * Change the value here to see the borders for debugging */ - if (!IS_IRONLAKE(dev)) { + if (!HAS_PCH_SPLIT(dev)) { I915_WRITE(BCLRPAT_A, 0); I915_WRITE(BCLRPAT_B, 0); } @@ -548,7 +556,7 @@ static void intel_lvds_prepare(struct drm_encoder *encoder) struct drm_i915_private *dev_priv = dev->dev_private; u32 reg; - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) reg = BLC_PWM_CPU_CTL; else reg = BLC_PWM_CTL; @@ -587,7 +595,7 @@ static void intel_lvds_mode_set(struct drm_encoder *encoder, * settings. */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) return; /* @@ -611,7 +619,7 @@ static const struct dmi_system_id bad_lid_status[] = { { .ident = "Samsung SX20S", .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_SYS_VENDOR, "Samsung Electronics"), DMI_MATCH(DMI_BOARD_NAME, "SX20S"), }, }, @@ -623,12 +631,26 @@ static const struct dmi_system_id bad_lid_status[] = { }, }, { + .ident = "Aspire 1810T", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Acer"), + DMI_MATCH(DMI_PRODUCT_NAME, "Aspire 1810T"), + }, + }, + { .ident = "PC-81005", .matches = { DMI_MATCH(DMI_SYS_VENDOR, "MALATA"), DMI_MATCH(DMI_PRODUCT_NAME, "PC-81005"), }, }, + { + .ident = "Clevo M5x0N", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "CLEVO Co."), + DMI_MATCH(DMI_BOARD_NAME, "M5x0N"), + }, + }, { } }; @@ -641,9 +663,16 @@ static const struct dmi_system_id bad_lid_status[] = { */ static enum drm_connector_status intel_lvds_detect(struct drm_connector *connector) { + struct drm_device *dev = connector->dev; enum drm_connector_status status = connector_status_connected; - if (!acpi_lid_open() && !dmi_check_system(bad_lid_status)) + /* ACPI lid methods were generally unreliable in this generation, so + * don't even bother. + */ + if (IS_GEN2(dev)) + return connector_status_connected; + + if (!dmi_check_system(bad_lid_status) && !acpi_lid_open()) status = connector_status_disconnected; return status; @@ -1006,7 +1035,7 @@ void intel_lvds_init(struct drm_device *dev) return; } - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { if ((I915_READ(PCH_LVDS) & LVDS_DETECTED) == 0) return; if (dev_priv->edp_support) { @@ -1109,7 +1138,7 @@ void intel_lvds_init(struct drm_device *dev) */ /* Ironlake: FIXME if still fail, not try pipe mode now */ - if (IS_IRONLAKE(dev)) + if (HAS_PCH_SPLIT(dev)) goto failed; lvds = I915_READ(LVDS); @@ -1130,7 +1159,7 @@ void intel_lvds_init(struct drm_device *dev) goto failed; out: - if (IS_IRONLAKE(dev)) { + if (HAS_PCH_SPLIT(dev)) { u32 pwm; /* make sure PWM is enabled */ pwm = I915_READ(BLC_PWM_CPU_CTL2); diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 2639591c72e..d355d1d527e 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -172,7 +172,7 @@ struct overlay_registers { #define OFC_UPDATE 0x1 #define OVERLAY_NONPHYSICAL(dev) (IS_G33(dev) || IS_I965G(dev)) -#define OVERLAY_EXISTS(dev) (!IS_G4X(dev) && !IS_IRONLAKE(dev)) +#define OVERLAY_EXISTS(dev) (!IS_G4X(dev) && !IS_IRONLAKE(dev) && !IS_GEN6(dev)) static struct overlay_registers *intel_overlay_map_regs_atomic(struct intel_overlay *overlay) @@ -199,16 +199,11 @@ static struct overlay_registers *intel_overlay_map_regs_atomic(struct intel_over static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay) { - struct drm_device *dev = overlay->dev; - drm_i915_private_t *dev_priv = dev->dev_private; - if (OVERLAY_NONPHYSICAL(overlay->dev)) io_mapping_unmap_atomic(overlay->virt_addr); overlay->virt_addr = NULL; - I915_READ(OVADD); /* flush wc cashes */ - return; } @@ -225,9 +220,7 @@ static int intel_overlay_on(struct intel_overlay *overlay) overlay->active = 1; overlay->hw_wedged = NEEDS_WAIT_FOR_FLIP; - BEGIN_LP_RING(6); - OUT_RING(MI_FLUSH); - OUT_RING(MI_NOOP); + BEGIN_LP_RING(4); OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_ON); OUT_RING(overlay->flip_addr | OFC_UPDATE); OUT_RING(MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP); @@ -267,9 +260,7 @@ static void intel_overlay_continue(struct intel_overlay *overlay, if (tmp & (1 << 17)) DRM_DEBUG("overlay underrun, DOVSTA: %x\n", tmp); - BEGIN_LP_RING(4); - OUT_RING(MI_FLUSH); - OUT_RING(MI_NOOP); + BEGIN_LP_RING(2); OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE); OUT_RING(flip_addr); ADVANCE_LP_RING(); @@ -338,9 +329,7 @@ static int intel_overlay_off(struct intel_overlay *overlay) /* wait for overlay to go idle */ overlay->hw_wedged = SWITCH_OFF_STAGE_1; - BEGIN_LP_RING(6); - OUT_RING(MI_FLUSH); - OUT_RING(MI_NOOP); + BEGIN_LP_RING(4); OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE); OUT_RING(flip_addr); OUT_RING(MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP); @@ -358,9 +347,7 @@ static int intel_overlay_off(struct intel_overlay *overlay) /* turn overlay off */ overlay->hw_wedged = SWITCH_OFF_STAGE_2; - BEGIN_LP_RING(6); - OUT_RING(MI_FLUSH); - OUT_RING(MI_NOOP); + BEGIN_LP_RING(4); OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_OFF); OUT_RING(flip_addr); OUT_RING(MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP); @@ -435,9 +422,7 @@ int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay, overlay->hw_wedged = SWITCH_OFF_STAGE_2; - BEGIN_LP_RING(6); - OUT_RING(MI_FLUSH); - OUT_RING(MI_NOOP); + BEGIN_LP_RING(4); OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_OFF); OUT_RING(flip_addr); OUT_RING(MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP); @@ -1179,7 +1164,7 @@ int intel_overlay_put_image(struct drm_device *dev, void *data, out_unlock: mutex_unlock(&dev->struct_mutex); mutex_unlock(&dev->mode_config.mutex); - drm_gem_object_unreference(new_bo); + drm_gem_object_unreference_unlocked(new_bo); kfree(params); return ret; diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index eaacfd0920d..48daee5c9c6 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -35,6 +35,7 @@ #include "i915_drm.h" #include "i915_drv.h" #include "intel_sdvo_regs.h" +#include <linux/dmi.h> static char *tv_format_names[] = { "NTSC_M" , "NTSC_J" , "NTSC_443", @@ -2283,6 +2284,25 @@ intel_sdvo_get_slave_addr(struct drm_device *dev, int output_device) return 0x72; } +static int intel_sdvo_bad_tv_callback(const struct dmi_system_id *id) +{ + DRM_DEBUG_KMS("Ignoring bad SDVO TV connector for %s\n", id->ident); + return 1; +} + +static struct dmi_system_id intel_sdvo_bad_tv[] = { + { + .callback = intel_sdvo_bad_tv_callback, + .ident = "IntelG45/ICH10R/DME1737", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "IBM CORPORATION"), + DMI_MATCH(DMI_PRODUCT_NAME, "4800784"), + }, + }, + + { } /* terminating entry */ +}; + static bool intel_sdvo_output_setup(struct intel_output *intel_output, uint16_t flags) { @@ -2323,7 +2343,8 @@ intel_sdvo_output_setup(struct intel_output *intel_output, uint16_t flags) (1 << INTEL_SDVO_NON_TV_CLONE_BIT) | (1 << INTEL_ANALOG_CLONE_BIT); } - } else if (flags & SDVO_OUTPUT_SVID0) { + } else if ((flags & SDVO_OUTPUT_SVID0) && + !dmi_check_system(intel_sdvo_bad_tv)) { sdvo_priv->controlled_output = SDVO_OUTPUT_SVID0; encoder->encoder_type = DRM_MODE_ENCODER_TVDAC; @@ -2345,6 +2366,14 @@ intel_sdvo_output_setup(struct intel_output *intel_output, uint16_t flags) connector->connector_type = DRM_MODE_CONNECTOR_VGA; intel_output->clone_mask = (1 << INTEL_SDVO_NON_TV_CLONE_BIT) | (1 << INTEL_ANALOG_CLONE_BIT); + } else if (flags & SDVO_OUTPUT_CVBS0) { + + sdvo_priv->controlled_output = SDVO_OUTPUT_CVBS0; + encoder->encoder_type = DRM_MODE_ENCODER_TVDAC; + connector->connector_type = DRM_MODE_CONNECTOR_SVIDEO; + sdvo_priv->is_tv = true; + intel_output->needs_tv_clock = true; + intel_output->clone_mask = 1 << INTEL_SDVO_TV_CLONE_BIT; } else if (flags & SDVO_OUTPUT_LVDS0) { sdvo_priv->controlled_output = SDVO_OUTPUT_LVDS0; diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile index 48c290b5da8..32db806f3b5 100644 --- a/drivers/gpu/drm/nouveau/Makefile +++ b/drivers/gpu/drm/nouveau/Makefile @@ -16,7 +16,7 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \ nv04_fifo.o nv10_fifo.o nv40_fifo.o nv50_fifo.o \ nv04_graph.o nv10_graph.o nv20_graph.o \ nv40_graph.o nv50_graph.o \ - nv40_grctx.o \ + nv40_grctx.o nv50_grctx.o \ nv04_instmem.o nv50_instmem.o \ nv50_crtc.o nv50_dac.o nv50_sor.o \ nv50_cursor.o nv50_display.o nv50_fbcon.o \ diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c index 1cf488247a1..0e0730a5313 100644 --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c @@ -11,6 +11,8 @@ #include "nouveau_drm.h" #include "nv50_display.h" +#include <linux/vga_switcheroo.h> + #define NOUVEAU_DSM_SUPPORTED 0x00 #define NOUVEAU_DSM_SUPPORTED_FUNCTIONS 0x00 @@ -28,31 +30,30 @@ #define NOUVEAU_DSM_POWER_SPEED 0x01 #define NOUVEAU_DSM_POWER_STAMINA 0x02 -static int nouveau_dsm(struct drm_device *dev, int func, int arg, int *result) -{ - static char muid[] = { - 0xA0, 0xA0, 0x95, 0x9D, 0x60, 0x00, 0x48, 0x4D, - 0xB3, 0x4D, 0x7E, 0x5F, 0xEA, 0x12, 0x9F, 0xD4, - }; +static struct nouveau_dsm_priv { + bool dsm_detected; + acpi_handle dhandle; + acpi_handle dsm_handle; +} nouveau_dsm_priv; + +static const char nouveau_dsm_muid[] = { + 0xA0, 0xA0, 0x95, 0x9D, 0x60, 0x00, 0x48, 0x4D, + 0xB3, 0x4D, 0x7E, 0x5F, 0xEA, 0x12, 0x9F, 0xD4, +}; - struct pci_dev *pdev = dev->pdev; - struct acpi_handle *handle; +static int nouveau_dsm(acpi_handle handle, int func, int arg, int *result) +{ struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL }; struct acpi_object_list input; union acpi_object params[4]; union acpi_object *obj; int err; - handle = DEVICE_ACPI_HANDLE(&pdev->dev); - - if (!handle) - return -ENODEV; - input.count = 4; input.pointer = params; params[0].type = ACPI_TYPE_BUFFER; - params[0].buffer.length = sizeof(muid); - params[0].buffer.pointer = (char *)muid; + params[0].buffer.length = sizeof(nouveau_dsm_muid); + params[0].buffer.pointer = (char *)nouveau_dsm_muid; params[1].type = ACPI_TYPE_INTEGER; params[1].integer.value = 0x00000102; params[2].type = ACPI_TYPE_INTEGER; @@ -62,7 +63,7 @@ static int nouveau_dsm(struct drm_device *dev, int func, int arg, int *result) err = acpi_evaluate_object(handle, "_DSM", &input, &output); if (err) { - NV_INFO(dev, "failed to evaluate _DSM: %d\n", err); + printk(KERN_INFO "failed to evaluate _DSM: %d\n", err); return err; } @@ -86,40 +87,119 @@ static int nouveau_dsm(struct drm_device *dev, int func, int arg, int *result) return 0; } -int nouveau_hybrid_setup(struct drm_device *dev) +static int nouveau_dsm_switch_mux(acpi_handle handle, int mux_id) { - int result; - - if (nouveau_dsm(dev, NOUVEAU_DSM_ACTIVE, NOUVEAU_DSM_ACTIVE_QUERY, - &result)) - return -ENODEV; - - NV_INFO(dev, "_DSM hardware status gave 0x%x\n", result); - - if (result & 0x1) { /* Stamina mode - disable the external GPU */ - nouveau_dsm(dev, NOUVEAU_DSM_LED, NOUVEAU_DSM_LED_STAMINA, - NULL); - nouveau_dsm(dev, NOUVEAU_DSM_POWER, NOUVEAU_DSM_POWER_STAMINA, - NULL); - } else { /* Ensure that the external GPU is enabled */ - nouveau_dsm(dev, NOUVEAU_DSM_LED, NOUVEAU_DSM_LED_SPEED, NULL); - nouveau_dsm(dev, NOUVEAU_DSM_POWER, NOUVEAU_DSM_POWER_SPEED, - NULL); - } + return nouveau_dsm(handle, NOUVEAU_DSM_LED, mux_id, NULL); +} + +static int nouveau_dsm_set_discrete_state(acpi_handle handle, enum vga_switcheroo_state state) +{ + int arg; + if (state == VGA_SWITCHEROO_ON) + arg = NOUVEAU_DSM_POWER_SPEED; + else + arg = NOUVEAU_DSM_POWER_STAMINA; + nouveau_dsm(handle, NOUVEAU_DSM_POWER, arg, NULL); + return 0; +} + +static int nouveau_dsm_switchto(enum vga_switcheroo_client_id id) +{ + if (id == VGA_SWITCHEROO_IGD) + return nouveau_dsm_switch_mux(nouveau_dsm_priv.dsm_handle, NOUVEAU_DSM_LED_STAMINA); + else + return nouveau_dsm_switch_mux(nouveau_dsm_priv.dsm_handle, NOUVEAU_DSM_LED_SPEED); +} +static int nouveau_dsm_power_state(enum vga_switcheroo_client_id id, + enum vga_switcheroo_state state) +{ + if (id == VGA_SWITCHEROO_IGD) + return 0; + + return nouveau_dsm_set_discrete_state(nouveau_dsm_priv.dsm_handle, state); +} + +static int nouveau_dsm_init(void) +{ return 0; } -bool nouveau_dsm_probe(struct drm_device *dev) +static int nouveau_dsm_get_client_id(struct pci_dev *pdev) { - int support = 0; + if (nouveau_dsm_priv.dhandle == DEVICE_ACPI_HANDLE(&pdev->dev)) + return VGA_SWITCHEROO_IGD; + else + return VGA_SWITCHEROO_DIS; +} + +static struct vga_switcheroo_handler nouveau_dsm_handler = { + .switchto = nouveau_dsm_switchto, + .power_state = nouveau_dsm_power_state, + .init = nouveau_dsm_init, + .get_client_id = nouveau_dsm_get_client_id, +}; - if (nouveau_dsm(dev, NOUVEAU_DSM_SUPPORTED, - NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &support)) +static bool nouveau_dsm_pci_probe(struct pci_dev *pdev) +{ + acpi_handle dhandle, nvidia_handle; + acpi_status status; + int ret; + uint32_t result; + + dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); + if (!dhandle) + return false; + status = acpi_get_handle(dhandle, "_DSM", &nvidia_handle); + if (ACPI_FAILURE(status)) { return false; + } - if (!support) + ret= nouveau_dsm(nvidia_handle, NOUVEAU_DSM_SUPPORTED, + NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &result); + if (ret < 0) return false; + nouveau_dsm_priv.dhandle = dhandle; + nouveau_dsm_priv.dsm_handle = nvidia_handle; return true; } + +static bool nouveau_dsm_detect(void) +{ + char acpi_method_name[255] = { 0 }; + struct acpi_buffer buffer = {sizeof(acpi_method_name), acpi_method_name}; + struct pci_dev *pdev = NULL; + int has_dsm = 0; + int vga_count = 0; + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) { + vga_count++; + + has_dsm |= (nouveau_dsm_pci_probe(pdev) == true); + } + + if (vga_count == 2 && has_dsm) { + acpi_get_name(nouveau_dsm_priv.dsm_handle, ACPI_FULL_PATHNAME, &buffer); + printk(KERN_INFO "VGA switcheroo: detected DSM switching method %s handle\n", + acpi_method_name); + nouveau_dsm_priv.dsm_detected = true; + return true; + } + return false; +} + +void nouveau_register_dsm_handler(void) +{ + bool r; + + r = nouveau_dsm_detect(); + if (!r) + return; + + vga_switcheroo_register_handler(&nouveau_dsm_handler); +} + +void nouveau_unregister_dsm_handler(void) +{ + vga_switcheroo_unregister_handler(); +} diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c index ba143972769..71247da17da 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.c +++ b/drivers/gpu/drm/nouveau/nouveau_bios.c @@ -310,63 +310,22 @@ valid_reg(struct nvbios *bios, uint32_t reg) struct drm_device *dev = bios->dev; /* C51 has misaligned regs on purpose. Marvellous */ - if (reg & 0x2 || (reg & 0x1 && dev_priv->VBIOS.pub.chip_version != 0x51)) { - NV_ERROR(dev, "========== misaligned reg 0x%08X ==========\n", - reg); - return 0; - } - /* - * Warn on C51 regs that have not been verified accessible in - * mmiotracing - */ - if (reg & 0x1 && dev_priv->VBIOS.pub.chip_version == 0x51 && + if (reg & 0x2 || + (reg & 0x1 && dev_priv->vbios.chip_version != 0x51)) + NV_ERROR(dev, "======= misaligned reg 0x%08X =======\n", reg); + + /* warn on C51 regs that haven't been verified accessible in tracing */ + if (reg & 0x1 && dev_priv->vbios.chip_version == 0x51 && reg != 0x130d && reg != 0x1311 && reg != 0x60081d) NV_WARN(dev, "=== C51 misaligned reg 0x%08X not verified ===\n", reg); - /* Trust the init scripts on G80 */ - if (dev_priv->card_type >= NV_50) - return 1; - - #define WITHIN(x, y, z) ((x >= y) && (x < y + z)) - if (WITHIN(reg, NV_PMC_OFFSET, NV_PMC_SIZE)) - return 1; - if (WITHIN(reg, NV_PBUS_OFFSET, NV_PBUS_SIZE)) - return 1; - if (WITHIN(reg, NV_PFIFO_OFFSET, NV_PFIFO_SIZE)) - return 1; - if (dev_priv->VBIOS.pub.chip_version >= 0x30 && - (WITHIN(reg, 0x4000, 0x600) || reg == 0x00004600)) - return 1; - if (dev_priv->VBIOS.pub.chip_version >= 0x40 && - WITHIN(reg, 0xc000, 0x48)) - return 1; - if (dev_priv->VBIOS.pub.chip_version >= 0x17 && reg == 0x0000d204) - return 1; - if (dev_priv->VBIOS.pub.chip_version >= 0x40) { - if (reg == 0x00011014 || reg == 0x00020328) - return 1; - if (WITHIN(reg, 0x88000, NV_PBUS_SIZE)) /* new PBUS */ - return 1; + if (reg >= (8*1024*1024)) { + NV_ERROR(dev, "=== reg 0x%08x out of mapped bounds ===\n", reg); + return 0; } - if (WITHIN(reg, NV_PFB_OFFSET, NV_PFB_SIZE)) - return 1; - if (WITHIN(reg, NV_PEXTDEV_OFFSET, NV_PEXTDEV_SIZE)) - return 1; - if (WITHIN(reg, NV_PCRTC0_OFFSET, NV_PCRTC0_SIZE * 2)) - return 1; - if (WITHIN(reg, NV_PRAMDAC0_OFFSET, NV_PRAMDAC0_SIZE * 2)) - return 1; - if (dev_priv->VBIOS.pub.chip_version >= 0x17 && reg == 0x0070fff0) - return 1; - if (dev_priv->VBIOS.pub.chip_version == 0x51 && - WITHIN(reg, NV_PRAMIN_OFFSET, NV_PRAMIN_SIZE)) - return 1; - #undef WITHIN - - NV_ERROR(dev, "========== unknown reg 0x%08X ==========\n", reg); - return 0; + return 1; } static bool @@ -461,7 +420,7 @@ bios_wr32(struct nvbios *bios, uint32_t reg, uint32_t data) LOG_OLD_VALUE(bios_rd32(bios, reg)); BIOSLOG(bios, " Write: Reg: 0x%08X, Data: 0x%08X\n", reg, data); - if (dev_priv->VBIOS.execute) { + if (dev_priv->vbios.execute) { still_alive(); nv_wr32(bios->dev, reg, data); } @@ -688,7 +647,7 @@ nv50_pll_set(struct drm_device *dev, uint32_t reg, uint32_t clk) reg0 = (reg0 & 0xfff8ffff) | (pll.log2P << 16); reg1 = (reg1 & 0xffff0000) | (pll.N1 << 8) | pll.M1; - if (dev_priv->VBIOS.execute) { + if (dev_priv->vbios.execute) { still_alive(); nv_wr32(dev, reg + 4, reg1); nv_wr32(dev, reg + 0, reg0); @@ -730,7 +689,7 @@ setPLL(struct nvbios *bios, uint32_t reg, uint32_t clk) static int dcb_entry_idx_from_crtchead(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; /* * For the results of this function to be correct, CR44 must have been @@ -741,7 +700,7 @@ static int dcb_entry_idx_from_crtchead(struct drm_device *dev) uint8_t dcb_entry = NVReadVgaCrtc5758(dev, bios->state.crtchead, 0); - if (dcb_entry > bios->bdcb.dcb.entries) { + if (dcb_entry > bios->dcb.entries) { NV_ERROR(dev, "CR58 doesn't have a valid DCB entry currently " "(%02X)\n", dcb_entry); dcb_entry = 0x7f; /* unused / invalid marker */ @@ -754,25 +713,26 @@ static struct nouveau_i2c_chan * init_i2c_device_find(struct drm_device *dev, int i2c_index) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct bios_parsed_dcb *bdcb = &dev_priv->VBIOS.bdcb; + struct dcb_table *dcb = &dev_priv->vbios.dcb; if (i2c_index == 0xff) { /* note: dcb_entry_idx_from_crtchead needs pre-script set-up */ int idx = dcb_entry_idx_from_crtchead(dev), shift = 0; - int default_indices = bdcb->i2c_default_indices; + int default_indices = dcb->i2c_default_indices; - if (idx != 0x7f && bdcb->dcb.entry[idx].i2c_upper_default) + if (idx != 0x7f && dcb->entry[idx].i2c_upper_default) shift = 4; i2c_index = (default_indices >> shift) & 0xf; } if (i2c_index == 0x80) /* g80+ */ - i2c_index = bdcb->i2c_default_indices & 0xf; + i2c_index = dcb->i2c_default_indices & 0xf; return nouveau_i2c_find(dev, i2c_index); } -static uint32_t get_tmds_index_reg(struct drm_device *dev, uint8_t mlv) +static uint32_t +get_tmds_index_reg(struct drm_device *dev, uint8_t mlv) { /* * For mlv < 0x80, it is an index into a table of TMDS base addresses. @@ -785,6 +745,7 @@ static uint32_t get_tmds_index_reg(struct drm_device *dev, uint8_t mlv) */ struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; const int pramdac_offset[13] = { 0, 0, 0x8, 0, 0x2000, 0, 0, 0, 0x2008, 0, 0, 0, 0x2000 }; const uint32_t pramdac_table[4] = { @@ -797,13 +758,12 @@ static uint32_t get_tmds_index_reg(struct drm_device *dev, uint8_t mlv) dcb_entry = dcb_entry_idx_from_crtchead(dev); if (dcb_entry == 0x7f) return 0; - dacoffset = pramdac_offset[ - dev_priv->VBIOS.bdcb.dcb.entry[dcb_entry].or]; + dacoffset = pramdac_offset[bios->dcb.entry[dcb_entry].or]; if (mlv == 0x81) dacoffset ^= 8; return 0x6808b0 + dacoffset; } else { - if (mlv > ARRAY_SIZE(pramdac_table)) { + if (mlv >= ARRAY_SIZE(pramdac_table)) { NV_ERROR(dev, "Magic Lookup Value too big (%02X)\n", mlv); return 0; @@ -1906,7 +1866,7 @@ init_compute_mem(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) struct drm_nouveau_private *dev_priv = bios->dev->dev_private; - if (dev_priv->card_type >= NV_50) + if (dev_priv->card_type >= NV_40) return 1; /* @@ -2615,19 +2575,19 @@ init_gpio(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) const uint32_t nv50_gpio_reg[4] = { 0xe104, 0xe108, 0xe280, 0xe284 }; const uint32_t nv50_gpio_ctl[2] = { 0xe100, 0xe28c }; - const uint8_t *gpio_table = &bios->data[bios->bdcb.gpio_table_ptr]; + const uint8_t *gpio_table = &bios->data[bios->dcb.gpio_table_ptr]; const uint8_t *gpio_entry; int i; if (!iexec->execute) return 1; - if (bios->bdcb.version != 0x40) { + if (bios->dcb.version != 0x40) { NV_ERROR(bios->dev, "DCB table not version 4.0\n"); return 0; } - if (!bios->bdcb.gpio_table_ptr) { + if (!bios->dcb.gpio_table_ptr) { NV_WARN(bios->dev, "Invalid pointer to INIT_8E table\n"); return 0; } @@ -3164,7 +3124,7 @@ run_digital_op_script(struct drm_device *dev, uint16_t scriptptr, struct dcb_entry *dcbent, int head, bool dl) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; struct init_exec iexec = {true, false}; NV_TRACE(dev, "0x%04X: Parsing digital output script table\n", @@ -3181,7 +3141,7 @@ run_digital_op_script(struct drm_device *dev, uint16_t scriptptr, static int call_lvds_manufacturer_script(struct drm_device *dev, struct dcb_entry *dcbent, int head, enum LVDS_script script) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint8_t sub = bios->data[bios->fp.xlated_entry + script] + (bios->fp.link_c_increment && dcbent->or & OUTPUT_C ? 1 : 0); uint16_t scriptofs = ROM16(bios->data[bios->init_script_tbls_ptr + sub * 2]); @@ -3196,16 +3156,25 @@ static int call_lvds_manufacturer_script(struct drm_device *dev, struct dcb_entr } #ifdef __powerpc__ /* Powerbook specific quirks */ - if (script == LVDS_RESET && ((dev->pci_device & 0xffff) == 0x0179 || (dev->pci_device & 0xffff) == 0x0329)) - nv_write_tmds(dev, dcbent->or, 0, 0x02, 0x72); - if ((dev->pci_device & 0xffff) == 0x0179 || (dev->pci_device & 0xffff) == 0x0189 || (dev->pci_device & 0xffff) == 0x0329) { - if (script == LVDS_PANEL_ON) { - bios_wr32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL, bios_rd32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL) | (1 << 31)); - bios_wr32(bios, NV_PCRTC_GPIO_EXT, bios_rd32(bios, NV_PCRTC_GPIO_EXT) | 1); - } - if (script == LVDS_PANEL_OFF) { - bios_wr32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL, bios_rd32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL) & ~(1 << 31)); - bios_wr32(bios, NV_PCRTC_GPIO_EXT, bios_rd32(bios, NV_PCRTC_GPIO_EXT) & ~3); + if ((dev->pci_device & 0xffff) == 0x0179 || + (dev->pci_device & 0xffff) == 0x0189 || + (dev->pci_device & 0xffff) == 0x0329) { + if (script == LVDS_RESET) { + nv_write_tmds(dev, dcbent->or, 0, 0x02, 0x72); + + } else if (script == LVDS_PANEL_ON) { + bios_wr32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL, + bios_rd32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL) + | (1 << 31)); + bios_wr32(bios, NV_PCRTC_GPIO_EXT, + bios_rd32(bios, NV_PCRTC_GPIO_EXT) | 1); + + } else if (script == LVDS_PANEL_OFF) { + bios_wr32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL, + bios_rd32(bios, NV_PBUS_DEBUG_DUALHEAD_CTL) + & ~(1 << 31)); + bios_wr32(bios, NV_PCRTC_GPIO_EXT, + bios_rd32(bios, NV_PCRTC_GPIO_EXT) & ~3); } } #endif @@ -3226,7 +3195,7 @@ static int run_lvds_table(struct drm_device *dev, struct dcb_entry *dcbent, int * of a list of pxclks and script pointers. */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; unsigned int outputset = (dcbent->or == 4) ? 1 : 0; uint16_t scriptptr = 0, clktable; uint8_t clktableptr = 0; @@ -3293,7 +3262,7 @@ int call_lvds_script(struct drm_device *dev, struct dcb_entry *dcbent, int head, */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint8_t lvds_ver = bios->data[bios->fp.lvdsmanufacturerpointer]; uint32_t sel_clk_binding, sel_clk; int ret; @@ -3427,7 +3396,7 @@ static int parse_fp_mode_table(struct drm_device *dev, struct nvbios *bios) #ifndef __powerpc__ NV_ERROR(dev, "Pointer to flat panel table invalid\n"); #endif - bios->pub.digital_min_front_porch = 0x4b; + bios->digital_min_front_porch = 0x4b; return 0; } @@ -3460,7 +3429,7 @@ static int parse_fp_mode_table(struct drm_device *dev, struct nvbios *bios) * fptable[4] is the minimum * RAMDAC_FP_HCRTC -> RAMDAC_FP_HSYNC_START gap */ - bios->pub.digital_min_front_porch = fptable[4]; + bios->digital_min_front_porch = fptable[4]; ofs = -7; break; default: @@ -3499,7 +3468,7 @@ static int parse_fp_mode_table(struct drm_device *dev, struct nvbios *bios) /* nv4x cards need both a strap value and fpindex of 0xf to use DDC */ if (lth.lvds_ver > 0x10) - bios->pub.fp_no_ddc = fpstrapping != 0xf || fpindex != 0xf; + bios->fp_no_ddc = fpstrapping != 0xf || fpindex != 0xf; /* * If either the strap or xlated fpindex value are 0xf there is no @@ -3523,7 +3492,7 @@ static int parse_fp_mode_table(struct drm_device *dev, struct nvbios *bios) bool nouveau_bios_fp_mode(struct drm_device *dev, struct drm_display_mode *mode) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint8_t *mode_entry = &bios->data[bios->fp.mode_ptr]; if (!mode) /* just checking whether we can produce a mode */ @@ -3594,11 +3563,11 @@ int nouveau_bios_parse_lvds_table(struct drm_device *dev, int pxclk, bool *dl, b * until later, when this function should be called with non-zero pxclk */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; int fpstrapping = get_fp_strap(dev, bios), lvdsmanufacturerindex = 0; struct lvdstableheader lth; uint16_t lvdsofs; - int ret, chip_version = bios->pub.chip_version; + int ret, chip_version = bios->chip_version; ret = parse_lvds_manufacturer_table_header(dev, bios, <h); if (ret) @@ -3714,7 +3683,7 @@ bios_output_config_match(struct drm_device *dev, struct dcb_entry *dcbent, uint16_t record, int record_len, int record_nr) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint32_t entry; uint16_t table; int i, v; @@ -3748,7 +3717,7 @@ nouveau_bios_dp_table(struct drm_device *dev, struct dcb_entry *dcbent, int *length) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint8_t *table; if (!bios->display.dp_table_ptr) { @@ -3757,7 +3726,7 @@ nouveau_bios_dp_table(struct drm_device *dev, struct dcb_entry *dcbent, } table = &bios->data[bios->display.dp_table_ptr]; - if (table[0] != 0x21) { + if (table[0] != 0x20 && table[0] != 0x21) { NV_ERROR(dev, "DisplayPort table version 0x%02x unknown\n", table[0]); return NULL; @@ -3797,8 +3766,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct init_exec iexec = {true, false}; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint8_t *table = &bios->data[bios->display.script_table_ptr]; uint8_t *otable = NULL; uint16_t script; @@ -3877,8 +3845,6 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } } - bios->display.output = dcbent; - if (pxclk == 0) { script = ROM16(otable[6]); if (!script) { @@ -3887,7 +3853,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } NV_TRACE(dev, "0x%04X: parsing output script 0\n", script); - parse_init_table(bios, script, &iexec); + nouveau_bios_run_init_table(dev, script, dcbent); } else if (pxclk == -1) { script = ROM16(otable[8]); @@ -3897,7 +3863,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } NV_TRACE(dev, "0x%04X: parsing output script 1\n", script); - parse_init_table(bios, script, &iexec); + nouveau_bios_run_init_table(dev, script, dcbent); } else if (pxclk == -2) { if (table[4] >= 12) @@ -3910,7 +3876,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } NV_TRACE(dev, "0x%04X: parsing output script 2\n", script); - parse_init_table(bios, script, &iexec); + nouveau_bios_run_init_table(dev, script, dcbent); } else if (pxclk > 0) { script = ROM16(otable[table[4] + i*6 + 2]); @@ -3922,7 +3888,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } NV_TRACE(dev, "0x%04X: parsing clock script 0\n", script); - parse_init_table(bios, script, &iexec); + nouveau_bios_run_init_table(dev, script, dcbent); } else if (pxclk < 0) { script = ROM16(otable[table[4] + i*6 + 4]); @@ -3934,7 +3900,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent, } NV_TRACE(dev, "0x%04X: parsing clock script 1\n", script); - parse_init_table(bios, script, &iexec); + nouveau_bios_run_init_table(dev, script, dcbent); } return 0; @@ -3953,8 +3919,8 @@ int run_tmds_table(struct drm_device *dev, struct dcb_entry *dcbent, int head, i */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; - int cv = bios->pub.chip_version; + struct nvbios *bios = &dev_priv->vbios; + int cv = bios->chip_version; uint16_t clktable = 0, scriptptr; uint32_t sel_clk_binding, sel_clk; @@ -4013,8 +3979,8 @@ int get_pll_limits(struct drm_device *dev, uint32_t limit_match, struct pll_lims */ struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; - int cv = bios->pub.chip_version, pllindex = 0; + struct nvbios *bios = &dev_priv->vbios; + int cv = bios->chip_version, pllindex = 0; uint8_t pll_lim_ver = 0, headerlen = 0, recordlen = 0, entries = 0; uint32_t crystal_strap_mask, crystal_straps; @@ -4367,7 +4333,7 @@ static void parse_bios_version(struct drm_device *dev, struct nvbios *bios, uint */ bios->major_version = bios->data[offset + 3]; - bios->pub.chip_version = bios->data[offset + 2]; + bios->chip_version = bios->data[offset + 2]; NV_TRACE(dev, "Bios version %02x.%02x.%02x.%02x\n", bios->data[offset + 3], bios->data[offset + 2], bios->data[offset + 1], bios->data[offset]); @@ -4437,7 +4403,7 @@ static int parse_bit_A_tbl_entry(struct drm_device *dev, struct nvbios *bios, st } /* First entry is normal dac, 2nd tv-out perhaps? */ - bios->pub.dactestval = ROM32(bios->data[load_table_ptr + headerlen]) & 0x3ff; + bios->dactestval = ROM32(bios->data[load_table_ptr + headerlen]) & 0x3ff; return 0; } @@ -4561,8 +4527,8 @@ static int parse_bit_i_tbl_entry(struct drm_device *dev, struct nvbios *bios, st return -ENOSYS; } - bios->pub.dactestval = ROM32(bios->data[daccmpoffset + dacheaderlen]); - bios->pub.tvdactestval = ROM32(bios->data[daccmpoffset + dacheaderlen + 4]); + bios->dactestval = ROM32(bios->data[daccmpoffset + dacheaderlen]); + bios->tvdactestval = ROM32(bios->data[daccmpoffset + dacheaderlen + 4]); return 0; } @@ -4831,11 +4797,11 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi uint16_t legacy_scripts_offset, legacy_i2c_offset; /* load needed defaults in case we can't parse this info */ - bios->bdcb.dcb.i2c[0].write = NV_CIO_CRE_DDC_WR__INDEX; - bios->bdcb.dcb.i2c[0].read = NV_CIO_CRE_DDC_STATUS__INDEX; - bios->bdcb.dcb.i2c[1].write = NV_CIO_CRE_DDC0_WR__INDEX; - bios->bdcb.dcb.i2c[1].read = NV_CIO_CRE_DDC0_STATUS__INDEX; - bios->pub.digital_min_front_porch = 0x4b; + bios->dcb.i2c[0].write = NV_CIO_CRE_DDC_WR__INDEX; + bios->dcb.i2c[0].read = NV_CIO_CRE_DDC_STATUS__INDEX; + bios->dcb.i2c[1].write = NV_CIO_CRE_DDC0_WR__INDEX; + bios->dcb.i2c[1].read = NV_CIO_CRE_DDC0_STATUS__INDEX; + bios->digital_min_front_porch = 0x4b; bios->fmaxvco = 256000; bios->fminvco = 128000; bios->fp.duallink_transition_clk = 90000; @@ -4942,10 +4908,10 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi bios->legacy.i2c_indices.crt = bios->data[legacy_i2c_offset]; bios->legacy.i2c_indices.tv = bios->data[legacy_i2c_offset + 1]; bios->legacy.i2c_indices.panel = bios->data[legacy_i2c_offset + 2]; - bios->bdcb.dcb.i2c[0].write = bios->data[legacy_i2c_offset + 4]; - bios->bdcb.dcb.i2c[0].read = bios->data[legacy_i2c_offset + 5]; - bios->bdcb.dcb.i2c[1].write = bios->data[legacy_i2c_offset + 6]; - bios->bdcb.dcb.i2c[1].read = bios->data[legacy_i2c_offset + 7]; + bios->dcb.i2c[0].write = bios->data[legacy_i2c_offset + 4]; + bios->dcb.i2c[0].read = bios->data[legacy_i2c_offset + 5]; + bios->dcb.i2c[1].write = bios->data[legacy_i2c_offset + 6]; + bios->dcb.i2c[1].read = bios->data[legacy_i2c_offset + 7]; if (bmplength > 74) { bios->fmaxvco = ROM32(bmp[67]); @@ -5019,7 +4985,8 @@ read_dcb_i2c_entry(struct drm_device *dev, int dcb_version, uint8_t *i2ctable, i else NV_WARN(dev, "DCB I2C table has more entries than indexable " - "(%d entries, max index 15)\n", i2ctable[2]); + "(%d entries, max %d)\n", i2ctable[2], + DCB_MAX_NUM_I2C_ENTRIES); entry_len = i2ctable[3]; /* [4] is i2c_default_indices, read in parse_dcb_table() */ } @@ -5035,8 +5002,8 @@ read_dcb_i2c_entry(struct drm_device *dev, int dcb_version, uint8_t *i2ctable, i if (index == 0xf) return 0; - if (index > i2c_entries) { - NV_ERROR(dev, "DCB I2C index too big (%d > %d)\n", + if (index >= i2c_entries) { + NV_ERROR(dev, "DCB I2C index too big (%d >= %d)\n", index, i2ctable[2]); return -ENOENT; } @@ -5071,7 +5038,7 @@ read_dcb_i2c_entry(struct drm_device *dev, int dcb_version, uint8_t *i2ctable, i static struct dcb_gpio_entry * new_gpio_entry(struct nvbios *bios) { - struct parsed_dcb_gpio *gpio = &bios->bdcb.gpio; + struct dcb_gpio_table *gpio = &bios->dcb.gpio; return &gpio->entry[gpio->entries++]; } @@ -5080,14 +5047,14 @@ struct dcb_gpio_entry * nouveau_bios_gpio_entry(struct drm_device *dev, enum dcb_gpio_tag tag) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; int i; - for (i = 0; i < bios->bdcb.gpio.entries; i++) { - if (bios->bdcb.gpio.entry[i].tag != tag) + for (i = 0; i < bios->dcb.gpio.entries; i++) { + if (bios->dcb.gpio.entry[i].tag != tag) continue; - return &bios->bdcb.gpio.entry[i]; + return &bios->dcb.gpio.entry[i]; } return NULL; @@ -5135,7 +5102,7 @@ static void parse_dcb_gpio_table(struct nvbios *bios) { struct drm_device *dev = bios->dev; - uint16_t gpio_table_ptr = bios->bdcb.gpio_table_ptr; + uint16_t gpio_table_ptr = bios->dcb.gpio_table_ptr; uint8_t *gpio_table = &bios->data[gpio_table_ptr]; int header_len = gpio_table[1], entries = gpio_table[2], @@ -5143,7 +5110,7 @@ parse_dcb_gpio_table(struct nvbios *bios) void (*parse_entry)(struct nvbios *, uint16_t) = NULL; int i; - if (bios->bdcb.version >= 0x40) { + if (bios->dcb.version >= 0x40) { if (gpio_table_ptr && entry_len != 4) { NV_WARN(dev, "Invalid DCB GPIO table entry length.\n"); return; @@ -5151,7 +5118,7 @@ parse_dcb_gpio_table(struct nvbios *bios) parse_entry = parse_dcb40_gpio_entry; - } else if (bios->bdcb.version >= 0x30) { + } else if (bios->dcb.version >= 0x30) { if (gpio_table_ptr && entry_len != 2) { NV_WARN(dev, "Invalid DCB GPIO table entry length.\n"); return; @@ -5159,7 +5126,7 @@ parse_dcb_gpio_table(struct nvbios *bios) parse_entry = parse_dcb30_gpio_entry; - } else if (bios->bdcb.version >= 0x22) { + } else if (bios->dcb.version >= 0x22) { /* * DCBs older than v3.0 don't really have a GPIO * table, instead they keep some GPIO info at fixed @@ -5193,30 +5160,67 @@ struct dcb_connector_table_entry * nouveau_bios_connector_entry(struct drm_device *dev, int index) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; struct dcb_connector_table_entry *cte; - if (index >= bios->bdcb.connector.entries) + if (index >= bios->dcb.connector.entries) return NULL; - cte = &bios->bdcb.connector.entry[index]; + cte = &bios->dcb.connector.entry[index]; if (cte->type == 0xff) return NULL; return cte; } +static enum dcb_connector_type +divine_connector_type(struct nvbios *bios, int index) +{ + struct dcb_table *dcb = &bios->dcb; + unsigned encoders = 0, type = DCB_CONNECTOR_NONE; + int i; + + for (i = 0; i < dcb->entries; i++) { + if (dcb->entry[i].connector == index) + encoders |= (1 << dcb->entry[i].type); + } + + if (encoders & (1 << OUTPUT_DP)) { + if (encoders & (1 << OUTPUT_TMDS)) + type = DCB_CONNECTOR_DP; + else + type = DCB_CONNECTOR_eDP; + } else + if (encoders & (1 << OUTPUT_TMDS)) { + if (encoders & (1 << OUTPUT_ANALOG)) + type = DCB_CONNECTOR_DVI_I; + else + type = DCB_CONNECTOR_DVI_D; + } else + if (encoders & (1 << OUTPUT_ANALOG)) { + type = DCB_CONNECTOR_VGA; + } else + if (encoders & (1 << OUTPUT_LVDS)) { + type = DCB_CONNECTOR_LVDS; + } else + if (encoders & (1 << OUTPUT_TV)) { + type = DCB_CONNECTOR_TV_0; + } + + return type; +} + static void parse_dcb_connector_table(struct nvbios *bios) { struct drm_device *dev = bios->dev; - struct dcb_connector_table *ct = &bios->bdcb.connector; + struct dcb_connector_table *ct = &bios->dcb.connector; struct dcb_connector_table_entry *cte; - uint8_t *conntab = &bios->data[bios->bdcb.connector_table_ptr]; + uint8_t *conntab = &bios->data[bios->dcb.connector_table_ptr]; uint8_t *entry; int i; - if (!bios->bdcb.connector_table_ptr) { + if (!bios->dcb.connector_table_ptr) { NV_DEBUG_KMS(dev, "No DCB connector table present\n"); return; } @@ -5238,6 +5242,7 @@ parse_dcb_connector_table(struct nvbios *bios) cte->entry = ROM16(entry[0]); else cte->entry = ROM32(entry[0]); + cte->type = (cte->entry & 0x000000ff) >> 0; cte->index = (cte->entry & 0x00000f00) >> 8; switch (cte->entry & 0x00033000) { @@ -5263,10 +5268,33 @@ parse_dcb_connector_table(struct nvbios *bios) NV_INFO(dev, " %d: 0x%08x: type 0x%02x idx %d tag 0x%02x\n", i, cte->entry, cte->type, cte->index, cte->gpio_tag); + + /* check for known types, fallback to guessing the type + * from attached encoders if we hit an unknown. + */ + switch (cte->type) { + case DCB_CONNECTOR_VGA: + case DCB_CONNECTOR_TV_0: + case DCB_CONNECTOR_TV_1: + case DCB_CONNECTOR_TV_3: + case DCB_CONNECTOR_DVI_I: + case DCB_CONNECTOR_DVI_D: + case DCB_CONNECTOR_LVDS: + case DCB_CONNECTOR_DP: + case DCB_CONNECTOR_eDP: + case DCB_CONNECTOR_HDMI_0: + case DCB_CONNECTOR_HDMI_1: + break; + default: + cte->type = divine_connector_type(bios, cte->index); + NV_WARN(dev, "unknown type, using 0x%02x", cte->type); + break; + } + } } -static struct dcb_entry *new_dcb_entry(struct parsed_dcb *dcb) +static struct dcb_entry *new_dcb_entry(struct dcb_table *dcb) { struct dcb_entry *entry = &dcb->entry[dcb->entries]; @@ -5276,7 +5304,7 @@ static struct dcb_entry *new_dcb_entry(struct parsed_dcb *dcb) return entry; } -static void fabricate_vga_output(struct parsed_dcb *dcb, int i2c, int heads) +static void fabricate_vga_output(struct dcb_table *dcb, int i2c, int heads) { struct dcb_entry *entry = new_dcb_entry(dcb); @@ -5287,7 +5315,7 @@ static void fabricate_vga_output(struct parsed_dcb *dcb, int i2c, int heads) /* "or" mostly unused in early gen crt modesetting, 0 is fine */ } -static void fabricate_dvi_i_output(struct parsed_dcb *dcb, bool twoHeads) +static void fabricate_dvi_i_output(struct dcb_table *dcb, bool twoHeads) { struct dcb_entry *entry = new_dcb_entry(dcb); @@ -5314,7 +5342,7 @@ static void fabricate_dvi_i_output(struct parsed_dcb *dcb, bool twoHeads) #endif } -static void fabricate_tv_output(struct parsed_dcb *dcb, bool twoHeads) +static void fabricate_tv_output(struct dcb_table *dcb, bool twoHeads) { struct dcb_entry *entry = new_dcb_entry(dcb); @@ -5325,13 +5353,13 @@ static void fabricate_tv_output(struct parsed_dcb *dcb, bool twoHeads) } static bool -parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, +parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb, uint32_t conn, uint32_t conf, struct dcb_entry *entry) { entry->type = conn & 0xf; entry->i2c_index = (conn >> 4) & 0xf; entry->heads = (conn >> 8) & 0xf; - if (bdcb->version >= 0x40) + if (dcb->version >= 0x40) entry->connector = (conn >> 12) & 0xf; entry->bus = (conn >> 16) & 0xf; entry->location = (conn >> 20) & 0x3; @@ -5349,7 +5377,7 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, * Although the rest of a CRT conf dword is usually * zeros, mac biosen have stuff there so we must mask */ - entry->crtconf.maxfreq = (bdcb->version < 0x30) ? + entry->crtconf.maxfreq = (dcb->version < 0x30) ? (conf & 0xffff) * 10 : (conf & 0xff) * 10000; break; @@ -5358,7 +5386,7 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, uint32_t mask; if (conf & 0x1) entry->lvdsconf.use_straps_for_mode = true; - if (bdcb->version < 0x22) { + if (dcb->version < 0x22) { mask = ~0xd; /* * The laptop in bug 14567 lies and claims to not use @@ -5382,7 +5410,7 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, * Until we even try to use these on G8x, it's * useless reporting unknown bits. They all are. */ - if (bdcb->version >= 0x40) + if (dcb->version >= 0x40) break; NV_ERROR(dev, "Unknown LVDS configuration bits, " @@ -5392,7 +5420,7 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, } case OUTPUT_TV: { - if (bdcb->version >= 0x30) + if (dcb->version >= 0x30) entry->tvconf.has_component_output = conf & (0x8 << 4); else entry->tvconf.has_component_output = false; @@ -5419,8 +5447,10 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, break; case 0xe: /* weird g80 mobile type that "nv" treats as a terminator */ - bdcb->dcb.entries--; + dcb->entries--; return false; + default: + break; } /* unsure what DCB version introduces this, 3.0? */ @@ -5431,55 +5461,52 @@ parse_dcb20_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, } static bool -parse_dcb15_entry(struct drm_device *dev, struct parsed_dcb *dcb, +parse_dcb15_entry(struct drm_device *dev, struct dcb_table *dcb, uint32_t conn, uint32_t conf, struct dcb_entry *entry) { - if (conn != 0xf0003f00 && conn != 0xf2247f10 && conn != 0xf2204001 && - conn != 0xf2204301 && conn != 0xf2204311 && conn != 0xf2208001 && - conn != 0xf2244001 && conn != 0xf2244301 && conn != 0xf2244311 && - conn != 0xf4204011 && conn != 0xf4208011 && conn != 0xf4248011 && - conn != 0xf2045ff2 && conn != 0xf2045f14 && conn != 0xf207df14 && - conn != 0xf2205004 && conn != 0xf2209004) { - NV_ERROR(dev, "Unknown DCB 1.5 entry, please report\n"); - - /* cause output setting to fail for !TV, so message is seen */ - if ((conn & 0xf) != 0x1) - dcb->entries = 0; - - return false; - } - /* most of the below is a "best guess" atm */ - entry->type = conn & 0xf; - if (entry->type == 2) - /* another way of specifying straps based lvds... */ + switch (conn & 0x0000000f) { + case 0: + entry->type = OUTPUT_ANALOG; + break; + case 1: + entry->type = OUTPUT_TV; + break; + case 2: + case 3: entry->type = OUTPUT_LVDS; - if (entry->type == 4) { /* digital */ - if (conn & 0x10) - entry->type = OUTPUT_LVDS; - else + break; + case 4: + switch ((conn & 0x000000f0) >> 4) { + case 0: entry->type = OUTPUT_TMDS; + break; + case 1: + entry->type = OUTPUT_LVDS; + break; + default: + NV_ERROR(dev, "Unknown DCB subtype 4/%d\n", + (conn & 0x000000f0) >> 4); + return false; + } + break; + default: + NV_ERROR(dev, "Unknown DCB type %d\n", conn & 0x0000000f); + return false; } - /* what's in bits 5-13? could be some encoder maker thing, in tv case */ - entry->i2c_index = (conn >> 14) & 0xf; - /* raw heads field is in range 0-1, so move to 1-2 */ - entry->heads = ((conn >> 18) & 0x7) + 1; - entry->location = (conn >> 21) & 0xf; - /* unused: entry->bus = (conn >> 25) & 0x7; */ - /* set or to be same as heads -- hopefully safe enough */ - entry->or = entry->heads; + + entry->i2c_index = (conn & 0x0003c000) >> 14; + entry->heads = ((conn & 0x001c0000) >> 18) + 1; + entry->or = entry->heads; /* same as heads, hopefully safe enough */ + entry->location = (conn & 0x01e00000) >> 21; + entry->bus = (conn & 0x0e000000) >> 25; entry->duallink_possible = false; switch (entry->type) { case OUTPUT_ANALOG: entry->crtconf.maxfreq = (conf & 0xffff) * 10; break; - case OUTPUT_LVDS: - /* - * This is probably buried in conn's unknown bits. - * This will upset EDID-ful models, if they exist - */ - entry->lvdsconf.use_straps_for_mode = true; - entry->lvdsconf.use_power_scripts = true; + case OUTPUT_TV: + entry->tvconf.has_component_output = false; break; case OUTPUT_TMDS: /* @@ -5488,35 +5515,39 @@ parse_dcb15_entry(struct drm_device *dev, struct parsed_dcb *dcb, */ fabricate_vga_output(dcb, entry->i2c_index, entry->heads); break; - case OUTPUT_TV: - entry->tvconf.has_component_output = false; + case OUTPUT_LVDS: + if ((conn & 0x00003f00) != 0x10) + entry->lvdsconf.use_straps_for_mode = true; + entry->lvdsconf.use_power_scripts = true; + break; + default: break; } return true; } -static bool parse_dcb_entry(struct drm_device *dev, struct bios_parsed_dcb *bdcb, +static bool parse_dcb_entry(struct drm_device *dev, struct dcb_table *dcb, uint32_t conn, uint32_t conf) { - struct dcb_entry *entry = new_dcb_entry(&bdcb->dcb); + struct dcb_entry *entry = new_dcb_entry(dcb); bool ret; - if (bdcb->version >= 0x20) - ret = parse_dcb20_entry(dev, bdcb, conn, conf, entry); + if (dcb->version >= 0x20) + ret = parse_dcb20_entry(dev, dcb, conn, conf, entry); else - ret = parse_dcb15_entry(dev, &bdcb->dcb, conn, conf, entry); + ret = parse_dcb15_entry(dev, dcb, conn, conf, entry); if (!ret) return ret; - read_dcb_i2c_entry(dev, bdcb->version, bdcb->i2c_table, - entry->i2c_index, &bdcb->dcb.i2c[entry->i2c_index]); + read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table, + entry->i2c_index, &dcb->i2c[entry->i2c_index]); return true; } static -void merge_like_dcb_entries(struct drm_device *dev, struct parsed_dcb *dcb) +void merge_like_dcb_entries(struct drm_device *dev, struct dcb_table *dcb) { /* * DCB v2.0 lists each output combination separately. @@ -5564,31 +5595,31 @@ void merge_like_dcb_entries(struct drm_device *dev, struct parsed_dcb *dcb) dcb->entries = newentries; } -static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool twoHeads) +static int +parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool twoHeads) { - struct bios_parsed_dcb *bdcb = &bios->bdcb; - struct parsed_dcb *dcb; - uint16_t dcbptr, i2ctabptr = 0; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct dcb_table *dcb = &bios->dcb; + uint16_t dcbptr = 0, i2ctabptr = 0; uint8_t *dcbtable; uint8_t headerlen = 0x4, entries = DCB_MAX_NUM_ENTRIES; bool configblock = true; int recordlength = 8, confofs = 4; int i; - dcb = bios->pub.dcb = &bdcb->dcb; - dcb->entries = 0; - /* get the offset from 0x36 */ - dcbptr = ROM16(bios->data[0x36]); + if (dev_priv->card_type > NV_04) { + dcbptr = ROM16(bios->data[0x36]); + if (dcbptr == 0x0000) + NV_WARN(dev, "No output data (DCB) found in BIOS\n"); + } + /* this situation likely means a really old card, pre DCB */ if (dcbptr == 0x0) { - NV_WARN(dev, "No output data (DCB) found in BIOS, " - "assuming a CRT output exists\n"); - /* this situation likely means a really old card, pre DCB */ + NV_INFO(dev, "Assuming a CRT output exists\n"); fabricate_vga_output(dcb, LEGACY_I2C_CRT, 1); - if (nv04_tv_identify(dev, - bios->legacy.i2c_indices.tv) >= 0) + if (nv04_tv_identify(dev, bios->legacy.i2c_indices.tv) >= 0) fabricate_tv_output(dcb, twoHeads); return 0; @@ -5597,21 +5628,21 @@ static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool two dcbtable = &bios->data[dcbptr]; /* get DCB version */ - bdcb->version = dcbtable[0]; + dcb->version = dcbtable[0]; NV_TRACE(dev, "Found Display Configuration Block version %d.%d\n", - bdcb->version >> 4, bdcb->version & 0xf); + dcb->version >> 4, dcb->version & 0xf); - if (bdcb->version >= 0x20) { /* NV17+ */ + if (dcb->version >= 0x20) { /* NV17+ */ uint32_t sig; - if (bdcb->version >= 0x30) { /* NV40+ */ + if (dcb->version >= 0x30) { /* NV40+ */ headerlen = dcbtable[1]; entries = dcbtable[2]; recordlength = dcbtable[3]; i2ctabptr = ROM16(dcbtable[4]); sig = ROM32(dcbtable[6]); - bdcb->gpio_table_ptr = ROM16(dcbtable[10]); - bdcb->connector_table_ptr = ROM16(dcbtable[20]); + dcb->gpio_table_ptr = ROM16(dcbtable[10]); + dcb->connector_table_ptr = ROM16(dcbtable[20]); } else { i2ctabptr = ROM16(dcbtable[2]); sig = ROM32(dcbtable[4]); @@ -5623,7 +5654,7 @@ static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool two "signature (%08X)\n", sig); return -EINVAL; } - } else if (bdcb->version >= 0x15) { /* some NV11 and NV20 */ + } else if (dcb->version >= 0x15) { /* some NV11 and NV20 */ char sig[8] = { 0 }; strncpy(sig, (char *)&dcbtable[-7], 7); @@ -5671,14 +5702,11 @@ static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool two if (!i2ctabptr) NV_WARN(dev, "No pointer to DCB I2C port table\n"); else { - bdcb->i2c_table = &bios->data[i2ctabptr]; - if (bdcb->version >= 0x30) - bdcb->i2c_default_indices = bdcb->i2c_table[4]; + dcb->i2c_table = &bios->data[i2ctabptr]; + if (dcb->version >= 0x30) + dcb->i2c_default_indices = dcb->i2c_table[4]; } - parse_dcb_gpio_table(bios); - parse_dcb_connector_table(bios); - if (entries > DCB_MAX_NUM_ENTRIES) entries = DCB_MAX_NUM_ENTRIES; @@ -5703,7 +5731,7 @@ static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool two NV_TRACEWARN(dev, "Raw DCB entry %d: %08x %08x\n", dcb->entries, connection, config); - if (!parse_dcb_entry(dev, bdcb, connection, config)) + if (!parse_dcb_entry(dev, dcb, connection, config)) break; } @@ -5711,18 +5739,22 @@ static int parse_dcb_table(struct drm_device *dev, struct nvbios *bios, bool two * apart for v2.1+ not being known for requiring merging, this * guarantees dcbent->index is the index of the entry in the rom image */ - if (bdcb->version < 0x21) + if (dcb->version < 0x21) merge_like_dcb_entries(dev, dcb); - return dcb->entries ? 0 : -ENXIO; + if (!dcb->entries) + return -ENXIO; + + parse_dcb_gpio_table(bios); + parse_dcb_connector_table(bios); + return 0; } static void fixup_legacy_connector(struct nvbios *bios) { - struct bios_parsed_dcb *bdcb = &bios->bdcb; - struct parsed_dcb *dcb = &bdcb->dcb; - int high = 0, i; + struct dcb_table *dcb = &bios->dcb; + int i, i2c, i2c_conn[DCB_MAX_NUM_I2C_ENTRIES] = { }; /* * DCB 3.0 also has the table in most cases, but there are some cards @@ -5730,9 +5762,11 @@ fixup_legacy_connector(struct nvbios *bios) * indices are all 0. We don't need the connector indices on pre-G80 * chips (yet?) so limit the use to DCB 4.0 and above. */ - if (bdcb->version >= 0x40) + if (dcb->version >= 0x40) return; + dcb->connector.entries = 0; + /* * No known connector info before v3.0, so make it up. the rule here * is: anything on the same i2c bus is considered to be on the same @@ -5740,37 +5774,38 @@ fixup_legacy_connector(struct nvbios *bios) * its own unique connector index. */ for (i = 0; i < dcb->entries; i++) { - if (dcb->entry[i].i2c_index == 0xf) - continue; - /* * Ignore the I2C index for on-chip TV-out, as there * are cards with bogus values (nv31m in bug 23212), * and it's otherwise useless. */ if (dcb->entry[i].type == OUTPUT_TV && - dcb->entry[i].location == DCB_LOC_ON_CHIP) { + dcb->entry[i].location == DCB_LOC_ON_CHIP) dcb->entry[i].i2c_index = 0xf; + i2c = dcb->entry[i].i2c_index; + + if (i2c_conn[i2c]) { + dcb->entry[i].connector = i2c_conn[i2c] - 1; continue; } - dcb->entry[i].connector = dcb->entry[i].i2c_index; - if (dcb->entry[i].connector > high) - high = dcb->entry[i].connector; + dcb->entry[i].connector = dcb->connector.entries++; + if (i2c != 0xf) + i2c_conn[i2c] = dcb->connector.entries; } - for (i = 0; i < dcb->entries; i++) { - if (dcb->entry[i].i2c_index != 0xf) - continue; - - dcb->entry[i].connector = ++high; + /* Fake the connector table as well as just connector indices */ + for (i = 0; i < dcb->connector.entries; i++) { + dcb->connector.entry[i].index = i; + dcb->connector.entry[i].type = divine_connector_type(bios, i); + dcb->connector.entry[i].gpio_tag = 0xff; } } static void fixup_legacy_i2c(struct nvbios *bios) { - struct parsed_dcb *dcb = &bios->bdcb.dcb; + struct dcb_table *dcb = &bios->dcb; int i; for (i = 0; i < dcb->entries; i++) { @@ -5856,7 +5891,7 @@ static int load_nv17_hw_sequencer_ucode(struct drm_device *dev, uint8_t *nouveau_bios_embedded_edid(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; const uint8_t edid_sig[] = { 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00 }; uint16_t offset = 0; @@ -5889,20 +5924,23 @@ nouveau_bios_run_init_table(struct drm_device *dev, uint16_t table, struct dcb_entry *dcbent) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; struct init_exec iexec = { true, false }; + mutex_lock(&bios->lock); bios->display.output = dcbent; parse_init_table(bios, table, &iexec); bios->display.output = NULL; + mutex_unlock(&bios->lock); } static bool NVInitVBIOS(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; memset(bios, 0, sizeof(struct nvbios)); + mutex_init(&bios->lock); bios->dev = dev; if (!NVShadowVBIOS(dev, bios->data)) @@ -5915,7 +5953,7 @@ static bool NVInitVBIOS(struct drm_device *dev) static int nouveau_parse_vbios_struct(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; const uint8_t bit_signature[] = { 0xff, 0xb8, 'B', 'I', 'T' }; const uint8_t bmp_signature[] = { 0xff, 0x7f, 'N', 'V', 0x0 }; int offset; @@ -5942,7 +5980,7 @@ int nouveau_run_vbios_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; int i, ret = 0; NVLockVgaCrtcs(dev, false); @@ -5973,9 +6011,9 @@ nouveau_run_vbios_init(struct drm_device *dev) } if (dev_priv->card_type >= NV_50) { - for (i = 0; i < bios->bdcb.dcb.entries; i++) { + for (i = 0; i < bios->dcb.entries; i++) { nouveau_bios_run_display_table(dev, - &bios->bdcb.dcb.entry[i], + &bios->dcb.entry[i], 0, 0); } } @@ -5989,11 +6027,11 @@ static void nouveau_bios_i2c_devices_takedown(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; struct dcb_i2c_entry *entry; int i; - entry = &bios->bdcb.dcb.i2c[0]; + entry = &bios->dcb.i2c[0]; for (i = 0; i < DCB_MAX_NUM_I2C_ENTRIES; i++, entry++) nouveau_i2c_fini(dev, entry); } @@ -6002,13 +6040,11 @@ int nouveau_bios_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint32_t saved_nv_pextdev_boot_0; bool was_locked; int ret; - dev_priv->vbios = &bios->pub; - if (!NVInitVBIOS(dev)) return -ENODEV; @@ -6050,10 +6086,8 @@ nouveau_bios_init(struct drm_device *dev) bios_wr32(bios, NV_PEXTDEV_BOOT_0, saved_nv_pextdev_boot_0); ret = nouveau_run_vbios_init(dev); - if (ret) { - dev_priv->vbios = NULL; + if (ret) return ret; - } /* feature_byte on BMP is poor, but init always sets CR4B */ was_locked = NVLockVgaCrtcs(dev, false); diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h index 058e98c76d8..9f688aa9a65 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.h +++ b/drivers/gpu/drm/nouveau/nouveau_bios.h @@ -34,9 +34,67 @@ #define DCB_LOC_ON_CHIP 0 +struct dcb_i2c_entry { + uint8_t port_type; + uint8_t read, write; + struct nouveau_i2c_chan *chan; +}; + +enum dcb_gpio_tag { + DCB_GPIO_TVDAC0 = 0xc, + DCB_GPIO_TVDAC1 = 0x2d, +}; + +struct dcb_gpio_entry { + enum dcb_gpio_tag tag; + int line; + bool invert; +}; + +struct dcb_gpio_table { + int entries; + struct dcb_gpio_entry entry[DCB_MAX_NUM_GPIO_ENTRIES]; +}; + +enum dcb_connector_type { + DCB_CONNECTOR_VGA = 0x00, + DCB_CONNECTOR_TV_0 = 0x10, + DCB_CONNECTOR_TV_1 = 0x11, + DCB_CONNECTOR_TV_3 = 0x13, + DCB_CONNECTOR_DVI_I = 0x30, + DCB_CONNECTOR_DVI_D = 0x31, + DCB_CONNECTOR_LVDS = 0x40, + DCB_CONNECTOR_DP = 0x46, + DCB_CONNECTOR_eDP = 0x47, + DCB_CONNECTOR_HDMI_0 = 0x60, + DCB_CONNECTOR_HDMI_1 = 0x61, + DCB_CONNECTOR_NONE = 0xff +}; + +struct dcb_connector_table_entry { + uint32_t entry; + enum dcb_connector_type type; + uint8_t index; + uint8_t gpio_tag; +}; + +struct dcb_connector_table { + int entries; + struct dcb_connector_table_entry entry[DCB_MAX_NUM_CONNECTOR_ENTRIES]; +}; + +enum dcb_type { + OUTPUT_ANALOG = 0, + OUTPUT_TV = 1, + OUTPUT_TMDS = 2, + OUTPUT_LVDS = 3, + OUTPUT_DP = 6, + OUTPUT_ANY = -1 +}; + struct dcb_entry { int index; /* may not be raw dcb index if merging has happened */ - uint8_t type; + enum dcb_type type; uint8_t i2c_index; uint8_t heads; uint8_t connector; @@ -71,69 +129,22 @@ struct dcb_entry { bool i2c_upper_default; }; -struct dcb_i2c_entry { - uint8_t port_type; - uint8_t read, write; - struct nouveau_i2c_chan *chan; -}; +struct dcb_table { + uint8_t version; -struct parsed_dcb { int entries; struct dcb_entry entry[DCB_MAX_NUM_ENTRIES]; - struct dcb_i2c_entry i2c[DCB_MAX_NUM_I2C_ENTRIES]; -}; - -enum dcb_gpio_tag { - DCB_GPIO_TVDAC0 = 0xc, - DCB_GPIO_TVDAC1 = 0x2d, -}; - -struct dcb_gpio_entry { - enum dcb_gpio_tag tag; - int line; - bool invert; -}; - -struct parsed_dcb_gpio { - int entries; - struct dcb_gpio_entry entry[DCB_MAX_NUM_GPIO_ENTRIES]; -}; - -struct dcb_connector_table_entry { - uint32_t entry; - uint8_t type; - uint8_t index; - uint8_t gpio_tag; -}; - -struct dcb_connector_table { - int entries; - struct dcb_connector_table_entry entry[DCB_MAX_NUM_CONNECTOR_ENTRIES]; -}; - -struct bios_parsed_dcb { - uint8_t version; - - struct parsed_dcb dcb; uint8_t *i2c_table; uint8_t i2c_default_indices; + struct dcb_i2c_entry i2c[DCB_MAX_NUM_I2C_ENTRIES]; uint16_t gpio_table_ptr; - struct parsed_dcb_gpio gpio; + struct dcb_gpio_table gpio; uint16_t connector_table_ptr; struct dcb_connector_table connector; }; -enum nouveau_encoder_type { - OUTPUT_ANALOG = 0, - OUTPUT_TV = 1, - OUTPUT_TMDS = 2, - OUTPUT_LVDS = 3, - OUTPUT_DP = 6, - OUTPUT_ANY = -1 -}; - enum nouveau_or { OUTPUT_A = (1 << 0), OUTPUT_B = (1 << 1), @@ -190,8 +201,8 @@ struct pll_lims { int refclk; }; -struct nouveau_bios_info { - struct parsed_dcb *dcb; +struct nvbios { + struct drm_device *dev; uint8_t chip_version; @@ -199,11 +210,8 @@ struct nouveau_bios_info { uint32_t tvdactestval; uint8_t digital_min_front_porch; bool fp_no_ddc; -}; -struct nvbios { - struct drm_device *dev; - struct nouveau_bios_info pub; + struct mutex lock; uint8_t data[NV_PROM_SIZE]; unsigned int length; @@ -232,7 +240,7 @@ struct nvbios { uint16_t some_script_ptr; /* BIT I + 14 */ uint16_t init96_tbl_ptr; /* BIT I + 16 */ - struct bios_parsed_dcb bdcb; + struct dcb_table dcb; struct { int crtchead; diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index e342a418d43..028719fddf7 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -65,8 +65,10 @@ nouveau_bo_fixup_align(struct drm_device *dev, /* * Some of the tile_flags have a periodic structure of N*4096 bytes, - * align to to that as well as the page size. Overallocate memory to - * avoid corruption of other buffer objects. + * align to to that as well as the page size. Align the size to the + * appropriate boundaries. This does imply that sizes are rounded up + * 3-7 pages, so be aware of this and do not waste memory by allocating + * many small buffers. */ if (dev_priv->card_type == NV_50) { uint32_t block_size = nouveau_mem_fb_amount(dev) >> 15; @@ -77,22 +79,20 @@ nouveau_bo_fixup_align(struct drm_device *dev, case 0x2800: case 0x4800: case 0x7a00: - *size = roundup(*size, block_size); if (is_power_of_2(block_size)) { - *size += 3 * block_size; for (i = 1; i < 10; i++) { *align = 12 * i * block_size; if (!(*align % 65536)) break; } } else { - *size += 6 * block_size; for (i = 1; i < 10; i++) { *align = 8 * i * block_size; if (!(*align % 65536)) break; } } + *size = roundup(*size, *align); break; default: break; @@ -469,6 +469,8 @@ nouveau_bo_move_accel_cleanup(struct nouveau_channel *chan, ret = ttm_bo_move_accel_cleanup(&nvbo->bo, fence, NULL, evict, no_wait, new_mem); + if (nvbo->channel && nvbo->channel != chan) + ret = nouveau_fence_wait(fence, NULL, false, false); nouveau_fence_unref((void *)&fence); return ret; } diff --git a/drivers/gpu/drm/nouveau/nouveau_calc.c b/drivers/gpu/drm/nouveau/nouveau_calc.c index ee2b84504d0..88f9bc0941e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_calc.c +++ b/drivers/gpu/drm/nouveau/nouveau_calc.c @@ -274,7 +274,7 @@ getMNP_single(struct drm_device *dev, struct pll_lims *pll_lim, int clk, * returns calculated clock */ struct drm_nouveau_private *dev_priv = dev->dev_private; - int cv = dev_priv->vbios->chip_version; + int cv = dev_priv->vbios.chip_version; int minvco = pll_lim->vco1.minfreq, maxvco = pll_lim->vco1.maxfreq; int minM = pll_lim->vco1.min_m, maxM = pll_lim->vco1.max_m; int minN = pll_lim->vco1.min_n, maxN = pll_lim->vco1.max_n; @@ -373,7 +373,7 @@ getMNP_double(struct drm_device *dev, struct pll_lims *pll_lim, int clk, * returns calculated clock */ struct drm_nouveau_private *dev_priv = dev->dev_private; - int chip_version = dev_priv->vbios->chip_version; + int chip_version = dev_priv->vbios.chip_version; int minvco1 = pll_lim->vco1.minfreq, maxvco1 = pll_lim->vco1.maxfreq; int minvco2 = pll_lim->vco2.minfreq, maxvco2 = pll_lim->vco2.maxfreq; int minU1 = pll_lim->vco1.min_inputfreq, minU2 = pll_lim->vco2.min_inputfreq; diff --git a/drivers/gpu/drm/nouveau/nouveau_channel.c b/drivers/gpu/drm/nouveau/nouveau_channel.c index 343d718a966..6dfb425cbae 100644 --- a/drivers/gpu/drm/nouveau/nouveau_channel.c +++ b/drivers/gpu/drm/nouveau/nouveau_channel.c @@ -35,22 +35,27 @@ nouveau_channel_pushbuf_ctxdma_init(struct nouveau_channel *chan) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_bo *pb = chan->pushbuf_bo; struct nouveau_gpuobj *pushbuf = NULL; - uint32_t start = pb->bo.mem.mm_node->start << PAGE_SHIFT; int ret; + if (dev_priv->card_type >= NV_50) { + ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY, 0, + dev_priv->vm_end, NV_DMA_ACCESS_RO, + NV_DMA_TARGET_AGP, &pushbuf); + chan->pushbuf_base = pb->bo.offset; + } else if (pb->bo.mem.mem_type == TTM_PL_TT) { ret = nouveau_gpuobj_gart_dma_new(chan, 0, dev_priv->gart_info.aper_size, NV_DMA_ACCESS_RO, &pushbuf, NULL); - chan->pushbuf_base = start; + chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT; } else if (dev_priv->card_type != NV_04) { ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY, 0, dev_priv->fb_available_size, NV_DMA_ACCESS_RO, NV_DMA_TARGET_VIDMEM, &pushbuf); - chan->pushbuf_base = start; + chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT; } else { /* NV04 cmdbuf hack, from original ddx.. not sure of it's * exact reason for existing :) PCI access to cmdbuf in @@ -61,7 +66,7 @@ nouveau_channel_pushbuf_ctxdma_init(struct nouveau_channel *chan) dev_priv->fb_available_size, NV_DMA_ACCESS_RO, NV_DMA_TARGET_PCI, &pushbuf); - chan->pushbuf_base = start; + chan->pushbuf_base = pb->bo.mem.mm_node->start << PAGE_SHIFT; } ret = nouveau_gpuobj_ref_add(dev, chan, 0, pushbuf, &chan->pushbuf); @@ -275,15 +280,23 @@ nouveau_channel_free(struct nouveau_channel *chan) */ nouveau_fence_fini(chan); - /* Ensure the channel is no longer active on the GPU */ + /* This will prevent pfifo from switching channels. */ pfifo->reassign(dev, false); - if (pgraph->channel(dev) == chan) { - pgraph->fifo_access(dev, false); + /* We want to give pgraph a chance to idle and get rid of all potential + * errors. We need to do this before the lock, otherwise the irq handler + * is unable to process them. + */ + if (pgraph->channel(dev) == chan) + nouveau_wait_for_idle(dev); + + spin_lock_irqsave(&dev_priv->context_switch_lock, flags); + + pgraph->fifo_access(dev, false); + if (pgraph->channel(dev) == chan) pgraph->unload_context(dev); - pgraph->fifo_access(dev, true); - } pgraph->destroy_context(chan); + pgraph->fifo_access(dev, true); if (pfifo->channel_id(dev) == chan->id) { pfifo->disable(dev); @@ -294,6 +307,8 @@ nouveau_channel_free(struct nouveau_channel *chan) pfifo->reassign(dev, true); + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); + /* Release the channel's resources */ nouveau_gpuobj_ref_del(dev, &chan->pushbuf); if (chan->pushbuf_bo) { @@ -370,6 +385,14 @@ nouveau_ioctl_fifo_alloc(struct drm_device *dev, void *data, return ret; init->channel = chan->id; + if (chan->dma.ib_max) + init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM | + NOUVEAU_GEM_DOMAIN_GART; + else if (chan->pushbuf_bo->bo.mem.mem_type == TTM_PL_VRAM) + init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM; + else + init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_GART; + init->subchan[0].handle = NvM2MF; if (dev_priv->card_type < NV_50) init->subchan[0].grclass = 0x0039; @@ -409,7 +432,6 @@ nouveau_ioctl_fifo_free(struct drm_device *dev, void *data, ***********************************/ struct drm_ioctl_desc nouveau_ioctls[] = { - DRM_IOCTL_DEF(DRM_NOUVEAU_CARD_INIT, nouveau_ioctl_card_init, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GETPARAM, nouveau_ioctl_getparam, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_SETPARAM, nouveau_ioctl_setparam, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF(DRM_NOUVEAU_CHANNEL_ALLOC, nouveau_ioctl_fifo_alloc, DRM_AUTH), @@ -419,13 +441,9 @@ struct drm_ioctl_desc nouveau_ioctls[] = { DRM_IOCTL_DEF(DRM_NOUVEAU_GPUOBJ_FREE, nouveau_ioctl_gpuobj_free, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_NEW, nouveau_gem_ioctl_new, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_PUSHBUF, nouveau_gem_ioctl_pushbuf, DRM_AUTH), - DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_PUSHBUF_CALL, nouveau_gem_ioctl_pushbuf_call, DRM_AUTH), - DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_PIN, nouveau_gem_ioctl_pin, DRM_AUTH), - DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_UNPIN, nouveau_gem_ioctl_unpin, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_CPU_PREP, nouveau_gem_ioctl_cpu_prep, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_CPU_FINI, nouveau_gem_ioctl_cpu_fini, DRM_AUTH), DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_INFO, nouveau_gem_ioctl_info, DRM_AUTH), - DRM_IOCTL_DEF(DRM_NOUVEAU_GEM_PUSHBUF_CALL2, nouveau_gem_ioctl_pushbuf_call2, DRM_AUTH), }; int nouveau_max_ioctl = DRM_ARRAY_SIZE(nouveau_ioctls); diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index 5a10deb8bdb..24327f468c4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -24,9 +24,12 @@ * */ +#include <acpi/button.h> + #include "drmP.h" #include "drm_edid.h" #include "drm_crtc_helper.h" + #include "nouveau_reg.h" #include "nouveau_drv.h" #include "nouveau_encoder.h" @@ -83,14 +86,17 @@ nouveau_encoder_connector_get(struct nouveau_encoder *encoder) static void nouveau_connector_destroy(struct drm_connector *drm_connector) { - struct nouveau_connector *connector = nouveau_connector(drm_connector); - struct drm_device *dev = connector->base.dev; - - NV_DEBUG_KMS(dev, "\n"); + struct nouveau_connector *nv_connector = + nouveau_connector(drm_connector); + struct drm_device *dev; - if (!connector) + if (!nv_connector) return; + dev = nv_connector->base.dev; + NV_DEBUG_KMS(dev, "\n"); + + kfree(nv_connector->edid); drm_sysfs_connector_remove(drm_connector); drm_connector_cleanup(drm_connector); kfree(drm_connector); @@ -212,7 +218,7 @@ nouveau_connector_set_encoder(struct drm_connector *connector, connector->interlace_allowed = true; } - if (connector->connector_type == DRM_MODE_CONNECTOR_DVII) { + if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) { drm_connector_property_set_value(connector, dev->mode_config.dvi_i_subconnector_property, nv_encoder->dcb->type == OUTPUT_TMDS ? @@ -230,11 +236,24 @@ nouveau_connector_detect(struct drm_connector *connector) struct nouveau_i2c_chan *i2c; int type, flags; - if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) + if (nv_connector->dcb->type == DCB_CONNECTOR_LVDS) nv_encoder = find_encoder_by_type(connector, OUTPUT_LVDS); if (nv_encoder && nv_connector->native_mode) { + unsigned status = connector_status_connected; + +#ifdef CONFIG_ACPI + if (!nouveau_ignorelid && !acpi_lid_open()) + status = connector_status_unknown; +#endif nouveau_connector_set_encoder(connector, nv_encoder); - return connector_status_connected; + return status; + } + + /* Cleanup the previous EDID block. */ + if (nv_connector->edid) { + drm_mode_connector_update_edid_property(connector, NULL); + kfree(nv_connector->edid); + nv_connector->edid = NULL; } i2c = nouveau_connector_ddc_detect(connector, &nv_encoder); @@ -247,7 +266,7 @@ nouveau_connector_detect(struct drm_connector *connector) if (!nv_connector->edid) { NV_ERROR(dev, "DDC responded, but no EDID for %s\n", drm_get_connector_name(connector)); - return connector_status_disconnected; + goto detect_analog; } if (nv_encoder->dcb->type == OUTPUT_DP && @@ -262,7 +281,7 @@ nouveau_connector_detect(struct drm_connector *connector) * same i2c channel so the value returned from ddc_detect * isn't necessarily correct. */ - if (connector->connector_type == DRM_MODE_CONNECTOR_DVII) { + if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) { if (nv_connector->edid->input & DRM_EDID_INPUT_DIGITAL) type = OUTPUT_TMDS; else @@ -281,6 +300,7 @@ nouveau_connector_detect(struct drm_connector *connector) return connector_status_connected; } +detect_analog: nv_encoder = find_encoder_by_type(connector, OUTPUT_ANALOG); if (!nv_encoder) nv_encoder = find_encoder_by_type(connector, OUTPUT_TV); @@ -303,11 +323,11 @@ nouveau_connector_detect(struct drm_connector *connector) static void nouveau_connector_force(struct drm_connector *connector) { - struct drm_device *dev = connector->dev; + struct nouveau_connector *nv_connector = nouveau_connector(connector); struct nouveau_encoder *nv_encoder; int type; - if (connector->connector_type == DRM_MODE_CONNECTOR_DVII) { + if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) { if (connector->force == DRM_FORCE_ON_DIGITAL) type = OUTPUT_TMDS; else @@ -317,7 +337,7 @@ nouveau_connector_force(struct drm_connector *connector) nv_encoder = find_encoder_by_type(connector, type); if (!nv_encoder) { - NV_ERROR(dev, "can't find encoder to force %s on!\n", + NV_ERROR(connector->dev, "can't find encoder to force %s on!\n", drm_get_connector_name(connector)); connector->status = connector_status_disconnected; return; @@ -351,7 +371,7 @@ nouveau_connector_set_property(struct drm_connector *connector, } /* LVDS always needs gpu scaling */ - if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS && + if (nv_connector->dcb->type == DCB_CONNECTOR_LVDS && value == DRM_MODE_SCALE_NONE) return -EINVAL; @@ -517,7 +537,7 @@ nouveau_connector_get_modes(struct drm_connector *connector) /* If we're not LVDS, destroy the previous native mode, the attached * monitor could have changed. */ - if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS && + if (nv_connector->dcb->type != DCB_CONNECTOR_LVDS && nv_connector->native_mode) { drm_mode_destroy(dev, nv_connector->native_mode); nv_connector->native_mode = NULL; @@ -545,7 +565,7 @@ nouveau_connector_get_modes(struct drm_connector *connector) ret = get_slave_funcs(nv_encoder)-> get_modes(to_drm_encoder(nv_encoder), connector); - if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) + if (nv_encoder->dcb->type == OUTPUT_LVDS) ret += nouveau_connector_scaler_modes_add(connector); return ret; @@ -595,6 +615,9 @@ nouveau_connector_mode_valid(struct drm_connector *connector, clock *= 3; break; + default: + BUG_ON(1); + return MODE_BAD; } if (clock < min_clock) @@ -662,7 +685,7 @@ nouveau_connector_create_lvds(struct drm_device *dev, /* Firstly try getting EDID over DDC, if allowed and I2C channel * is available. */ - if (!dev_priv->VBIOS.pub.fp_no_ddc && nv_encoder->dcb->i2c_index < 0xf) + if (!dev_priv->vbios.fp_no_ddc && nv_encoder->dcb->i2c_index < 0xf) i2c = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index); if (i2c) { @@ -677,7 +700,7 @@ nouveau_connector_create_lvds(struct drm_device *dev, */ if (!nv_connector->edid && nouveau_bios_fp_mode(dev, &native) && (nv_encoder->dcb->lvdsconf.use_straps_for_mode || - dev_priv->VBIOS.pub.fp_no_ddc)) { + dev_priv->vbios.fp_no_ddc)) { nv_connector->native_mode = drm_mode_duplicate(dev, &native); goto out; } @@ -686,9 +709,13 @@ nouveau_connector_create_lvds(struct drm_device *dev, * stored for the panel stored in them. */ if (!nv_connector->edid && !nv_connector->native_mode && - !dev_priv->VBIOS.pub.fp_no_ddc) { - nv_connector->edid = + !dev_priv->vbios.fp_no_ddc) { + struct edid *edid = (struct edid *)nouveau_bios_embedded_edid(dev); + if (edid) { + nv_connector->edid = kmalloc(EDID_LENGTH, GFP_KERNEL); + *(nv_connector->edid) = *edid; + } } if (!nv_connector->edid) @@ -717,46 +744,66 @@ out: } int -nouveau_connector_create(struct drm_device *dev, int index, int type) +nouveau_connector_create(struct drm_device *dev, + struct dcb_connector_table_entry *dcb) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_connector *nv_connector = NULL; struct drm_connector *connector; struct drm_encoder *encoder; - int ret; + int ret, type; NV_DEBUG_KMS(dev, "\n"); - nv_connector = kzalloc(sizeof(*nv_connector), GFP_KERNEL); - if (!nv_connector) - return -ENOMEM; - nv_connector->dcb = nouveau_bios_connector_entry(dev, index); - connector = &nv_connector->base; - - switch (type) { - case DRM_MODE_CONNECTOR_VGA: + switch (dcb->type) { + case DCB_CONNECTOR_NONE: + return 0; + case DCB_CONNECTOR_VGA: NV_INFO(dev, "Detected a VGA connector\n"); + type = DRM_MODE_CONNECTOR_VGA; break; - case DRM_MODE_CONNECTOR_DVID: - NV_INFO(dev, "Detected a DVI-D connector\n"); + case DCB_CONNECTOR_TV_0: + case DCB_CONNECTOR_TV_1: + case DCB_CONNECTOR_TV_3: + NV_INFO(dev, "Detected a TV connector\n"); + type = DRM_MODE_CONNECTOR_TV; break; - case DRM_MODE_CONNECTOR_DVII: + case DCB_CONNECTOR_DVI_I: NV_INFO(dev, "Detected a DVI-I connector\n"); + type = DRM_MODE_CONNECTOR_DVII; break; - case DRM_MODE_CONNECTOR_LVDS: - NV_INFO(dev, "Detected a LVDS connector\n"); + case DCB_CONNECTOR_DVI_D: + NV_INFO(dev, "Detected a DVI-D connector\n"); + type = DRM_MODE_CONNECTOR_DVID; break; - case DRM_MODE_CONNECTOR_TV: - NV_INFO(dev, "Detected a TV connector\n"); + case DCB_CONNECTOR_HDMI_0: + case DCB_CONNECTOR_HDMI_1: + NV_INFO(dev, "Detected a HDMI connector\n"); + type = DRM_MODE_CONNECTOR_HDMIA; break; - case DRM_MODE_CONNECTOR_DisplayPort: + case DCB_CONNECTOR_LVDS: + NV_INFO(dev, "Detected a LVDS connector\n"); + type = DRM_MODE_CONNECTOR_LVDS; + break; + case DCB_CONNECTOR_DP: NV_INFO(dev, "Detected a DisplayPort connector\n"); + type = DRM_MODE_CONNECTOR_DisplayPort; break; - default: - NV_ERROR(dev, "Unknown connector, this is not good.\n"); + case DCB_CONNECTOR_eDP: + NV_INFO(dev, "Detected an eDP connector\n"); + type = DRM_MODE_CONNECTOR_eDP; break; + default: + NV_ERROR(dev, "unknown connector type: 0x%02x!!\n", dcb->type); + return -EINVAL; } + nv_connector = kzalloc(sizeof(*nv_connector), GFP_KERNEL); + if (!nv_connector) + return -ENOMEM; + nv_connector->dcb = dcb; + connector = &nv_connector->base; + /* defaults, will get overridden in detect() */ connector->interlace_allowed = false; connector->doublescan_allowed = false; @@ -764,55 +811,65 @@ nouveau_connector_create(struct drm_device *dev, int index, int type) drm_connector_init(dev, connector, &nouveau_connector_funcs, type); drm_connector_helper_add(connector, &nouveau_connector_helper_funcs); + /* attach encoders */ + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + + if (nv_encoder->dcb->connector != dcb->index) + continue; + + if (get_slave_funcs(nv_encoder)) + get_slave_funcs(nv_encoder)->create_resources(encoder, connector); + + drm_mode_connector_attach_encoder(connector, encoder); + } + + if (!connector->encoder_ids[0]) { + NV_WARN(dev, " no encoders, ignoring\n"); + drm_connector_cleanup(connector); + kfree(connector); + return 0; + } + /* Init DVI-I specific properties */ - if (type == DRM_MODE_CONNECTOR_DVII) { + if (dcb->type == DCB_CONNECTOR_DVI_I) { drm_mode_create_dvi_i_properties(dev); drm_connector_attach_property(connector, dev->mode_config.dvi_i_subconnector_property, 0); drm_connector_attach_property(connector, dev->mode_config.dvi_i_select_subconnector_property, 0); } - if (type != DRM_MODE_CONNECTOR_LVDS) + if (dcb->type != DCB_CONNECTOR_LVDS) nv_connector->use_dithering = false; - if (type == DRM_MODE_CONNECTOR_DVID || - type == DRM_MODE_CONNECTOR_DVII || - type == DRM_MODE_CONNECTOR_LVDS || - type == DRM_MODE_CONNECTOR_DisplayPort) { - nv_connector->scaling_mode = DRM_MODE_SCALE_FULLSCREEN; - - drm_connector_attach_property(connector, dev->mode_config.scaling_mode_property, - nv_connector->scaling_mode); - drm_connector_attach_property(connector, dev->mode_config.dithering_mode_property, - nv_connector->use_dithering ? DRM_MODE_DITHERING_ON - : DRM_MODE_DITHERING_OFF); - - } else { - nv_connector->scaling_mode = DRM_MODE_SCALE_NONE; - - if (type == DRM_MODE_CONNECTOR_VGA && - dev_priv->card_type >= NV_50) { + switch (dcb->type) { + case DCB_CONNECTOR_VGA: + if (dev_priv->card_type >= NV_50) { drm_connector_attach_property(connector, dev->mode_config.scaling_mode_property, nv_connector->scaling_mode); } - } - - /* attach encoders */ - list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { - struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); - - if (nv_encoder->dcb->connector != index) - continue; - - if (get_slave_funcs(nv_encoder)) - get_slave_funcs(nv_encoder)->create_resources(encoder, connector); + /* fall-through */ + case DCB_CONNECTOR_TV_0: + case DCB_CONNECTOR_TV_1: + case DCB_CONNECTOR_TV_3: + nv_connector->scaling_mode = DRM_MODE_SCALE_NONE; + break; + default: + nv_connector->scaling_mode = DRM_MODE_SCALE_FULLSCREEN; - drm_mode_connector_attach_encoder(connector, encoder); + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + nv_connector->scaling_mode); + drm_connector_attach_property(connector, + dev->mode_config.dithering_mode_property, + nv_connector->use_dithering ? + DRM_MODE_DITHERING_ON : DRM_MODE_DITHERING_OFF); + break; } drm_sysfs_connector_add(connector); - if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) { + if (dcb->type == DCB_CONNECTOR_LVDS) { ret = nouveau_connector_create_lvds(dev, connector); if (ret) { connector->funcs->destroy(connector); diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h index 728b8090e5f..4ef38abc2d9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.h +++ b/drivers/gpu/drm/nouveau/nouveau_connector.h @@ -49,6 +49,7 @@ static inline struct nouveau_connector *nouveau_connector( return container_of(con, struct nouveau_connector, base); } -int nouveau_connector_create(struct drm_device *dev, int i2c_index, int type); +int nouveau_connector_create(struct drm_device *, + struct dcb_connector_table_entry *); #endif /* __NOUVEAU_CONNECTOR_H__ */ diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c index d79db3698f1..8ff9ef5d4b4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c +++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c @@ -47,12 +47,23 @@ nouveau_debugfs_channel_info(struct seq_file *m, void *data) seq_printf(m, " cur: 0x%08x\n", chan->dma.cur << 2); seq_printf(m, " put: 0x%08x\n", chan->dma.put << 2); seq_printf(m, " free: 0x%08x\n", chan->dma.free << 2); + if (chan->dma.ib_max) { + seq_printf(m, " ib max: 0x%08x\n", chan->dma.ib_max); + seq_printf(m, " ib put: 0x%08x\n", chan->dma.ib_put); + seq_printf(m, " ib free: 0x%08x\n", chan->dma.ib_free); + } seq_printf(m, "gpu fifo state:\n"); seq_printf(m, " get: 0x%08x\n", nvchan_rd32(chan, chan->user_get)); seq_printf(m, " put: 0x%08x\n", nvchan_rd32(chan, chan->user_put)); + if (chan->dma.ib_max) { + seq_printf(m, " ib get: 0x%08x\n", + nvchan_rd32(chan, 0x88)); + seq_printf(m, " ib put: 0x%08x\n", + nvchan_rd32(chan, 0x8c)); + } seq_printf(m, "last fence : %d\n", chan->fence.sequence); seq_printf(m, "last signalled: %d\n", chan->fence.sequence_ack); @@ -133,9 +144,22 @@ nouveau_debugfs_memory_info(struct seq_file *m, void *data) return 0; } +static int +nouveau_debugfs_vbios_image(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_nouveau_private *dev_priv = node->minor->dev->dev_private; + int i; + + for (i = 0; i < dev_priv->vbios.length; i++) + seq_printf(m, "%c", dev_priv->vbios.data[i]); + return 0; +} + static struct drm_info_list nouveau_debugfs_list[] = { { "chipset", nouveau_debugfs_chipset_info, 0, NULL }, { "memory", nouveau_debugfs_memory_info, 0, NULL }, + { "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL }, }; #define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list) diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c index dfc94391d71..cf1c5c0a0ab 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.c +++ b/drivers/gpu/drm/nouveau/nouveau_display.c @@ -39,11 +39,8 @@ nouveau_user_framebuffer_destroy(struct drm_framebuffer *drm_fb) if (drm_fb->fbdev) nouveau_fbcon_remove(dev, drm_fb); - if (fb->nvbo) { - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(fb->nvbo->gem); - mutex_unlock(&dev->struct_mutex); - } + if (fb->nvbo) + drm_gem_object_unreference_unlocked(fb->nvbo->gem); drm_framebuffer_cleanup(drm_fb); kfree(fb); diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c index 7afbe8b40d5..c8482a108a7 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.c +++ b/drivers/gpu/drm/nouveau/nouveau_dma.c @@ -32,7 +32,22 @@ void nouveau_dma_pre_init(struct nouveau_channel *chan) { - chan->dma.max = (chan->pushbuf_bo->bo.mem.size >> 2) - 2; + struct drm_nouveau_private *dev_priv = chan->dev->dev_private; + struct nouveau_bo *pushbuf = chan->pushbuf_bo; + + if (dev_priv->card_type == NV_50) { + const int ib_size = pushbuf->bo.mem.size / 2; + + chan->dma.ib_base = (pushbuf->bo.mem.size - ib_size) >> 2; + chan->dma.ib_max = (ib_size / 8) - 1; + chan->dma.ib_put = 0; + chan->dma.ib_free = chan->dma.ib_max - chan->dma.ib_put; + + chan->dma.max = (pushbuf->bo.mem.size - ib_size) >> 2; + } else { + chan->dma.max = (pushbuf->bo.mem.size >> 2) - 2; + } + chan->dma.put = 0; chan->dma.cur = chan->dma.put; chan->dma.free = chan->dma.max - chan->dma.cur; @@ -126,37 +141,66 @@ OUT_RINGp(struct nouveau_channel *chan, const void *data, unsigned nr_dwords) chan->dma.cur += nr_dwords; } -static inline bool -READ_GET(struct nouveau_channel *chan, uint32_t *get) +/* Fetch and adjust GPU GET pointer + * + * Returns: + * value >= 0, the adjusted GET pointer + * -EINVAL if GET pointer currently outside main push buffer + * -EBUSY if timeout exceeded + */ +static inline int +READ_GET(struct nouveau_channel *chan, uint32_t *prev_get, uint32_t *timeout) { uint32_t val; val = nvchan_rd32(chan, chan->user_get); - if (val < chan->pushbuf_base || - val > chan->pushbuf_base + (chan->dma.max << 2)) { - /* meaningless to dma_wait() except to know whether the - * GPU has stalled or not - */ - *get = val; - return false; + + /* reset counter as long as GET is still advancing, this is + * to avoid misdetecting a GPU lockup if the GPU happens to + * just be processing an operation that takes a long time + */ + if (val != *prev_get) { + *prev_get = val; + *timeout = 0; } - *get = (val - chan->pushbuf_base) >> 2; - return true; + if ((++*timeout & 0xff) == 0) { + DRM_UDELAY(1); + if (*timeout > 100000) + return -EBUSY; + } + + if (val < chan->pushbuf_base || + val > chan->pushbuf_base + (chan->dma.max << 2)) + return -EINVAL; + + return (val - chan->pushbuf_base) >> 2; } -int -nouveau_dma_wait(struct nouveau_channel *chan, int size) +void +nv50_dma_push(struct nouveau_channel *chan, struct nouveau_bo *bo, + int delta, int length) { - uint32_t get, prev_get = 0, cnt = 0; - bool get_valid; + struct nouveau_bo *pb = chan->pushbuf_bo; + uint64_t offset = bo->bo.offset + delta; + int ip = (chan->dma.ib_put * 2) + chan->dma.ib_base; - while (chan->dma.free < size) { - /* reset counter as long as GET is still advancing, this is - * to avoid misdetecting a GPU lockup if the GPU happens to - * just be processing an operation that takes a long time - */ - get_valid = READ_GET(chan, &get); + BUG_ON(chan->dma.ib_free < 1); + nouveau_bo_wr32(pb, ip++, lower_32_bits(offset)); + nouveau_bo_wr32(pb, ip++, upper_32_bits(offset) | length << 8); + + chan->dma.ib_put = (chan->dma.ib_put + 1) & chan->dma.ib_max; + nvchan_wr32(chan, 0x8c, chan->dma.ib_put); + chan->dma.ib_free--; +} + +static int +nv50_dma_push_wait(struct nouveau_channel *chan, int count) +{ + uint32_t cnt = 0, prev_get = 0; + + while (chan->dma.ib_free < count) { + uint32_t get = nvchan_rd32(chan, 0x88); if (get != prev_get) { prev_get = get; cnt = 0; @@ -168,6 +212,71 @@ nouveau_dma_wait(struct nouveau_channel *chan, int size) return -EBUSY; } + chan->dma.ib_free = get - chan->dma.ib_put; + if (chan->dma.ib_free <= 0) + chan->dma.ib_free += chan->dma.ib_max + 1; + } + + return 0; +} + +static int +nv50_dma_wait(struct nouveau_channel *chan, int slots, int count) +{ + uint32_t cnt = 0, prev_get = 0; + int ret; + + ret = nv50_dma_push_wait(chan, slots + 1); + if (unlikely(ret)) + return ret; + + while (chan->dma.free < count) { + int get = READ_GET(chan, &prev_get, &cnt); + if (unlikely(get < 0)) { + if (get == -EINVAL) + continue; + + return get; + } + + if (get <= chan->dma.cur) { + chan->dma.free = chan->dma.max - chan->dma.cur; + if (chan->dma.free >= count) + break; + + FIRE_RING(chan); + do { + get = READ_GET(chan, &prev_get, &cnt); + if (unlikely(get < 0)) { + if (get == -EINVAL) + continue; + return get; + } + } while (get == 0); + chan->dma.cur = 0; + chan->dma.put = 0; + } + + chan->dma.free = get - chan->dma.cur - 1; + } + + return 0; +} + +int +nouveau_dma_wait(struct nouveau_channel *chan, int slots, int size) +{ + uint32_t prev_get = 0, cnt = 0; + int get; + + if (chan->dma.ib_max) + return nv50_dma_wait(chan, slots, size); + + while (chan->dma.free < size) { + get = READ_GET(chan, &prev_get, &cnt); + if (unlikely(get == -EBUSY)) + return -EBUSY; + /* loop until we have a usable GET pointer. the value * we read from the GPU may be outside the main ring if * PFIFO is processing a buffer called from the main ring, @@ -177,7 +286,7 @@ nouveau_dma_wait(struct nouveau_channel *chan, int size) * from the SKIPS area, so the code below doesn't have to deal * with some fun corner cases. */ - if (!get_valid || get < NOUVEAU_DMA_SKIPS) + if (unlikely(get == -EINVAL) || get < NOUVEAU_DMA_SKIPS) continue; if (get <= chan->dma.cur) { @@ -203,6 +312,19 @@ nouveau_dma_wait(struct nouveau_channel *chan, int size) * after processing the currently pending commands. */ OUT_RING(chan, chan->pushbuf_base | 0x20000000); + + /* wait for GET to depart from the skips area. + * prevents writing GET==PUT and causing a race + * condition that causes us to think the GPU is + * idle when it's not. + */ + do { + get = READ_GET(chan, &prev_get, &cnt); + if (unlikely(get == -EBUSY)) + return -EBUSY; + if (unlikely(get == -EINVAL)) + continue; + } while (get <= NOUVEAU_DMA_SKIPS); WRITE_PUT(NOUVEAU_DMA_SKIPS); /* we're now submitting commands at the start of diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.h b/drivers/gpu/drm/nouveau/nouveau_dma.h index dabfd655f93..8b05c15866d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.h +++ b/drivers/gpu/drm/nouveau/nouveau_dma.h @@ -31,6 +31,9 @@ #define NOUVEAU_DMA_DEBUG 0 #endif +void nv50_dma_push(struct nouveau_channel *, struct nouveau_bo *, + int delta, int length); + /* * There's a hw race condition where you can't jump to your PUT offset, * to avoid this we jump to offset + SKIPS and fill the difference with @@ -96,13 +99,11 @@ enum { static __must_check inline int RING_SPACE(struct nouveau_channel *chan, int size) { - if (chan->dma.free < size) { - int ret; + int ret; - ret = nouveau_dma_wait(chan, size); - if (ret) - return ret; - } + ret = nouveau_dma_wait(chan, 1, size); + if (ret) + return ret; chan->dma.free -= size; return 0; @@ -146,7 +147,13 @@ FIRE_RING(struct nouveau_channel *chan) return; chan->accel_done = true; - WRITE_PUT(chan->dma.cur); + if (chan->dma.ib_max) { + nv50_dma_push(chan, chan->pushbuf_bo, chan->dma.put << 2, + (chan->dma.cur - chan->dma.put) << 2); + } else { + WRITE_PUT(chan->dma.cur); + } + chan->dma.put = chan->dma.cur; } diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c index 9e2926c4857..f954ad93e81 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dp.c +++ b/drivers/gpu/drm/nouveau/nouveau_dp.c @@ -490,7 +490,8 @@ nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr, if (!nv_wait(NV50_AUXCH_CTRL(index), 0x00010000, 0x00000000)) { NV_ERROR(dev, "expected bit 16 == 0, got 0x%08x\n", nv_rd32(dev, NV50_AUXCH_CTRL(index))); - return -EBUSY; + ret = -EBUSY; + goto out; } udelay(400); @@ -502,6 +503,11 @@ nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr, } if (cmd & 1) { + if ((stat & NV50_AUXCH_STAT_COUNT) != data_nr) { + ret = -EREMOTEIO; + goto out; + } + for (i = 0; i < 4; i++) { data32[i] = nv_rd32(dev, NV50_AUXCH_DATA_IN(index, i)); NV_DEBUG_KMS(dev, "rd %d: 0x%08x\n", i, data32[i]); diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c index 06eb993e088..30cc09e8a70 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.c +++ b/drivers/gpu/drm/nouveau/nouveau_drv.c @@ -56,7 +56,7 @@ int nouveau_vram_pushbuf; module_param_named(vram_pushbuf, nouveau_vram_pushbuf, int, 0400); MODULE_PARM_DESC(vram_notify, "Force DMA notifiers to be in VRAM"); -int nouveau_vram_notify; +int nouveau_vram_notify = 1; module_param_named(vram_notify, nouveau_vram_notify, int, 0400); MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (>=GeForce 8)"); @@ -71,6 +71,18 @@ MODULE_PARM_DESC(uscript_tmds, "TMDS output script table ID (>=GeForce 8)"); int nouveau_uscript_tmds = -1; module_param_named(uscript_tmds, nouveau_uscript_tmds, int, 0400); +MODULE_PARM_DESC(ignorelid, "Ignore ACPI lid status"); +int nouveau_ignorelid = 0; +module_param_named(ignorelid, nouveau_ignorelid, int, 0400); + +MODULE_PARM_DESC(noaccel, "Disable all acceleration"); +int nouveau_noaccel = 0; +module_param_named(noaccel, nouveau_noaccel, int, 0400); + +MODULE_PARM_DESC(nofbaccel, "Disable fbcon acceleration"); +int nouveau_nofbaccel = 0; +module_param_named(nofbaccel, nouveau_nofbaccel, int, 0400); + MODULE_PARM_DESC(tv_norm, "Default TV norm.\n" "\t\tSupported: PAL, PAL-M, PAL-N, PAL-Nc, NTSC-M, NTSC-J,\n" "\t\t\thd480i, hd480p, hd576i, hd576p, hd720p, hd1080i.\n" @@ -123,7 +135,7 @@ nouveau_pci_remove(struct pci_dev *pdev) drm_put_dev(dev); } -static int +int nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state) { struct drm_device *dev = pci_get_drvdata(pdev); @@ -221,7 +233,7 @@ out_abort: return ret; } -static int +int nouveau_pci_resume(struct pci_dev *pdev) { struct drm_device *dev = pci_get_drvdata(pdev); @@ -390,8 +402,10 @@ static int __init nouveau_init(void) nouveau_modeset = 1; } - if (nouveau_modeset == 1) + if (nouveau_modeset == 1) { driver.driver_features |= DRIVER_MODESET; + nouveau_register_dsm_handler(); + } return drm_init(&driver); } @@ -399,6 +413,7 @@ static int __init nouveau_init(void) static void __exit nouveau_exit(void) { drm_exit(&driver); + nouveau_unregister_dsm_handler(); } module_init(nouveau_init); diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 026419fe879..5f8d987af36 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -34,7 +34,7 @@ #define DRIVER_MAJOR 0 #define DRIVER_MINOR 0 -#define DRIVER_PATCHLEVEL 15 +#define DRIVER_PATCHLEVEL 16 #define NOUVEAU_FAMILY 0x0000FFFF #define NOUVEAU_FLAGS 0xFFFF0000 @@ -83,6 +83,7 @@ struct nouveau_bo { struct drm_file *reserved_by; struct list_head entry; int pbbo_index; + bool validate_mapped; struct nouveau_channel *channel; @@ -239,6 +240,11 @@ struct nouveau_channel { int cur; int put; /* access via pushbuf_bo */ + + int ib_base; + int ib_max; + int ib_free; + int ib_put; } dma; uint32_t sw_subchannel[8]; @@ -509,6 +515,8 @@ struct drm_nouveau_private { void __iomem *ramin; uint32_t ramin_size; + struct nouveau_bo *vga_ram; + struct workqueue_struct *wq; struct work_struct irq_work; @@ -531,6 +539,9 @@ struct drm_nouveau_private { struct nouveau_engine engine; struct nouveau_channel *channel; + /* For PFIFO and PGRAPH. */ + spinlock_t context_switch_lock; + /* RAMIN configuration, RAMFC, RAMHT and RAMRO offsets */ struct nouveau_gpuobj *ramht; uint32_t ramin_rsvd_vram; @@ -581,6 +592,7 @@ struct drm_nouveau_private { uint64_t vm_end; struct nouveau_gpuobj *vm_vram_pt[NV50_VM_VRAM_NR]; int vm_vram_pt_nr; + uint64_t vram_sys_base; /* the mtrr covering the FB */ int fb_mtrr; @@ -593,8 +605,7 @@ struct drm_nouveau_private { struct list_head gpuobj_list; - struct nvbios VBIOS; - struct nouveau_bios_info *vbios; + struct nvbios vbios; struct nv04_mode_state mode_reg; struct nv04_mode_state saved_reg; @@ -611,7 +622,6 @@ struct drm_nouveau_private { } susres; struct backlight_device *backlight; - bool acpi_dsm; struct nouveau_channel *evo; @@ -675,6 +685,12 @@ extern char *nouveau_tv_norm; extern int nouveau_reg_debug; extern char *nouveau_vbios; extern int nouveau_ctxfw; +extern int nouveau_ignorelid; +extern int nouveau_nofbaccel; +extern int nouveau_noaccel; + +extern int nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state); +extern int nouveau_pci_resume(struct pci_dev *pdev); /* nouveau_state.c */ extern void nouveau_preclose(struct drm_device *dev, struct drm_file *); @@ -690,12 +706,6 @@ extern bool nouveau_wait_until(struct drm_device *, uint64_t timeout, uint32_t reg, uint32_t mask, uint32_t val); extern bool nouveau_wait_for_idle(struct drm_device *); extern int nouveau_card_init(struct drm_device *); -extern int nouveau_ioctl_card_init(struct drm_device *, void *data, - struct drm_file *); -extern int nouveau_ioctl_suspend(struct drm_device *, void *data, - struct drm_file *); -extern int nouveau_ioctl_resume(struct drm_device *, void *data, - struct drm_file *); /* nouveau_mem.c */ extern int nouveau_mem_init_heap(struct mem_block **, uint64_t start, @@ -839,21 +849,15 @@ nouveau_debugfs_channel_fini(struct nouveau_channel *chan) /* nouveau_dma.c */ extern void nouveau_dma_pre_init(struct nouveau_channel *); extern int nouveau_dma_init(struct nouveau_channel *); -extern int nouveau_dma_wait(struct nouveau_channel *, int size); +extern int nouveau_dma_wait(struct nouveau_channel *, int slots, int size); /* nouveau_acpi.c */ -#ifdef CONFIG_ACPI -extern int nouveau_hybrid_setup(struct drm_device *dev); -extern bool nouveau_dsm_probe(struct drm_device *dev); +#if defined(CONFIG_ACPI) +void nouveau_register_dsm_handler(void); +void nouveau_unregister_dsm_handler(void); #else -static inline int nouveau_hybrid_setup(struct drm_device *dev) -{ - return 0; -} -static inline bool nouveau_dsm_probe(struct drm_device *dev) -{ - return false; -} +static inline void nouveau_register_dsm_handler(void) {} +static inline void nouveau_unregister_dsm_handler(void) {} #endif /* nouveau_backlight.c */ @@ -1021,6 +1025,7 @@ extern void nv50_graph_destroy_context(struct nouveau_channel *); extern int nv50_graph_load_context(struct nouveau_channel *); extern int nv50_graph_unload_context(struct drm_device *); extern void nv50_graph_context_switch(struct drm_device *); +extern int nv50_grctx_init(struct nouveau_grctx *); /* nouveau_grctx.c */ extern int nouveau_grctx_prog_load(struct drm_device *); @@ -1146,16 +1151,6 @@ extern int nouveau_gem_ioctl_new(struct drm_device *, void *, struct drm_file *); extern int nouveau_gem_ioctl_pushbuf(struct drm_device *, void *, struct drm_file *); -extern int nouveau_gem_ioctl_pushbuf_call(struct drm_device *, void *, - struct drm_file *); -extern int nouveau_gem_ioctl_pushbuf_call2(struct drm_device *, void *, - struct drm_file *); -extern int nouveau_gem_ioctl_pin(struct drm_device *, void *, - struct drm_file *); -extern int nouveau_gem_ioctl_unpin(struct drm_device *, void *, - struct drm_file *); -extern int nouveau_gem_ioctl_tile(struct drm_device *, void *, - struct drm_file *); extern int nouveau_gem_ioctl_cpu_prep(struct drm_device *, void *, struct drm_file *); extern int nouveau_gem_ioctl_cpu_fini(struct drm_device *, void *, diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index 0b05c869e0e..68cedd9194f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -36,6 +36,7 @@ #include <linux/fb.h> #include <linux/init.h> #include <linux/screen_info.h> +#include <linux/vga_switcheroo.h> #include "drmP.h" #include "drm.h" @@ -107,6 +108,34 @@ static struct fb_ops nouveau_fbcon_ops = { .fb_setcmap = drm_fb_helper_setcmap, }; +static struct fb_ops nv04_fbcon_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_setcolreg = drm_fb_helper_setcolreg, + .fb_fillrect = nv04_fbcon_fillrect, + .fb_copyarea = nv04_fbcon_copyarea, + .fb_imageblit = nv04_fbcon_imageblit, + .fb_sync = nouveau_fbcon_sync, + .fb_pan_display = drm_fb_helper_pan_display, + .fb_blank = drm_fb_helper_blank, + .fb_setcmap = drm_fb_helper_setcmap, +}; + +static struct fb_ops nv50_fbcon_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_setcolreg = drm_fb_helper_setcolreg, + .fb_fillrect = nv50_fbcon_fillrect, + .fb_copyarea = nv50_fbcon_copyarea, + .fb_imageblit = nv50_fbcon_imageblit, + .fb_sync = nouveau_fbcon_sync, + .fb_pan_display = drm_fb_helper_pan_display, + .fb_blank = drm_fb_helper_blank, + .fb_setcmap = drm_fb_helper_setcmap, +}; + static void nouveau_fbcon_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, u16 blue, int regno) { @@ -267,8 +296,12 @@ nouveau_fbcon_create(struct drm_device *dev, uint32_t fb_width, dev_priv->fbdev_info = info; strcpy(info->fix.id, "nouveaufb"); - info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_COPYAREA | - FBINFO_HWACCEL_FILLRECT | FBINFO_HWACCEL_IMAGEBLIT; + if (nouveau_nofbaccel) + info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_DISABLED; + else + info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_COPYAREA | + FBINFO_HWACCEL_FILLRECT | + FBINFO_HWACCEL_IMAGEBLIT; info->fbops = &nouveau_fbcon_ops; info->fix.smem_start = dev->mode_config.fb_base + nvbo->bo.offset - dev_priv->vm_vram_base; @@ -316,13 +349,15 @@ nouveau_fbcon_create(struct drm_device *dev, uint32_t fb_width, par->nouveau_fb = nouveau_fb; par->dev = dev; - if (dev_priv->channel) { + if (dev_priv->channel && !nouveau_nofbaccel) { switch (dev_priv->card_type) { case NV_50: nv50_fbcon_accel_init(info); + info->fbops = &nv50_fbcon_ops; break; default: nv04_fbcon_accel_init(info); + info->fbops = &nv04_fbcon_ops; break; }; } @@ -336,6 +371,7 @@ nouveau_fbcon_create(struct drm_device *dev, uint32_t fb_width, nvbo->bo.offset, nvbo); mutex_unlock(&dev->struct_mutex); + vga_switcheroo_client_fb_set(dev->pdev, info); return 0; out_unref: @@ -367,10 +403,8 @@ nouveau_fbcon_remove(struct drm_device *dev, struct drm_framebuffer *fb) unregister_framebuffer(info); nouveau_bo_unmap(nouveau_fb->nvbo); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(nouveau_fb->nvbo->gem); + drm_gem_object_unreference_unlocked(nouveau_fb->nvbo->gem); nouveau_fb->nvbo = NULL; - mutex_unlock(&dev->struct_mutex); if (par) drm_fb_helper_free(&par->helper); framebuffer_release(info); diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.h b/drivers/gpu/drm/nouveau/nouveau_fbcon.h index 462e0b87b4b..f9c34e1a8c1 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.h +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.h @@ -40,7 +40,13 @@ int nouveau_fbcon_remove(struct drm_device *dev, struct drm_framebuffer *fb); void nouveau_fbcon_restore(void); void nouveau_fbcon_zfill(struct drm_device *dev); +void nv04_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region); +void nv04_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect); +void nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image); int nv04_fbcon_accel_init(struct fb_info *info); +void nv50_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect); +void nv50_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region); +void nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image); int nv50_fbcon_accel_init(struct fb_info *info); void nouveau_fbcon_gpu_lockup(struct fb_info *info); diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 2009db2426c..0d22f66f1c7 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -167,12 +167,10 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data, ret = drm_gem_handle_create(file_priv, nvbo->gem, &req->info.handle); out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_handle_unreference(nvbo->gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_handle_unreference_unlocked(nvbo->gem); if (ret) - drm_gem_object_unreference(nvbo->gem); + drm_gem_object_unreference_unlocked(nvbo->gem); return ret; } @@ -243,6 +241,11 @@ validate_fini_list(struct list_head *list, struct nouveau_fence *fence) nouveau_fence_unref((void *)&prev_fence); } + if (unlikely(nvbo->validate_mapped)) { + ttm_bo_kunmap(&nvbo->kmap); + nvbo->validate_mapped = false; + } + list_del(&nvbo->entry); nvbo->reserved_by = NULL; ttm_bo_unreserve(&nvbo->bo); @@ -302,11 +305,14 @@ retry: if (ret == -EAGAIN) ret = ttm_bo_wait_unreserved(&nvbo->bo, false); drm_gem_object_unreference(gem); - if (ret) + if (ret) { + NV_ERROR(dev, "fail reserve\n"); return ret; + } goto retry; } + b->user_priv = (uint64_t)(unsigned long)nvbo; nvbo->reserved_by = file_priv; nvbo->pbbo_index = i; if ((b->valid_domains & NOUVEAU_GEM_DOMAIN_VRAM) && @@ -321,6 +327,7 @@ retry: else { NV_ERROR(dev, "invalid valid domains: 0x%08x\n", b->valid_domains); + list_add_tail(&nvbo->entry, &op->both_list); validate_fini(op, NULL); return -EINVAL; } @@ -335,8 +342,10 @@ retry: } ret = ttm_bo_wait_cpu(&nvbo->bo, false); - if (ret) + if (ret) { + NV_ERROR(dev, "fail wait_cpu\n"); return ret; + } goto retry; } } @@ -350,6 +359,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list, { struct drm_nouveau_gem_pushbuf_bo __user *upbbo = (void __force __user *)(uintptr_t)user_pbbo_ptr; + struct drm_device *dev = chan->dev; struct nouveau_bo *nvbo; int ret, relocs = 0; @@ -361,39 +371,46 @@ validate_list(struct nouveau_channel *chan, struct list_head *list, spin_lock(&nvbo->bo.lock); ret = ttm_bo_wait(&nvbo->bo, false, false, false); spin_unlock(&nvbo->bo.lock); - if (unlikely(ret)) + if (unlikely(ret)) { + NV_ERROR(dev, "fail wait other chan\n"); return ret; + } } ret = nouveau_gem_set_domain(nvbo->gem, b->read_domains, b->write_domains, b->valid_domains); - if (unlikely(ret)) + if (unlikely(ret)) { + NV_ERROR(dev, "fail set_domain\n"); return ret; + } nvbo->channel = chan; ret = ttm_bo_validate(&nvbo->bo, &nvbo->placement, false, false); nvbo->channel = NULL; - if (unlikely(ret)) + if (unlikely(ret)) { + NV_ERROR(dev, "fail ttm_validate\n"); return ret; + } - if (nvbo->bo.offset == b->presumed_offset && + if (nvbo->bo.offset == b->presumed.offset && ((nvbo->bo.mem.mem_type == TTM_PL_VRAM && - b->presumed_domain & NOUVEAU_GEM_DOMAIN_VRAM) || + b->presumed.domain & NOUVEAU_GEM_DOMAIN_VRAM) || (nvbo->bo.mem.mem_type == TTM_PL_TT && - b->presumed_domain & NOUVEAU_GEM_DOMAIN_GART))) + b->presumed.domain & NOUVEAU_GEM_DOMAIN_GART))) continue; if (nvbo->bo.mem.mem_type == TTM_PL_TT) - b->presumed_domain = NOUVEAU_GEM_DOMAIN_GART; + b->presumed.domain = NOUVEAU_GEM_DOMAIN_GART; else - b->presumed_domain = NOUVEAU_GEM_DOMAIN_VRAM; - b->presumed_offset = nvbo->bo.offset; - b->presumed_ok = 0; + b->presumed.domain = NOUVEAU_GEM_DOMAIN_VRAM; + b->presumed.offset = nvbo->bo.offset; + b->presumed.valid = 0; relocs++; - if (DRM_COPY_TO_USER(&upbbo[nvbo->pbbo_index], b, sizeof(*b))) + if (DRM_COPY_TO_USER(&upbbo[nvbo->pbbo_index].presumed, + &b->presumed, sizeof(b->presumed))) return -EFAULT; } @@ -407,6 +424,7 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan, uint64_t user_buffers, int nr_buffers, struct validate_op *op, int *apply_relocs) { + struct drm_device *dev = chan->dev; int ret, relocs = 0; INIT_LIST_HEAD(&op->vram_list); @@ -417,11 +435,14 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan, return 0; ret = validate_init(chan, file_priv, pbbo, nr_buffers, op); - if (unlikely(ret)) + if (unlikely(ret)) { + NV_ERROR(dev, "validate_init\n"); return ret; + } ret = validate_list(chan, &op->vram_list, pbbo, user_buffers); if (unlikely(ret < 0)) { + NV_ERROR(dev, "validate vram_list\n"); validate_fini(op, NULL); return ret; } @@ -429,6 +450,7 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan, ret = validate_list(chan, &op->gart_list, pbbo, user_buffers); if (unlikely(ret < 0)) { + NV_ERROR(dev, "validate gart_list\n"); validate_fini(op, NULL); return ret; } @@ -436,6 +458,7 @@ nouveau_gem_pushbuf_validate(struct nouveau_channel *chan, ret = validate_list(chan, &op->both_list, pbbo, user_buffers); if (unlikely(ret < 0)) { + NV_ERROR(dev, "validate both_list\n"); validate_fini(op, NULL); return ret; } @@ -464,58 +487,82 @@ u_memcpya(uint64_t user, unsigned nmemb, unsigned size) } static int -nouveau_gem_pushbuf_reloc_apply(struct nouveau_channel *chan, int nr_bo, - struct drm_nouveau_gem_pushbuf_bo *bo, - int nr_relocs, uint64_t ptr_relocs, - int nr_dwords, int first_dword, - uint32_t *pushbuf, bool is_iomem) +nouveau_gem_pushbuf_reloc_apply(struct drm_device *dev, + struct drm_nouveau_gem_pushbuf *req, + struct drm_nouveau_gem_pushbuf_bo *bo) { struct drm_nouveau_gem_pushbuf_reloc *reloc = NULL; - struct drm_device *dev = chan->dev; - int ret = 0, i; + int ret = 0; + unsigned i; - reloc = u_memcpya(ptr_relocs, nr_relocs, sizeof(*reloc)); + reloc = u_memcpya(req->relocs, req->nr_relocs, sizeof(*reloc)); if (IS_ERR(reloc)) return PTR_ERR(reloc); - for (i = 0; i < nr_relocs; i++) { + for (i = 0; i < req->nr_relocs; i++) { struct drm_nouveau_gem_pushbuf_reloc *r = &reloc[i]; struct drm_nouveau_gem_pushbuf_bo *b; + struct nouveau_bo *nvbo; uint32_t data; - if (r->bo_index >= nr_bo || r->reloc_index < first_dword || - r->reloc_index >= first_dword + nr_dwords) { - NV_ERROR(dev, "Bad relocation %d\n", i); - NV_ERROR(dev, " bo: %d max %d\n", r->bo_index, nr_bo); - NV_ERROR(dev, " id: %d max %d\n", r->reloc_index, nr_dwords); + if (unlikely(r->bo_index > req->nr_buffers)) { + NV_ERROR(dev, "reloc bo index invalid\n"); ret = -EINVAL; break; } b = &bo[r->bo_index]; - if (b->presumed_ok) + if (b->presumed.valid) continue; + if (unlikely(r->reloc_bo_index > req->nr_buffers)) { + NV_ERROR(dev, "reloc container bo index invalid\n"); + ret = -EINVAL; + break; + } + nvbo = (void *)(unsigned long)bo[r->reloc_bo_index].user_priv; + + if (unlikely(r->reloc_bo_offset + 4 > + nvbo->bo.mem.num_pages << PAGE_SHIFT)) { + NV_ERROR(dev, "reloc outside of bo\n"); + ret = -EINVAL; + break; + } + + if (!nvbo->kmap.virtual) { + ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages, + &nvbo->kmap); + if (ret) { + NV_ERROR(dev, "failed kmap for reloc\n"); + break; + } + nvbo->validate_mapped = true; + } + if (r->flags & NOUVEAU_GEM_RELOC_LOW) - data = b->presumed_offset + r->data; + data = b->presumed.offset + r->data; else if (r->flags & NOUVEAU_GEM_RELOC_HIGH) - data = (b->presumed_offset + r->data) >> 32; + data = (b->presumed.offset + r->data) >> 32; else data = r->data; if (r->flags & NOUVEAU_GEM_RELOC_OR) { - if (b->presumed_domain == NOUVEAU_GEM_DOMAIN_GART) + if (b->presumed.domain == NOUVEAU_GEM_DOMAIN_GART) data |= r->tor; else data |= r->vor; } - if (is_iomem) - iowrite32_native(data, (void __force __iomem *) - &pushbuf[r->reloc_index]); - else - pushbuf[r->reloc_index] = data; + spin_lock(&nvbo->bo.lock); + ret = ttm_bo_wait(&nvbo->bo, false, false, false); + spin_unlock(&nvbo->bo.lock); + if (ret) { + NV_ERROR(dev, "reloc wait_idle failed: %d\n", ret); + break; + } + + nouveau_bo_wr32(nvbo, r->reloc_bo_offset >> 2, data); } kfree(reloc); @@ -526,127 +573,50 @@ int nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data, struct drm_file *file_priv) { + struct drm_nouveau_private *dev_priv = dev->dev_private; struct drm_nouveau_gem_pushbuf *req = data; - struct drm_nouveau_gem_pushbuf_bo *bo = NULL; + struct drm_nouveau_gem_pushbuf_push *push; + struct drm_nouveau_gem_pushbuf_bo *bo; struct nouveau_channel *chan; struct validate_op op; - struct nouveau_fence* fence = 0; - uint32_t *pushbuf = NULL; - int ret = 0, do_reloc = 0, i; + struct nouveau_fence *fence = 0; + int i, j, ret = 0, do_reloc = 0; NOUVEAU_CHECK_INITIALISED_WITH_RETURN; NOUVEAU_GET_USER_CHANNEL_WITH_RETURN(req->channel, file_priv, chan); - if (req->nr_dwords >= chan->dma.max || - req->nr_buffers > NOUVEAU_GEM_MAX_BUFFERS || - req->nr_relocs > NOUVEAU_GEM_MAX_RELOCS) { - NV_ERROR(dev, "Pushbuf config exceeds limits:\n"); - NV_ERROR(dev, " dwords : %d max %d\n", req->nr_dwords, - chan->dma.max - 1); - NV_ERROR(dev, " buffers: %d max %d\n", req->nr_buffers, - NOUVEAU_GEM_MAX_BUFFERS); - NV_ERROR(dev, " relocs : %d max %d\n", req->nr_relocs, - NOUVEAU_GEM_MAX_RELOCS); - return -EINVAL; - } - - pushbuf = u_memcpya(req->dwords, req->nr_dwords, sizeof(uint32_t)); - if (IS_ERR(pushbuf)) - return PTR_ERR(pushbuf); - - bo = u_memcpya(req->buffers, req->nr_buffers, sizeof(*bo)); - if (IS_ERR(bo)) { - kfree(pushbuf); - return PTR_ERR(bo); - } - - mutex_lock(&dev->struct_mutex); - - /* Validate buffer list */ - ret = nouveau_gem_pushbuf_validate(chan, file_priv, bo, req->buffers, - req->nr_buffers, &op, &do_reloc); - if (ret) - goto out; - - /* Apply any relocations that are required */ - if (do_reloc) { - ret = nouveau_gem_pushbuf_reloc_apply(chan, req->nr_buffers, - bo, req->nr_relocs, - req->relocs, - req->nr_dwords, 0, - pushbuf, false); - if (ret) - goto out; - } - - /* Emit push buffer to the hw - */ - ret = RING_SPACE(chan, req->nr_dwords); - if (ret) - goto out; - - OUT_RINGp(chan, pushbuf, req->nr_dwords); + req->vram_available = dev_priv->fb_aper_free; + req->gart_available = dev_priv->gart_info.aper_free; + if (unlikely(req->nr_push == 0)) + goto out_next; - ret = nouveau_fence_new(chan, &fence, true); - if (ret) { - NV_ERROR(dev, "error fencing pushbuf: %d\n", ret); - WIND_RING(chan); - goto out; + if (unlikely(req->nr_push > NOUVEAU_GEM_MAX_PUSH)) { + NV_ERROR(dev, "pushbuf push count exceeds limit: %d max %d\n", + req->nr_push, NOUVEAU_GEM_MAX_PUSH); + return -EINVAL; } - if (nouveau_gem_pushbuf_sync(chan)) { - ret = nouveau_fence_wait(fence, NULL, false, false); - if (ret) { - for (i = 0; i < req->nr_dwords; i++) - NV_ERROR(dev, "0x%08x\n", pushbuf[i]); - NV_ERROR(dev, "^^ above push buffer is fail :(\n"); - } + if (unlikely(req->nr_buffers > NOUVEAU_GEM_MAX_BUFFERS)) { + NV_ERROR(dev, "pushbuf bo count exceeds limit: %d max %d\n", + req->nr_buffers, NOUVEAU_GEM_MAX_BUFFERS); + return -EINVAL; } -out: - validate_fini(&op, fence); - nouveau_fence_unref((void**)&fence); - mutex_unlock(&dev->struct_mutex); - kfree(pushbuf); - kfree(bo); - return ret; -} - -#define PUSHBUF_CAL (dev_priv->card_type >= NV_20) - -int -nouveau_gem_ioctl_pushbuf_call(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct drm_nouveau_gem_pushbuf_call *req = data; - struct drm_nouveau_gem_pushbuf_bo *bo = NULL; - struct nouveau_channel *chan; - struct drm_gem_object *gem; - struct nouveau_bo *pbbo; - struct validate_op op; - struct nouveau_fence* fence = 0; - int i, ret = 0, do_reloc = 0; - - NOUVEAU_CHECK_INITIALISED_WITH_RETURN; - NOUVEAU_GET_USER_CHANNEL_WITH_RETURN(req->channel, file_priv, chan); - - if (unlikely(req->handle == 0)) - goto out_next; - - if (req->nr_buffers > NOUVEAU_GEM_MAX_BUFFERS || - req->nr_relocs > NOUVEAU_GEM_MAX_RELOCS) { - NV_ERROR(dev, "Pushbuf config exceeds limits:\n"); - NV_ERROR(dev, " buffers: %d max %d\n", req->nr_buffers, - NOUVEAU_GEM_MAX_BUFFERS); - NV_ERROR(dev, " relocs : %d max %d\n", req->nr_relocs, - NOUVEAU_GEM_MAX_RELOCS); + if (unlikely(req->nr_relocs > NOUVEAU_GEM_MAX_RELOCS)) { + NV_ERROR(dev, "pushbuf reloc count exceeds limit: %d max %d\n", + req->nr_relocs, NOUVEAU_GEM_MAX_RELOCS); return -EINVAL; } + push = u_memcpya(req->push, req->nr_push, sizeof(*push)); + if (IS_ERR(push)) + return PTR_ERR(push); + bo = u_memcpya(req->buffers, req->nr_buffers, sizeof(*bo)); - if (IS_ERR(bo)) + if (IS_ERR(bo)) { + kfree(push); return PTR_ERR(bo); + } mutex_lock(&dev->struct_mutex); @@ -658,110 +628,84 @@ nouveau_gem_ioctl_pushbuf_call(struct drm_device *dev, void *data, goto out; } - /* Validate DMA push buffer */ - gem = drm_gem_object_lookup(dev, file_priv, req->handle); - if (!gem) { - NV_ERROR(dev, "Unknown pb handle 0x%08x\n", req->handle); - ret = -EINVAL; - goto out; - } - pbbo = nouveau_gem_object(gem); - - ret = ttm_bo_reserve(&pbbo->bo, false, false, true, - chan->fence.sequence); - if (ret) { - NV_ERROR(dev, "resv pb: %d\n", ret); - drm_gem_object_unreference(gem); - goto out; - } - - nouveau_bo_placement_set(pbbo, 1 << chan->pushbuf_bo->bo.mem.mem_type); - ret = ttm_bo_validate(&pbbo->bo, &pbbo->placement, false, false); - if (ret) { - NV_ERROR(dev, "validate pb: %d\n", ret); - ttm_bo_unreserve(&pbbo->bo); - drm_gem_object_unreference(gem); - goto out; - } - - list_add_tail(&pbbo->entry, &op.both_list); - - /* If presumed return address doesn't match, we need to map the - * push buffer and fix it.. - */ - if (!PUSHBUF_CAL) { - uint32_t retaddy; - - if (chan->dma.free < 4 + NOUVEAU_DMA_SKIPS) { - ret = nouveau_dma_wait(chan, 4 + NOUVEAU_DMA_SKIPS); - if (ret) { - NV_ERROR(dev, "jmp_space: %d\n", ret); - goto out; - } - } - - retaddy = chan->pushbuf_base + ((chan->dma.cur + 2) << 2); - retaddy |= 0x20000000; - if (retaddy != req->suffix0) { - req->suffix0 = retaddy; - do_reloc = 1; - } - } - /* Apply any relocations that are required */ if (do_reloc) { - void *pbvirt; - bool is_iomem; - ret = ttm_bo_kmap(&pbbo->bo, 0, pbbo->bo.mem.num_pages, - &pbbo->kmap); + ret = nouveau_gem_pushbuf_reloc_apply(dev, req, bo); if (ret) { - NV_ERROR(dev, "kmap pb: %d\n", ret); + NV_ERROR(dev, "reloc apply: %d\n", ret); goto out; } + } - pbvirt = ttm_kmap_obj_virtual(&pbbo->kmap, &is_iomem); - ret = nouveau_gem_pushbuf_reloc_apply(chan, req->nr_buffers, bo, - req->nr_relocs, - req->relocs, - req->nr_dwords, - req->offset / 4, - pbvirt, is_iomem); - - if (!PUSHBUF_CAL) { - nouveau_bo_wr32(pbbo, - req->offset / 4 + req->nr_dwords - 2, - req->suffix0); - } - - ttm_bo_kunmap(&pbbo->kmap); + if (chan->dma.ib_max) { + ret = nouveau_dma_wait(chan, req->nr_push + 1, 6); if (ret) { - NV_ERROR(dev, "reloc apply: %d\n", ret); + NV_INFO(dev, "nv50cal_space: %d\n", ret); goto out; } - } - if (PUSHBUF_CAL) { - ret = RING_SPACE(chan, 2); + for (i = 0; i < req->nr_push; i++) { + struct nouveau_bo *nvbo = (void *)(unsigned long) + bo[push[i].bo_index].user_priv; + + nv50_dma_push(chan, nvbo, push[i].offset, + push[i].length); + } + } else + if (dev_priv->card_type >= NV_20) { + ret = RING_SPACE(chan, req->nr_push * 2); if (ret) { NV_ERROR(dev, "cal_space: %d\n", ret); goto out; } - OUT_RING(chan, ((pbbo->bo.mem.mm_node->start << PAGE_SHIFT) + - req->offset) | 2); - OUT_RING(chan, 0); + + for (i = 0; i < req->nr_push; i++) { + struct nouveau_bo *nvbo = (void *)(unsigned long) + bo[push[i].bo_index].user_priv; + struct drm_mm_node *mem = nvbo->bo.mem.mm_node; + + OUT_RING(chan, ((mem->start << PAGE_SHIFT) + + push[i].offset) | 2); + OUT_RING(chan, 0); + } } else { - ret = RING_SPACE(chan, 2 + NOUVEAU_DMA_SKIPS); + ret = RING_SPACE(chan, req->nr_push * (2 + NOUVEAU_DMA_SKIPS)); if (ret) { NV_ERROR(dev, "jmp_space: %d\n", ret); goto out; } - OUT_RING(chan, ((pbbo->bo.mem.mm_node->start << PAGE_SHIFT) + - req->offset) | 0x20000000); - OUT_RING(chan, 0); - /* Space the jumps apart with NOPs. */ - for (i = 0; i < NOUVEAU_DMA_SKIPS; i++) + for (i = 0; i < req->nr_push; i++) { + struct nouveau_bo *nvbo = (void *)(unsigned long) + bo[push[i].bo_index].user_priv; + struct drm_mm_node *mem = nvbo->bo.mem.mm_node; + uint32_t cmd; + + cmd = chan->pushbuf_base + ((chan->dma.cur + 2) << 2); + cmd |= 0x20000000; + if (unlikely(cmd != req->suffix0)) { + if (!nvbo->kmap.virtual) { + ret = ttm_bo_kmap(&nvbo->bo, 0, + nvbo->bo.mem. + num_pages, + &nvbo->kmap); + if (ret) { + WIND_RING(chan); + goto out; + } + nvbo->validate_mapped = true; + } + + nouveau_bo_wr32(nvbo, (push[i].offset + + push[i].length - 8) / 4, cmd); + } + + OUT_RING(chan, ((mem->start << PAGE_SHIFT) + + push[i].offset) | 0x20000000); OUT_RING(chan, 0); + for (j = 0; j < NOUVEAU_DMA_SKIPS; j++) + OUT_RING(chan, 0); + } } ret = nouveau_fence_new(chan, &fence, true); @@ -776,9 +720,14 @@ out: nouveau_fence_unref((void**)&fence); mutex_unlock(&dev->struct_mutex); kfree(bo); + kfree(push); out_next: - if (PUSHBUF_CAL) { + if (chan->dma.ib_max) { + req->suffix0 = 0x00000000; + req->suffix1 = 0x00000000; + } else + if (dev_priv->card_type >= NV_20) { req->suffix0 = 0x00020000; req->suffix1 = 0x00000000; } else { @@ -790,19 +739,6 @@ out_next: return ret; } -int -nouveau_gem_ioctl_pushbuf_call2(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct drm_nouveau_gem_pushbuf_call *req = data; - - req->vram_available = dev_priv->fb_aper_free; - req->gart_available = dev_priv->gart_info.aper_free; - - return nouveau_gem_ioctl_pushbuf_call(dev, data, file_priv); -} - static inline uint32_t domain_to_ttm(struct nouveau_bo *nvbo, uint32_t domain) { @@ -817,74 +753,6 @@ domain_to_ttm(struct nouveau_bo *nvbo, uint32_t domain) } int -nouveau_gem_ioctl_pin(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct drm_nouveau_gem_pin *req = data; - struct drm_gem_object *gem; - struct nouveau_bo *nvbo; - int ret = 0; - - NOUVEAU_CHECK_INITIALISED_WITH_RETURN; - - if (drm_core_check_feature(dev, DRIVER_MODESET)) { - NV_ERROR(dev, "pin only allowed without kernel modesetting\n"); - return -EINVAL; - } - - if (!DRM_SUSER(DRM_CURPROC)) - return -EPERM; - - gem = drm_gem_object_lookup(dev, file_priv, req->handle); - if (!gem) - return -EINVAL; - nvbo = nouveau_gem_object(gem); - - ret = nouveau_bo_pin(nvbo, domain_to_ttm(nvbo, req->domain)); - if (ret) - goto out; - - req->offset = nvbo->bo.offset; - if (nvbo->bo.mem.mem_type == TTM_PL_TT) - req->domain = NOUVEAU_GEM_DOMAIN_GART; - else - req->domain = NOUVEAU_GEM_DOMAIN_VRAM; - -out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); - - return ret; -} - -int -nouveau_gem_ioctl_unpin(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - struct drm_nouveau_gem_pin *req = data; - struct drm_gem_object *gem; - int ret; - - NOUVEAU_CHECK_INITIALISED_WITH_RETURN; - - if (drm_core_check_feature(dev, DRIVER_MODESET)) - return -EINVAL; - - gem = drm_gem_object_lookup(dev, file_priv, req->handle); - if (!gem) - return -EINVAL; - - ret = nouveau_bo_unpin(nouveau_gem_object(gem)); - - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); - - return ret; -} - -int nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -911,7 +779,9 @@ nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void *data, } if (req->flags & NOUVEAU_GEM_CPU_PREP_NOBLOCK) { + spin_lock(&nvbo->bo.lock); ret = ttm_bo_wait(&nvbo->bo, false, false, no_wait); + spin_unlock(&nvbo->bo.lock); } else { ret = ttm_bo_synccpu_write_grab(&nvbo->bo, no_wait); if (ret == 0) @@ -919,9 +789,7 @@ nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void *data, } out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gem); return ret; } @@ -949,9 +817,7 @@ nouveau_gem_ioctl_cpu_fini(struct drm_device *dev, void *data, ret = 0; out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gem); return ret; } @@ -970,9 +836,7 @@ nouveau_gem_ioctl_info(struct drm_device *dev, void *data, return -EINVAL; ret = nouveau_gem_info(gem, req); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gem); return ret; } diff --git a/drivers/gpu/drm/nouveau/nouveau_grctx.c b/drivers/gpu/drm/nouveau/nouveau_grctx.c index 419f4c2b3b8..c7ebec69674 100644 --- a/drivers/gpu/drm/nouveau/nouveau_grctx.c +++ b/drivers/gpu/drm/nouveau/nouveau_grctx.c @@ -97,8 +97,8 @@ nouveau_grctx_prog_load(struct drm_device *dev) } pgraph->ctxvals = kmalloc(fw->size, GFP_KERNEL); - if (!pgraph->ctxprog) { - NV_ERROR(dev, "OOM copying ctxprog\n"); + if (!pgraph->ctxvals) { + NV_ERROR(dev, "OOM copying ctxvals\n"); release_firmware(fw); nouveau_grctx_fini(dev); return -ENOMEM; diff --git a/drivers/gpu/drm/nouveau/nouveau_hw.c b/drivers/gpu/drm/nouveau/nouveau_hw.c index dc46792a5c9..7855b35effc 100644 --- a/drivers/gpu/drm/nouveau/nouveau_hw.c +++ b/drivers/gpu/drm/nouveau/nouveau_hw.c @@ -160,7 +160,7 @@ static void setPLL_single(struct drm_device *dev, uint32_t reg, struct nouveau_pll_vals *pv) { struct drm_nouveau_private *dev_priv = dev->dev_private; - int chip_version = dev_priv->vbios->chip_version; + int chip_version = dev_priv->vbios.chip_version; uint32_t oldpll = NVReadRAMDAC(dev, 0, reg); int oldN = (oldpll >> 8) & 0xff, oldM = oldpll & 0xff; uint32_t pll = (oldpll & 0xfff80000) | pv->log2P << 16 | pv->NM1; @@ -216,7 +216,7 @@ setPLL_double_highregs(struct drm_device *dev, uint32_t reg1, struct nouveau_pll_vals *pv) { struct drm_nouveau_private *dev_priv = dev->dev_private; - int chip_version = dev_priv->vbios->chip_version; + int chip_version = dev_priv->vbios.chip_version; bool nv3035 = chip_version == 0x30 || chip_version == 0x35; uint32_t reg2 = reg1 + ((reg1 == NV_RAMDAC_VPLL2) ? 0x5c : 0x70); uint32_t oldpll1 = NVReadRAMDAC(dev, 0, reg1); @@ -374,7 +374,7 @@ nouveau_hw_setpll(struct drm_device *dev, uint32_t reg1, struct nouveau_pll_vals *pv) { struct drm_nouveau_private *dev_priv = dev->dev_private; - int cv = dev_priv->vbios->chip_version; + int cv = dev_priv->vbios.chip_version; if (cv == 0x30 || cv == 0x31 || cv == 0x35 || cv == 0x36 || cv >= 0x40) { diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c b/drivers/gpu/drm/nouveau/nouveau_i2c.c index 70e994d2812..88583e7bf65 100644 --- a/drivers/gpu/drm/nouveau/nouveau_i2c.c +++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c @@ -254,16 +254,16 @@ struct nouveau_i2c_chan * nouveau_i2c_find(struct drm_device *dev, int index) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; - if (index > DCB_MAX_NUM_I2C_ENTRIES) + if (index >= DCB_MAX_NUM_I2C_ENTRIES) return NULL; - if (!bios->bdcb.dcb.i2c[index].chan) { - if (nouveau_i2c_init(dev, &bios->bdcb.dcb.i2c[index], index)) + if (!bios->dcb.i2c[index].chan) { + if (nouveau_i2c_init(dev, &bios->dcb.i2c[index], index)) return NULL; } - return bios->bdcb.dcb.i2c[index].chan; + return bios->dcb.i2c[index].chan; } diff --git a/drivers/gpu/drm/nouveau/nouveau_irq.c b/drivers/gpu/drm/nouveau/nouveau_irq.c index 919a619ca7f..95220ddebb4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_irq.c +++ b/drivers/gpu/drm/nouveau/nouveau_irq.c @@ -211,6 +211,20 @@ nouveau_fifo_irq_handler(struct drm_device *dev) get + 4); } + if (status & NV_PFIFO_INTR_SEMAPHORE) { + uint32_t sem; + + status &= ~NV_PFIFO_INTR_SEMAPHORE; + nv_wr32(dev, NV03_PFIFO_INTR_0, + NV_PFIFO_INTR_SEMAPHORE); + + sem = nv_rd32(dev, NV10_PFIFO_CACHE1_SEMAPHORE); + nv_wr32(dev, NV10_PFIFO_CACHE1_SEMAPHORE, sem | 0x1); + + nv_wr32(dev, NV03_PFIFO_CACHE1_GET, get + 4); + nv_wr32(dev, NV04_PFIFO_CACHE1_PULL0, 1); + } + if (status) { NV_INFO(dev, "PFIFO_INTR 0x%08x - Ch %d\n", status, chid); @@ -483,6 +497,13 @@ nouveau_pgraph_intr_error(struct drm_device *dev, uint32_t nsource) if (nsource & NV03_PGRAPH_NSOURCE_ILLEGAL_MTHD) { if (nouveau_pgraph_intr_swmthd(dev, &trap)) unhandled = 1; + } else if (nsource & NV03_PGRAPH_NSOURCE_DMA_VTX_PROTECTION) { + uint32_t v = nv_rd32(dev, 0x402000); + nv_wr32(dev, 0x402000, v); + + /* dump the error anyway for now: it's useful for + Gallium development */ + unhandled = 1; } else { unhandled = 1; } @@ -559,86 +580,99 @@ nouveau_pgraph_irq_handler(struct drm_device *dev) static void nv50_pgraph_irq_handler(struct drm_device *dev) { - uint32_t status, nsource; + uint32_t status; - status = nv_rd32(dev, NV03_PGRAPH_INTR); - nsource = nv_rd32(dev, NV03_PGRAPH_NSOURCE); + while ((status = nv_rd32(dev, NV03_PGRAPH_INTR))) { + uint32_t nsource = nv_rd32(dev, NV03_PGRAPH_NSOURCE); - if (status & 0x00000001) { - nouveau_pgraph_intr_notify(dev, nsource); - status &= ~0x00000001; - nv_wr32(dev, NV03_PGRAPH_INTR, 0x00000001); - } + if (status & 0x00000001) { + nouveau_pgraph_intr_notify(dev, nsource); + status &= ~0x00000001; + nv_wr32(dev, NV03_PGRAPH_INTR, 0x00000001); + } - if (status & 0x00000010) { - nouveau_pgraph_intr_error(dev, nsource | - NV03_PGRAPH_NSOURCE_ILLEGAL_MTHD); + if (status & 0x00000010) { + nouveau_pgraph_intr_error(dev, nsource | + NV03_PGRAPH_NSOURCE_ILLEGAL_MTHD); - status &= ~0x00000010; - nv_wr32(dev, NV03_PGRAPH_INTR, 0x00000010); - } + status &= ~0x00000010; + nv_wr32(dev, NV03_PGRAPH_INTR, 0x00000010); + } - if (status & 0x00001000) { - nv_wr32(dev, 0x400500, 0x00000000); - nv_wr32(dev, NV03_PGRAPH_INTR, NV_PGRAPH_INTR_CONTEXT_SWITCH); - nv_wr32(dev, NV40_PGRAPH_INTR_EN, nv_rd32(dev, - NV40_PGRAPH_INTR_EN) & ~NV_PGRAPH_INTR_CONTEXT_SWITCH); - nv_wr32(dev, 0x400500, 0x00010001); + if (status & 0x00001000) { + nv_wr32(dev, 0x400500, 0x00000000); + nv_wr32(dev, NV03_PGRAPH_INTR, + NV_PGRAPH_INTR_CONTEXT_SWITCH); + nv_wr32(dev, NV40_PGRAPH_INTR_EN, nv_rd32(dev, + NV40_PGRAPH_INTR_EN) & + ~NV_PGRAPH_INTR_CONTEXT_SWITCH); + nv_wr32(dev, 0x400500, 0x00010001); - nv50_graph_context_switch(dev); + nv50_graph_context_switch(dev); - status &= ~NV_PGRAPH_INTR_CONTEXT_SWITCH; - } + status &= ~NV_PGRAPH_INTR_CONTEXT_SWITCH; + } - if (status & 0x00100000) { - nouveau_pgraph_intr_error(dev, nsource | - NV03_PGRAPH_NSOURCE_DATA_ERROR); + if (status & 0x00100000) { + nouveau_pgraph_intr_error(dev, nsource | + NV03_PGRAPH_NSOURCE_DATA_ERROR); - status &= ~0x00100000; - nv_wr32(dev, NV03_PGRAPH_INTR, 0x00100000); - } + status &= ~0x00100000; + nv_wr32(dev, NV03_PGRAPH_INTR, 0x00100000); + } - if (status & 0x00200000) { - int r; - - nouveau_pgraph_intr_error(dev, nsource | - NV03_PGRAPH_NSOURCE_PROTECTION_ERROR); - - NV_ERROR(dev, "magic set 1:\n"); - for (r = 0x408900; r <= 0x408910; r += 4) - NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, nv_rd32(dev, r)); - nv_wr32(dev, 0x408900, nv_rd32(dev, 0x408904) | 0xc0000000); - for (r = 0x408e08; r <= 0x408e24; r += 4) - NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, nv_rd32(dev, r)); - nv_wr32(dev, 0x408e08, nv_rd32(dev, 0x408e08) | 0xc0000000); - - NV_ERROR(dev, "magic set 2:\n"); - for (r = 0x409900; r <= 0x409910; r += 4) - NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, nv_rd32(dev, r)); - nv_wr32(dev, 0x409900, nv_rd32(dev, 0x409904) | 0xc0000000); - for (r = 0x409e08; r <= 0x409e24; r += 4) - NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, nv_rd32(dev, r)); - nv_wr32(dev, 0x409e08, nv_rd32(dev, 0x409e08) | 0xc0000000); - - status &= ~0x00200000; - nv_wr32(dev, NV03_PGRAPH_NSOURCE, nsource); - nv_wr32(dev, NV03_PGRAPH_INTR, 0x00200000); - } + if (status & 0x00200000) { + int r; + + nouveau_pgraph_intr_error(dev, nsource | + NV03_PGRAPH_NSOURCE_PROTECTION_ERROR); + + NV_ERROR(dev, "magic set 1:\n"); + for (r = 0x408900; r <= 0x408910; r += 4) + NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, + nv_rd32(dev, r)); + nv_wr32(dev, 0x408900, + nv_rd32(dev, 0x408904) | 0xc0000000); + for (r = 0x408e08; r <= 0x408e24; r += 4) + NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, + nv_rd32(dev, r)); + nv_wr32(dev, 0x408e08, + nv_rd32(dev, 0x408e08) | 0xc0000000); + + NV_ERROR(dev, "magic set 2:\n"); + for (r = 0x409900; r <= 0x409910; r += 4) + NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, + nv_rd32(dev, r)); + nv_wr32(dev, 0x409900, + nv_rd32(dev, 0x409904) | 0xc0000000); + for (r = 0x409e08; r <= 0x409e24; r += 4) + NV_ERROR(dev, "\t0x%08x: 0x%08x\n", r, + nv_rd32(dev, r)); + nv_wr32(dev, 0x409e08, + nv_rd32(dev, 0x409e08) | 0xc0000000); + + status &= ~0x00200000; + nv_wr32(dev, NV03_PGRAPH_NSOURCE, nsource); + nv_wr32(dev, NV03_PGRAPH_INTR, 0x00200000); + } - if (status) { - NV_INFO(dev, "Unhandled PGRAPH_INTR - 0x%08x\n", status); - nv_wr32(dev, NV03_PGRAPH_INTR, status); - } + if (status) { + NV_INFO(dev, "Unhandled PGRAPH_INTR - 0x%08x\n", + status); + nv_wr32(dev, NV03_PGRAPH_INTR, status); + } - { - const int isb = (1 << 16) | (1 << 0); + { + const int isb = (1 << 16) | (1 << 0); - if ((nv_rd32(dev, 0x400500) & isb) != isb) - nv_wr32(dev, 0x400500, nv_rd32(dev, 0x400500) | isb); - nv_wr32(dev, 0x400824, nv_rd32(dev, 0x400824) & ~(1 << 31)); + if ((nv_rd32(dev, 0x400500) & isb) != isb) + nv_wr32(dev, 0x400500, + nv_rd32(dev, 0x400500) | isb); + } } nv_wr32(dev, NV03_PMC_INTR_0, NV_PMC_INTR_0_PGRAPH_PENDING); + nv_wr32(dev, 0x400824, nv_rd32(dev, 0x400824) & ~(1 << 31)); } static void @@ -657,11 +691,14 @@ nouveau_irq_handler(DRM_IRQ_ARGS) struct drm_device *dev = (struct drm_device *)arg; struct drm_nouveau_private *dev_priv = dev->dev_private; uint32_t status, fbdev_flags = 0; + unsigned long flags; status = nv_rd32(dev, NV03_PMC_INTR_0); if (!status) return IRQ_NONE; + spin_lock_irqsave(&dev_priv->context_switch_lock, flags); + if (dev_priv->fbdev_info) { fbdev_flags = dev_priv->fbdev_info->flags; dev_priv->fbdev_info->flags |= FBINFO_HWACCEL_DISABLED; @@ -699,5 +736,7 @@ nouveau_irq_handler(DRM_IRQ_ARGS) if (dev_priv->fbdev_info) dev_priv->fbdev_info->flags = fbdev_flags; + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); + return IRQ_HANDLED; } diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index fb9bdd6edf1..2dc09dbd817 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -285,53 +285,50 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, uint32_t flags, uint64_t phys) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpuobj **pgt; - unsigned psz, pfl, pages; - - if (virt >= dev_priv->vm_gart_base && - (virt + size) < (dev_priv->vm_gart_base + dev_priv->vm_gart_size)) { - psz = 12; - pgt = &dev_priv->gart_info.sg_ctxdma; - pfl = 0x21; - virt -= dev_priv->vm_gart_base; - } else - if (virt >= dev_priv->vm_vram_base && - (virt + size) < (dev_priv->vm_vram_base + dev_priv->vm_vram_size)) { - psz = 16; - pgt = dev_priv->vm_vram_pt; - pfl = 0x01; - virt -= dev_priv->vm_vram_base; - } else { - NV_ERROR(dev, "Invalid address: 0x%16llx-0x%16llx\n", - virt, virt + size - 1); - return -EINVAL; - } + struct nouveau_gpuobj *pgt; + unsigned block; + int i; - pages = size >> psz; + virt = ((virt - dev_priv->vm_vram_base) >> 16) << 1; + size = (size >> 16) << 1; + + phys |= ((uint64_t)flags << 32); + phys |= 1; + if (dev_priv->vram_sys_base) { + phys += dev_priv->vram_sys_base; + phys |= 0x30; + } dev_priv->engine.instmem.prepare_access(dev, true); - if (flags & 0x80000000) { - while (pages--) { - struct nouveau_gpuobj *pt = pgt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> psz) << 1; + while (size) { + unsigned offset_h = upper_32_bits(phys); + unsigned offset_l = lower_32_bits(phys); + unsigned pte, end; + + for (i = 7; i >= 0; i--) { + block = 1 << (i + 1); + if (size >= block && !(virt & (block - 1))) + break; + } + offset_l |= (i << 7); - nv_wo32(dev, pt, pte++, 0x00000000); - nv_wo32(dev, pt, pte++, 0x00000000); + phys += block << 15; + size -= block; - virt += (1 << psz); - } - } else { - while (pages--) { - struct nouveau_gpuobj *pt = pgt[virt >> 29]; - unsigned pte = ((virt & 0x1fffffffULL) >> psz) << 1; - unsigned offset_h = upper_32_bits(phys) & 0xff; - unsigned offset_l = lower_32_bits(phys); + while (block) { + pgt = dev_priv->vm_vram_pt[virt >> 14]; + pte = virt & 0x3ffe; - nv_wo32(dev, pt, pte++, offset_l | pfl); - nv_wo32(dev, pt, pte++, offset_h | flags); + end = pte + block; + if (end > 16384) + end = 16384; + block -= (end - pte); + virt += (end - pte); - phys += (1 << psz); - virt += (1 << psz); + while (pte < end) { + nv_wo32(dev, pgt, pte++, offset_l); + nv_wo32(dev, pgt, pte++, offset_h); + } } } dev_priv->engine.instmem.finish_access(dev); @@ -356,7 +353,41 @@ nv50_mem_vm_bind_linear(struct drm_device *dev, uint64_t virt, uint32_t size, void nv50_mem_vm_unbind(struct drm_device *dev, uint64_t virt, uint32_t size) { - nv50_mem_vm_bind_linear(dev, virt, size, 0x80000000, 0); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpuobj *pgt; + unsigned pages, pte, end; + + virt -= dev_priv->vm_vram_base; + pages = (size >> 16) << 1; + + dev_priv->engine.instmem.prepare_access(dev, true); + while (pages) { + pgt = dev_priv->vm_vram_pt[virt >> 29]; + pte = (virt & 0x1ffe0000ULL) >> 15; + + end = pte + pages; + if (end > 16384) + end = 16384; + pages -= (end - pte); + virt += (end - pte) << 15; + + while (pte < end) + nv_wo32(dev, pgt, pte++, 0); + } + dev_priv->engine.instmem.finish_access(dev); + + nv_wr32(dev, 0x100c80, 0x00050001); + if (!nv_wait(0x100c80, 0x00000001, 0x00000000)) { + NV_ERROR(dev, "timeout: (0x100c80 & 1) == 0 (2)\n"); + NV_ERROR(dev, "0x100c80 = 0x%08x\n", nv_rd32(dev, 0x100c80)); + return; + } + + nv_wr32(dev, 0x100c80, 0x00000001); + if (!nv_wait(0x100c80, 0x00000001, 0x00000000)) { + NV_ERROR(dev, "timeout: (0x100c80 & 1) == 0 (2)\n"); + NV_ERROR(dev, "0x100c80 = 0x%08x\n", nv_rd32(dev, 0x100c80)); + } } /* @@ -383,9 +414,8 @@ void nouveau_mem_close(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - if (dev_priv->ttm.bdev.man[TTM_PL_PRIV0].has_type) - ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_PRIV0); - ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM); + nouveau_bo_unpin(dev_priv->vga_ram); + nouveau_bo_ref(NULL, &dev_priv->vga_ram); ttm_bo_device_release(&dev_priv->ttm.bdev); @@ -622,6 +652,15 @@ nouveau_mem_init(struct drm_device *dev) return ret; } + ret = nouveau_bo_new(dev, NULL, 256*1024, 0, TTM_PL_FLAG_VRAM, + 0, 0, true, true, &dev_priv->vga_ram); + if (ret == 0) + ret = nouveau_bo_pin(dev_priv->vga_ram, TTM_PL_FLAG_VRAM); + if (ret) { + NV_WARN(dev, "failed to reserve VGA memory\n"); + nouveau_bo_ref(NULL, &dev_priv->vga_ram); + } + /* GART */ #if !defined(__powerpc__) && !defined(__ia64__) if (drm_device_is_agp(dev) && dev->agp) { @@ -653,6 +692,7 @@ nouveau_mem_init(struct drm_device *dev) dev_priv->fb_mtrr = drm_mtrr_add(drm_get_resource_start(dev, 1), drm_get_resource_len(dev, 1), DRM_MTRR_WC); + return 0; } diff --git a/drivers/gpu/drm/nouveau/nouveau_notifier.c b/drivers/gpu/drm/nouveau/nouveau_notifier.c index 6c66a34b634..9537f3e3011 100644 --- a/drivers/gpu/drm/nouveau/nouveau_notifier.c +++ b/drivers/gpu/drm/nouveau/nouveau_notifier.c @@ -34,15 +34,20 @@ nouveau_notifier_init_channel(struct nouveau_channel *chan) { struct drm_device *dev = chan->dev; struct nouveau_bo *ntfy = NULL; + uint32_t flags; int ret; - ret = nouveau_gem_new(dev, NULL, PAGE_SIZE, 0, nouveau_vram_notify ? - TTM_PL_FLAG_VRAM : TTM_PL_FLAG_TT, + if (nouveau_vram_notify) + flags = TTM_PL_FLAG_VRAM; + else + flags = TTM_PL_FLAG_TT; + + ret = nouveau_gem_new(dev, NULL, PAGE_SIZE, 0, flags, 0, 0x0000, false, true, &ntfy); if (ret) return ret; - ret = nouveau_bo_pin(ntfy, TTM_PL_FLAG_VRAM); + ret = nouveau_bo_pin(ntfy, flags); if (ret) goto out_err; @@ -56,11 +61,8 @@ nouveau_notifier_init_channel(struct nouveau_channel *chan) chan->notifier_bo = ntfy; out_err: - if (ret) { - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(ntfy->gem); - mutex_unlock(&dev->struct_mutex); - } + if (ret) + drm_gem_object_unreference_unlocked(ntfy->gem); return ret; } @@ -76,8 +78,8 @@ nouveau_notifier_takedown_channel(struct nouveau_channel *chan) nouveau_bo_unmap(chan->notifier_bo); mutex_lock(&dev->struct_mutex); nouveau_bo_unpin(chan->notifier_bo); - drm_gem_object_unreference(chan->notifier_bo->gem); mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(chan->notifier_bo->gem); nouveau_mem_takedown(&chan->notifier_heap); } @@ -128,6 +130,8 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle, target = NV_DMA_TARGET_PCI; } else { target = NV_DMA_TARGET_AGP; + if (dev_priv->card_type >= NV_50) + offset += dev_priv->vm_gart_base; } } else { NV_ERROR(dev, "Bad DMA target, mem_type %d!\n", diff --git a/drivers/gpu/drm/nouveau/nouveau_object.c b/drivers/gpu/drm/nouveau/nouveau_object.c index 6c2cf81716d..e7c100ba63a 100644 --- a/drivers/gpu/drm/nouveau/nouveau_object.c +++ b/drivers/gpu/drm/nouveau/nouveau_object.c @@ -885,11 +885,12 @@ int nouveau_gpuobj_sw_new(struct nouveau_channel *chan, int class, struct nouveau_gpuobj **gpuobj_ret) { - struct drm_nouveau_private *dev_priv = chan->dev->dev_private; + struct drm_nouveau_private *dev_priv; struct nouveau_gpuobj *gpuobj; if (!chan || !gpuobj_ret || *gpuobj_ret != NULL) return -EINVAL; + dev_priv = chan->dev->dev_private; gpuobj = kzalloc(sizeof(*gpuobj), GFP_KERNEL); if (!gpuobj) diff --git a/drivers/gpu/drm/nouveau/nouveau_reg.h b/drivers/gpu/drm/nouveau/nouveau_reg.h index 251f1b3b38b..aa9b310e41b 100644 --- a/drivers/gpu/drm/nouveau/nouveau_reg.h +++ b/drivers/gpu/drm/nouveau/nouveau_reg.h @@ -99,6 +99,7 @@ * the card will hang early on in the X init process. */ # define NV_PMC_ENABLE_UNK13 (1<<13) +#define NV40_PMC_GRAPH_UNITS 0x00001540 #define NV40_PMC_BACKLIGHT 0x000015f0 # define NV40_PMC_BACKLIGHT_MASK 0x001f0000 #define NV40_PMC_1700 0x00001700 diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c index 4c7f1e403e8..ed1590577b6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c +++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c @@ -54,11 +54,12 @@ static void nouveau_sgdma_clear(struct ttm_backend *be) { struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; - struct drm_device *dev = nvbe->dev; - - NV_DEBUG(nvbe->dev, "\n"); + struct drm_device *dev; if (nvbe && nvbe->pages) { + dev = nvbe->dev; + NV_DEBUG(dev, "\n"); + if (nvbe->bound) be->func->unbind(be); diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c index 09b9a46dfc0..eb8f084d5f5 100644 --- a/drivers/gpu/drm/nouveau/nouveau_state.c +++ b/drivers/gpu/drm/nouveau/nouveau_state.c @@ -29,6 +29,7 @@ #include "drm_sarea.h" #include "drm_crtc_helper.h" #include <linux/vgaarb.h> +#include <linux/vga_switcheroo.h> #include "nouveau_drv.h" #include "nouveau_drm.h" @@ -310,6 +311,14 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) static unsigned int nouveau_vga_set_decode(void *priv, bool state) { + struct drm_device *dev = priv; + struct drm_nouveau_private *dev_priv = dev->dev_private; + + if (dev_priv->chipset >= 0x40) + nv_wr32(dev, 0x88054, state); + else + nv_wr32(dev, 0x1854, state); + if (state) return VGA_RSRC_LEGACY_IO | VGA_RSRC_LEGACY_MEM | VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM; @@ -363,6 +372,30 @@ out_err: return ret; } +static void nouveau_switcheroo_set_state(struct pci_dev *pdev, + enum vga_switcheroo_state state) +{ + pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; + if (state == VGA_SWITCHEROO_ON) { + printk(KERN_ERR "VGA switcheroo: switched nouveau on\n"); + nouveau_pci_resume(pdev); + } else { + printk(KERN_ERR "VGA switcheroo: switched nouveau off\n"); + nouveau_pci_suspend(pdev, pmm); + } +} + +static bool nouveau_switcheroo_can_switch(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + bool can_switch; + + spin_lock(&dev->count_lock); + can_switch = (dev->open_count == 0); + spin_unlock(&dev->count_lock); + return can_switch; +} + int nouveau_card_init(struct drm_device *dev) { @@ -376,6 +409,8 @@ nouveau_card_init(struct drm_device *dev) return 0; vga_client_register(dev->pdev, dev, NULL, nouveau_vga_set_decode); + vga_switcheroo_register_client(dev->pdev, nouveau_switcheroo_set_state, + nouveau_switcheroo_can_switch); /* Initialise internal driver API hooks */ ret = nouveau_init_engine_ptrs(dev); @@ -383,6 +418,7 @@ nouveau_card_init(struct drm_device *dev) goto out; engine = &dev_priv->engine; dev_priv->init_state = NOUVEAU_CARD_INIT_FAILED; + spin_lock_init(&dev_priv->context_switch_lock); /* Parse BIOS tables / Run init tables if card not POSTed */ if (drm_core_check_feature(dev, DRIVER_MODESET)) { @@ -427,15 +463,19 @@ nouveau_card_init(struct drm_device *dev) if (ret) goto out_timer; - /* PGRAPH */ - ret = engine->graph.init(dev); - if (ret) - goto out_fb; + if (nouveau_noaccel) + engine->graph.accel_blocked = true; + else { + /* PGRAPH */ + ret = engine->graph.init(dev); + if (ret) + goto out_fb; - /* PFIFO */ - ret = engine->fifo.init(dev); - if (ret) - goto out_graph; + /* PFIFO */ + ret = engine->fifo.init(dev); + if (ret) + goto out_graph; + } /* this call irq_preinstall, register irq handler and * call irq_postinstall @@ -479,9 +519,11 @@ nouveau_card_init(struct drm_device *dev) out_irq: drm_irq_uninstall(dev); out_fifo: - engine->fifo.takedown(dev); + if (!nouveau_noaccel) + engine->fifo.takedown(dev); out_graph: - engine->graph.takedown(dev); + if (!nouveau_noaccel) + engine->graph.takedown(dev); out_fb: engine->fb.takedown(dev); out_timer: @@ -518,13 +560,16 @@ static void nouveau_card_takedown(struct drm_device *dev) dev_priv->channel = NULL; } - engine->fifo.takedown(dev); - engine->graph.takedown(dev); + if (!nouveau_noaccel) { + engine->fifo.takedown(dev); + engine->graph.takedown(dev); + } engine->fb.takedown(dev); engine->timer.takedown(dev); engine->mc.takedown(dev); mutex_lock(&dev->struct_mutex); + ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM); ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_TT); mutex_unlock(&dev->struct_mutex); nouveau_sgdma_takedown(dev); @@ -600,11 +645,6 @@ int nouveau_load(struct drm_device *dev, unsigned long flags) NV_DEBUG(dev, "vendor: 0x%X device: 0x%X class: 0x%X\n", dev->pci_vendor, dev->pci_device, dev->pdev->class); - dev_priv->acpi_dsm = nouveau_dsm_probe(dev); - - if (dev_priv->acpi_dsm) - nouveau_hybrid_setup(dev); - dev_priv->wq = create_workqueue("nouveau"); if (!dev_priv->wq) return -EINVAL; @@ -759,13 +799,6 @@ int nouveau_unload(struct drm_device *dev) return 0; } -int -nouveau_ioctl_card_init(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - return nouveau_card_init(dev); -} - int nouveau_ioctl_getparam(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -816,6 +849,15 @@ int nouveau_ioctl_getparam(struct drm_device *dev, void *data, case NOUVEAU_GETPARAM_VM_VRAM_BASE: getparam->value = dev_priv->vm_vram_base; break; + case NOUVEAU_GETPARAM_GRAPH_UNITS: + /* NV40 and NV50 versions are quite different, but register + * address is the same. User is supposed to know the card + * family anyway... */ + if (dev_priv->chipset >= 0x40) { + getparam->value = nv_rd32(dev, NV40_PMC_GRAPH_UNITS); + break; + } + /* FALLTHRU */ default: NV_ERROR(dev, "unknown parameter %lld\n", getparam->param); return -EINVAL; diff --git a/drivers/gpu/drm/nouveau/nv04_crtc.c b/drivers/gpu/drm/nouveau/nv04_crtc.c index d2f143ed97c..a1d1ebb073d 100644 --- a/drivers/gpu/drm/nouveau/nv04_crtc.c +++ b/drivers/gpu/drm/nouveau/nv04_crtc.c @@ -926,9 +926,7 @@ nv04_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv, nv_crtc->cursor.set_offset(nv_crtc, nv_crtc->cursor.offset); nv_crtc->cursor.show(nv_crtc, true); out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gem); return ret; } diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/nv04_dac.c index d0e038d2894..1cb19e3acb5 100644 --- a/drivers/gpu/drm/nouveau/nv04_dac.c +++ b/drivers/gpu/drm/nouveau/nv04_dac.c @@ -119,7 +119,7 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector) { struct drm_device *dev = encoder->dev; - uint8_t saved_seq1, saved_pi, saved_rpc1; + uint8_t saved_seq1, saved_pi, saved_rpc1, saved_cr_mode; uint8_t saved_palette0[3], saved_palette_mask; uint32_t saved_rtest_ctrl, saved_rgen_ctrl; int i; @@ -135,6 +135,9 @@ static enum drm_connector_status nv04_dac_detect(struct drm_encoder *encoder, /* only implemented for head A for now */ NVSetOwner(dev, 0); + saved_cr_mode = NVReadVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX); + NVWriteVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX, saved_cr_mode | 0x80); + saved_seq1 = NVReadVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX); NVWriteVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX, saved_seq1 & ~0x20); @@ -203,6 +206,7 @@ out: NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_PIXEL_INDEX, saved_pi); NVWriteVgaCrtc(dev, 0, NV_CIO_CRE_RPC1_INDEX, saved_rpc1); NVWriteVgaSeq(dev, 0, NV_VIO_SR_CLOCK_INDEX, saved_seq1); + NVWriteVgaCrtc(dev, 0, NV_CIO_CR_MODE_INDEX, saved_cr_mode); if (blue == 0x18) { NV_INFO(dev, "Load detected on head A\n"); @@ -226,13 +230,13 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder) if (dcb->type == OUTPUT_TV) { testval = RGB_TEST_DATA(0xa0, 0xa0, 0xa0); - if (dev_priv->vbios->tvdactestval) - testval = dev_priv->vbios->tvdactestval; + if (dev_priv->vbios.tvdactestval) + testval = dev_priv->vbios.tvdactestval; } else { testval = RGB_TEST_DATA(0x140, 0x140, 0x140); /* 0x94050140 */ - if (dev_priv->vbios->dactestval) - testval = dev_priv->vbios->dactestval; + if (dev_priv->vbios.dactestval) + testval = dev_priv->vbios.dactestval; } saved_rtest_ctrl = NVReadRAMDAC(dev, 0, NV_PRAMDAC_TEST_CONTROL + regoffset); diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/nv04_dfp.c index 483f875bdb6..41634d4752f 100644 --- a/drivers/gpu/drm/nouveau/nv04_dfp.c +++ b/drivers/gpu/drm/nouveau/nv04_dfp.c @@ -269,10 +269,10 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder, regp->fp_horiz_regs[FP_TOTAL] = output_mode->htotal - 1; if (!nv_gf4_disp_arch(dev) || (output_mode->hsync_start - output_mode->hdisplay) >= - dev_priv->vbios->digital_min_front_porch) + dev_priv->vbios.digital_min_front_porch) regp->fp_horiz_regs[FP_CRTC] = output_mode->hdisplay; else - regp->fp_horiz_regs[FP_CRTC] = output_mode->hsync_start - dev_priv->vbios->digital_min_front_porch - 1; + regp->fp_horiz_regs[FP_CRTC] = output_mode->hsync_start - dev_priv->vbios.digital_min_front_porch - 1; regp->fp_horiz_regs[FP_SYNC_START] = output_mode->hsync_start - 1; regp->fp_horiz_regs[FP_SYNC_END] = output_mode->hsync_end - 1; regp->fp_horiz_regs[FP_VALID_START] = output_mode->hskew; diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c index ef77215fa5b..c7898b4f6df 100644 --- a/drivers/gpu/drm/nouveau/nv04_display.c +++ b/drivers/gpu/drm/nouveau/nv04_display.c @@ -93,10 +93,9 @@ int nv04_display_create(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct parsed_dcb *dcb = dev_priv->vbios->dcb; + struct dcb_table *dcb = &dev_priv->vbios.dcb; struct drm_encoder *encoder; struct drm_crtc *crtc; - uint16_t connector[16] = { 0 }; int i, ret; NV_DEBUG_KMS(dev, "\n"); @@ -154,52 +153,10 @@ nv04_display_create(struct drm_device *dev) if (ret) continue; - - connector[dcbent->connector] |= (1 << dcbent->type); } - for (i = 0; i < dcb->entries; i++) { - struct dcb_entry *dcbent = &dcb->entry[i]; - uint16_t encoders; - int type; - - encoders = connector[dcbent->connector]; - if (!(encoders & (1 << dcbent->type))) - continue; - connector[dcbent->connector] = 0; - - switch (dcbent->type) { - case OUTPUT_ANALOG: - if (!MULTIPLE_ENCODERS(encoders)) - type = DRM_MODE_CONNECTOR_VGA; - else - type = DRM_MODE_CONNECTOR_DVII; - break; - case OUTPUT_TMDS: - if (!MULTIPLE_ENCODERS(encoders)) - type = DRM_MODE_CONNECTOR_DVID; - else - type = DRM_MODE_CONNECTOR_DVII; - break; - case OUTPUT_LVDS: - type = DRM_MODE_CONNECTOR_LVDS; -#if 0 - /* don't create i2c adapter when lvds ddc not allowed */ - if (dcbent->lvdsconf.use_straps_for_mode || - dev_priv->vbios->fp_no_ddc) - i2c_index = 0xf; -#endif - break; - case OUTPUT_TV: - type = DRM_MODE_CONNECTOR_TV; - break; - default: - type = DRM_MODE_CONNECTOR_Unknown; - continue; - } - - nouveau_connector_create(dev, dcbent->connector, type); - } + for (i = 0; i < dcb->connector.entries; i++) + nouveau_connector_create(dev, &dcb->connector.entry[i]); /* Save previous state */ NVLockVgaCrtcs(dev, false); diff --git a/drivers/gpu/drm/nouveau/nv04_fbcon.c b/drivers/gpu/drm/nouveau/nv04_fbcon.c index d910873c136..3da90c2c4e6 100644 --- a/drivers/gpu/drm/nouveau/nv04_fbcon.c +++ b/drivers/gpu/drm/nouveau/nv04_fbcon.c @@ -27,7 +27,7 @@ #include "nouveau_dma.h" #include "nouveau_fbcon.h" -static void +void nv04_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region) { struct nouveau_fbcon_par *par = info->par; @@ -54,7 +54,7 @@ nv04_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region) FIRE_RING(chan); } -static void +void nv04_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect) { struct nouveau_fbcon_par *par = info->par; @@ -88,7 +88,7 @@ nv04_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect) FIRE_RING(chan); } -static void +void nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) { struct nouveau_fbcon_par *par = info->par; @@ -118,7 +118,7 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) return; } - width = (image->width + 31) & ~31; + width = ALIGN(image->width, 32); dsize = (width * image->height) >> 5; if (info->fix.visual == FB_VISUAL_TRUECOLOR || @@ -307,9 +307,6 @@ nv04_fbcon_accel_init(struct fb_info *info) FIRE_RING(chan); - info->fbops->fb_fillrect = nv04_fbcon_fillrect; - info->fbops->fb_copyarea = nv04_fbcon_copyarea; - info->fbops->fb_imageblit = nv04_fbcon_imageblit; return 0; } diff --git a/drivers/gpu/drm/nouveau/nv04_fifo.c b/drivers/gpu/drm/nouveau/nv04_fifo.c index f31347b8c9b..66fe55983b6 100644 --- a/drivers/gpu/drm/nouveau/nv04_fifo.c +++ b/drivers/gpu/drm/nouveau/nv04_fifo.c @@ -117,6 +117,7 @@ nv04_fifo_create_context(struct nouveau_channel *chan) { struct drm_device *dev = chan->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; + unsigned long flags; int ret; ret = nouveau_gpuobj_new_fake(dev, NV04_RAMFC(chan->id), ~0, @@ -127,6 +128,8 @@ nv04_fifo_create_context(struct nouveau_channel *chan) if (ret) return ret; + spin_lock_irqsave(&dev_priv->context_switch_lock, flags); + /* Setup initial state */ dev_priv->engine.instmem.prepare_access(dev, true); RAMFC_WR(DMA_PUT, chan->pushbuf_base); @@ -144,6 +147,8 @@ nv04_fifo_create_context(struct nouveau_channel *chan) /* enable the fifo dma operation */ nv_wr32(dev, NV04_PFIFO_MODE, nv_rd32(dev, NV04_PFIFO_MODE) | (1 << chan->id)); + + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); return 0; } diff --git a/drivers/gpu/drm/nouveau/nv04_instmem.c b/drivers/gpu/drm/nouveau/nv04_instmem.c index a20c206625a..a3b9563a6f6 100644 --- a/drivers/gpu/drm/nouveau/nv04_instmem.c +++ b/drivers/gpu/drm/nouveau/nv04_instmem.c @@ -30,7 +30,7 @@ nv04_instmem_determine_amount(struct drm_device *dev) * of vram. For now, only reserve a small piece until we know * more about what each chipset requires. */ - switch (dev_priv->chipset & 0xf0) { + switch (dev_priv->chipset) { case 0x40: case 0x47: case 0x49: diff --git a/drivers/gpu/drm/nouveau/nv04_tv.c b/drivers/gpu/drm/nouveau/nv04_tv.c index 9c63099e9c4..c4e3404337d 100644 --- a/drivers/gpu/drm/nouveau/nv04_tv.c +++ b/drivers/gpu/drm/nouveau/nv04_tv.c @@ -262,7 +262,7 @@ int nv04_tv_create(struct drm_device *dev, struct dcb_entry *entry) nv_encoder->or = ffs(entry->or) - 1; /* Run the slave-specific initialization */ - adap = &dev_priv->vbios->dcb->i2c[i2c_index].chan->adapter; + adap = &dev_priv->vbios.dcb.i2c[i2c_index].chan->adapter; was_locked = NVLockVgaCrtcs(dev, false); diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c index 58b917c3341..74c880374fb 100644 --- a/drivers/gpu/drm/nouveau/nv17_tv.c +++ b/drivers/gpu/drm/nouveau/nv17_tv.c @@ -45,8 +45,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder) #define RGB_TEST_DATA(r, g, b) (r << 0 | g << 10 | b << 20) testval = RGB_TEST_DATA(0x82, 0xeb, 0x82); - if (dev_priv->vbios->tvdactestval) - testval = dev_priv->vbios->tvdactestval; + if (dev_priv->vbios.tvdactestval) + testval = dev_priv->vbios.tvdactestval; dacclk = NVReadRAMDAC(dev, 0, NV_PRAMDAC_DACCLK + regoffset); head = (dacclk & 0x100) >> 8; @@ -367,7 +367,7 @@ static void nv17_tv_prepare(struct drm_encoder *encoder) !enc->crtc && nv04_dfp_get_bound_head(dev, dcb) == head) { nv04_dfp_bind_head(dev, dcb, head ^ 1, - dev_priv->VBIOS.fp.dual_link); + dev_priv->vbios.fp.dual_link); } } @@ -579,6 +579,8 @@ static void nv17_tv_restore(struct drm_encoder *encoder) nouveau_encoder(encoder)->restore.output); nv17_tv_state_load(dev, &to_tv_enc(encoder)->saved_state); + + nouveau_encoder(encoder)->last_dpms = NV_DPMS_CLEARED; } static int nv17_tv_create_resources(struct drm_encoder *encoder, diff --git a/drivers/gpu/drm/nouveau/nv40_fifo.c b/drivers/gpu/drm/nouveau/nv40_fifo.c index b4f19ccb8b4..6b2ef4a9fce 100644 --- a/drivers/gpu/drm/nouveau/nv40_fifo.c +++ b/drivers/gpu/drm/nouveau/nv40_fifo.c @@ -37,6 +37,7 @@ nv40_fifo_create_context(struct nouveau_channel *chan) struct drm_device *dev = chan->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; uint32_t fc = NV40_RAMFC(chan->id); + unsigned long flags; int ret; ret = nouveau_gpuobj_new_fake(dev, NV40_RAMFC(chan->id), ~0, @@ -45,6 +46,8 @@ nv40_fifo_create_context(struct nouveau_channel *chan) if (ret) return ret; + spin_lock_irqsave(&dev_priv->context_switch_lock, flags); + dev_priv->engine.instmem.prepare_access(dev, true); nv_wi32(dev, fc + 0, chan->pushbuf_base); nv_wi32(dev, fc + 4, chan->pushbuf_base); @@ -63,6 +66,8 @@ nv40_fifo_create_context(struct nouveau_channel *chan) /* enable the fifo dma operation */ nv_wr32(dev, NV04_PFIFO_MODE, nv_rd32(dev, NV04_PFIFO_MODE) | (1 << chan->id)); + + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); return 0; } diff --git a/drivers/gpu/drm/nouveau/nv50_crtc.c b/drivers/gpu/drm/nouveau/nv50_crtc.c index 118d3285fd8..cfabeb974a5 100644 --- a/drivers/gpu/drm/nouveau/nv50_crtc.c +++ b/drivers/gpu/drm/nouveau/nv50_crtc.c @@ -298,14 +298,17 @@ nv50_crtc_set_clock(struct drm_device *dev, int head, int pclk) static void nv50_crtc_destroy(struct drm_crtc *crtc) { - struct drm_device *dev = crtc->dev; - struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); - - NV_DEBUG_KMS(dev, "\n"); + struct drm_device *dev; + struct nouveau_crtc *nv_crtc; if (!crtc) return; + dev = crtc->dev; + nv_crtc = nouveau_crtc(crtc); + + NV_DEBUG_KMS(dev, "\n"); + drm_crtc_cleanup(&nv_crtc->base); nv50_cursor_fini(nv_crtc); @@ -355,9 +358,7 @@ nv50_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv, nv_crtc->cursor.show(nv_crtc, true); out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gem); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gem); return ret; } @@ -432,6 +433,7 @@ nv50_crtc_prepare(struct drm_crtc *crtc) struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); struct drm_device *dev = crtc->dev; struct drm_encoder *encoder; + uint32_t dac = 0, sor = 0; NV_DEBUG_KMS(dev, "index %d\n", nv_crtc->index); @@ -439,9 +441,28 @@ nv50_crtc_prepare(struct drm_crtc *crtc) list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); - if (drm_helper_encoder_in_use(encoder)) + if (!drm_helper_encoder_in_use(encoder)) continue; + if (nv_encoder->dcb->type == OUTPUT_ANALOG || + nv_encoder->dcb->type == OUTPUT_TV) + dac |= (1 << nv_encoder->or); + else + sor |= (1 << nv_encoder->or); + } + + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + + if (nv_encoder->dcb->type == OUTPUT_ANALOG || + nv_encoder->dcb->type == OUTPUT_TV) { + if (dac & (1 << nv_encoder->or)) + continue; + } else { + if (sor & (1 << nv_encoder->or)) + continue; + } + nv_encoder->disconnect(nv_encoder); } diff --git a/drivers/gpu/drm/nouveau/nv50_dac.c b/drivers/gpu/drm/nouveau/nv50_dac.c index f08f042a8e1..1fd9537beff 100644 --- a/drivers/gpu/drm/nouveau/nv50_dac.c +++ b/drivers/gpu/drm/nouveau/nv50_dac.c @@ -79,8 +79,8 @@ nv50_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector) } /* Use bios provided value if possible. */ - if (dev_priv->vbios->dactestval) { - load_pattern = dev_priv->vbios->dactestval; + if (dev_priv->vbios.dactestval) { + load_pattern = dev_priv->vbios.dactestval; NV_DEBUG_KMS(dev, "Using bios provided load_pattern of %d\n", load_pattern); } else { diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c index 90f0bf59fbc..61a89f2dc55 100644 --- a/drivers/gpu/drm/nouveau/nv50_display.c +++ b/drivers/gpu/drm/nouveau/nv50_display.c @@ -370,9 +370,7 @@ nv50_display_init(struct drm_device *dev) struct nouveau_connector *conn = nouveau_connector(connector); struct dcb_gpio_entry *gpio; - if (connector->connector_type != DRM_MODE_CONNECTOR_DVII && - connector->connector_type != DRM_MODE_CONNECTOR_DVID && - connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) + if (conn->dcb->gpio_tag == 0xff) continue; gpio = nouveau_bios_gpio_entry(dev, conn->dcb->gpio_tag); @@ -465,8 +463,7 @@ static int nv50_display_disable(struct drm_device *dev) int nv50_display_create(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct parsed_dcb *dcb = dev_priv->vbios->dcb; - uint32_t connector[16] = {}; + struct dcb_table *dcb = &dev_priv->vbios.dcb; int ret, i; NV_DEBUG_KMS(dev, "\n"); @@ -522,44 +519,13 @@ int nv50_display_create(struct drm_device *dev) NV_WARN(dev, "DCB encoder %d unknown\n", entry->type); continue; } - - connector[entry->connector] |= (1 << entry->type); } - /* It appears that DCB 3.0+ VBIOS has a connector table, however, - * I'm not 100% certain how to decode it correctly yet so just - * look at what encoders are present on each connector index and - * attempt to derive the connector type from that. - */ - for (i = 0 ; i < dcb->entries; i++) { - struct dcb_entry *entry = &dcb->entry[i]; - uint16_t encoders; - int type; - - encoders = connector[entry->connector]; - if (!(encoders & (1 << entry->type))) + for (i = 0 ; i < dcb->connector.entries; i++) { + if (i != 0 && dcb->connector.entry[i].index == + dcb->connector.entry[i - 1].index) continue; - connector[entry->connector] = 0; - - if (encoders & (1 << OUTPUT_DP)) { - type = DRM_MODE_CONNECTOR_DisplayPort; - } else if (encoders & (1 << OUTPUT_TMDS)) { - if (encoders & (1 << OUTPUT_ANALOG)) - type = DRM_MODE_CONNECTOR_DVII; - else - type = DRM_MODE_CONNECTOR_DVID; - } else if (encoders & (1 << OUTPUT_ANALOG)) { - type = DRM_MODE_CONNECTOR_VGA; - } else if (encoders & (1 << OUTPUT_LVDS)) { - type = DRM_MODE_CONNECTOR_LVDS; - } else { - type = DRM_MODE_CONNECTOR_Unknown; - } - - if (type == DRM_MODE_CONNECTOR_Unknown) - continue; - - nouveau_connector_create(dev, entry->connector, type); + nouveau_connector_create(dev, &dcb->connector.entry[i]); } ret = nv50_display_init(dev); @@ -667,8 +633,8 @@ nv50_display_irq_head(struct drm_device *dev, int *phead, return -1; } - for (i = 0; i < dev_priv->vbios->dcb->entries; i++) { - struct dcb_entry *dcbent = &dev_priv->vbios->dcb->entry[i]; + for (i = 0; i < dev_priv->vbios.dcb.entries; i++) { + struct dcb_entry *dcbent = &dev_priv->vbios.dcb.entry[i]; if (dcbent->type != type) continue; @@ -692,7 +658,7 @@ nv50_display_script_select(struct drm_device *dev, struct dcb_entry *dcbent, struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_connector *nv_connector = NULL; struct drm_encoder *encoder; - struct nvbios *bios = &dev_priv->VBIOS; + struct nvbios *bios = &dev_priv->vbios; uint32_t mc, script = 0, or; list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { @@ -710,7 +676,7 @@ nv50_display_script_select(struct drm_device *dev, struct dcb_entry *dcbent, switch (dcbent->type) { case OUTPUT_LVDS: script = (mc >> 8) & 0xf; - if (bios->pub.fp_no_ddc) { + if (bios->fp_no_ddc) { if (bios->fp.dual_link) script |= 0x0100; if (bios->fp.if_is_24bit) diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c index e4f279ee61c..993c7126fbd 100644 --- a/drivers/gpu/drm/nouveau/nv50_fbcon.c +++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c @@ -3,7 +3,7 @@ #include "nouveau_dma.h" #include "nouveau_fbcon.h" -static void +void nv50_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect) { struct nouveau_fbcon_par *par = info->par; @@ -46,7 +46,7 @@ nv50_fbcon_fillrect(struct fb_info *info, const struct fb_fillrect *rect) FIRE_RING(chan); } -static void +void nv50_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region) { struct nouveau_fbcon_par *par = info->par; @@ -81,7 +81,7 @@ nv50_fbcon_copyarea(struct fb_info *info, const struct fb_copyarea *region) FIRE_RING(chan); } -static void +void nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) { struct nouveau_fbcon_par *par = info->par; @@ -109,7 +109,7 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) return; } - width = (image->width + 31) & ~31; + width = ALIGN(image->width, 32); dwords = (width * image->height) >> 5; BEGIN_RING(chan, NvSub2D, 0x0814, 2); @@ -262,9 +262,6 @@ nv50_fbcon_accel_init(struct fb_info *info) OUT_RING(chan, info->fix.smem_start - dev_priv->fb_phys + dev_priv->vm_vram_base); - info->fbops->fb_fillrect = nv50_fbcon_fillrect; - info->fbops->fb_copyarea = nv50_fbcon_copyarea; - info->fbops->fb_imageblit = nv50_fbcon_imageblit; return 0; } diff --git a/drivers/gpu/drm/nouveau/nv50_fifo.c b/drivers/gpu/drm/nouveau/nv50_fifo.c index 39caf167587..e20c0e2474f 100644 --- a/drivers/gpu/drm/nouveau/nv50_fifo.c +++ b/drivers/gpu/drm/nouveau/nv50_fifo.c @@ -243,6 +243,7 @@ nv50_fifo_create_context(struct nouveau_channel *chan) struct drm_device *dev = chan->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *ramfc = NULL; + unsigned long flags; int ret; NV_DEBUG(dev, "ch%d\n", chan->id); @@ -272,25 +273,27 @@ nv50_fifo_create_context(struct nouveau_channel *chan) return ret; ramfc = chan->ramfc->gpuobj; - ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, 4096, 256, + ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, 4096, 1024, 0, &chan->cache); if (ret) return ret; } + spin_lock_irqsave(&dev_priv->context_switch_lock, flags); + dev_priv->engine.instmem.prepare_access(dev, true); - nv_wo32(dev, ramfc, 0x08/4, chan->pushbuf_base); - nv_wo32(dev, ramfc, 0x10/4, chan->pushbuf_base); nv_wo32(dev, ramfc, 0x48/4, chan->pushbuf->instance >> 4); nv_wo32(dev, ramfc, 0x80/4, (0xc << 24) | (chan->ramht->instance >> 4)); - nv_wo32(dev, ramfc, 0x3c/4, 0x00086078); nv_wo32(dev, ramfc, 0x44/4, 0x2101ffff); nv_wo32(dev, ramfc, 0x60/4, 0x7fffffff); nv_wo32(dev, ramfc, 0x40/4, 0x00000000); nv_wo32(dev, ramfc, 0x7c/4, 0x30000001); nv_wo32(dev, ramfc, 0x78/4, 0x00000000); - nv_wo32(dev, ramfc, 0x4c/4, 0xffffffff); + nv_wo32(dev, ramfc, 0x3c/4, 0x403f6078); + nv_wo32(dev, ramfc, 0x50/4, chan->pushbuf_base + + chan->dma.ib_base * 4); + nv_wo32(dev, ramfc, 0x54/4, drm_order(chan->dma.ib_max + 1) << 16); if (!IS_G80) { nv_wo32(dev, chan->ramin->gpuobj, 0, chan->id); @@ -306,10 +309,12 @@ nv50_fifo_create_context(struct nouveau_channel *chan) ret = nv50_fifo_channel_enable(dev, chan->id, false); if (ret) { NV_ERROR(dev, "error enabling ch%d: %d\n", chan->id, ret); + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); nouveau_gpuobj_ref_del(dev, &chan->ramfc); return ret; } + spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); return 0; } @@ -317,17 +322,20 @@ void nv50_fifo_destroy_context(struct nouveau_channel *chan) { struct drm_device *dev = chan->dev; + struct nouveau_gpuobj_ref *ramfc = chan->ramfc; NV_DEBUG(dev, "ch%d\n", chan->id); - nouveau_gpuobj_ref_del(dev, &chan->ramfc); - nouveau_gpuobj_ref_del(dev, &chan->cache); - + /* This will ensure the channel is seen as disabled. */ + chan->ramfc = NULL; nv50_fifo_channel_disable(dev, chan->id, false); /* Dummy channel, also used on ch 127 */ if (chan->id == 0) nv50_fifo_channel_disable(dev, 127, false); + + nouveau_gpuobj_ref_del(dev, &ramfc); + nouveau_gpuobj_ref_del(dev, &chan->cache); } int diff --git a/drivers/gpu/drm/nouveau/nv50_graph.c b/drivers/gpu/drm/nouveau/nv50_graph.c index ca79f32be44..857a09671a3 100644 --- a/drivers/gpu/drm/nouveau/nv50_graph.c +++ b/drivers/gpu/drm/nouveau/nv50_graph.c @@ -28,30 +28,7 @@ #include "drm.h" #include "nouveau_drv.h" -MODULE_FIRMWARE("nouveau/nv50.ctxprog"); -MODULE_FIRMWARE("nouveau/nv50.ctxvals"); -MODULE_FIRMWARE("nouveau/nv84.ctxprog"); -MODULE_FIRMWARE("nouveau/nv84.ctxvals"); -MODULE_FIRMWARE("nouveau/nv86.ctxprog"); -MODULE_FIRMWARE("nouveau/nv86.ctxvals"); -MODULE_FIRMWARE("nouveau/nv92.ctxprog"); -MODULE_FIRMWARE("nouveau/nv92.ctxvals"); -MODULE_FIRMWARE("nouveau/nv94.ctxprog"); -MODULE_FIRMWARE("nouveau/nv94.ctxvals"); -MODULE_FIRMWARE("nouveau/nv96.ctxprog"); -MODULE_FIRMWARE("nouveau/nv96.ctxvals"); -MODULE_FIRMWARE("nouveau/nv98.ctxprog"); -MODULE_FIRMWARE("nouveau/nv98.ctxvals"); -MODULE_FIRMWARE("nouveau/nva0.ctxprog"); -MODULE_FIRMWARE("nouveau/nva0.ctxvals"); -MODULE_FIRMWARE("nouveau/nva5.ctxprog"); -MODULE_FIRMWARE("nouveau/nva5.ctxvals"); -MODULE_FIRMWARE("nouveau/nva8.ctxprog"); -MODULE_FIRMWARE("nouveau/nva8.ctxvals"); -MODULE_FIRMWARE("nouveau/nvaa.ctxprog"); -MODULE_FIRMWARE("nouveau/nvaa.ctxvals"); -MODULE_FIRMWARE("nouveau/nvac.ctxprog"); -MODULE_FIRMWARE("nouveau/nvac.ctxvals"); +#include "nouveau_grctx.h" #define IS_G80 ((dev_priv->chipset & 0xf0) == 0x50) @@ -84,7 +61,7 @@ nv50_graph_init_regs__nv(struct drm_device *dev) nv_wr32(dev, 0x400804, 0xc0000000); nv_wr32(dev, 0x406800, 0xc0000000); nv_wr32(dev, 0x400c04, 0xc0000000); - nv_wr32(dev, 0x401804, 0xc0000000); + nv_wr32(dev, 0x401800, 0xc0000000); nv_wr32(dev, 0x405018, 0xc0000000); nv_wr32(dev, 0x402000, 0xc0000000); @@ -111,9 +88,34 @@ nv50_graph_init_ctxctl(struct drm_device *dev) NV_DEBUG(dev, "\n"); - nouveau_grctx_prog_load(dev); - if (!dev_priv->engine.graph.ctxprog) - dev_priv->engine.graph.accel_blocked = true; + if (nouveau_ctxfw) { + nouveau_grctx_prog_load(dev); + dev_priv->engine.graph.grctx_size = 0x70000; + } + if (!dev_priv->engine.graph.ctxprog) { + struct nouveau_grctx ctx = {}; + uint32_t *cp = kmalloc(512 * 4, GFP_KERNEL); + int i; + if (!cp) { + NV_ERROR(dev, "Couldn't alloc ctxprog! Disabling acceleration.\n"); + dev_priv->engine.graph.accel_blocked = true; + return 0; + } + ctx.dev = dev; + ctx.mode = NOUVEAU_GRCTX_PROG; + ctx.data = cp; + ctx.ctxprog_max = 512; + if (!nv50_grctx_init(&ctx)) { + dev_priv->engine.graph.grctx_size = ctx.ctxvals_pos * 4; + + nv_wr32(dev, NV40_PGRAPH_CTXCTL_UCODE_INDEX, 0); + for (i = 0; i < ctx.ctxprog_len; i++) + nv_wr32(dev, NV40_PGRAPH_CTXCTL_UCODE_DATA, cp[i]); + } else { + dev_priv->engine.graph.accel_blocked = true; + } + kfree(cp); + } nv_wr32(dev, 0x400320, 4); nv_wr32(dev, NV40_PGRAPH_CTXCTL_CUR, 0); @@ -165,6 +167,12 @@ nv50_graph_channel(struct drm_device *dev) uint32_t inst; int i; + /* Be sure we're not in the middle of a context switch or bad things + * will happen, such as unloading the wrong pgraph context. + */ + if (!nv_wait(0x400300, 0x00000001, 0x00000000)) + NV_ERROR(dev, "Ctxprog is still running\n"); + inst = nv_rd32(dev, NV50_PGRAPH_CTXCTL_CUR); if (!(inst & NV50_PGRAPH_CTXCTL_CUR_LOADED)) return NULL; @@ -187,13 +195,13 @@ nv50_graph_create_context(struct nouveau_channel *chan) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *ramin = chan->ramin->gpuobj; struct nouveau_gpuobj *ctx; - uint32_t grctx_size = 0x70000; + struct nouveau_pgraph_engine *pgraph = &dev_priv->engine.graph; int hdr, ret; NV_DEBUG(dev, "ch%d\n", chan->id); - ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, grctx_size, 0x1000, - NVOBJ_FLAG_ZERO_ALLOC | + ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pgraph->grctx_size, + 0x1000, NVOBJ_FLAG_ZERO_ALLOC | NVOBJ_FLAG_ZERO_FREE, &chan->ramin_grctx); if (ret) return ret; @@ -203,7 +211,7 @@ nv50_graph_create_context(struct nouveau_channel *chan) dev_priv->engine.instmem.prepare_access(dev, true); nv_wo32(dev, ramin, (hdr + 0x00)/4, 0x00190002); nv_wo32(dev, ramin, (hdr + 0x04)/4, chan->ramin_grctx->instance + - grctx_size - 1); + pgraph->grctx_size - 1); nv_wo32(dev, ramin, (hdr + 0x08)/4, chan->ramin_grctx->instance); nv_wo32(dev, ramin, (hdr + 0x0c)/4, 0); nv_wo32(dev, ramin, (hdr + 0x10)/4, 0); @@ -211,7 +219,15 @@ nv50_graph_create_context(struct nouveau_channel *chan) dev_priv->engine.instmem.finish_access(dev); dev_priv->engine.instmem.prepare_access(dev, true); - nouveau_grctx_vals_load(dev, ctx); + if (!pgraph->ctxprog) { + struct nouveau_grctx ctx = {}; + ctx.dev = chan->dev; + ctx.mode = NOUVEAU_GRCTX_VALS; + ctx.data = chan->ramin_grctx->gpuobj; + nv50_grctx_init(&ctx); + } else { + nouveau_grctx_vals_load(dev, ctx); + } nv_wo32(dev, ctx, 0x00000/4, chan->ramin->instance >> 12); if ((dev_priv->chipset & 0xf0) == 0xa0) nv_wo32(dev, ctx, 0x00004/4, 0x00000000); @@ -275,19 +291,18 @@ nv50_graph_load_context(struct nouveau_channel *chan) int nv50_graph_unload_context(struct drm_device *dev) { - uint32_t inst, fifo = nv_rd32(dev, 0x400500); + uint32_t inst; inst = nv_rd32(dev, NV50_PGRAPH_CTXCTL_CUR); if (!(inst & NV50_PGRAPH_CTXCTL_CUR_LOADED)) return 0; inst &= NV50_PGRAPH_CTXCTL_CUR_INSTANCE; - nv_wr32(dev, 0x400500, fifo & ~1); + nouveau_wait_for_idle(dev); nv_wr32(dev, 0x400784, inst); nv_wr32(dev, 0x400824, nv_rd32(dev, 0x400824) | 0x20); nv_wr32(dev, 0x400304, nv_rd32(dev, 0x400304) | 0x01); nouveau_wait_for_idle(dev); - nv_wr32(dev, 0x400500, fifo); nv_wr32(dev, NV50_PGRAPH_CTXCTL_CUR, inst); return 0; diff --git a/drivers/gpu/drm/nouveau/nv50_grctx.c b/drivers/gpu/drm/nouveau/nv50_grctx.c new file mode 100644 index 00000000000..d105fcd42ca --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv50_grctx.c @@ -0,0 +1,2367 @@ +/* + * Copyright 2009 Marcin KoÅ›cielnicki + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#define CP_FLAG_CLEAR 0 +#define CP_FLAG_SET 1 +#define CP_FLAG_SWAP_DIRECTION ((0 * 32) + 0) +#define CP_FLAG_SWAP_DIRECTION_LOAD 0 +#define CP_FLAG_SWAP_DIRECTION_SAVE 1 +#define CP_FLAG_UNK01 ((0 * 32) + 1) +#define CP_FLAG_UNK01_CLEAR 0 +#define CP_FLAG_UNK01_SET 1 +#define CP_FLAG_UNK03 ((0 * 32) + 3) +#define CP_FLAG_UNK03_CLEAR 0 +#define CP_FLAG_UNK03_SET 1 +#define CP_FLAG_USER_SAVE ((0 * 32) + 5) +#define CP_FLAG_USER_SAVE_NOT_PENDING 0 +#define CP_FLAG_USER_SAVE_PENDING 1 +#define CP_FLAG_USER_LOAD ((0 * 32) + 6) +#define CP_FLAG_USER_LOAD_NOT_PENDING 0 +#define CP_FLAG_USER_LOAD_PENDING 1 +#define CP_FLAG_UNK0B ((0 * 32) + 0xb) +#define CP_FLAG_UNK0B_CLEAR 0 +#define CP_FLAG_UNK0B_SET 1 +#define CP_FLAG_UNK1D ((0 * 32) + 0x1d) +#define CP_FLAG_UNK1D_CLEAR 0 +#define CP_FLAG_UNK1D_SET 1 +#define CP_FLAG_UNK20 ((1 * 32) + 0) +#define CP_FLAG_UNK20_CLEAR 0 +#define CP_FLAG_UNK20_SET 1 +#define CP_FLAG_STATUS ((2 * 32) + 0) +#define CP_FLAG_STATUS_BUSY 0 +#define CP_FLAG_STATUS_IDLE 1 +#define CP_FLAG_AUTO_SAVE ((2 * 32) + 4) +#define CP_FLAG_AUTO_SAVE_NOT_PENDING 0 +#define CP_FLAG_AUTO_SAVE_PENDING 1 +#define CP_FLAG_AUTO_LOAD ((2 * 32) + 5) +#define CP_FLAG_AUTO_LOAD_NOT_PENDING 0 +#define CP_FLAG_AUTO_LOAD_PENDING 1 +#define CP_FLAG_XFER ((2 * 32) + 11) +#define CP_FLAG_XFER_IDLE 0 +#define CP_FLAG_XFER_BUSY 1 +#define CP_FLAG_NEWCTX ((2 * 32) + 12) +#define CP_FLAG_NEWCTX_BUSY 0 +#define CP_FLAG_NEWCTX_DONE 1 +#define CP_FLAG_ALWAYS ((2 * 32) + 13) +#define CP_FLAG_ALWAYS_FALSE 0 +#define CP_FLAG_ALWAYS_TRUE 1 + +#define CP_CTX 0x00100000 +#define CP_CTX_COUNT 0x000f0000 +#define CP_CTX_COUNT_SHIFT 16 +#define CP_CTX_REG 0x00003fff +#define CP_LOAD_SR 0x00200000 +#define CP_LOAD_SR_VALUE 0x000fffff +#define CP_BRA 0x00400000 +#define CP_BRA_IP 0x0001ff00 +#define CP_BRA_IP_SHIFT 8 +#define CP_BRA_IF_CLEAR 0x00000080 +#define CP_BRA_FLAG 0x0000007f +#define CP_WAIT 0x00500000 +#define CP_WAIT_SET 0x00000080 +#define CP_WAIT_FLAG 0x0000007f +#define CP_SET 0x00700000 +#define CP_SET_1 0x00000080 +#define CP_SET_FLAG 0x0000007f +#define CP_NEWCTX 0x00600004 +#define CP_NEXT_TO_SWAP 0x00600005 +#define CP_SET_CONTEXT_POINTER 0x00600006 +#define CP_SET_XFER_POINTER 0x00600007 +#define CP_ENABLE 0x00600009 +#define CP_END 0x0060000c +#define CP_NEXT_TO_CURRENT 0x0060000d +#define CP_DISABLE1 0x0090ffff +#define CP_DISABLE2 0x0091ffff +#define CP_XFER_1 0x008000ff +#define CP_XFER_2 0x008800ff +#define CP_SEEK_1 0x00c000ff +#define CP_SEEK_2 0x00c800ff + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_grctx.h" + +/* + * This code deals with PGRAPH contexts on NV50 family cards. Like NV40, it's + * the GPU itself that does context-switching, but it needs a special + * microcode to do it. And it's the driver's task to supply this microcode, + * further known as ctxprog, as well as the initial context values, known + * as ctxvals. + * + * Without ctxprog, you cannot switch contexts. Not even in software, since + * the majority of context [xfer strands] isn't accessible directly. You're + * stuck with a single channel, and you also suffer all the problems resulting + * from missing ctxvals, since you cannot load them. + * + * Without ctxvals, you're stuck with PGRAPH's default context. It's enough to + * run 2d operations, but trying to utilise 3d or CUDA will just lock you up, + * since you don't have... some sort of needed setup. + * + * Nouveau will just disable acceleration if not given ctxprog + ctxvals, since + * it's too much hassle to handle no-ctxprog as a special case. + */ + +/* + * How ctxprogs work. + * + * The ctxprog is written in its own kind of microcode, with very small and + * crappy set of available commands. You upload it to a small [512 insns] + * area of memory on PGRAPH, and it'll be run when PFIFO wants PGRAPH to + * switch channel. or when the driver explicitely requests it. Stuff visible + * to ctxprog consists of: PGRAPH MMIO registers, PGRAPH context strands, + * the per-channel context save area in VRAM [known as ctxvals or grctx], + * 4 flags registers, a scratch register, two grctx pointers, plus many + * random poorly-understood details. + * + * When ctxprog runs, it's supposed to check what operations are asked of it, + * save old context if requested, optionally reset PGRAPH and switch to the + * new channel, and load the new context. Context consists of three major + * parts: subset of MMIO registers and two "xfer areas". + */ + +/* TODO: + * - document unimplemented bits compared to nvidia + * - NVAx: make a TP subroutine, use it. + * - use 0x4008fc instead of 0x1540? + */ + +enum cp_label { + cp_check_load = 1, + cp_setup_auto_load, + cp_setup_load, + cp_setup_save, + cp_swap_state, + cp_prepare_exit, + cp_exit, +}; + +static void nv50_graph_construct_mmio(struct nouveau_grctx *ctx); +static void nv50_graph_construct_xfer1(struct nouveau_grctx *ctx); +static void nv50_graph_construct_xfer2(struct nouveau_grctx *ctx); + +/* Main function: construct the ctxprog skeleton, call the other functions. */ + +int +nv50_grctx_init(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + + switch (dev_priv->chipset) { + case 0x50: + case 0x84: + case 0x86: + case 0x92: + case 0x94: + case 0x96: + case 0x98: + case 0xa0: + case 0xa5: + case 0xa8: + case 0xaa: + case 0xac: + break; + default: + NV_ERROR(ctx->dev, "I don't know how to make a ctxprog for " + "your NV%x card.\n", dev_priv->chipset); + NV_ERROR(ctx->dev, "Disabling acceleration. Please contact " + "the devs.\n"); + return -ENOSYS; + } + /* decide whether we're loading/unloading the context */ + cp_bra (ctx, AUTO_SAVE, PENDING, cp_setup_save); + cp_bra (ctx, USER_SAVE, PENDING, cp_setup_save); + + cp_name(ctx, cp_check_load); + cp_bra (ctx, AUTO_LOAD, PENDING, cp_setup_auto_load); + cp_bra (ctx, USER_LOAD, PENDING, cp_setup_load); + cp_bra (ctx, ALWAYS, TRUE, cp_exit); + + /* setup for context load */ + cp_name(ctx, cp_setup_auto_load); + cp_out (ctx, CP_DISABLE1); + cp_out (ctx, CP_DISABLE2); + cp_out (ctx, CP_ENABLE); + cp_out (ctx, CP_NEXT_TO_SWAP); + cp_set (ctx, UNK01, SET); + cp_name(ctx, cp_setup_load); + cp_out (ctx, CP_NEWCTX); + cp_wait(ctx, NEWCTX, BUSY); + cp_set (ctx, UNK1D, CLEAR); + cp_set (ctx, SWAP_DIRECTION, LOAD); + cp_bra (ctx, UNK0B, SET, cp_prepare_exit); + cp_bra (ctx, ALWAYS, TRUE, cp_swap_state); + + /* setup for context save */ + cp_name(ctx, cp_setup_save); + cp_set (ctx, UNK1D, SET); + cp_wait(ctx, STATUS, BUSY); + cp_set (ctx, UNK01, SET); + cp_set (ctx, SWAP_DIRECTION, SAVE); + + /* general PGRAPH state */ + cp_name(ctx, cp_swap_state); + cp_set (ctx, UNK03, SET); + cp_pos (ctx, 0x00004/4); + cp_ctx (ctx, 0x400828, 1); /* needed. otherwise, flickering happens. */ + cp_pos (ctx, 0x00100/4); + nv50_graph_construct_mmio(ctx); + nv50_graph_construct_xfer1(ctx); + nv50_graph_construct_xfer2(ctx); + + cp_bra (ctx, SWAP_DIRECTION, SAVE, cp_check_load); + + cp_set (ctx, UNK20, SET); + cp_set (ctx, SWAP_DIRECTION, SAVE); /* no idea why this is needed, but fixes at least one lockup. */ + cp_lsr (ctx, ctx->ctxvals_base); + cp_out (ctx, CP_SET_XFER_POINTER); + cp_lsr (ctx, 4); + cp_out (ctx, CP_SEEK_1); + cp_out (ctx, CP_XFER_1); + cp_wait(ctx, XFER, BUSY); + + /* pre-exit state updates */ + cp_name(ctx, cp_prepare_exit); + cp_set (ctx, UNK01, CLEAR); + cp_set (ctx, UNK03, CLEAR); + cp_set (ctx, UNK1D, CLEAR); + + cp_bra (ctx, USER_SAVE, PENDING, cp_exit); + cp_out (ctx, CP_NEXT_TO_CURRENT); + + cp_name(ctx, cp_exit); + cp_set (ctx, USER_SAVE, NOT_PENDING); + cp_set (ctx, USER_LOAD, NOT_PENDING); + cp_out (ctx, CP_END); + ctx->ctxvals_pos += 0x400; /* padding... no idea why you need it */ + + return 0; +} + +/* + * Constructs MMIO part of ctxprog and ctxvals. Just a matter of knowing which + * registers to save/restore and the default values for them. + */ + +static void +nv50_graph_construct_mmio(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int i, j; + int offset, base; + uint32_t units = nv_rd32 (ctx->dev, 0x1540); + + /* 0800 */ + cp_ctx(ctx, 0x400808, 7); + gr_def(ctx, 0x400814, 0x00000030); + cp_ctx(ctx, 0x400834, 0x32); + if (dev_priv->chipset == 0x50) { + gr_def(ctx, 0x400834, 0xff400040); + gr_def(ctx, 0x400838, 0xfff00080); + gr_def(ctx, 0x40083c, 0xfff70090); + gr_def(ctx, 0x400840, 0xffe806a8); + } + gr_def(ctx, 0x400844, 0x00000002); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + gr_def(ctx, 0x400894, 0x00001000); + gr_def(ctx, 0x4008e8, 0x00000003); + gr_def(ctx, 0x4008ec, 0x00001000); + if (dev_priv->chipset == 0x50) + cp_ctx(ctx, 0x400908, 0xb); + else if (dev_priv->chipset < 0xa0) + cp_ctx(ctx, 0x400908, 0xc); + else + cp_ctx(ctx, 0x400908, 0xe); + + if (dev_priv->chipset >= 0xa0) + cp_ctx(ctx, 0x400b00, 0x1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + cp_ctx(ctx, 0x400b10, 0x1); + gr_def(ctx, 0x400b10, 0x0001629d); + cp_ctx(ctx, 0x400b20, 0x1); + gr_def(ctx, 0x400b20, 0x0001629d); + } + + /* 0C00 */ + cp_ctx(ctx, 0x400c08, 0x2); + gr_def(ctx, 0x400c08, 0x0000fe0c); + + /* 1000 */ + if (dev_priv->chipset < 0xa0) { + cp_ctx(ctx, 0x401008, 0x4); + gr_def(ctx, 0x401014, 0x00001000); + } else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa) { + cp_ctx(ctx, 0x401008, 0x5); + gr_def(ctx, 0x401018, 0x00001000); + } else { + cp_ctx(ctx, 0x401008, 0x5); + gr_def(ctx, 0x401018, 0x00004000); + } + + /* 1400 */ + cp_ctx(ctx, 0x401400, 0x8); + cp_ctx(ctx, 0x401424, 0x3); + if (dev_priv->chipset == 0x50) + gr_def(ctx, 0x40142c, 0x0001fd87); + else + gr_def(ctx, 0x40142c, 0x00000187); + cp_ctx(ctx, 0x401540, 0x5); + gr_def(ctx, 0x401550, 0x00001018); + + /* 1800 */ + cp_ctx(ctx, 0x401814, 0x1); + gr_def(ctx, 0x401814, 0x000000ff); + if (dev_priv->chipset == 0x50) { + cp_ctx(ctx, 0x40181c, 0xe); + gr_def(ctx, 0x401850, 0x00000004); + } else if (dev_priv->chipset < 0xa0) { + cp_ctx(ctx, 0x40181c, 0xf); + gr_def(ctx, 0x401854, 0x00000004); + } else { + cp_ctx(ctx, 0x40181c, 0x13); + gr_def(ctx, 0x401864, 0x00000004); + } + + /* 1C00 */ + cp_ctx(ctx, 0x401c00, 0x1); + switch (dev_priv->chipset) { + case 0x50: + gr_def(ctx, 0x401c00, 0x0001005f); + break; + case 0x84: + case 0x86: + case 0x94: + gr_def(ctx, 0x401c00, 0x044d00df); + break; + case 0x92: + case 0x96: + case 0x98: + case 0xa0: + case 0xaa: + case 0xac: + gr_def(ctx, 0x401c00, 0x042500df); + break; + case 0xa5: + case 0xa8: + gr_def(ctx, 0x401c00, 0x142500df); + break; + } + + /* 2400 */ + cp_ctx(ctx, 0x402400, 0x1); + if (dev_priv->chipset == 0x50) + cp_ctx(ctx, 0x402408, 0x1); + else + cp_ctx(ctx, 0x402408, 0x2); + gr_def(ctx, 0x402408, 0x00000600); + + /* 2800 */ + cp_ctx(ctx, 0x402800, 0x1); + if (dev_priv->chipset == 0x50) + gr_def(ctx, 0x402800, 0x00000006); + + /* 2C00 */ + cp_ctx(ctx, 0x402c08, 0x6); + if (dev_priv->chipset != 0x50) + gr_def(ctx, 0x402c14, 0x01000000); + gr_def(ctx, 0x402c18, 0x000000ff); + if (dev_priv->chipset == 0x50) + cp_ctx(ctx, 0x402ca0, 0x1); + else + cp_ctx(ctx, 0x402ca0, 0x2); + if (dev_priv->chipset < 0xa0) + gr_def(ctx, 0x402ca0, 0x00000400); + else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa) + gr_def(ctx, 0x402ca0, 0x00000800); + else + gr_def(ctx, 0x402ca0, 0x00000400); + cp_ctx(ctx, 0x402cac, 0x4); + + /* 3000 */ + cp_ctx(ctx, 0x403004, 0x1); + gr_def(ctx, 0x403004, 0x00000001); + + /* 3404 */ + if (dev_priv->chipset >= 0xa0) { + cp_ctx(ctx, 0x403404, 0x1); + gr_def(ctx, 0x403404, 0x00000001); + } + + /* 5000 */ + cp_ctx(ctx, 0x405000, 0x1); + switch (dev_priv->chipset) { + case 0x50: + gr_def(ctx, 0x405000, 0x00300080); + break; + case 0x84: + case 0xa0: + case 0xa5: + case 0xa8: + case 0xaa: + case 0xac: + gr_def(ctx, 0x405000, 0x000e0080); + break; + case 0x86: + case 0x92: + case 0x94: + case 0x96: + case 0x98: + gr_def(ctx, 0x405000, 0x00000080); + break; + } + cp_ctx(ctx, 0x405014, 0x1); + gr_def(ctx, 0x405014, 0x00000004); + cp_ctx(ctx, 0x40501c, 0x1); + cp_ctx(ctx, 0x405024, 0x1); + cp_ctx(ctx, 0x40502c, 0x1); + + /* 5400 or maybe 4800 */ + if (dev_priv->chipset == 0x50) { + offset = 0x405400; + cp_ctx(ctx, 0x405400, 0xea); + } else if (dev_priv->chipset < 0x94) { + offset = 0x405400; + cp_ctx(ctx, 0x405400, 0xcb); + } else if (dev_priv->chipset < 0xa0) { + offset = 0x405400; + cp_ctx(ctx, 0x405400, 0xcc); + } else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + offset = 0x404800; + cp_ctx(ctx, 0x404800, 0xda); + } else { + offset = 0x405400; + cp_ctx(ctx, 0x405400, 0xd4); + } + gr_def(ctx, offset + 0x0c, 0x00000002); + gr_def(ctx, offset + 0x10, 0x00000001); + if (dev_priv->chipset >= 0x94) + offset += 4; + gr_def(ctx, offset + 0x1c, 0x00000001); + gr_def(ctx, offset + 0x20, 0x00000100); + gr_def(ctx, offset + 0x38, 0x00000002); + gr_def(ctx, offset + 0x3c, 0x00000001); + gr_def(ctx, offset + 0x40, 0x00000001); + gr_def(ctx, offset + 0x50, 0x00000001); + gr_def(ctx, offset + 0x54, 0x003fffff); + gr_def(ctx, offset + 0x58, 0x00001fff); + gr_def(ctx, offset + 0x60, 0x00000001); + gr_def(ctx, offset + 0x64, 0x00000001); + gr_def(ctx, offset + 0x6c, 0x00000001); + gr_def(ctx, offset + 0x70, 0x00000001); + gr_def(ctx, offset + 0x74, 0x00000001); + gr_def(ctx, offset + 0x78, 0x00000004); + gr_def(ctx, offset + 0x7c, 0x00000001); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + offset += 4; + gr_def(ctx, offset + 0x80, 0x00000001); + gr_def(ctx, offset + 0x84, 0x00000001); + gr_def(ctx, offset + 0x88, 0x00000007); + gr_def(ctx, offset + 0x8c, 0x00000001); + gr_def(ctx, offset + 0x90, 0x00000007); + gr_def(ctx, offset + 0x94, 0x00000001); + gr_def(ctx, offset + 0x98, 0x00000001); + gr_def(ctx, offset + 0x9c, 0x00000001); + if (dev_priv->chipset == 0x50) { + gr_def(ctx, offset + 0xb0, 0x00000001); + gr_def(ctx, offset + 0xb4, 0x00000001); + gr_def(ctx, offset + 0xbc, 0x00000001); + gr_def(ctx, offset + 0xc0, 0x0000000a); + gr_def(ctx, offset + 0xd0, 0x00000040); + gr_def(ctx, offset + 0xd8, 0x00000002); + gr_def(ctx, offset + 0xdc, 0x00000100); + gr_def(ctx, offset + 0xe0, 0x00000001); + gr_def(ctx, offset + 0xe4, 0x00000100); + gr_def(ctx, offset + 0x100, 0x00000001); + gr_def(ctx, offset + 0x124, 0x00000004); + gr_def(ctx, offset + 0x13c, 0x00000001); + gr_def(ctx, offset + 0x140, 0x00000100); + gr_def(ctx, offset + 0x148, 0x00000001); + gr_def(ctx, offset + 0x154, 0x00000100); + gr_def(ctx, offset + 0x158, 0x00000001); + gr_def(ctx, offset + 0x15c, 0x00000100); + gr_def(ctx, offset + 0x164, 0x00000001); + gr_def(ctx, offset + 0x170, 0x00000100); + gr_def(ctx, offset + 0x174, 0x00000001); + gr_def(ctx, offset + 0x17c, 0x00000001); + gr_def(ctx, offset + 0x188, 0x00000002); + gr_def(ctx, offset + 0x190, 0x00000001); + gr_def(ctx, offset + 0x198, 0x00000001); + gr_def(ctx, offset + 0x1ac, 0x00000003); + offset += 0xd0; + } else { + gr_def(ctx, offset + 0xb0, 0x00000001); + gr_def(ctx, offset + 0xb4, 0x00000100); + gr_def(ctx, offset + 0xbc, 0x00000001); + gr_def(ctx, offset + 0xc8, 0x00000100); + gr_def(ctx, offset + 0xcc, 0x00000001); + gr_def(ctx, offset + 0xd0, 0x00000100); + gr_def(ctx, offset + 0xd8, 0x00000001); + gr_def(ctx, offset + 0xe4, 0x00000100); + } + gr_def(ctx, offset + 0xf8, 0x00000004); + gr_def(ctx, offset + 0xfc, 0x00000070); + gr_def(ctx, offset + 0x100, 0x00000080); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + offset += 4; + gr_def(ctx, offset + 0x114, 0x0000000c); + if (dev_priv->chipset == 0x50) + offset -= 4; + gr_def(ctx, offset + 0x11c, 0x00000008); + gr_def(ctx, offset + 0x120, 0x00000014); + if (dev_priv->chipset == 0x50) { + gr_def(ctx, offset + 0x124, 0x00000026); + offset -= 0x18; + } else { + gr_def(ctx, offset + 0x128, 0x00000029); + gr_def(ctx, offset + 0x12c, 0x00000027); + gr_def(ctx, offset + 0x130, 0x00000026); + gr_def(ctx, offset + 0x134, 0x00000008); + gr_def(ctx, offset + 0x138, 0x00000004); + gr_def(ctx, offset + 0x13c, 0x00000027); + } + gr_def(ctx, offset + 0x148, 0x00000001); + gr_def(ctx, offset + 0x14c, 0x00000002); + gr_def(ctx, offset + 0x150, 0x00000003); + gr_def(ctx, offset + 0x154, 0x00000004); + gr_def(ctx, offset + 0x158, 0x00000005); + gr_def(ctx, offset + 0x15c, 0x00000006); + gr_def(ctx, offset + 0x160, 0x00000007); + gr_def(ctx, offset + 0x164, 0x00000001); + gr_def(ctx, offset + 0x1a8, 0x000000cf); + if (dev_priv->chipset == 0x50) + offset -= 4; + gr_def(ctx, offset + 0x1d8, 0x00000080); + gr_def(ctx, offset + 0x1dc, 0x00000004); + gr_def(ctx, offset + 0x1e0, 0x00000004); + if (dev_priv->chipset == 0x50) + offset -= 4; + else + gr_def(ctx, offset + 0x1e4, 0x00000003); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + gr_def(ctx, offset + 0x1ec, 0x00000003); + offset += 8; + } + gr_def(ctx, offset + 0x1e8, 0x00000001); + if (dev_priv->chipset == 0x50) + offset -= 4; + gr_def(ctx, offset + 0x1f4, 0x00000012); + gr_def(ctx, offset + 0x1f8, 0x00000010); + gr_def(ctx, offset + 0x1fc, 0x0000000c); + gr_def(ctx, offset + 0x200, 0x00000001); + gr_def(ctx, offset + 0x210, 0x00000004); + gr_def(ctx, offset + 0x214, 0x00000002); + gr_def(ctx, offset + 0x218, 0x00000004); + if (dev_priv->chipset >= 0xa0) + offset += 4; + gr_def(ctx, offset + 0x224, 0x003fffff); + gr_def(ctx, offset + 0x228, 0x00001fff); + if (dev_priv->chipset == 0x50) + offset -= 0x20; + else if (dev_priv->chipset >= 0xa0) { + gr_def(ctx, offset + 0x250, 0x00000001); + gr_def(ctx, offset + 0x254, 0x00000001); + gr_def(ctx, offset + 0x258, 0x00000002); + offset += 0x10; + } + gr_def(ctx, offset + 0x250, 0x00000004); + gr_def(ctx, offset + 0x254, 0x00000014); + gr_def(ctx, offset + 0x258, 0x00000001); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + offset += 4; + gr_def(ctx, offset + 0x264, 0x00000002); + if (dev_priv->chipset >= 0xa0) + offset += 8; + gr_def(ctx, offset + 0x270, 0x00000001); + gr_def(ctx, offset + 0x278, 0x00000002); + gr_def(ctx, offset + 0x27c, 0x00001000); + if (dev_priv->chipset == 0x50) + offset -= 0xc; + else { + gr_def(ctx, offset + 0x280, 0x00000e00); + gr_def(ctx, offset + 0x284, 0x00001000); + gr_def(ctx, offset + 0x288, 0x00001e00); + } + gr_def(ctx, offset + 0x290, 0x00000001); + gr_def(ctx, offset + 0x294, 0x00000001); + gr_def(ctx, offset + 0x298, 0x00000001); + gr_def(ctx, offset + 0x29c, 0x00000001); + gr_def(ctx, offset + 0x2a0, 0x00000001); + gr_def(ctx, offset + 0x2b0, 0x00000200); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + gr_def(ctx, offset + 0x2b4, 0x00000200); + offset += 4; + } + if (dev_priv->chipset < 0xa0) { + gr_def(ctx, offset + 0x2b8, 0x00000001); + gr_def(ctx, offset + 0x2bc, 0x00000070); + gr_def(ctx, offset + 0x2c0, 0x00000080); + gr_def(ctx, offset + 0x2cc, 0x00000001); + gr_def(ctx, offset + 0x2d0, 0x00000070); + gr_def(ctx, offset + 0x2d4, 0x00000080); + } else { + gr_def(ctx, offset + 0x2b8, 0x00000001); + gr_def(ctx, offset + 0x2bc, 0x000000f0); + gr_def(ctx, offset + 0x2c0, 0x000000ff); + gr_def(ctx, offset + 0x2cc, 0x00000001); + gr_def(ctx, offset + 0x2d0, 0x000000f0); + gr_def(ctx, offset + 0x2d4, 0x000000ff); + gr_def(ctx, offset + 0x2dc, 0x00000009); + offset += 4; + } + gr_def(ctx, offset + 0x2e4, 0x00000001); + gr_def(ctx, offset + 0x2e8, 0x000000cf); + gr_def(ctx, offset + 0x2f0, 0x00000001); + gr_def(ctx, offset + 0x300, 0x000000cf); + gr_def(ctx, offset + 0x308, 0x00000002); + gr_def(ctx, offset + 0x310, 0x00000001); + gr_def(ctx, offset + 0x318, 0x00000001); + gr_def(ctx, offset + 0x320, 0x000000cf); + gr_def(ctx, offset + 0x324, 0x000000cf); + gr_def(ctx, offset + 0x328, 0x00000001); + + /* 6000? */ + if (dev_priv->chipset == 0x50) + cp_ctx(ctx, 0x4063e0, 0x1); + + /* 6800 */ + if (dev_priv->chipset < 0x90) { + cp_ctx(ctx, 0x406814, 0x2b); + gr_def(ctx, 0x406818, 0x00000f80); + gr_def(ctx, 0x406860, 0x007f0080); + gr_def(ctx, 0x40689c, 0x007f0080); + } else { + cp_ctx(ctx, 0x406814, 0x4); + if (dev_priv->chipset == 0x98) + gr_def(ctx, 0x406818, 0x00000f80); + else + gr_def(ctx, 0x406818, 0x00001f80); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + gr_def(ctx, 0x40681c, 0x00000030); + cp_ctx(ctx, 0x406830, 0x3); + } + + /* 7000: per-ROP group state */ + for (i = 0; i < 8; i++) { + if (units & (1<<(i+16))) { + cp_ctx(ctx, 0x407000 + (i<<8), 3); + if (dev_priv->chipset == 0x50) + gr_def(ctx, 0x407000 + (i<<8), 0x1b74f820); + else if (dev_priv->chipset != 0xa5) + gr_def(ctx, 0x407000 + (i<<8), 0x3b74f821); + else + gr_def(ctx, 0x407000 + (i<<8), 0x7b74f821); + gr_def(ctx, 0x407004 + (i<<8), 0x89058001); + + if (dev_priv->chipset == 0x50) { + cp_ctx(ctx, 0x407010 + (i<<8), 1); + } else if (dev_priv->chipset < 0xa0) { + cp_ctx(ctx, 0x407010 + (i<<8), 2); + gr_def(ctx, 0x407010 + (i<<8), 0x00001000); + gr_def(ctx, 0x407014 + (i<<8), 0x0000001f); + } else { + cp_ctx(ctx, 0x407010 + (i<<8), 3); + gr_def(ctx, 0x407010 + (i<<8), 0x00001000); + if (dev_priv->chipset != 0xa5) + gr_def(ctx, 0x407014 + (i<<8), 0x000000ff); + else + gr_def(ctx, 0x407014 + (i<<8), 0x000001ff); + } + + cp_ctx(ctx, 0x407080 + (i<<8), 4); + if (dev_priv->chipset != 0xa5) + gr_def(ctx, 0x407080 + (i<<8), 0x027c10fa); + else + gr_def(ctx, 0x407080 + (i<<8), 0x827c10fa); + if (dev_priv->chipset == 0x50) + gr_def(ctx, 0x407084 + (i<<8), 0x000000c0); + else + gr_def(ctx, 0x407084 + (i<<8), 0x400000c0); + gr_def(ctx, 0x407088 + (i<<8), 0xb7892080); + + if (dev_priv->chipset < 0xa0) + cp_ctx(ctx, 0x407094 + (i<<8), 1); + else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) + cp_ctx(ctx, 0x407094 + (i<<8), 3); + else { + cp_ctx(ctx, 0x407094 + (i<<8), 4); + gr_def(ctx, 0x4070a0 + (i<<8), 1); + } + } + } + + cp_ctx(ctx, 0x407c00, 0x3); + if (dev_priv->chipset < 0x90) + gr_def(ctx, 0x407c00, 0x00010040); + else if (dev_priv->chipset < 0xa0) + gr_def(ctx, 0x407c00, 0x00390040); + else + gr_def(ctx, 0x407c00, 0x003d0040); + gr_def(ctx, 0x407c08, 0x00000022); + if (dev_priv->chipset >= 0xa0) { + cp_ctx(ctx, 0x407c10, 0x3); + cp_ctx(ctx, 0x407c20, 0x1); + cp_ctx(ctx, 0x407c2c, 0x1); + } + + if (dev_priv->chipset < 0xa0) { + cp_ctx(ctx, 0x407d00, 0x9); + } else { + cp_ctx(ctx, 0x407d00, 0x15); + } + if (dev_priv->chipset == 0x98) + gr_def(ctx, 0x407d08, 0x00380040); + else { + if (dev_priv->chipset < 0x90) + gr_def(ctx, 0x407d08, 0x00010040); + else if (dev_priv->chipset < 0xa0) + gr_def(ctx, 0x407d08, 0x00390040); + else + gr_def(ctx, 0x407d08, 0x003d0040); + gr_def(ctx, 0x407d0c, 0x00000022); + } + + /* 8000+: per-TP state */ + for (i = 0; i < 10; i++) { + if (units & (1<<i)) { + if (dev_priv->chipset < 0xa0) + base = 0x408000 + (i<<12); + else + base = 0x408000 + (i<<11); + if (dev_priv->chipset < 0xa0) + offset = base + 0xc00; + else + offset = base + 0x80; + cp_ctx(ctx, offset + 0x00, 1); + gr_def(ctx, offset + 0x00, 0x0000ff0a); + cp_ctx(ctx, offset + 0x08, 1); + + /* per-MP state */ + for (j = 0; j < (dev_priv->chipset < 0xa0 ? 2 : 4); j++) { + if (!(units & (1 << (j+24)))) continue; + if (dev_priv->chipset < 0xa0) + offset = base + 0x200 + (j<<7); + else + offset = base + 0x100 + (j<<7); + cp_ctx(ctx, offset, 0x20); + gr_def(ctx, offset + 0x00, 0x01800000); + gr_def(ctx, offset + 0x04, 0x00160000); + gr_def(ctx, offset + 0x08, 0x01800000); + gr_def(ctx, offset + 0x18, 0x0003ffff); + switch (dev_priv->chipset) { + case 0x50: + gr_def(ctx, offset + 0x1c, 0x00080000); + break; + case 0x84: + gr_def(ctx, offset + 0x1c, 0x00880000); + break; + case 0x86: + gr_def(ctx, offset + 0x1c, 0x008c0000); + break; + case 0x92: + case 0x96: + case 0x98: + gr_def(ctx, offset + 0x1c, 0x118c0000); + break; + case 0x94: + gr_def(ctx, offset + 0x1c, 0x10880000); + break; + case 0xa0: + case 0xa5: + gr_def(ctx, offset + 0x1c, 0x310c0000); + break; + case 0xa8: + case 0xaa: + case 0xac: + gr_def(ctx, offset + 0x1c, 0x300c0000); + break; + } + gr_def(ctx, offset + 0x40, 0x00010401); + if (dev_priv->chipset == 0x50) + gr_def(ctx, offset + 0x48, 0x00000040); + else + gr_def(ctx, offset + 0x48, 0x00000078); + gr_def(ctx, offset + 0x50, 0x000000bf); + gr_def(ctx, offset + 0x58, 0x00001210); + if (dev_priv->chipset == 0x50) + gr_def(ctx, offset + 0x5c, 0x00000080); + else + gr_def(ctx, offset + 0x5c, 0x08000080); + if (dev_priv->chipset >= 0xa0) + gr_def(ctx, offset + 0x68, 0x0000003e); + } + + if (dev_priv->chipset < 0xa0) + cp_ctx(ctx, base + 0x300, 0x4); + else + cp_ctx(ctx, base + 0x300, 0x5); + if (dev_priv->chipset == 0x50) + gr_def(ctx, base + 0x304, 0x00007070); + else if (dev_priv->chipset < 0xa0) + gr_def(ctx, base + 0x304, 0x00027070); + else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) + gr_def(ctx, base + 0x304, 0x01127070); + else + gr_def(ctx, base + 0x304, 0x05127070); + + if (dev_priv->chipset < 0xa0) + cp_ctx(ctx, base + 0x318, 1); + else + cp_ctx(ctx, base + 0x320, 1); + if (dev_priv->chipset == 0x50) + gr_def(ctx, base + 0x318, 0x0003ffff); + else if (dev_priv->chipset < 0xa0) + gr_def(ctx, base + 0x318, 0x03ffffff); + else + gr_def(ctx, base + 0x320, 0x07ffffff); + + if (dev_priv->chipset < 0xa0) + cp_ctx(ctx, base + 0x324, 5); + else + cp_ctx(ctx, base + 0x328, 4); + + if (dev_priv->chipset < 0xa0) { + cp_ctx(ctx, base + 0x340, 9); + offset = base + 0x340; + } else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) { + cp_ctx(ctx, base + 0x33c, 0xb); + offset = base + 0x344; + } else { + cp_ctx(ctx, base + 0x33c, 0xd); + offset = base + 0x344; + } + gr_def(ctx, offset + 0x0, 0x00120407); + gr_def(ctx, offset + 0x4, 0x05091507); + if (dev_priv->chipset == 0x84) + gr_def(ctx, offset + 0x8, 0x05100202); + else + gr_def(ctx, offset + 0x8, 0x05010202); + gr_def(ctx, offset + 0xc, 0x00030201); + + cp_ctx(ctx, base + 0x400, 2); + gr_def(ctx, base + 0x404, 0x00000040); + cp_ctx(ctx, base + 0x40c, 2); + gr_def(ctx, base + 0x40c, 0x0d0c0b0a); + gr_def(ctx, base + 0x410, 0x00141210); + + if (dev_priv->chipset < 0xa0) + offset = base + 0x800; + else + offset = base + 0x500; + cp_ctx(ctx, offset, 6); + gr_def(ctx, offset + 0x0, 0x000001f0); + gr_def(ctx, offset + 0x4, 0x00000001); + gr_def(ctx, offset + 0x8, 0x00000003); + if (dev_priv->chipset == 0x50 || dev_priv->chipset >= 0xaa) + gr_def(ctx, offset + 0xc, 0x00008000); + gr_def(ctx, offset + 0x14, 0x00039e00); + cp_ctx(ctx, offset + 0x1c, 2); + if (dev_priv->chipset == 0x50) + gr_def(ctx, offset + 0x1c, 0x00000040); + else + gr_def(ctx, offset + 0x1c, 0x00000100); + gr_def(ctx, offset + 0x20, 0x00003800); + + if (dev_priv->chipset >= 0xa0) { + cp_ctx(ctx, base + 0x54c, 2); + if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) + gr_def(ctx, base + 0x54c, 0x003fe006); + else + gr_def(ctx, base + 0x54c, 0x003fe007); + gr_def(ctx, base + 0x550, 0x003fe000); + } + + if (dev_priv->chipset < 0xa0) + offset = base + 0xa00; + else + offset = base + 0x680; + cp_ctx(ctx, offset, 1); + gr_def(ctx, offset, 0x00404040); + + if (dev_priv->chipset < 0xa0) + offset = base + 0xe00; + else + offset = base + 0x700; + cp_ctx(ctx, offset, 2); + if (dev_priv->chipset < 0xa0) + gr_def(ctx, offset, 0x0077f005); + else if (dev_priv->chipset == 0xa5) + gr_def(ctx, offset, 0x6cf7f007); + else if (dev_priv->chipset == 0xa8) + gr_def(ctx, offset, 0x6cfff007); + else if (dev_priv->chipset == 0xac) + gr_def(ctx, offset, 0x0cfff007); + else + gr_def(ctx, offset, 0x0cf7f007); + if (dev_priv->chipset == 0x50) + gr_def(ctx, offset + 0x4, 0x00007fff); + else if (dev_priv->chipset < 0xa0) + gr_def(ctx, offset + 0x4, 0x003f7fff); + else + gr_def(ctx, offset + 0x4, 0x02bf7fff); + cp_ctx(ctx, offset + 0x2c, 1); + if (dev_priv->chipset == 0x50) { + cp_ctx(ctx, offset + 0x50, 9); + gr_def(ctx, offset + 0x54, 0x000003ff); + gr_def(ctx, offset + 0x58, 0x00000003); + gr_def(ctx, offset + 0x5c, 0x00000003); + gr_def(ctx, offset + 0x60, 0x000001ff); + gr_def(ctx, offset + 0x64, 0x0000001f); + gr_def(ctx, offset + 0x68, 0x0000000f); + gr_def(ctx, offset + 0x6c, 0x0000000f); + } else if(dev_priv->chipset < 0xa0) { + cp_ctx(ctx, offset + 0x50, 1); + cp_ctx(ctx, offset + 0x70, 1); + } else { + cp_ctx(ctx, offset + 0x50, 1); + cp_ctx(ctx, offset + 0x60, 5); + } + } + } +} + +/* + * xfer areas. These are a pain. + * + * There are 2 xfer areas: the first one is big and contains all sorts of + * stuff, the second is small and contains some per-TP context. + * + * Each area is split into 8 "strands". The areas, when saved to grctx, + * are made of 8-word blocks. Each block contains a single word from + * each strand. The strands are independent of each other, their + * addresses are unrelated to each other, and data in them is closely + * packed together. The strand layout varies a bit between cards: here + * and there, a single word is thrown out in the middle and the whole + * strand is offset by a bit from corresponding one on another chipset. + * For this reason, addresses of stuff in strands are almost useless. + * Knowing sequence of stuff and size of gaps between them is much more + * useful, and that's how we build the strands in our generator. + * + * NVA0 takes this mess to a whole new level by cutting the old strands + * into a few dozen pieces [known as genes], rearranging them randomly, + * and putting them back together to make new strands. Hopefully these + * genes correspond more or less directly to the same PGRAPH subunits + * as in 400040 register. + * + * The most common value in default context is 0, and when the genes + * are separated by 0's, gene bounduaries are quite speculative... + * some of them can be clearly deduced, others can be guessed, and yet + * others won't be resolved without figuring out the real meaning of + * given ctxval. For the same reason, ending point of each strand + * is unknown. Except for strand 0, which is the longest strand and + * its end corresponds to end of the whole xfer. + * + * An unsolved mystery is the seek instruction: it takes an argument + * in bits 8-18, and that argument is clearly the place in strands to + * seek to... but the offsets don't seem to correspond to offsets as + * seen in grctx. Perhaps there's another, real, not randomly-changing + * addressing in strands, and the xfer insn just happens to skip over + * the unused bits? NV10-NV30 PIPE comes to mind... + * + * As far as I know, there's no way to access the xfer areas directly + * without the help of ctxprog. + */ + +static inline void +xf_emit(struct nouveau_grctx *ctx, int num, uint32_t val) { + int i; + if (val && ctx->mode == NOUVEAU_GRCTX_VALS) + for (i = 0; i < num; i++) + nv_wo32(ctx->dev, ctx->data, ctx->ctxvals_pos + (i << 3), val); + ctx->ctxvals_pos += num << 3; +} + +/* Gene declarations... */ + +static void nv50_graph_construct_gene_m2mf(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk1(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk2(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk3(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk4(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk5(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk6(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk7(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk8(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk9(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_unk10(struct nouveau_grctx *ctx); +static void nv50_graph_construct_gene_ropc(struct nouveau_grctx *ctx); +static void nv50_graph_construct_xfer_tp(struct nouveau_grctx *ctx); + +static void +nv50_graph_construct_xfer1(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int i; + int offset; + int size = 0; + uint32_t units = nv_rd32 (ctx->dev, 0x1540); + + offset = (ctx->ctxvals_pos+0x3f)&~0x3f; + ctx->ctxvals_base = offset; + + if (dev_priv->chipset < 0xa0) { + /* Strand 0 */ + ctx->ctxvals_pos = offset; + switch (dev_priv->chipset) { + case 0x50: + xf_emit(ctx, 0x99, 0); + break; + case 0x84: + case 0x86: + xf_emit(ctx, 0x384, 0); + break; + case 0x92: + case 0x94: + case 0x96: + case 0x98: + xf_emit(ctx, 0x380, 0); + break; + } + nv50_graph_construct_gene_m2mf (ctx); + switch (dev_priv->chipset) { + case 0x50: + case 0x84: + case 0x86: + case 0x98: + xf_emit(ctx, 0x4c4, 0); + break; + case 0x92: + case 0x94: + case 0x96: + xf_emit(ctx, 0x984, 0); + break; + } + nv50_graph_construct_gene_unk5(ctx); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 0xa, 0); + else + xf_emit(ctx, 0xb, 0); + nv50_graph_construct_gene_unk4(ctx); + nv50_graph_construct_gene_unk3(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 1 */ + ctx->ctxvals_pos = offset + 0x1; + nv50_graph_construct_gene_unk6(ctx); + nv50_graph_construct_gene_unk7(ctx); + nv50_graph_construct_gene_unk8(ctx); + switch (dev_priv->chipset) { + case 0x50: + case 0x92: + xf_emit(ctx, 0xfb, 0); + break; + case 0x84: + xf_emit(ctx, 0xd3, 0); + break; + case 0x94: + case 0x96: + xf_emit(ctx, 0xab, 0); + break; + case 0x86: + case 0x98: + xf_emit(ctx, 0x6b, 0); + break; + } + xf_emit(ctx, 2, 0x4e3bfdf); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 0xb, 0); + xf_emit(ctx, 2, 0x4e3bfdf); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 2 */ + ctx->ctxvals_pos = offset + 0x2; + switch (dev_priv->chipset) { + case 0x50: + case 0x92: + xf_emit(ctx, 0xa80, 0); + break; + case 0x84: + xf_emit(ctx, 0xa7e, 0); + break; + case 0x94: + case 0x96: + xf_emit(ctx, 0xa7c, 0); + break; + case 0x86: + case 0x98: + xf_emit(ctx, 0xa7a, 0); + break; + } + xf_emit(ctx, 1, 0x3fffff); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x1fff); + xf_emit(ctx, 0xe, 0); + nv50_graph_construct_gene_unk9(ctx); + nv50_graph_construct_gene_unk2(ctx); + nv50_graph_construct_gene_unk1(ctx); + nv50_graph_construct_gene_unk10(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 3: per-ROP group state */ + ctx->ctxvals_pos = offset + 3; + for (i = 0; i < 6; i++) + if (units & (1 << (i + 16))) + nv50_graph_construct_gene_ropc(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strands 4-7: per-TP state */ + for (i = 0; i < 4; i++) { + ctx->ctxvals_pos = offset + 4 + i; + if (units & (1 << (2 * i))) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << (2 * i + 1))) + nv50_graph_construct_xfer_tp(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + } + } else { + /* Strand 0 */ + ctx->ctxvals_pos = offset; + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x385, 0); + else + xf_emit(ctx, 0x384, 0); + nv50_graph_construct_gene_m2mf(ctx); + xf_emit(ctx, 0x950, 0); + nv50_graph_construct_gene_unk10(ctx); + xf_emit(ctx, 1, 0x0fac6881); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 1, 1); + xf_emit(ctx, 3, 0); + } + nv50_graph_construct_gene_unk8(ctx); + if (dev_priv->chipset == 0xa0) + xf_emit(ctx, 0x189, 0); + else if (dev_priv->chipset < 0xa8) + xf_emit(ctx, 0x99, 0); + else if (dev_priv->chipset == 0xaa) + xf_emit(ctx, 0x65, 0); + else + xf_emit(ctx, 0x6d, 0); + nv50_graph_construct_gene_unk9(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 1 */ + ctx->ctxvals_pos = offset + 1; + nv50_graph_construct_gene_unk1(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 2 */ + ctx->ctxvals_pos = offset + 2; + if (dev_priv->chipset == 0xa0) { + nv50_graph_construct_gene_unk2(ctx); + } + xf_emit(ctx, 0x36, 0); + nv50_graph_construct_gene_unk5(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 3 */ + ctx->ctxvals_pos = offset + 3; + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + nv50_graph_construct_gene_unk6(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 4 */ + ctx->ctxvals_pos = offset + 4; + if (dev_priv->chipset == 0xa0) + xf_emit(ctx, 0xa80, 0); + else + xf_emit(ctx, 0xa7a, 0); + xf_emit(ctx, 1, 0x3fffff); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x1fff); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 5 */ + ctx->ctxvals_pos = offset + 5; + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 0xb, 0); + xf_emit(ctx, 2, 0x4e3bfdf); + xf_emit(ctx, 3, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 0x4e3bfdf); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 0); + for (i = 0; i < 8; i++) + if (units & (1<<(i+16))) + nv50_graph_construct_gene_ropc(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 6 */ + ctx->ctxvals_pos = offset + 6; + nv50_graph_construct_gene_unk3(ctx); + xf_emit(ctx, 0xb, 0); + nv50_graph_construct_gene_unk4(ctx); + nv50_graph_construct_gene_unk7(ctx); + if (units & (1 << 0)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 1)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 2)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 3)) + nv50_graph_construct_xfer_tp(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 7 */ + ctx->ctxvals_pos = offset + 7; + if (dev_priv->chipset == 0xa0) { + if (units & (1 << 4)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 5)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 6)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 7)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 8)) + nv50_graph_construct_xfer_tp(ctx); + if (units & (1 << 9)) + nv50_graph_construct_xfer_tp(ctx); + } else { + nv50_graph_construct_gene_unk2(ctx); + } + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + } + + ctx->ctxvals_pos = offset + size * 8; + ctx->ctxvals_pos = (ctx->ctxvals_pos+0x3f)&~0x3f; + cp_lsr (ctx, offset); + cp_out (ctx, CP_SET_XFER_POINTER); + cp_lsr (ctx, size); + cp_out (ctx, CP_SEEK_1); + cp_out (ctx, CP_XFER_1); + cp_wait(ctx, XFER, BUSY); +} + +/* + * non-trivial demagiced parts of ctx init go here + */ + +static void +nv50_graph_construct_gene_m2mf(struct nouveau_grctx *ctx) +{ + /* m2mf state */ + xf_emit (ctx, 1, 0); /* DMA_NOTIFY instance >> 4 */ + xf_emit (ctx, 1, 0); /* DMA_BUFFER_IN instance >> 4 */ + xf_emit (ctx, 1, 0); /* DMA_BUFFER_OUT instance >> 4 */ + xf_emit (ctx, 1, 0); /* OFFSET_IN */ + xf_emit (ctx, 1, 0); /* OFFSET_OUT */ + xf_emit (ctx, 1, 0); /* PITCH_IN */ + xf_emit (ctx, 1, 0); /* PITCH_OUT */ + xf_emit (ctx, 1, 0); /* LINE_LENGTH */ + xf_emit (ctx, 1, 0); /* LINE_COUNT */ + xf_emit (ctx, 1, 0x21); /* FORMAT: bits 0-4 INPUT_INC, bits 5-9 OUTPUT_INC */ + xf_emit (ctx, 1, 1); /* LINEAR_IN */ + xf_emit (ctx, 1, 0x2); /* TILING_MODE_IN: bits 0-2 y tiling, bits 3-5 z tiling */ + xf_emit (ctx, 1, 0x100); /* TILING_PITCH_IN */ + xf_emit (ctx, 1, 0x100); /* TILING_HEIGHT_IN */ + xf_emit (ctx, 1, 1); /* TILING_DEPTH_IN */ + xf_emit (ctx, 1, 0); /* TILING_POSITION_IN_Z */ + xf_emit (ctx, 1, 0); /* TILING_POSITION_IN */ + xf_emit (ctx, 1, 1); /* LINEAR_OUT */ + xf_emit (ctx, 1, 0x2); /* TILING_MODE_OUT: bits 0-2 y tiling, bits 3-5 z tiling */ + xf_emit (ctx, 1, 0x100); /* TILING_PITCH_OUT */ + xf_emit (ctx, 1, 0x100); /* TILING_HEIGHT_OUT */ + xf_emit (ctx, 1, 1); /* TILING_DEPTH_OUT */ + xf_emit (ctx, 1, 0); /* TILING_POSITION_OUT_Z */ + xf_emit (ctx, 1, 0); /* TILING_POSITION_OUT */ + xf_emit (ctx, 1, 0); /* OFFSET_IN_HIGH */ + xf_emit (ctx, 1, 0); /* OFFSET_OUT_HIGH */ +} + +static void +nv50_graph_construct_gene_unk1(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* end of area 2 on pre-NVA0, area 1 on NVAx */ + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0x3ff); + else + xf_emit(ctx, 1, 0x7ff); + switch (dev_priv->chipset) { + case 0x50: + case 0x86: + case 0x98: + case 0xaa: + case 0xac: + xf_emit(ctx, 0x542, 0); + break; + case 0x84: + case 0x92: + case 0x94: + case 0x96: + xf_emit(ctx, 0x942, 0); + break; + case 0xa0: + xf_emit(ctx, 0x2042, 0); + break; + case 0xa5: + case 0xa8: + xf_emit(ctx, 0x842, 0); + break; + } + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x27); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x26); + xf_emit(ctx, 3, 0); +} + +static void +nv50_graph_construct_gene_unk10(struct nouveau_grctx *ctx) +{ + /* end of area 2 on pre-NVA0, area 1 on NVAx */ + xf_emit(ctx, 0x10, 0x04000000); + xf_emit(ctx, 0x24, 0); + xf_emit(ctx, 2, 0x04e3bfdf); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x1fe21); +} + +static void +nv50_graph_construct_gene_unk2(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* middle of area 2 on pre-NVA0, beginning of area 2 on NVA0, area 7 on >NVA0 */ + if (dev_priv->chipset != 0x50) { + xf_emit(ctx, 5, 0); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x804); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0x8100c12); + } + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x10); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 3, 0); + else + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x804); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x1a); + if (dev_priv->chipset != 0x50) + xf_emit(ctx, 1, 0x7f); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 6, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0x3ff); + else + xf_emit(ctx, 1, 0x7ff); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 0x38, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 0x38, 0); + xf_emit(ctx, 2, 0x88); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 0x16, 0); + xf_emit(ctx, 1, 0x26); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x3f800000); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 4, 0); + else + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x1a); + xf_emit(ctx, 1, 0x10); + if (dev_priv->chipset != 0x50) + xf_emit(ctx, 0x28, 0); + else + xf_emit(ctx, 0x25, 0); + xf_emit(ctx, 1, 0x52); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x26); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x1a); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x00ffff00); + xf_emit(ctx, 1, 0); +} + +static void +nv50_graph_construct_gene_unk3(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* end of area 0 on pre-NVA0, beginning of area 6 on NVAx */ + xf_emit(ctx, 1, 0x3f); + xf_emit(ctx, 0xa, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 0x04000000); + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 4); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 0x10, 0); + else + xf_emit(ctx, 0x11, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x1001); + xf_emit(ctx, 4, 0xffff); + xf_emit(ctx, 0x20, 0); + xf_emit(ctx, 0x10, 0x3f800000); + xf_emit(ctx, 1, 0x10); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0); + else + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 3); + xf_emit(ctx, 2, 0); +} + +static void +nv50_graph_construct_gene_unk4(struct nouveau_grctx *ctx) +{ + /* middle of area 0 on pre-NVA0, middle of area 6 on NVAx */ + xf_emit(ctx, 2, 0x04000000); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 1, 0); +} + +static void +nv50_graph_construct_gene_unk5(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* middle of area 0 on pre-NVA0 [after m2mf], end of area 2 on NVAx */ + xf_emit(ctx, 2, 4); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x1c4d, 0); + else + xf_emit(ctx, 0x1c4b, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0x8100c12); + if (dev_priv->chipset != 0x50) + xf_emit(ctx, 1, 3); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0x80c14); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 1, 0x27); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x3c1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x16, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 1, 0); +} + +static void +nv50_graph_construct_gene_unk6(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* beginning of area 1 on pre-NVA0 [after m2mf], area 3 on NVAx */ + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0xf); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 8, 0); + else + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x20); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x11, 0); + else if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 0xf, 0); + else + xf_emit(ctx, 0xe, 0); + xf_emit(ctx, 1, 0x1a); + xf_emit(ctx, 0xd, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 8); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0x3ff); + else + xf_emit(ctx, 1, 0x7ff); + if (dev_priv->chipset == 0xa8) + xf_emit(ctx, 1, 0x1e00); + xf_emit(ctx, 0xc, 0); + xf_emit(ctx, 1, 0xf); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 0x125, 0); + else if (dev_priv->chipset < 0xa0) + xf_emit(ctx, 0x126, 0); + else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa) + xf_emit(ctx, 0x124, 0); + else + xf_emit(ctx, 0x1f7, 0); + xf_emit(ctx, 1, 0xf); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 3, 0); + else + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0xa1, 0); + else + xf_emit(ctx, 0x5a, 0); + xf_emit(ctx, 1, 0xf); + if (dev_priv->chipset < 0xa0) + xf_emit(ctx, 0x834, 0); + else if (dev_priv->chipset == 0xa0) + xf_emit(ctx, 0x1873, 0); + else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x8ba, 0); + else + xf_emit(ctx, 0x833, 0); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 0xf, 0); +} + +static void +nv50_graph_construct_gene_unk7(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* middle of area 1 on pre-NVA0 [after m2mf], middle of area 6 on NVAx */ + xf_emit(ctx, 2, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 2, 1); + else + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0x100); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 8); + xf_emit(ctx, 5, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 3, 1); + xf_emit(ctx, 1, 0xcf); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 6, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 3, 1); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x15); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x4444480); + xf_emit(ctx, 0x37, 0); +} + +static void +nv50_graph_construct_gene_unk8(struct nouveau_grctx *ctx) +{ + /* middle of area 1 on pre-NVA0 [after m2mf], middle of area 0 on NVAx */ + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x100); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x10001); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x10001); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x10001); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 2); +} + +static void +nv50_graph_construct_gene_unk9(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + /* middle of area 2 on pre-NVA0 [after m2mf], end of area 0 on NVAx */ + xf_emit(ctx, 1, 0x3f800000); + xf_emit(ctx, 6, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0x1a); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x12, 0); + xf_emit(ctx, 1, 0x00ffff00); + xf_emit(ctx, 6, 0); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 0xf, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 3); + else if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 0x04000000); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 5); + xf_emit(ctx, 1, 0x52); + if (dev_priv->chipset == 0x50) { + xf_emit(ctx, 0x13, 0); + } else { + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x11, 0); + else + xf_emit(ctx, 0x10, 0); + } + xf_emit(ctx, 0x10, 0x3f800000); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 0x26, 0); + xf_emit(ctx, 1, 0x8100c12); + xf_emit(ctx, 1, 5); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 4, 0xffff); + if (dev_priv->chipset != 0x50) + xf_emit(ctx, 1, 3); + if (dev_priv->chipset < 0xa0) + xf_emit(ctx, 0x1f, 0); + else if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0xc, 0); + else + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x00ffff00); + xf_emit(ctx, 1, 0x1a); + if (dev_priv->chipset != 0x50) { + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 3); + } + if (dev_priv->chipset < 0xa0) + xf_emit(ctx, 0x26, 0); + else + xf_emit(ctx, 0x3c, 0); + xf_emit(ctx, 1, 0x102); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 4, 4); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 8, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0x3ff); + else + xf_emit(ctx, 1, 0x7ff); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x102); + xf_emit(ctx, 9, 0); + xf_emit(ctx, 4, 4); + xf_emit(ctx, 0x2c, 0); +} + +static void +nv50_graph_construct_gene_ropc(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int magic2; + if (dev_priv->chipset == 0x50) { + magic2 = 0x00003e60; + } else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) { + magic2 = 0x001ffe67; + } else { + magic2 = 0x00087e67; + } + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, magic2); + xf_emit(ctx, 4, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 7, 0); + if (dev_priv->chipset >= 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 0x15); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 4, 0); + if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x92 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa0) { + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0x400); + xf_emit(ctx, 1, 0x300); + xf_emit(ctx, 1, 0x1001); + if (dev_priv->chipset != 0xa0) { + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 0); + else + xf_emit(ctx, 1, 0x15); + } + xf_emit(ctx, 3, 0); + } + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x13, 0); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 0x10, 0); + xf_emit(ctx, 0x10, 0x3f800000); + xf_emit(ctx, 0x19, 0); + xf_emit(ctx, 1, 0x10); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x3f); + xf_emit(ctx, 6, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset >= 0xa0) { + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x1001); + xf_emit(ctx, 0xb, 0); + } else { + xf_emit(ctx, 0xc, 0); + } + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x11); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 4, 0); + else + xf_emit(ctx, 6, 0); + xf_emit(ctx, 3, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, magic2); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x0fac6881); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 1, 0); + xf_emit(ctx, 0x18, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 5, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x16, 0); + } else { + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 0x1b, 0); + else + xf_emit(ctx, 0x15, 0); + } + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 1); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 4, 0); + else + xf_emit(ctx, 3, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 0x10, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 0x10, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 3, 0); + } + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x5b, 0); +} + +static void +nv50_graph_construct_xfer_tp_x1(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int magic3; + if (dev_priv->chipset == 0x50) + magic3 = 0x1000; + else if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa8) + magic3 = 0x1e00; + else + magic3 = 0; + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 4); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0x24, 0); + else if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 0x14, 0); + else + xf_emit(ctx, 0x15, 0); + xf_emit(ctx, 2, 4); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 1, 0x03020100); + else + xf_emit(ctx, 1, 0x00608080); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 2, 4); + xf_emit(ctx, 1, 0x80); + if (magic3) + xf_emit(ctx, 1, magic3); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 0x24, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0x03020100); + xf_emit(ctx, 1, 3); + if (magic3) + xf_emit(ctx, 1, magic3); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 3); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 4); + if (dev_priv->chipset == 0x94 || dev_priv->chipset == 0x96) + xf_emit(ctx, 0x1024, 0); + else if (dev_priv->chipset < 0xa0) + xf_emit(ctx, 0xa24, 0); + else if (dev_priv->chipset == 0xa0 || dev_priv->chipset >= 0xaa) + xf_emit(ctx, 0x214, 0); + else + xf_emit(ctx, 0x414, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 3); + xf_emit(ctx, 2, 0); +} + +static void +nv50_graph_construct_xfer_tp_x2(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int magic1, magic2; + if (dev_priv->chipset == 0x50) { + magic1 = 0x3ff; + magic2 = 0x00003e60; + } else if (dev_priv->chipset <= 0xa0 || dev_priv->chipset >= 0xaa) { + magic1 = 0x7ff; + magic2 = 0x001ffe67; + } else { + magic1 = 0x7ff; + magic2 = 0x00087e67; + } + xf_emit(ctx, 3, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0xc, 0); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 0xb, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 4, 0xffff); + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 5, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 1, 3); + xf_emit(ctx, 1, 0); + } else if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0xa, 0); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 1, 0); + xf_emit(ctx, 0x18, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 1, 0); + } + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 3, 0xcf); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0xa, 0); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, magic2); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x11); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 2, 1); + else + xf_emit(ctx, 1, 1); + if(dev_priv->chipset == 0x50) + xf_emit(ctx, 1, 0); + else + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 5, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, magic1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x28, 0); + xf_emit(ctx, 8, 8); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 8, 0x400); + xf_emit(ctx, 8, 0x300); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x20); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 0x100); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x40); + xf_emit(ctx, 1, 0x100); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 3); + xf_emit(ctx, 4, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, magic2); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 9, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x400); + xf_emit(ctx, 1, 0x300); + xf_emit(ctx, 1, 0x1001); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 4, 0); + else + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 1, 0xf); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 0x15, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 3, 0); + } else + xf_emit(ctx, 0x17, 0); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 1, 0x0fac6881); + xf_emit(ctx, 1, magic2); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 3, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 2, 1); + else + xf_emit(ctx, 1, 1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 2, 0); + else if (dev_priv->chipset != 0x50) + xf_emit(ctx, 1, 0); +} + +static void +nv50_graph_construct_xfer_tp_x3(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 2, 0); + else + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0x2a712488); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x4085c000); + xf_emit(ctx, 1, 0x40); + xf_emit(ctx, 1, 0x100); + xf_emit(ctx, 1, 0x10100); + xf_emit(ctx, 1, 0x02800000); +} + +static void +nv50_graph_construct_xfer_tp_x4(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + xf_emit(ctx, 2, 0x04e3bfdf); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x00ffff00); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 2, 1); + else + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 0x00ffff00); + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0x30201000); + xf_emit(ctx, 1, 0x70605040); + xf_emit(ctx, 1, 0xb8a89888); + xf_emit(ctx, 1, 0xf8e8d8c8); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x1a); +} + +static void +nv50_graph_construct_xfer_tp_x5(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 0xfac6881); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 2, 0); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 0xb, 0); + else + xf_emit(ctx, 0xa, 0); + xf_emit(ctx, 8, 1); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0xfac6881); + xf_emit(ctx, 1, 0xf); + xf_emit(ctx, 7, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 1); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 6, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 6, 0); + } else { + xf_emit(ctx, 0xb, 0); + } +} + +static void +nv50_graph_construct_xfer_tp(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + if (dev_priv->chipset < 0xa0) { + nv50_graph_construct_xfer_tp_x1(ctx); + nv50_graph_construct_xfer_tp_x2(ctx); + nv50_graph_construct_xfer_tp_x3(ctx); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 0xf, 0); + else + xf_emit(ctx, 0x12, 0); + nv50_graph_construct_xfer_tp_x4(ctx); + } else { + nv50_graph_construct_xfer_tp_x3(ctx); + if (dev_priv->chipset < 0xaa) + xf_emit(ctx, 0xc, 0); + else + xf_emit(ctx, 0xa, 0); + nv50_graph_construct_xfer_tp_x2(ctx); + nv50_graph_construct_xfer_tp_x5(ctx); + nv50_graph_construct_xfer_tp_x4(ctx); + nv50_graph_construct_xfer_tp_x1(ctx); + } +} + +static void +nv50_graph_construct_xfer_tp2(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int i, mpcnt; + if (dev_priv->chipset == 0x98 || dev_priv->chipset == 0xaa) + mpcnt = 1; + else if (dev_priv->chipset < 0xa0 || dev_priv->chipset >= 0xa8) + mpcnt = 2; + else + mpcnt = 3; + for (i = 0; i < mpcnt; i++) { + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x80); + xf_emit(ctx, 1, 0x80007004); + xf_emit(ctx, 1, 0x04000400); + if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 1, 0xc0); + xf_emit(ctx, 1, 0x1000); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset == 0x86 || dev_priv->chipset == 0x98 || dev_priv->chipset >= 0xa8) { + xf_emit(ctx, 1, 0xe00); + xf_emit(ctx, 1, 0x1e00); + } + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 2, 0x1000); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 2); + if (dev_priv->chipset >= 0xaa) + xf_emit(ctx, 0xb, 0); + else if (dev_priv->chipset >= 0xa0) + xf_emit(ctx, 0xc, 0); + else + xf_emit(ctx, 0xa, 0); + } + xf_emit(ctx, 1, 0x08100c12); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset >= 0xa0) { + xf_emit(ctx, 1, 0x1fe21); + } + xf_emit(ctx, 5, 0); + xf_emit(ctx, 4, 0xffff); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 2, 0x10001); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 0x1fe21); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 1); + xf_emit(ctx, 4, 0); + xf_emit(ctx, 1, 0x08100c12); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 8, 0); + xf_emit(ctx, 1, 0xfac6881); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) + xf_emit(ctx, 1, 3); + xf_emit(ctx, 3, 0); + xf_emit(ctx, 1, 4); + xf_emit(ctx, 9, 0); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 2, 1); + xf_emit(ctx, 1, 2); + xf_emit(ctx, 3, 1); + xf_emit(ctx, 1, 0); + if (dev_priv->chipset > 0xa0 && dev_priv->chipset < 0xaa) { + xf_emit(ctx, 8, 2); + xf_emit(ctx, 0x10, 1); + xf_emit(ctx, 8, 2); + xf_emit(ctx, 0x18, 1); + xf_emit(ctx, 3, 0); + } + xf_emit(ctx, 1, 4); + if (dev_priv->chipset == 0x50) + xf_emit(ctx, 0x3a0, 0); + else if (dev_priv->chipset < 0x94) + xf_emit(ctx, 0x3a2, 0); + else if (dev_priv->chipset == 0x98 || dev_priv->chipset == 0xaa) + xf_emit(ctx, 0x39f, 0); + else + xf_emit(ctx, 0x3a3, 0); + xf_emit(ctx, 1, 0x11); + xf_emit(ctx, 1, 0); + xf_emit(ctx, 1, 1); + xf_emit(ctx, 0x2d, 0); +} + +static void +nv50_graph_construct_xfer2(struct nouveau_grctx *ctx) +{ + struct drm_nouveau_private *dev_priv = ctx->dev->dev_private; + int i; + uint32_t offset; + uint32_t units = nv_rd32 (ctx->dev, 0x1540); + int size = 0; + + offset = (ctx->ctxvals_pos+0x3f)&~0x3f; + + if (dev_priv->chipset < 0xa0) { + for (i = 0; i < 8; i++) { + ctx->ctxvals_pos = offset + i; + if (i == 0) + xf_emit(ctx, 1, 0x08100c12); + if (units & (1 << i)) + nv50_graph_construct_xfer_tp2(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + } + } else { + /* Strand 0: TPs 0, 1 */ + ctx->ctxvals_pos = offset; + xf_emit(ctx, 1, 0x08100c12); + if (units & (1 << 0)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 1)) + nv50_graph_construct_xfer_tp2(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 0: TPs 2, 3 */ + ctx->ctxvals_pos = offset + 1; + if (units & (1 << 2)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 3)) + nv50_graph_construct_xfer_tp2(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 0: TPs 4, 5, 6 */ + ctx->ctxvals_pos = offset + 2; + if (units & (1 << 4)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 5)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 6)) + nv50_graph_construct_xfer_tp2(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + + /* Strand 0: TPs 7, 8, 9 */ + ctx->ctxvals_pos = offset + 3; + if (units & (1 << 7)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 8)) + nv50_graph_construct_xfer_tp2(ctx); + if (units & (1 << 9)) + nv50_graph_construct_xfer_tp2(ctx); + if ((ctx->ctxvals_pos-offset)/8 > size) + size = (ctx->ctxvals_pos-offset)/8; + } + ctx->ctxvals_pos = offset + size * 8; + ctx->ctxvals_pos = (ctx->ctxvals_pos+0x3f)&~0x3f; + cp_lsr (ctx, offset); + cp_out (ctx, CP_SET_XFER_POINTER); + cp_lsr (ctx, size); + cp_out (ctx, CP_SEEK_2); + cp_out (ctx, CP_XFER_2); + cp_wait(ctx, XFER, BUSY); +} diff --git a/drivers/gpu/drm/nouveau/nv50_instmem.c b/drivers/gpu/drm/nouveau/nv50_instmem.c index 94400f777e7..de1f5b0062c 100644 --- a/drivers/gpu/drm/nouveau/nv50_instmem.c +++ b/drivers/gpu/drm/nouveau/nv50_instmem.c @@ -76,6 +76,11 @@ nv50_instmem_init(struct drm_device *dev) for (i = 0x1700; i <= 0x1710; i += 4) priv->save1700[(i-0x1700)/4] = nv_rd32(dev, i); + if (dev_priv->chipset == 0xaa || dev_priv->chipset == 0xac) + dev_priv->vram_sys_base = nv_rd32(dev, 0x100e10) << 12; + else + dev_priv->vram_sys_base = 0; + /* Reserve the last MiB of VRAM, we should probably try to avoid * setting up the below tables over the top of the VBIOS image at * some point. @@ -172,16 +177,28 @@ nv50_instmem_init(struct drm_device *dev) * We map the entire fake channel into the start of the PRAMIN BAR */ ret = nouveau_gpuobj_new_ref(dev, chan, NULL, 0, pt_size, 0x1000, - 0, &priv->pramin_pt); + 0, &priv->pramin_pt); if (ret) return ret; - for (i = 0, v = c_offset; i < pt_size; i += 8, v += 0x1000) { - if (v < (c_offset + c_size)) - BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, v | 1); - else - BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, 0x00000009); + v = c_offset | 1; + if (dev_priv->vram_sys_base) { + v += dev_priv->vram_sys_base; + v |= 0x30; + } + + i = 0; + while (v < dev_priv->vram_sys_base + c_offset + c_size) { + BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, v); + BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, 0x00000000); + v += 0x1000; + i += 8; + } + + while (i < pt_size) { + BAR0_WI32(priv->pramin_pt->gpuobj, i + 0, 0x00000000); BAR0_WI32(priv->pramin_pt->gpuobj, i + 4, 0x00000000); + i += 8; } BAR0_WI32(chan->vm_pd, 0x00, priv->pramin_pt->instance | 0x63); @@ -373,7 +390,7 @@ nv50_instmem_populate(struct drm_device *dev, struct nouveau_gpuobj *gpuobj, if (gpuobj->im_backing) return -EINVAL; - *sz = (*sz + (NV50_INSTMEM_PAGE_SIZE-1)) & ~(NV50_INSTMEM_PAGE_SIZE-1); + *sz = ALIGN(*sz, NV50_INSTMEM_PAGE_SIZE); if (*sz == 0) return -EINVAL; @@ -416,7 +433,9 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nv50_instmem_priv *priv = dev_priv->engine.instmem.priv; - uint32_t pte, pte_end, vram; + struct nouveau_gpuobj *pramin_pt = priv->pramin_pt->gpuobj; + uint32_t pte, pte_end; + uint64_t vram; if (!gpuobj->im_backing || !gpuobj->im_pramin || gpuobj->im_bound) return -EINVAL; @@ -424,20 +443,24 @@ nv50_instmem_bind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) NV_DEBUG(dev, "st=0x%0llx sz=0x%0llx\n", gpuobj->im_pramin->start, gpuobj->im_pramin->size); - pte = (gpuobj->im_pramin->start >> 12) << 3; - pte_end = ((gpuobj->im_pramin->size >> 12) << 3) + pte; + pte = (gpuobj->im_pramin->start >> 12) << 1; + pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte; vram = gpuobj->im_backing_start; NV_DEBUG(dev, "pramin=0x%llx, pte=%d, pte_end=%d\n", gpuobj->im_pramin->start, pte, pte_end); NV_DEBUG(dev, "first vram page: 0x%08x\n", gpuobj->im_backing_start); + vram |= 1; + if (dev_priv->vram_sys_base) { + vram += dev_priv->vram_sys_base; + vram |= 0x30; + } + dev_priv->engine.instmem.prepare_access(dev, true); while (pte < pte_end) { - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 0)/4, vram | 1); - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 4)/4, 0x00000000); - - pte += 8; + nv_wo32(dev, pramin_pt, pte++, lower_32_bits(vram)); + nv_wo32(dev, pramin_pt, pte++, upper_32_bits(vram)); vram += NV50_INSTMEM_PAGE_SIZE; } dev_priv->engine.instmem.finish_access(dev); @@ -470,14 +493,13 @@ nv50_instmem_unbind(struct drm_device *dev, struct nouveau_gpuobj *gpuobj) if (gpuobj->im_bound == 0) return -EINVAL; - pte = (gpuobj->im_pramin->start >> 12) << 3; - pte_end = ((gpuobj->im_pramin->size >> 12) << 3) + pte; + pte = (gpuobj->im_pramin->start >> 12) << 1; + pte_end = ((gpuobj->im_pramin->size >> 12) << 1) + pte; dev_priv->engine.instmem.prepare_access(dev, true); while (pte < pte_end) { - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 0)/4, 0x00000009); - nv_wo32(dev, priv->pramin_pt->gpuobj, (pte + 4)/4, 0x00000000); - pte += 8; + nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000); + nv_wo32(dev, priv->pramin_pt->gpuobj, pte++, 0x00000000); } dev_priv->engine.instmem.finish_access(dev); diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c index e395c16d30f..c2fff543b06 100644 --- a/drivers/gpu/drm/nouveau/nv50_sor.c +++ b/drivers/gpu/drm/nouveau/nv50_sor.c @@ -90,11 +90,25 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode) { struct drm_device *dev = encoder->dev; struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct drm_encoder *enc; uint32_t val; int or = nv_encoder->or; NV_DEBUG_KMS(dev, "or %d mode %d\n", or, mode); + nv_encoder->last_dpms = mode; + list_for_each_entry(enc, &dev->mode_config.encoder_list, head) { + struct nouveau_encoder *nvenc = nouveau_encoder(enc); + + if (nvenc == nv_encoder || + nvenc->disconnect != nv50_sor_disconnect || + nvenc->dcb->or != nv_encoder->dcb->or) + continue; + + if (nvenc->last_dpms == DRM_MODE_DPMS_ON) + return; + } + /* wait for it to be done */ if (!nv_wait(NV50_PDISPLAY_SOR_DPMS_CTRL(or), NV50_PDISPLAY_SOR_DPMS_CTRL_PENDING, 0)) { diff --git a/drivers/gpu/drm/radeon/Kconfig b/drivers/gpu/drm/radeon/Kconfig index 5982321be4d..1c02d23f6fc 100644 --- a/drivers/gpu/drm/radeon/Kconfig +++ b/drivers/gpu/drm/radeon/Kconfig @@ -1,10 +1,14 @@ config DRM_RADEON_KMS - bool "Enable modesetting on radeon by default" + bool "Enable modesetting on radeon by default - NEW DRIVER" depends on DRM_RADEON help - Choose this option if you want kernel modesetting enabled by default, - and you have a new enough userspace to support this. Running old - userspaces with this enabled will cause pain. + Choose this option if you want kernel modesetting enabled by default. + + This is a completely new driver. It's only part of the existing drm + for compatibility reasons. It requires an entirely different graphics + stack above it and works very differently from the old drm stack. + i.e. don't enable this unless you know what you are doing it may + cause issues or bugs compared to the previous userspace driver stack. When kernel modesetting is enabled the IOCTL of radeon/drm driver are considered as invalid and an error message is printed diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index 1cc7b937b1e..ed38262d998 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -30,6 +30,9 @@ $(obj)/r420_reg_safe.h: $(src)/reg_srcs/r420 $(obj)/mkregtable $(obj)/rs600_reg_safe.h: $(src)/reg_srcs/rs600 $(obj)/mkregtable $(call if_changed,mkregtable) +$(obj)/r600_reg_safe.h: $(src)/reg_srcs/r600 $(obj)/mkregtable + $(call if_changed,mkregtable) + $(obj)/r100.o: $(obj)/r100_reg_safe.h $(obj)/rn50_reg_safe.h $(obj)/r200.o: $(obj)/r200_reg_safe.h @@ -42,6 +45,8 @@ $(obj)/r420.o: $(obj)/r420_reg_safe.h $(obj)/rs600.o: $(obj)/rs600_reg_safe.h +$(obj)/r600_cs.o: $(obj)/r600_reg_safe.h + radeon-y := radeon_drv.o radeon_cp.o radeon_state.o radeon_mem.o \ radeon_irq.o r300_cmdbuf.o r600_cp.o # add KMS driver @@ -54,8 +59,10 @@ radeon-y += radeon_device.o radeon_kms.o \ radeon_cs.o radeon_bios.o radeon_benchmark.o r100.o r300.o r420.o \ rs400.o rs600.o rs690.o rv515.o r520.o r600.o rv770.o radeon_test.o \ r200.o radeon_legacy_tv.o r600_cs.o r600_blit.o r600_blit_shaders.o \ - r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o + r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \ + evergreen.o radeon-$(CONFIG_COMPAT) += radeon_ioc32.o +radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o obj-$(CONFIG_DRM_RADEON)+= radeon.o diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c index 388140a7e65..d75788feac6 100644 --- a/drivers/gpu/drm/radeon/atom.c +++ b/drivers/gpu/drm/radeon/atom.c @@ -24,6 +24,7 @@ #include <linux/module.h> #include <linux/sched.h> +#include <asm/unaligned.h> #define ATOM_DEBUG @@ -212,7 +213,9 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr, case ATOM_ARG_PS: idx = U8(*ptr); (*ptr)++; - val = le32_to_cpu(ctx->ps[idx]); + /* get_unaligned_le32 avoids unaligned accesses from atombios + * tables, noticed on a DEC Alpha. */ + val = get_unaligned_le32((u32 *)&ctx->ps[idx]); if (print) DEBUG("PS[0x%02X,0x%04X]", idx, val); break; @@ -246,6 +249,9 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr, case ATOM_WS_ATTRIBUTES: val = gctx->io_attr; break; + case ATOM_WS_REGPTR: + val = gctx->reg_block; + break; default: val = ctx->ws[idx]; } @@ -385,6 +391,32 @@ static uint32_t atom_get_src(atom_exec_context *ctx, uint8_t attr, int *ptr) return atom_get_src_int(ctx, attr, ptr, NULL, 1); } +static uint32_t atom_get_src_direct(atom_exec_context *ctx, uint8_t align, int *ptr) +{ + uint32_t val = 0xCDCDCDCD; + + switch (align) { + case ATOM_SRC_DWORD: + val = U32(*ptr); + (*ptr) += 4; + break; + case ATOM_SRC_WORD0: + case ATOM_SRC_WORD8: + case ATOM_SRC_WORD16: + val = U16(*ptr); + (*ptr) += 2; + break; + case ATOM_SRC_BYTE0: + case ATOM_SRC_BYTE8: + case ATOM_SRC_BYTE16: + case ATOM_SRC_BYTE24: + val = U8(*ptr); + (*ptr)++; + break; + } + return val; +} + static uint32_t atom_get_dst(atom_exec_context *ctx, int arg, uint8_t attr, int *ptr, uint32_t *saved, int print) { @@ -482,6 +514,9 @@ static void atom_put_dst(atom_exec_context *ctx, int arg, uint8_t attr, case ATOM_WS_ATTRIBUTES: gctx->io_attr = val; break; + case ATOM_WS_REGPTR: + gctx->reg_block = val; + break; default: ctx->ws[idx] = val; } @@ -608,7 +643,7 @@ static void atom_op_delay(atom_exec_context *ctx, int *ptr, int arg) uint8_t count = U8((*ptr)++); SDEBUG(" count: %d\n", count); if (arg == ATOM_UNIT_MICROSEC) - schedule_timeout_uninterruptible(usecs_to_jiffies(count)); + udelay(count); else schedule_timeout_uninterruptible(msecs_to_jiffies(count)); } @@ -677,7 +712,7 @@ static void atom_op_mask(atom_exec_context *ctx, int *ptr, int arg) SDEBUG(" dst: "); dst = atom_get_dst(ctx, arg, attr, ptr, &saved, 1); SDEBUG(" src1: "); - src1 = atom_get_src(ctx, attr, ptr); + src1 = atom_get_src_direct(ctx, ((attr >> 3) & 7), ptr); SDEBUG(" src2: "); src2 = atom_get_src(ctx, attr, ptr); dst &= src1; @@ -809,7 +844,7 @@ static void atom_op_setregblock(atom_exec_context *ctx, int *ptr, int arg) SDEBUG(" base: 0x%04X\n", ctx->ctx->reg_block); } -static void atom_op_shl(atom_exec_context *ctx, int *ptr, int arg) +static void atom_op_shift_left(atom_exec_context *ctx, int *ptr, int arg) { uint8_t attr = U8((*ptr)++), shift; uint32_t saved, dst; @@ -818,14 +853,14 @@ static void atom_op_shl(atom_exec_context *ctx, int *ptr, int arg) attr |= atom_def_dst[attr >> 3] << 6; SDEBUG(" dst: "); dst = atom_get_dst(ctx, arg, attr, ptr, &saved, 1); - shift = U8((*ptr)++); + shift = atom_get_src_direct(ctx, ATOM_SRC_BYTE0, ptr); SDEBUG(" shift: %d\n", shift); dst <<= shift; SDEBUG(" dst: "); atom_put_dst(ctx, arg, attr, &dptr, dst, saved); } -static void atom_op_shr(atom_exec_context *ctx, int *ptr, int arg) +static void atom_op_shift_right(atom_exec_context *ctx, int *ptr, int arg) { uint8_t attr = U8((*ptr)++), shift; uint32_t saved, dst; @@ -834,7 +869,35 @@ static void atom_op_shr(atom_exec_context *ctx, int *ptr, int arg) attr |= atom_def_dst[attr >> 3] << 6; SDEBUG(" dst: "); dst = atom_get_dst(ctx, arg, attr, ptr, &saved, 1); - shift = U8((*ptr)++); + shift = atom_get_src_direct(ctx, ATOM_SRC_BYTE0, ptr); + SDEBUG(" shift: %d\n", shift); + dst >>= shift; + SDEBUG(" dst: "); + atom_put_dst(ctx, arg, attr, &dptr, dst, saved); +} + +static void atom_op_shl(atom_exec_context *ctx, int *ptr, int arg) +{ + uint8_t attr = U8((*ptr)++), shift; + uint32_t saved, dst; + int dptr = *ptr; + SDEBUG(" dst: "); + dst = atom_get_dst(ctx, arg, attr, ptr, &saved, 1); + shift = atom_get_src(ctx, attr, ptr); + SDEBUG(" shift: %d\n", shift); + dst <<= shift; + SDEBUG(" dst: "); + atom_put_dst(ctx, arg, attr, &dptr, dst, saved); +} + +static void atom_op_shr(atom_exec_context *ctx, int *ptr, int arg) +{ + uint8_t attr = U8((*ptr)++), shift; + uint32_t saved, dst; + int dptr = *ptr; + SDEBUG(" dst: "); + dst = atom_get_dst(ctx, arg, attr, ptr, &saved, 1); + shift = atom_get_src(ctx, attr, ptr); SDEBUG(" shift: %d\n", shift); dst >>= shift; SDEBUG(" dst: "); @@ -937,18 +1000,18 @@ static struct { atom_op_or, ATOM_ARG_FB}, { atom_op_or, ATOM_ARG_PLL}, { atom_op_or, ATOM_ARG_MC}, { - atom_op_shl, ATOM_ARG_REG}, { - atom_op_shl, ATOM_ARG_PS}, { - atom_op_shl, ATOM_ARG_WS}, { - atom_op_shl, ATOM_ARG_FB}, { - atom_op_shl, ATOM_ARG_PLL}, { - atom_op_shl, ATOM_ARG_MC}, { - atom_op_shr, ATOM_ARG_REG}, { - atom_op_shr, ATOM_ARG_PS}, { - atom_op_shr, ATOM_ARG_WS}, { - atom_op_shr, ATOM_ARG_FB}, { - atom_op_shr, ATOM_ARG_PLL}, { - atom_op_shr, ATOM_ARG_MC}, { + atom_op_shift_left, ATOM_ARG_REG}, { + atom_op_shift_left, ATOM_ARG_PS}, { + atom_op_shift_left, ATOM_ARG_WS}, { + atom_op_shift_left, ATOM_ARG_FB}, { + atom_op_shift_left, ATOM_ARG_PLL}, { + atom_op_shift_left, ATOM_ARG_MC}, { + atom_op_shift_right, ATOM_ARG_REG}, { + atom_op_shift_right, ATOM_ARG_PS}, { + atom_op_shift_right, ATOM_ARG_WS}, { + atom_op_shift_right, ATOM_ARG_FB}, { + atom_op_shift_right, ATOM_ARG_PLL}, { + atom_op_shift_right, ATOM_ARG_MC}, { atom_op_mul, ATOM_ARG_REG}, { atom_op_mul, ATOM_ARG_PS}, { atom_op_mul, ATOM_ARG_WS}, { @@ -1058,8 +1121,6 @@ static void atom_execute_table_locked(struct atom_context *ctx, int index, uint3 SDEBUG(">> execute %04X (len %d, WS %d, PS %d)\n", base, len, ws, ps); - /* reset reg block */ - ctx->reg_block = 0; ectx.ctx = ctx; ectx.ps_shift = ps / 4; ectx.start = base; @@ -1096,6 +1157,12 @@ static void atom_execute_table_locked(struct atom_context *ctx, int index, uint3 void atom_execute_table(struct atom_context *ctx, int index, uint32_t * params) { mutex_lock(&ctx->mutex); + /* reset reg block */ + ctx->reg_block = 0; + /* reset fb window */ + ctx->fb_base = 0; + /* reset io mode */ + ctx->io_mode = ATOM_IO_MM; atom_execute_table_locked(ctx, index, params); mutex_unlock(&ctx->mutex); } diff --git a/drivers/gpu/drm/radeon/atom.h b/drivers/gpu/drm/radeon/atom.h index 47fd943f6d1..bc73781423a 100644 --- a/drivers/gpu/drm/radeon/atom.h +++ b/drivers/gpu/drm/radeon/atom.h @@ -91,6 +91,7 @@ #define ATOM_WS_AND_MASK 0x45 #define ATOM_WS_FB_WINDOW 0x46 #define ATOM_WS_ATTRIBUTES 0x47 +#define ATOM_WS_REGPTR 0x48 #define ATOM_IIO_NOP 0 #define ATOM_IIO_START 1 diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h index 91ad0d1c1b1..6732b5dd8ff 100644 --- a/drivers/gpu/drm/radeon/atombios.h +++ b/drivers/gpu/drm/radeon/atombios.h @@ -1,5 +1,5 @@ /* - * Copyright 2006-2007 Advanced Micro Devices, Inc. + * Copyright 2006-2007 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -20,10 +20,12 @@ * OTHER DEALINGS IN THE SOFTWARE. */ -/****************************************************************************/ + +/****************************************************************************/ /*Portion I: Definitions shared between VBIOS and Driver */ /****************************************************************************/ + #ifndef _ATOMBIOS_H #define _ATOMBIOS_H @@ -40,39 +42,46 @@ #endif #ifdef _H2INC -#ifndef ULONG -typedef unsigned long ULONG; -#endif + #ifndef ULONG + typedef unsigned long ULONG; + #endif -#ifndef UCHAR -typedef unsigned char UCHAR; -#endif + #ifndef UCHAR + typedef unsigned char UCHAR; + #endif -#ifndef USHORT -typedef unsigned short USHORT; -#endif + #ifndef USHORT + typedef unsigned short USHORT; + #endif #endif - -#define ATOM_DAC_A 0 + +#define ATOM_DAC_A 0 #define ATOM_DAC_B 1 #define ATOM_EXT_DAC 2 #define ATOM_CRTC1 0 #define ATOM_CRTC2 1 +#define ATOM_CRTC3 2 +#define ATOM_CRTC4 3 +#define ATOM_CRTC5 4 +#define ATOM_CRTC6 5 +#define ATOM_CRTC_INVALID 0xFF #define ATOM_DIGA 0 #define ATOM_DIGB 1 #define ATOM_PPLL1 0 #define ATOM_PPLL2 1 +#define ATOM_DCPLL 2 +#define ATOM_PPLL_INVALID 0xFF #define ATOM_SCALER1 0 #define ATOM_SCALER2 1 -#define ATOM_SCALER_DISABLE 0 -#define ATOM_SCALER_CENTER 1 -#define ATOM_SCALER_EXPANSION 2 -#define ATOM_SCALER_MULTI_EX 3 +#define ATOM_SCALER_DISABLE 0 +#define ATOM_SCALER_CENTER 1 +#define ATOM_SCALER_EXPANSION 2 +#define ATOM_SCALER_MULTI_EX 3 #define ATOM_DISABLE 0 #define ATOM_ENABLE 1 @@ -82,6 +91,7 @@ typedef unsigned short USHORT; #define ATOM_LCD_SELFTEST_START (ATOM_DISABLE+5) #define ATOM_LCD_SELFTEST_STOP (ATOM_ENABLE+5) #define ATOM_ENCODER_INIT (ATOM_DISABLE+7) +#define ATOM_GET_STATUS (ATOM_DISABLE+8) #define ATOM_BLANKING 1 #define ATOM_BLANKING_OFF 0 @@ -114,7 +124,7 @@ typedef unsigned short USHORT; #define ATOM_DAC2_CV ATOM_DAC1_CV #define ATOM_DAC2_NTSC ATOM_DAC1_NTSC #define ATOM_DAC2_PAL ATOM_DAC1_PAL - + #define ATOM_PM_ON 0 #define ATOM_PM_STANDBY 1 #define ATOM_PM_SUSPEND 2 @@ -134,6 +144,7 @@ typedef unsigned short USHORT; #define ATOM_PANEL_MISC_TEMPORAL 0x00000040 #define ATOM_PANEL_MISC_API_ENABLED 0x00000080 + #define MEMTYPE_DDR1 "DDR1" #define MEMTYPE_DDR2 "DDR2" #define MEMTYPE_DDR3 "DDR3" @@ -145,19 +156,19 @@ typedef unsigned short USHORT; /* Maximum size of that FireGL flag string */ -#define ATOM_FIREGL_FLAG_STRING "FGL" /* Flag used to enable FireGL Support */ -#define ATOM_MAX_SIZE_OF_FIREGL_FLAG_STRING 3 /* sizeof( ATOM_FIREGL_FLAG_STRING ) */ +#define ATOM_FIREGL_FLAG_STRING "FGL" //Flag used to enable FireGL Support +#define ATOM_MAX_SIZE_OF_FIREGL_FLAG_STRING 3 //sizeof( ATOM_FIREGL_FLAG_STRING ) -#define ATOM_FAKE_DESKTOP_STRING "DSK" /* Flag used to enable mobile ASIC on Desktop */ -#define ATOM_MAX_SIZE_OF_FAKE_DESKTOP_STRING ATOM_MAX_SIZE_OF_FIREGL_FLAG_STRING +#define ATOM_FAKE_DESKTOP_STRING "DSK" //Flag used to enable mobile ASIC on Desktop +#define ATOM_MAX_SIZE_OF_FAKE_DESKTOP_STRING ATOM_MAX_SIZE_OF_FIREGL_FLAG_STRING -#define ATOM_M54T_FLAG_STRING "M54T" /* Flag used to enable M54T Support */ -#define ATOM_MAX_SIZE_OF_M54T_FLAG_STRING 4 /* sizeof( ATOM_M54T_FLAG_STRING ) */ +#define ATOM_M54T_FLAG_STRING "M54T" //Flag used to enable M54T Support +#define ATOM_MAX_SIZE_OF_M54T_FLAG_STRING 4 //sizeof( ATOM_M54T_FLAG_STRING ) #define HW_ASSISTED_I2C_STATUS_FAILURE 2 #define HW_ASSISTED_I2C_STATUS_SUCCESS 1 -#pragma pack(1) /* BIOS data must use byte aligment */ +#pragma pack(1) /* BIOS data must use byte aligment */ /* Define offset to location of ROM header. */ @@ -165,367 +176,410 @@ typedef unsigned short USHORT; #define OFFSET_TO_ATOM_ROM_IMAGE_SIZE 0x00000002L #define OFFSET_TO_ATOMBIOS_ASIC_BUS_MEM_TYPE 0x94 -#define MAXSIZE_OF_ATOMBIOS_ASIC_BUS_MEM_TYPE 20 /* including the terminator 0x0! */ +#define MAXSIZE_OF_ATOMBIOS_ASIC_BUS_MEM_TYPE 20 /* including the terminator 0x0! */ #define OFFSET_TO_GET_ATOMBIOS_STRINGS_NUMBER 0x002f #define OFFSET_TO_GET_ATOMBIOS_STRINGS_START 0x006e /* Common header for all ROM Data tables. - Every table pointed _ATOM_MASTER_DATA_TABLE has this common header. + Every table pointed _ATOM_MASTER_DATA_TABLE has this common header. And the pointer actually points to this header. */ -typedef struct _ATOM_COMMON_TABLE_HEADER { - USHORT usStructureSize; - UCHAR ucTableFormatRevision; /*Change it when the Parser is not backward compatible */ - UCHAR ucTableContentRevision; /*Change it only when the table needs to change but the firmware */ - /*Image can't be updated, while Driver needs to carry the new table! */ -} ATOM_COMMON_TABLE_HEADER; - -typedef struct _ATOM_ROM_HEADER { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR uaFirmWareSignature[4]; /*Signature to distinguish between Atombios and non-atombios, - atombios should init it as "ATOM", don't change the position */ - USHORT usBiosRuntimeSegmentAddress; - USHORT usProtectedModeInfoOffset; - USHORT usConfigFilenameOffset; - USHORT usCRC_BlockOffset; - USHORT usBIOS_BootupMessageOffset; - USHORT usInt10Offset; - USHORT usPciBusDevInitCode; - USHORT usIoBaseAddress; - USHORT usSubsystemVendorID; - USHORT usSubsystemID; - USHORT usPCI_InfoOffset; - USHORT usMasterCommandTableOffset; /*Offset for SW to get all command table offsets, Don't change the position */ - USHORT usMasterDataTableOffset; /*Offset for SW to get all data table offsets, Don't change the position */ - UCHAR ucExtendedFunctionCode; - UCHAR ucReserved; -} ATOM_ROM_HEADER; +typedef struct _ATOM_COMMON_TABLE_HEADER +{ + USHORT usStructureSize; + UCHAR ucTableFormatRevision; /*Change it when the Parser is not backward compatible */ + UCHAR ucTableContentRevision; /*Change it only when the table needs to change but the firmware */ + /*Image can't be updated, while Driver needs to carry the new table! */ +}ATOM_COMMON_TABLE_HEADER; + +typedef struct _ATOM_ROM_HEADER +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR uaFirmWareSignature[4]; /*Signature to distinguish between Atombios and non-atombios, + atombios should init it as "ATOM", don't change the position */ + USHORT usBiosRuntimeSegmentAddress; + USHORT usProtectedModeInfoOffset; + USHORT usConfigFilenameOffset; + USHORT usCRC_BlockOffset; + USHORT usBIOS_BootupMessageOffset; + USHORT usInt10Offset; + USHORT usPciBusDevInitCode; + USHORT usIoBaseAddress; + USHORT usSubsystemVendorID; + USHORT usSubsystemID; + USHORT usPCI_InfoOffset; + USHORT usMasterCommandTableOffset; /*Offset for SW to get all command table offsets, Don't change the position */ + USHORT usMasterDataTableOffset; /*Offset for SW to get all data table offsets, Don't change the position */ + UCHAR ucExtendedFunctionCode; + UCHAR ucReserved; +}ATOM_ROM_HEADER; /*==============================Command Table Portion==================================== */ #ifdef UEFI_BUILD -#define UTEMP USHORT -#define USHORT void* + #define UTEMP USHORT + #define USHORT void* #endif -typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES { - USHORT ASIC_Init; /* Function Table, used by various SW components,latest version 1.1 */ - USHORT GetDisplaySurfaceSize; /* Atomic Table, Used by Bios when enabling HW ICON */ - USHORT ASIC_RegistersInit; /* Atomic Table, indirectly used by various SW components,called from ASIC_Init */ - USHORT VRAM_BlockVenderDetection; /* Atomic Table, used only by Bios */ - USHORT DIGxEncoderControl; /* Only used by Bios */ - USHORT MemoryControllerInit; /* Atomic Table, indirectly used by various SW components,called from ASIC_Init */ - USHORT EnableCRTCMemReq; /* Function Table,directly used by various SW components,latest version 2.1 */ - USHORT MemoryParamAdjust; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock if needed */ - USHORT DVOEncoderControl; /* Function Table,directly used by various SW components,latest version 1.2 */ - USHORT GPIOPinControl; /* Atomic Table, only used by Bios */ - USHORT SetEngineClock; /*Function Table,directly used by various SW components,latest version 1.1 */ - USHORT SetMemoryClock; /* Function Table,directly used by various SW components,latest version 1.1 */ - USHORT SetPixelClock; /*Function Table,directly used by various SW components,latest version 1.2 */ - USHORT DynamicClockGating; /* Atomic Table, indirectly used by various SW components,called from ASIC_Init */ - USHORT ResetMemoryDLL; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT ResetMemoryDevice; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT MemoryPLLInit; - USHORT AdjustDisplayPll; /* only used by Bios */ - USHORT AdjustMemoryController; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT EnableASIC_StaticPwrMgt; /* Atomic Table, only used by Bios */ - USHORT ASIC_StaticPwrMgtStatusChange; /* Obsolete, only used by Bios */ - USHORT DAC_LoadDetection; /* Atomic Table, directly used by various SW components,latest version 1.2 */ - USHORT LVTMAEncoderControl; /* Atomic Table,directly used by various SW components,latest version 1.3 */ - USHORT LCD1OutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT DAC1EncoderControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT DAC2EncoderControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT DVOOutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT CV1OutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT GetConditionalGoldenSetting; /* only used by Bios */ - USHORT TVEncoderControl; /* Function Table,directly used by various SW components,latest version 1.1 */ - USHORT TMDSAEncoderControl; /* Atomic Table, directly used by various SW components,latest version 1.3 */ - USHORT LVDSEncoderControl; /* Atomic Table, directly used by various SW components,latest version 1.3 */ - USHORT TV1OutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT EnableScaler; /* Atomic Table, used only by Bios */ - USHORT BlankCRTC; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT EnableCRTC; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT GetPixelClock; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT EnableVGA_Render; /* Function Table,directly used by various SW components,latest version 1.1 */ - USHORT EnableVGA_Access; /* Obsolete , only used by Bios */ - USHORT SetCRTC_Timing; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT SetCRTC_OverScan; /* Atomic Table, used by various SW components,latest version 1.1 */ - USHORT SetCRTC_Replication; /* Atomic Table, used only by Bios */ - USHORT SelectCRTC_Source; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT EnableGraphSurfaces; /* Atomic Table, used only by Bios */ - USHORT UpdateCRTC_DoubleBufferRegisters; - USHORT LUT_AutoFill; /* Atomic Table, only used by Bios */ - USHORT EnableHW_IconCursor; /* Atomic Table, only used by Bios */ - USHORT GetMemoryClock; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT GetEngineClock; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT SetCRTC_UsingDTDTiming; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT ExternalEncoderControl; /* Atomic Table, directly used by various SW components,latest version 2.1 */ - USHORT LVTMAOutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT VRAM_BlockDetectionByStrap; /* Atomic Table, used only by Bios */ - USHORT MemoryCleanUp; /* Atomic Table, only used by Bios */ - USHORT ProcessI2cChannelTransaction; /* Function Table,only used by Bios */ - USHORT WriteOneByteToHWAssistedI2C; /* Function Table,indirectly used by various SW components */ - USHORT ReadHWAssistedI2CStatus; /* Atomic Table, indirectly used by various SW components */ - USHORT SpeedFanControl; /* Function Table,indirectly used by various SW components,called from ASIC_Init */ - USHORT PowerConnectorDetection; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT MC_Synchronization; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT ComputeMemoryEnginePLL; /* Atomic Table, indirectly used by various SW components,called from SetMemory/EngineClock */ - USHORT MemoryRefreshConversion; /* Atomic Table, indirectly used by various SW components,called from SetMemory or SetEngineClock */ - USHORT VRAM_GetCurrentInfoBlock; /* Atomic Table, used only by Bios */ - USHORT DynamicMemorySettings; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT MemoryTraining; /* Atomic Table, used only by Bios */ - USHORT EnableSpreadSpectrumOnPPLL; /* Atomic Table, directly used by various SW components,latest version 1.2 */ - USHORT TMDSAOutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT SetVoltage; /* Function Table,directly and/or indirectly used by various SW components,latest version 1.1 */ - USHORT DAC1OutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT DAC2OutputControl; /* Atomic Table, directly used by various SW components,latest version 1.1 */ - USHORT SetupHWAssistedI2CStatus; /* Function Table,only used by Bios, obsolete soon.Switch to use "ReadEDIDFromHWAssistedI2C" */ - USHORT ClockSource; /* Atomic Table, indirectly used by various SW components,called from ASIC_Init */ - USHORT MemoryDeviceInit; /* Atomic Table, indirectly used by various SW components,called from SetMemoryClock */ - USHORT EnableYUV; /* Atomic Table, indirectly used by various SW components,called from EnableVGARender */ - USHORT DIG1EncoderControl; /* Atomic Table,directly used by various SW components,latest version 1.1 */ - USHORT DIG2EncoderControl; /* Atomic Table,directly used by various SW components,latest version 1.1 */ - USHORT DIG1TransmitterControl; /* Atomic Table,directly used by various SW components,latest version 1.1 */ - USHORT DIG2TransmitterControl; /* Atomic Table,directly used by various SW components,latest version 1.1 */ - USHORT ProcessAuxChannelTransaction; /* Function Table,only used by Bios */ - USHORT DPEncoderService; /* Function Table,only used by Bios */ -} ATOM_MASTER_LIST_OF_COMMAND_TABLES; - -/* For backward compatible */ +typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{ + USHORT ASIC_Init; //Function Table, used by various SW components,latest version 1.1 + USHORT GetDisplaySurfaceSize; //Atomic Table, Used by Bios when enabling HW ICON + USHORT ASIC_RegistersInit; //Atomic Table, indirectly used by various SW components,called from ASIC_Init + USHORT VRAM_BlockVenderDetection; //Atomic Table, used only by Bios + USHORT DIGxEncoderControl; //Only used by Bios + USHORT MemoryControllerInit; //Atomic Table, indirectly used by various SW components,called from ASIC_Init + USHORT EnableCRTCMemReq; //Function Table,directly used by various SW components,latest version 2.1 + USHORT MemoryParamAdjust; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock if needed + USHORT DVOEncoderControl; //Function Table,directly used by various SW components,latest version 1.2 + USHORT GPIOPinControl; //Atomic Table, only used by Bios + USHORT SetEngineClock; //Function Table,directly used by various SW components,latest version 1.1 + USHORT SetMemoryClock; //Function Table,directly used by various SW components,latest version 1.1 + USHORT SetPixelClock; //Function Table,directly used by various SW components,latest version 1.2 + USHORT DynamicClockGating; //Atomic Table, indirectly used by various SW components,called from ASIC_Init + USHORT ResetMemoryDLL; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT ResetMemoryDevice; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT MemoryPLLInit; + USHORT AdjustDisplayPll; //only used by Bios + USHORT AdjustMemoryController; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT EnableASIC_StaticPwrMgt; //Atomic Table, only used by Bios + USHORT ASIC_StaticPwrMgtStatusChange; //Obsolete , only used by Bios + USHORT DAC_LoadDetection; //Atomic Table, directly used by various SW components,latest version 1.2 + USHORT LVTMAEncoderControl; //Atomic Table,directly used by various SW components,latest version 1.3 + USHORT LCD1OutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT DAC1EncoderControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT DAC2EncoderControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT DVOOutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT CV1OutputControl; //Atomic Table, Atomic Table, Obsolete from Ry6xx, use DAC2 Output instead + USHORT GetConditionalGoldenSetting; //only used by Bios + USHORT TVEncoderControl; //Function Table,directly used by various SW components,latest version 1.1 + USHORT TMDSAEncoderControl; //Atomic Table, directly used by various SW components,latest version 1.3 + USHORT LVDSEncoderControl; //Atomic Table, directly used by various SW components,latest version 1.3 + USHORT TV1OutputControl; //Atomic Table, Obsolete from Ry6xx, use DAC2 Output instead + USHORT EnableScaler; //Atomic Table, used only by Bios + USHORT BlankCRTC; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT EnableCRTC; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT GetPixelClock; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT EnableVGA_Render; //Function Table,directly used by various SW components,latest version 1.1 + USHORT GetSCLKOverMCLKRatio; //Atomic Table, only used by Bios + USHORT SetCRTC_Timing; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT SetCRTC_OverScan; //Atomic Table, used by various SW components,latest version 1.1 + USHORT SetCRTC_Replication; //Atomic Table, used only by Bios + USHORT SelectCRTC_Source; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT EnableGraphSurfaces; //Atomic Table, used only by Bios + USHORT UpdateCRTC_DoubleBufferRegisters; + USHORT LUT_AutoFill; //Atomic Table, only used by Bios + USHORT EnableHW_IconCursor; //Atomic Table, only used by Bios + USHORT GetMemoryClock; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT GetEngineClock; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT SetCRTC_UsingDTDTiming; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT ExternalEncoderControl; //Atomic Table, directly used by various SW components,latest version 2.1 + USHORT LVTMAOutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT VRAM_BlockDetectionByStrap; //Atomic Table, used only by Bios + USHORT MemoryCleanUp; //Atomic Table, only used by Bios + USHORT ProcessI2cChannelTransaction; //Function Table,only used by Bios + USHORT WriteOneByteToHWAssistedI2C; //Function Table,indirectly used by various SW components + USHORT ReadHWAssistedI2CStatus; //Atomic Table, indirectly used by various SW components + USHORT SpeedFanControl; //Function Table,indirectly used by various SW components,called from ASIC_Init + USHORT PowerConnectorDetection; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT MC_Synchronization; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT ComputeMemoryEnginePLL; //Atomic Table, indirectly used by various SW components,called from SetMemory/EngineClock + USHORT MemoryRefreshConversion; //Atomic Table, indirectly used by various SW components,called from SetMemory or SetEngineClock + USHORT VRAM_GetCurrentInfoBlock; //Atomic Table, used only by Bios + USHORT DynamicMemorySettings; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT MemoryTraining; //Atomic Table, used only by Bios + USHORT EnableSpreadSpectrumOnPPLL; //Atomic Table, directly used by various SW components,latest version 1.2 + USHORT TMDSAOutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT SetVoltage; //Function Table,directly and/or indirectly used by various SW components,latest version 1.1 + USHORT DAC1OutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT DAC2OutputControl; //Atomic Table, directly used by various SW components,latest version 1.1 + USHORT SetupHWAssistedI2CStatus; //Function Table,only used by Bios, obsolete soon.Switch to use "ReadEDIDFromHWAssistedI2C" + USHORT ClockSource; //Atomic Table, indirectly used by various SW components,called from ASIC_Init + USHORT MemoryDeviceInit; //Atomic Table, indirectly used by various SW components,called from SetMemoryClock + USHORT EnableYUV; //Atomic Table, indirectly used by various SW components,called from EnableVGARender + USHORT DIG1EncoderControl; //Atomic Table,directly used by various SW components,latest version 1.1 + USHORT DIG2EncoderControl; //Atomic Table,directly used by various SW components,latest version 1.1 + USHORT DIG1TransmitterControl; //Atomic Table,directly used by various SW components,latest version 1.1 + USHORT DIG2TransmitterControl; //Atomic Table,directly used by various SW components,latest version 1.1 + USHORT ProcessAuxChannelTransaction; //Function Table,only used by Bios + USHORT DPEncoderService; //Function Table,only used by Bios +}ATOM_MASTER_LIST_OF_COMMAND_TABLES; + +// For backward compatible #define ReadEDIDFromHWAssistedI2C ProcessI2cChannelTransaction #define UNIPHYTransmitterControl DIG1TransmitterControl #define LVTMATransmitterControl DIG2TransmitterControl #define SetCRTC_DPM_State GetConditionalGoldenSetting #define SetUniphyInstance ASIC_StaticPwrMgtStatusChange +#define HPDInterruptService ReadHWAssistedI2CStatus +#define EnableVGA_Access GetSCLKOverMCLKRatio -typedef struct _ATOM_MASTER_COMMAND_TABLE { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_MASTER_LIST_OF_COMMAND_TABLES ListOfCommandTables; -} ATOM_MASTER_COMMAND_TABLE; - -/****************************************************************************/ -/* Structures used in every command table */ -/****************************************************************************/ -typedef struct _ATOM_TABLE_ATTRIBUTE { +typedef struct _ATOM_MASTER_COMMAND_TABLE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_MASTER_LIST_OF_COMMAND_TABLES ListOfCommandTables; +}ATOM_MASTER_COMMAND_TABLE; + +/****************************************************************************/ +// Structures used in every command table +/****************************************************************************/ +typedef struct _ATOM_TABLE_ATTRIBUTE +{ #if ATOM_BIG_ENDIAN - USHORT UpdatedByUtility:1; /* [15]=Table updated by utility flag */ - USHORT PS_SizeInBytes:7; /* [14:8]=Size of parameter space in Bytes (multiple of a dword), */ - USHORT WS_SizeInBytes:8; /* [7:0]=Size of workspace in Bytes (in multiple of a dword), */ + USHORT UpdatedByUtility:1; //[15]=Table updated by utility flag + USHORT PS_SizeInBytes:7; //[14:8]=Size of parameter space in Bytes (multiple of a dword), + USHORT WS_SizeInBytes:8; //[7:0]=Size of workspace in Bytes (in multiple of a dword), #else - USHORT WS_SizeInBytes:8; /* [7:0]=Size of workspace in Bytes (in multiple of a dword), */ - USHORT PS_SizeInBytes:7; /* [14:8]=Size of parameter space in Bytes (multiple of a dword), */ - USHORT UpdatedByUtility:1; /* [15]=Table updated by utility flag */ + USHORT WS_SizeInBytes:8; //[7:0]=Size of workspace in Bytes (in multiple of a dword), + USHORT PS_SizeInBytes:7; //[14:8]=Size of parameter space in Bytes (multiple of a dword), + USHORT UpdatedByUtility:1; //[15]=Table updated by utility flag #endif -} ATOM_TABLE_ATTRIBUTE; - -typedef union _ATOM_TABLE_ATTRIBUTE_ACCESS { - ATOM_TABLE_ATTRIBUTE sbfAccess; - USHORT susAccess; -} ATOM_TABLE_ATTRIBUTE_ACCESS; +}ATOM_TABLE_ATTRIBUTE; -/****************************************************************************/ -/* Common header for all command tables. */ -/* Every table pointed by _ATOM_MASTER_COMMAND_TABLE has this common header. */ -/* And the pointer actually points to this header. */ -/****************************************************************************/ -typedef struct _ATOM_COMMON_ROM_COMMAND_TABLE_HEADER { - ATOM_COMMON_TABLE_HEADER CommonHeader; - ATOM_TABLE_ATTRIBUTE TableAttribute; -} ATOM_COMMON_ROM_COMMAND_TABLE_HEADER; +typedef union _ATOM_TABLE_ATTRIBUTE_ACCESS +{ + ATOM_TABLE_ATTRIBUTE sbfAccess; + USHORT susAccess; +}ATOM_TABLE_ATTRIBUTE_ACCESS; + +/****************************************************************************/ +// Common header for all command tables. +// Every table pointed by _ATOM_MASTER_COMMAND_TABLE has this common header. +// And the pointer actually points to this header. +/****************************************************************************/ +typedef struct _ATOM_COMMON_ROM_COMMAND_TABLE_HEADER +{ + ATOM_COMMON_TABLE_HEADER CommonHeader; + ATOM_TABLE_ATTRIBUTE TableAttribute; +}ATOM_COMMON_ROM_COMMAND_TABLE_HEADER; -/****************************************************************************/ -/* Structures used by ComputeMemoryEnginePLLTable */ -/****************************************************************************/ +/****************************************************************************/ +// Structures used by ComputeMemoryEnginePLLTable +/****************************************************************************/ #define COMPUTE_MEMORY_PLL_PARAM 1 #define COMPUTE_ENGINE_PLL_PARAM 2 -typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS { - ULONG ulClock; /* When returen, it's the re-calculated clock based on given Fb_div Post_Div and ref_div */ - UCHAR ucAction; /* 0:reserved //1:Memory //2:Engine */ - UCHAR ucReserved; /* may expand to return larger Fbdiv later */ - UCHAR ucFbDiv; /* return value */ - UCHAR ucPostDiv; /* return value */ -} COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS; - -typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V2 { - ULONG ulClock; /* When return, [23:0] return real clock */ - UCHAR ucAction; /* 0:reserved;COMPUTE_MEMORY_PLL_PARAM:Memory;COMPUTE_ENGINE_PLL_PARAM:Engine. it return ref_div to be written to register */ - USHORT usFbDiv; /* return Feedback value to be written to register */ - UCHAR ucPostDiv; /* return post div to be written to register */ -} COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V2; +typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS +{ + ULONG ulClock; //When returen, it's the re-calculated clock based on given Fb_div Post_Div and ref_div + UCHAR ucAction; //0:reserved //1:Memory //2:Engine + UCHAR ucReserved; //may expand to return larger Fbdiv later + UCHAR ucFbDiv; //return value + UCHAR ucPostDiv; //return value +}COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS; + +typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V2 +{ + ULONG ulClock; //When return, [23:0] return real clock + UCHAR ucAction; //0:reserved;COMPUTE_MEMORY_PLL_PARAM:Memory;COMPUTE_ENGINE_PLL_PARAM:Engine. it return ref_div to be written to register + USHORT usFbDiv; //return Feedback value to be written to register + UCHAR ucPostDiv; //return post div to be written to register +}COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V2; #define COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_PS_ALLOCATION COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS -#define SET_CLOCK_FREQ_MASK 0x00FFFFFF /* Clock change tables only take bit [23:0] as the requested clock value */ -#define USE_NON_BUS_CLOCK_MASK 0x01000000 /* Applicable to both memory and engine clock change, when set, it uses another clock as the temporary clock (engine uses memory and vice versa) */ -#define USE_MEMORY_SELF_REFRESH_MASK 0x02000000 /* Only applicable to memory clock change, when set, using memory self refresh during clock transition */ -#define SKIP_INTERNAL_MEMORY_PARAMETER_CHANGE 0x04000000 /* Only applicable to memory clock change, when set, the table will skip predefined internal memory parameter change */ -#define FIRST_TIME_CHANGE_CLOCK 0x08000000 /* Applicable to both memory and engine clock change,when set, it means this is 1st time to change clock after ASIC bootup */ -#define SKIP_SW_PROGRAM_PLL 0x10000000 /* Applicable to both memory and engine clock change, when set, it means the table will not program SPLL/MPLL */ + +#define SET_CLOCK_FREQ_MASK 0x00FFFFFF //Clock change tables only take bit [23:0] as the requested clock value +#define USE_NON_BUS_CLOCK_MASK 0x01000000 //Applicable to both memory and engine clock change, when set, it uses another clock as the temporary clock (engine uses memory and vice versa) +#define USE_MEMORY_SELF_REFRESH_MASK 0x02000000 //Only applicable to memory clock change, when set, using memory self refresh during clock transition +#define SKIP_INTERNAL_MEMORY_PARAMETER_CHANGE 0x04000000 //Only applicable to memory clock change, when set, the table will skip predefined internal memory parameter change +#define FIRST_TIME_CHANGE_CLOCK 0x08000000 //Applicable to both memory and engine clock change,when set, it means this is 1st time to change clock after ASIC bootup +#define SKIP_SW_PROGRAM_PLL 0x10000000 //Applicable to both memory and engine clock change, when set, it means the table will not program SPLL/MPLL #define USE_SS_ENABLED_PIXEL_CLOCK USE_NON_BUS_CLOCK_MASK -#define b3USE_NON_BUS_CLOCK_MASK 0x01 /* Applicable to both memory and engine clock change, when set, it uses another clock as the temporary clock (engine uses memory and vice versa) */ -#define b3USE_MEMORY_SELF_REFRESH 0x02 /* Only applicable to memory clock change, when set, using memory self refresh during clock transition */ -#define b3SKIP_INTERNAL_MEMORY_PARAMETER_CHANGE 0x04 /* Only applicable to memory clock change, when set, the table will skip predefined internal memory parameter change */ -#define b3FIRST_TIME_CHANGE_CLOCK 0x08 /* Applicable to both memory and engine clock change,when set, it means this is 1st time to change clock after ASIC bootup */ -#define b3SKIP_SW_PROGRAM_PLL 0x10 /* Applicable to both memory and engine clock change, when set, it means the table will not program SPLL/MPLL */ +#define b3USE_NON_BUS_CLOCK_MASK 0x01 //Applicable to both memory and engine clock change, when set, it uses another clock as the temporary clock (engine uses memory and vice versa) +#define b3USE_MEMORY_SELF_REFRESH 0x02 //Only applicable to memory clock change, when set, using memory self refresh during clock transition +#define b3SKIP_INTERNAL_MEMORY_PARAMETER_CHANGE 0x04 //Only applicable to memory clock change, when set, the table will skip predefined internal memory parameter change +#define b3FIRST_TIME_CHANGE_CLOCK 0x08 //Applicable to both memory and engine clock change,when set, it means this is 1st time to change clock after ASIC bootup +#define b3SKIP_SW_PROGRAM_PLL 0x10 //Applicable to both memory and engine clock change, when set, it means the table will not program SPLL/MPLL -typedef struct _ATOM_COMPUTE_CLOCK_FREQ { +typedef struct _ATOM_COMPUTE_CLOCK_FREQ +{ #if ATOM_BIG_ENDIAN - ULONG ulComputeClockFlag:8; /* =1: COMPUTE_MEMORY_PLL_PARAM, =2: COMPUTE_ENGINE_PLL_PARAM */ - ULONG ulClockFreq:24; /* in unit of 10kHz */ + ULONG ulComputeClockFlag:8; // =1: COMPUTE_MEMORY_PLL_PARAM, =2: COMPUTE_ENGINE_PLL_PARAM + ULONG ulClockFreq:24; // in unit of 10kHz #else - ULONG ulClockFreq:24; /* in unit of 10kHz */ - ULONG ulComputeClockFlag:8; /* =1: COMPUTE_MEMORY_PLL_PARAM, =2: COMPUTE_ENGINE_PLL_PARAM */ + ULONG ulClockFreq:24; // in unit of 10kHz + ULONG ulComputeClockFlag:8; // =1: COMPUTE_MEMORY_PLL_PARAM, =2: COMPUTE_ENGINE_PLL_PARAM #endif -} ATOM_COMPUTE_CLOCK_FREQ; - -typedef struct _ATOM_S_MPLL_FB_DIVIDER { - USHORT usFbDivFrac; - USHORT usFbDiv; -} ATOM_S_MPLL_FB_DIVIDER; +}ATOM_COMPUTE_CLOCK_FREQ; -typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 { - union { - ATOM_COMPUTE_CLOCK_FREQ ulClock; /* Input Parameter */ - ATOM_S_MPLL_FB_DIVIDER ulFbDiv; /* Output Parameter */ - }; - UCHAR ucRefDiv; /* Output Parameter */ - UCHAR ucPostDiv; /* Output Parameter */ - UCHAR ucCntlFlag; /* Output Parameter */ - UCHAR ucReserved; -} COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3; +typedef struct _ATOM_S_MPLL_FB_DIVIDER +{ + USHORT usFbDivFrac; + USHORT usFbDiv; +}ATOM_S_MPLL_FB_DIVIDER; -/* ucCntlFlag */ +typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 +{ + union + { + ATOM_COMPUTE_CLOCK_FREQ ulClock; //Input Parameter + ATOM_S_MPLL_FB_DIVIDER ulFbDiv; //Output Parameter + }; + UCHAR ucRefDiv; //Output Parameter + UCHAR ucPostDiv; //Output Parameter + UCHAR ucCntlFlag; //Output Parameter + UCHAR ucReserved; +}COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3; + +// ucCntlFlag #define ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN 1 #define ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE 2 #define ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE 4 +#define ATOM_PLL_CNTL_FLAG_SPLL_ISPARE_9 8 -typedef struct _DYNAMICE_MEMORY_SETTINGS_PARAMETER { - ATOM_COMPUTE_CLOCK_FREQ ulClock; - ULONG ulReserved[2]; -} DYNAMICE_MEMORY_SETTINGS_PARAMETER; - -typedef struct _DYNAMICE_ENGINE_SETTINGS_PARAMETER { - ATOM_COMPUTE_CLOCK_FREQ ulClock; - ULONG ulMemoryClock; - ULONG ulReserved; -} DYNAMICE_ENGINE_SETTINGS_PARAMETER; - -/****************************************************************************/ -/* Structures used by SetEngineClockTable */ -/****************************************************************************/ -typedef struct _SET_ENGINE_CLOCK_PARAMETERS { - ULONG ulTargetEngineClock; /* In 10Khz unit */ -} SET_ENGINE_CLOCK_PARAMETERS; -typedef struct _SET_ENGINE_CLOCK_PS_ALLOCATION { - ULONG ulTargetEngineClock; /* In 10Khz unit */ - COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_PS_ALLOCATION sReserved; -} SET_ENGINE_CLOCK_PS_ALLOCATION; +// V4 are only used for APU which PLL outside GPU +typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4 +{ +#if ATOM_BIG_ENDIAN + ULONG ucPostDiv; //return parameter: post divider which is used to program to register directly + ULONG ulClock:24; //Input= target clock, output = actual clock +#else + ULONG ulClock:24; //Input= target clock, output = actual clock + ULONG ucPostDiv; //return parameter: post divider which is used to program to register directly +#endif +}COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4; -/****************************************************************************/ -/* Structures used by SetMemoryClockTable */ -/****************************************************************************/ -typedef struct _SET_MEMORY_CLOCK_PARAMETERS { - ULONG ulTargetMemoryClock; /* In 10Khz unit */ -} SET_MEMORY_CLOCK_PARAMETERS; +typedef struct _DYNAMICE_MEMORY_SETTINGS_PARAMETER +{ + ATOM_COMPUTE_CLOCK_FREQ ulClock; + ULONG ulReserved[2]; +}DYNAMICE_MEMORY_SETTINGS_PARAMETER; -typedef struct _SET_MEMORY_CLOCK_PS_ALLOCATION { - ULONG ulTargetMemoryClock; /* In 10Khz unit */ - COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_PS_ALLOCATION sReserved; -} SET_MEMORY_CLOCK_PS_ALLOCATION; +typedef struct _DYNAMICE_ENGINE_SETTINGS_PARAMETER +{ + ATOM_COMPUTE_CLOCK_FREQ ulClock; + ULONG ulMemoryClock; + ULONG ulReserved; +}DYNAMICE_ENGINE_SETTINGS_PARAMETER; + +/****************************************************************************/ +// Structures used by SetEngineClockTable +/****************************************************************************/ +typedef struct _SET_ENGINE_CLOCK_PARAMETERS +{ + ULONG ulTargetEngineClock; //In 10Khz unit +}SET_ENGINE_CLOCK_PARAMETERS; -/****************************************************************************/ -/* Structures used by ASIC_Init.ctb */ -/****************************************************************************/ -typedef struct _ASIC_INIT_PARAMETERS { - ULONG ulDefaultEngineClock; /* In 10Khz unit */ - ULONG ulDefaultMemoryClock; /* In 10Khz unit */ -} ASIC_INIT_PARAMETERS; +typedef struct _SET_ENGINE_CLOCK_PS_ALLOCATION +{ + ULONG ulTargetEngineClock; //In 10Khz unit + COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_PS_ALLOCATION sReserved; +}SET_ENGINE_CLOCK_PS_ALLOCATION; + +/****************************************************************************/ +// Structures used by SetMemoryClockTable +/****************************************************************************/ +typedef struct _SET_MEMORY_CLOCK_PARAMETERS +{ + ULONG ulTargetMemoryClock; //In 10Khz unit +}SET_MEMORY_CLOCK_PARAMETERS; -typedef struct _ASIC_INIT_PS_ALLOCATION { - ASIC_INIT_PARAMETERS sASICInitClocks; - SET_ENGINE_CLOCK_PS_ALLOCATION sReserved; /* Caller doesn't need to init this structure */ -} ASIC_INIT_PS_ALLOCATION; +typedef struct _SET_MEMORY_CLOCK_PS_ALLOCATION +{ + ULONG ulTargetMemoryClock; //In 10Khz unit + COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_PS_ALLOCATION sReserved; +}SET_MEMORY_CLOCK_PS_ALLOCATION; + +/****************************************************************************/ +// Structures used by ASIC_Init.ctb +/****************************************************************************/ +typedef struct _ASIC_INIT_PARAMETERS +{ + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit +}ASIC_INIT_PARAMETERS; -/****************************************************************************/ -/* Structure used by DynamicClockGatingTable.ctb */ -/****************************************************************************/ -typedef struct _DYNAMIC_CLOCK_GATING_PARAMETERS { - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucPadding[3]; -} DYNAMIC_CLOCK_GATING_PARAMETERS; +typedef struct _ASIC_INIT_PS_ALLOCATION +{ + ASIC_INIT_PARAMETERS sASICInitClocks; + SET_ENGINE_CLOCK_PS_ALLOCATION sReserved; //Caller doesn't need to init this structure +}ASIC_INIT_PS_ALLOCATION; + +/****************************************************************************/ +// Structure used by DynamicClockGatingTable.ctb +/****************************************************************************/ +typedef struct _DYNAMIC_CLOCK_GATING_PARAMETERS +{ + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + UCHAR ucPadding[3]; +}DYNAMIC_CLOCK_GATING_PARAMETERS; #define DYNAMIC_CLOCK_GATING_PS_ALLOCATION DYNAMIC_CLOCK_GATING_PARAMETERS -/****************************************************************************/ -/* Structure used by EnableASIC_StaticPwrMgtTable.ctb */ -/****************************************************************************/ -typedef struct _ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS { - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucPadding[3]; -} ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS; +/****************************************************************************/ +// Structure used by EnableASIC_StaticPwrMgtTable.ctb +/****************************************************************************/ +typedef struct _ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS +{ + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + UCHAR ucPadding[3]; +}ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS; #define ENABLE_ASIC_STATIC_PWR_MGT_PS_ALLOCATION ENABLE_ASIC_STATIC_PWR_MGT_PARAMETERS -/****************************************************************************/ -/* Structures used by DAC_LoadDetectionTable.ctb */ -/****************************************************************************/ -typedef struct _DAC_LOAD_DETECTION_PARAMETERS { - USHORT usDeviceID; /* {ATOM_DEVICE_CRTx_SUPPORT,ATOM_DEVICE_TVx_SUPPORT,ATOM_DEVICE_CVx_SUPPORT} */ - UCHAR ucDacType; /* {ATOM_DAC_A,ATOM_DAC_B, ATOM_EXT_DAC} */ - UCHAR ucMisc; /* Valid only when table revision =1.3 and above */ -} DAC_LOAD_DETECTION_PARAMETERS; +/****************************************************************************/ +// Structures used by DAC_LoadDetectionTable.ctb +/****************************************************************************/ +typedef struct _DAC_LOAD_DETECTION_PARAMETERS +{ + USHORT usDeviceID; //{ATOM_DEVICE_CRTx_SUPPORT,ATOM_DEVICE_TVx_SUPPORT,ATOM_DEVICE_CVx_SUPPORT} + UCHAR ucDacType; //{ATOM_DAC_A,ATOM_DAC_B, ATOM_EXT_DAC} + UCHAR ucMisc; //Valid only when table revision =1.3 and above +}DAC_LOAD_DETECTION_PARAMETERS; -/* DAC_LOAD_DETECTION_PARAMETERS.ucMisc */ +// DAC_LOAD_DETECTION_PARAMETERS.ucMisc #define DAC_LOAD_MISC_YPrPb 0x01 -typedef struct _DAC_LOAD_DETECTION_PS_ALLOCATION { - DAC_LOAD_DETECTION_PARAMETERS sDacload; - ULONG Reserved[2]; /* Don't set this one, allocation for EXT DAC */ -} DAC_LOAD_DETECTION_PS_ALLOCATION; - -/****************************************************************************/ -/* Structures used by DAC1EncoderControlTable.ctb and DAC2EncoderControlTable.ctb */ -/****************************************************************************/ -typedef struct _DAC_ENCODER_CONTROL_PARAMETERS { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - UCHAR ucDacStandard; /* See definition of ATOM_DACx_xxx, For DEC3.0, bit 7 used as internal flag to indicate DAC2 (==1) or DAC1 (==0) */ - UCHAR ucAction; /* 0: turn off encoder */ - /* 1: setup and turn on encoder */ - /* 7: ATOM_ENCODER_INIT Initialize DAC */ -} DAC_ENCODER_CONTROL_PARAMETERS; +typedef struct _DAC_LOAD_DETECTION_PS_ALLOCATION +{ + DAC_LOAD_DETECTION_PARAMETERS sDacload; + ULONG Reserved[2];// Don't set this one, allocation for EXT DAC +}DAC_LOAD_DETECTION_PS_ALLOCATION; + +/****************************************************************************/ +// Structures used by DAC1EncoderControlTable.ctb and DAC2EncoderControlTable.ctb +/****************************************************************************/ +typedef struct _DAC_ENCODER_CONTROL_PARAMETERS +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + UCHAR ucDacStandard; // See definition of ATOM_DACx_xxx, For DEC3.0, bit 7 used as internal flag to indicate DAC2 (==1) or DAC1 (==0) + UCHAR ucAction; // 0: turn off encoder + // 1: setup and turn on encoder + // 7: ATOM_ENCODER_INIT Initialize DAC +}DAC_ENCODER_CONTROL_PARAMETERS; #define DAC_ENCODER_CONTROL_PS_ALLOCATION DAC_ENCODER_CONTROL_PARAMETERS -/****************************************************************************/ -/* Structures used by DIG1EncoderControlTable */ -/* DIG2EncoderControlTable */ -/* ExternalEncoderControlTable */ -/****************************************************************************/ -typedef struct _DIG_ENCODER_CONTROL_PARAMETERS { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - UCHAR ucConfig; - /* [2] Link Select: */ - /* =0: PHY linkA if bfLane<3 */ - /* =1: PHY linkB if bfLanes<3 */ - /* =0: PHY linkA+B if bfLanes=3 */ - /* [3] Transmitter Sel */ - /* =0: UNIPHY or PCIEPHY */ - /* =1: LVTMA */ - UCHAR ucAction; /* =0: turn off encoder */ - /* =1: turn on encoder */ - UCHAR ucEncoderMode; - /* =0: DP encoder */ - /* =1: LVDS encoder */ - /* =2: DVI encoder */ - /* =3: HDMI encoder */ - /* =4: SDVO encoder */ - UCHAR ucLaneNum; /* how many lanes to enable */ - UCHAR ucReserved[2]; -} DIG_ENCODER_CONTROL_PARAMETERS; +/****************************************************************************/ +// Structures used by DIG1EncoderControlTable +// DIG2EncoderControlTable +// ExternalEncoderControlTable +/****************************************************************************/ +typedef struct _DIG_ENCODER_CONTROL_PARAMETERS +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + UCHAR ucConfig; + // [2] Link Select: + // =0: PHY linkA if bfLane<3 + // =1: PHY linkB if bfLanes<3 + // =0: PHY linkA+B if bfLanes=3 + // [3] Transmitter Sel + // =0: UNIPHY or PCIEPHY + // =1: LVTMA + UCHAR ucAction; // =0: turn off encoder + // =1: turn on encoder + UCHAR ucEncoderMode; + // =0: DP encoder + // =1: LVDS encoder + // =2: DVI encoder + // =3: HDMI encoder + // =4: SDVO encoder + UCHAR ucLaneNum; // how many lanes to enable + UCHAR ucReserved[2]; +}DIG_ENCODER_CONTROL_PARAMETERS; #define DIG_ENCODER_CONTROL_PS_ALLOCATION DIG_ENCODER_CONTROL_PARAMETERS #define EXTERNAL_ENCODER_CONTROL_PARAMETER DIG_ENCODER_CONTROL_PARAMETERS -/* ucConfig */ +//ucConfig #define ATOM_ENCODER_CONFIG_DPLINKRATE_MASK 0x01 #define ATOM_ENCODER_CONFIG_DPLINKRATE_1_62GHZ 0x00 #define ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ 0x01 @@ -539,52 +593,57 @@ typedef struct _DIG_ENCODER_CONTROL_PARAMETERS { #define ATOM_ENCODER_CONFIG_LVTMA 0x08 #define ATOM_ENCODER_CONFIG_TRANSMITTER1 0x00 #define ATOM_ENCODER_CONFIG_TRANSMITTER2 0x08 -#define ATOM_ENCODER_CONFIG_DIGB 0x80 /* VBIOS Internal use, outside SW should set this bit=0 */ -/* ucAction */ -/* ATOM_ENABLE: Enable Encoder */ -/* ATOM_DISABLE: Disable Encoder */ +#define ATOM_ENCODER_CONFIG_DIGB 0x80 // VBIOS Internal use, outside SW should set this bit=0 +// ucAction +// ATOM_ENABLE: Enable Encoder +// ATOM_DISABLE: Disable Encoder -/* ucEncoderMode */ +//ucEncoderMode #define ATOM_ENCODER_MODE_DP 0 #define ATOM_ENCODER_MODE_LVDS 1 #define ATOM_ENCODER_MODE_DVI 2 #define ATOM_ENCODER_MODE_HDMI 3 #define ATOM_ENCODER_MODE_SDVO 4 +#define ATOM_ENCODER_MODE_DP_AUDIO 5 #define ATOM_ENCODER_MODE_TV 13 #define ATOM_ENCODER_MODE_CV 14 #define ATOM_ENCODER_MODE_CRT 15 -typedef struct _ATOM_DIG_ENCODER_CONFIG_V2 { +typedef struct _ATOM_DIG_ENCODER_CONFIG_V2 +{ #if ATOM_BIG_ENDIAN - UCHAR ucReserved1:2; - UCHAR ucTransmitterSel:2; /* =0: UniphyAB, =1: UniphyCD =2: UniphyEF */ - UCHAR ucLinkSel:1; /* =0: linkA/C/E =1: linkB/D/F */ - UCHAR ucReserved:1; - UCHAR ucDPLinkRate:1; /* =0: 1.62Ghz, =1: 2.7Ghz */ + UCHAR ucReserved1:2; + UCHAR ucTransmitterSel:2; // =0: UniphyAB, =1: UniphyCD =2: UniphyEF + UCHAR ucLinkSel:1; // =0: linkA/C/E =1: linkB/D/F + UCHAR ucReserved:1; + UCHAR ucDPLinkRate:1; // =0: 1.62Ghz, =1: 2.7Ghz #else - UCHAR ucDPLinkRate:1; /* =0: 1.62Ghz, =1: 2.7Ghz */ - UCHAR ucReserved:1; - UCHAR ucLinkSel:1; /* =0: linkA/C/E =1: linkB/D/F */ - UCHAR ucTransmitterSel:2; /* =0: UniphyAB, =1: UniphyCD =2: UniphyEF */ - UCHAR ucReserved1:2; + UCHAR ucDPLinkRate:1; // =0: 1.62Ghz, =1: 2.7Ghz + UCHAR ucReserved:1; + UCHAR ucLinkSel:1; // =0: linkA/C/E =1: linkB/D/F + UCHAR ucTransmitterSel:2; // =0: UniphyAB, =1: UniphyCD =2: UniphyEF + UCHAR ucReserved1:2; #endif -} ATOM_DIG_ENCODER_CONFIG_V2; +}ATOM_DIG_ENCODER_CONFIG_V2; -typedef struct _DIG_ENCODER_CONTROL_PARAMETERS_V2 { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - ATOM_DIG_ENCODER_CONFIG_V2 acConfig; - UCHAR ucAction; - UCHAR ucEncoderMode; - /* =0: DP encoder */ - /* =1: LVDS encoder */ - /* =2: DVI encoder */ - /* =3: HDMI encoder */ - /* =4: SDVO encoder */ - UCHAR ucLaneNum; /* how many lanes to enable */ - UCHAR ucReserved[2]; -} DIG_ENCODER_CONTROL_PARAMETERS_V2; -/* ucConfig */ +typedef struct _DIG_ENCODER_CONTROL_PARAMETERS_V2 +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + ATOM_DIG_ENCODER_CONFIG_V2 acConfig; + UCHAR ucAction; + UCHAR ucEncoderMode; + // =0: DP encoder + // =1: LVDS encoder + // =2: DVI encoder + // =3: HDMI encoder + // =4: SDVO encoder + UCHAR ucLaneNum; // how many lanes to enable + UCHAR ucStatus; // = DP_LINK_TRAINING_COMPLETE or DP_LINK_TRAINING_INCOMPLETE, only used by VBIOS with command ATOM_ENCODER_CMD_QUERY_DP_LINK_TRAINING_STATUS + UCHAR ucReserved; +}DIG_ENCODER_CONTROL_PARAMETERS_V2; + +//ucConfig #define ATOM_ENCODER_CONFIG_V2_DPLINKRATE_MASK 0x01 #define ATOM_ENCODER_CONFIG_V2_DPLINKRATE_1_62GHZ 0x00 #define ATOM_ENCODER_CONFIG_V2_DPLINKRATE_2_70GHZ 0x01 @@ -596,58 +655,122 @@ typedef struct _DIG_ENCODER_CONTROL_PARAMETERS_V2 { #define ATOM_ENCODER_CONFIG_V2_TRANSMITTER2 0x08 #define ATOM_ENCODER_CONFIG_V2_TRANSMITTER3 0x10 -/****************************************************************************/ -/* Structures used by UNIPHYTransmitterControlTable */ -/* LVTMATransmitterControlTable */ -/* DVOOutputControlTable */ -/****************************************************************************/ -typedef struct _ATOM_DP_VS_MODE { - UCHAR ucLaneSel; - UCHAR ucLaneSet; -} ATOM_DP_VS_MODE; - -typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS { - union { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - USHORT usInitInfo; /* when init uniphy,lower 8bit is used for connector type defined in objectid.h */ - ATOM_DP_VS_MODE asMode; /* DP Voltage swing mode */ +// ucAction: +// ATOM_DISABLE +// ATOM_ENABLE +#define ATOM_ENCODER_CMD_DP_LINK_TRAINING_START 0x08 +#define ATOM_ENCODER_CMD_DP_LINK_TRAINING_PATTERN1 0x09 +#define ATOM_ENCODER_CMD_DP_LINK_TRAINING_PATTERN2 0x0a +#define ATOM_ENCODER_CMD_DP_LINK_TRAINING_COMPLETE 0x0b +#define ATOM_ENCODER_CMD_DP_VIDEO_OFF 0x0c +#define ATOM_ENCODER_CMD_DP_VIDEO_ON 0x0d +#define ATOM_ENCODER_CMD_QUERY_DP_LINK_TRAINING_STATUS 0x0e +#define ATOM_ENCODER_CMD_SETUP 0x0f + +// ucStatus +#define ATOM_ENCODER_STATUS_LINK_TRAINING_COMPLETE 0x10 +#define ATOM_ENCODER_STATUS_LINK_TRAINING_INCOMPLETE 0x00 + +// Following function ENABLE sub-function will be used by driver when TMDS/HDMI/LVDS is used, disable function will be used by driver +typedef struct _ATOM_DIG_ENCODER_CONFIG_V3 +{ +#if ATOM_BIG_ENDIAN + UCHAR ucReserved1:1; + UCHAR ucDigSel:3; // =0: DIGA/B/C/D/E/F + UCHAR ucReserved:3; + UCHAR ucDPLinkRate:1; // =0: 1.62Ghz, =1: 2.7Ghz +#else + UCHAR ucDPLinkRate:1; // =0: 1.62Ghz, =1: 2.7Ghz + UCHAR ucReserved:3; + UCHAR ucDigSel:3; // =0: DIGA/B/C/D/E/F + UCHAR ucReserved1:1; +#endif +}ATOM_DIG_ENCODER_CONFIG_V3; + +#define ATOM_ENCODER_CONFIG_V3_ENCODER_SEL 0x70 + + +typedef struct _DIG_ENCODER_CONTROL_PARAMETERS_V3 +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + ATOM_DIG_ENCODER_CONFIG_V3 acConfig; + UCHAR ucAction; + UCHAR ucEncoderMode; + // =0: DP encoder + // =1: LVDS encoder + // =2: DVI encoder + // =3: HDMI encoder + // =4: SDVO encoder + // =5: DP audio + UCHAR ucLaneNum; // how many lanes to enable + UCHAR ucBitPerColor; // only valid for DP mode when ucAction = ATOM_ENCODER_CMD_SETUP + UCHAR ucReserved; +}DIG_ENCODER_CONTROL_PARAMETERS_V3; + + +// define ucBitPerColor: +#define PANEL_BPC_UNDEFINE 0x00 +#define PANEL_6BIT_PER_COLOR 0x01 +#define PANEL_8BIT_PER_COLOR 0x02 +#define PANEL_10BIT_PER_COLOR 0x03 +#define PANEL_12BIT_PER_COLOR 0x04 +#define PANEL_16BIT_PER_COLOR 0x05 + +/****************************************************************************/ +// Structures used by UNIPHYTransmitterControlTable +// LVTMATransmitterControlTable +// DVOOutputControlTable +/****************************************************************************/ +typedef struct _ATOM_DP_VS_MODE +{ + UCHAR ucLaneSel; + UCHAR ucLaneSet; +}ATOM_DP_VS_MODE; + +typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS +{ + union + { + USHORT usPixelClock; // in 10KHz; for bios convenient + USHORT usInitInfo; // when init uniphy,lower 8bit is used for connector type defined in objectid.h + ATOM_DP_VS_MODE asMode; // DP Voltage swing mode }; - UCHAR ucConfig; - /* [0]=0: 4 lane Link, */ - /* =1: 8 lane Link ( Dual Links TMDS ) */ - /* [1]=0: InCoherent mode */ - /* =1: Coherent Mode */ - /* [2] Link Select: */ - /* =0: PHY linkA if bfLane<3 */ - /* =1: PHY linkB if bfLanes<3 */ - /* =0: PHY linkA+B if bfLanes=3 */ - /* [5:4]PCIE lane Sel */ - /* =0: lane 0~3 or 0~7 */ - /* =1: lane 4~7 */ - /* =2: lane 8~11 or 8~15 */ - /* =3: lane 12~15 */ - UCHAR ucAction; /* =0: turn off encoder */ - /* =1: turn on encoder */ - UCHAR ucReserved[4]; -} DIG_TRANSMITTER_CONTROL_PARAMETERS; - -#define DIG_TRANSMITTER_CONTROL_PS_ALLOCATION DIG_TRANSMITTER_CONTROL_PARAMETERS - -/* ucInitInfo */ -#define ATOM_TRAMITTER_INITINFO_CONNECTOR_MASK 0x00ff - -/* ucConfig */ + UCHAR ucConfig; + // [0]=0: 4 lane Link, + // =1: 8 lane Link ( Dual Links TMDS ) + // [1]=0: InCoherent mode + // =1: Coherent Mode + // [2] Link Select: + // =0: PHY linkA if bfLane<3 + // =1: PHY linkB if bfLanes<3 + // =0: PHY linkA+B if bfLanes=3 + // [5:4]PCIE lane Sel + // =0: lane 0~3 or 0~7 + // =1: lane 4~7 + // =2: lane 8~11 or 8~15 + // =3: lane 12~15 + UCHAR ucAction; // =0: turn off encoder + // =1: turn on encoder + UCHAR ucReserved[4]; +}DIG_TRANSMITTER_CONTROL_PARAMETERS; + +#define DIG_TRANSMITTER_CONTROL_PS_ALLOCATION DIG_TRANSMITTER_CONTROL_PARAMETERS + +//ucInitInfo +#define ATOM_TRAMITTER_INITINFO_CONNECTOR_MASK 0x00ff + +//ucConfig #define ATOM_TRANSMITTER_CONFIG_8LANE_LINK 0x01 #define ATOM_TRANSMITTER_CONFIG_COHERENT 0x02 #define ATOM_TRANSMITTER_CONFIG_LINK_SEL_MASK 0x04 #define ATOM_TRANSMITTER_CONFIG_LINKA 0x00 #define ATOM_TRANSMITTER_CONFIG_LINKB 0x04 -#define ATOM_TRANSMITTER_CONFIG_LINKA_B 0x00 +#define ATOM_TRANSMITTER_CONFIG_LINKA_B 0x00 #define ATOM_TRANSMITTER_CONFIG_LINKB_A 0x04 -#define ATOM_TRANSMITTER_CONFIG_ENCODER_SEL_MASK 0x08 /* only used when ATOM_TRANSMITTER_ACTION_ENABLE */ -#define ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER 0x00 /* only used when ATOM_TRANSMITTER_ACTION_ENABLE */ -#define ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER 0x08 /* only used when ATOM_TRANSMITTER_ACTION_ENABLE */ +#define ATOM_TRANSMITTER_CONFIG_ENCODER_SEL_MASK 0x08 // only used when ATOM_TRANSMITTER_ACTION_ENABLE +#define ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER 0x00 // only used when ATOM_TRANSMITTER_ACTION_ENABLE +#define ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER 0x08 // only used when ATOM_TRANSMITTER_ACTION_ENABLE #define ATOM_TRANSMITTER_CONFIG_CLKSRC_MASK 0x30 #define ATOM_TRANSMITTER_CONFIG_CLKSRC_PPLL 0x00 @@ -661,7 +784,7 @@ typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS { #define ATOM_TRANSMITTER_CONFIG_LANE_8_15 0x80 #define ATOM_TRANSMITTER_CONFIG_LANE_12_15 0xc0 -/* ucAction */ +//ucAction #define ATOM_TRANSMITTER_ACTION_DISABLE 0 #define ATOM_TRANSMITTER_ACTION_ENABLE 1 #define ATOM_TRANSMITTER_ACTION_LCD_BLOFF 2 @@ -674,93 +797,168 @@ typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS { #define ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT 9 #define ATOM_TRANSMITTER_ACTION_SETUP 10 #define ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH 11 +#define ATOM_TRANSMITTER_ACTION_POWER_ON 12 +#define ATOM_TRANSMITTER_ACTION_POWER_OFF 13 -/* Following are used for DigTransmitterControlTable ver1.2 */ -typedef struct _ATOM_DIG_TRANSMITTER_CONFIG_V2 { +// Following are used for DigTransmitterControlTable ver1.2 +typedef struct _ATOM_DIG_TRANSMITTER_CONFIG_V2 +{ #if ATOM_BIG_ENDIAN - UCHAR ucTransmitterSel:2; /* bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) */ - /* =1 Dig Transmitter 2 ( Uniphy CD ) */ - /* =2 Dig Transmitter 3 ( Uniphy EF ) */ - UCHAR ucReserved:1; - UCHAR fDPConnector:1; /* bit4=0: DP connector =1: None DP connector */ - UCHAR ucEncoderSel:1; /* bit3=0: Data/Clk path source from DIGA( DIG inst0 ). =1: Data/clk path source from DIGB ( DIG inst1 ) */ - UCHAR ucLinkSel:1; /* bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E */ - /* =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F */ - - UCHAR fCoherentMode:1; /* bit1=1: Coherent Mode ( for DVI/HDMI mode ) */ - UCHAR fDualLinkConnector:1; /* bit0=1: Dual Link DVI connector */ + UCHAR ucTransmitterSel:2; //bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) + // =1 Dig Transmitter 2 ( Uniphy CD ) + // =2 Dig Transmitter 3 ( Uniphy EF ) + UCHAR ucReserved:1; + UCHAR fDPConnector:1; //bit4=0: DP connector =1: None DP connector + UCHAR ucEncoderSel:1; //bit3=0: Data/Clk path source from DIGA( DIG inst0 ). =1: Data/clk path source from DIGB ( DIG inst1 ) + UCHAR ucLinkSel:1; //bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E + // =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F + + UCHAR fCoherentMode:1; //bit1=1: Coherent Mode ( for DVI/HDMI mode ) + UCHAR fDualLinkConnector:1; //bit0=1: Dual Link DVI connector #else - UCHAR fDualLinkConnector:1; /* bit0=1: Dual Link DVI connector */ - UCHAR fCoherentMode:1; /* bit1=1: Coherent Mode ( for DVI/HDMI mode ) */ - UCHAR ucLinkSel:1; /* bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E */ - /* =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F */ - UCHAR ucEncoderSel:1; /* bit3=0: Data/Clk path source from DIGA( DIG inst0 ). =1: Data/clk path source from DIGB ( DIG inst1 ) */ - UCHAR fDPConnector:1; /* bit4=0: DP connector =1: None DP connector */ - UCHAR ucReserved:1; - UCHAR ucTransmitterSel:2; /* bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) */ - /* =1 Dig Transmitter 2 ( Uniphy CD ) */ - /* =2 Dig Transmitter 3 ( Uniphy EF ) */ + UCHAR fDualLinkConnector:1; //bit0=1: Dual Link DVI connector + UCHAR fCoherentMode:1; //bit1=1: Coherent Mode ( for DVI/HDMI mode ) + UCHAR ucLinkSel:1; //bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E + // =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F + UCHAR ucEncoderSel:1; //bit3=0: Data/Clk path source from DIGA( DIG inst0 ). =1: Data/clk path source from DIGB ( DIG inst1 ) + UCHAR fDPConnector:1; //bit4=0: DP connector =1: None DP connector + UCHAR ucReserved:1; + UCHAR ucTransmitterSel:2; //bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) + // =1 Dig Transmitter 2 ( Uniphy CD ) + // =2 Dig Transmitter 3 ( Uniphy EF ) #endif -} ATOM_DIG_TRANSMITTER_CONFIG_V2; +}ATOM_DIG_TRANSMITTER_CONFIG_V2; -/* ucConfig */ -/* Bit0 */ +//ucConfig +//Bit0 #define ATOM_TRANSMITTER_CONFIG_V2_DUAL_LINK_CONNECTOR 0x01 -/* Bit1 */ +//Bit1 #define ATOM_TRANSMITTER_CONFIG_V2_COHERENT 0x02 -/* Bit2 */ +//Bit2 #define ATOM_TRANSMITTER_CONFIG_V2_LINK_SEL_MASK 0x04 -#define ATOM_TRANSMITTER_CONFIG_V2_LINKA 0x00 +#define ATOM_TRANSMITTER_CONFIG_V2_LINKA 0x00 #define ATOM_TRANSMITTER_CONFIG_V2_LINKB 0x04 -/* Bit3 */ +// Bit3 #define ATOM_TRANSMITTER_CONFIG_V2_ENCODER_SEL_MASK 0x08 -#define ATOM_TRANSMITTER_CONFIG_V2_DIG1_ENCODER 0x00 /* only used when ucAction == ATOM_TRANSMITTER_ACTION_ENABLE or ATOM_TRANSMITTER_ACTION_SETUP */ -#define ATOM_TRANSMITTER_CONFIG_V2_DIG2_ENCODER 0x08 /* only used when ucAction == ATOM_TRANSMITTER_ACTION_ENABLE or ATOM_TRANSMITTER_ACTION_SETUP */ +#define ATOM_TRANSMITTER_CONFIG_V2_DIG1_ENCODER 0x00 // only used when ucAction == ATOM_TRANSMITTER_ACTION_ENABLE or ATOM_TRANSMITTER_ACTION_SETUP +#define ATOM_TRANSMITTER_CONFIG_V2_DIG2_ENCODER 0x08 // only used when ucAction == ATOM_TRANSMITTER_ACTION_ENABLE or ATOM_TRANSMITTER_ACTION_SETUP -/* Bit4 */ +// Bit4 #define ATOM_TRASMITTER_CONFIG_V2_DP_CONNECTOR 0x10 -/* Bit7:6 */ +// Bit7:6 #define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER_SEL_MASK 0xC0 -#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER1 0x00 /* AB */ -#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER2 0x40 /* CD */ -#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER3 0x80 /* EF */ - -typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 { - union { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - USHORT usInitInfo; /* when init uniphy,lower 8bit is used for connector type defined in objectid.h */ - ATOM_DP_VS_MODE asMode; /* DP Voltage swing mode */ +#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER1 0x00 //AB +#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER2 0x40 //CD +#define ATOM_TRANSMITTER_CONFIG_V2_TRANSMITTER3 0x80 //EF + +typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 +{ + union + { + USHORT usPixelClock; // in 10KHz; for bios convenient + USHORT usInitInfo; // when init uniphy,lower 8bit is used for connector type defined in objectid.h + ATOM_DP_VS_MODE asMode; // DP Voltage swing mode }; - ATOM_DIG_TRANSMITTER_CONFIG_V2 acConfig; - UCHAR ucAction; /* define as ATOM_TRANSMITER_ACTION_XXX */ - UCHAR ucReserved[4]; -} DIG_TRANSMITTER_CONTROL_PARAMETERS_V2; + ATOM_DIG_TRANSMITTER_CONFIG_V2 acConfig; + UCHAR ucAction; // define as ATOM_TRANSMITER_ACTION_XXX + UCHAR ucReserved[4]; +}DIG_TRANSMITTER_CONTROL_PARAMETERS_V2; -/****************************************************************************/ -/* Structures used by DAC1OuputControlTable */ -/* DAC2OuputControlTable */ -/* LVTMAOutputControlTable (Before DEC30) */ -/* TMDSAOutputControlTable (Before DEC30) */ -/****************************************************************************/ -typedef struct _DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS { - UCHAR ucAction; /* Possible input:ATOM_ENABLE||ATOMDISABLE */ - /* When the display is LCD, in addition to above: */ - /* ATOM_LCD_BLOFF|| ATOM_LCD_BLON ||ATOM_LCD_BL_BRIGHTNESS_CONTROL||ATOM_LCD_SELFTEST_START|| */ - /* ATOM_LCD_SELFTEST_STOP */ +typedef struct _ATOM_DIG_TRANSMITTER_CONFIG_V3 +{ +#if ATOM_BIG_ENDIAN + UCHAR ucTransmitterSel:2; //bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) + // =1 Dig Transmitter 2 ( Uniphy CD ) + // =2 Dig Transmitter 3 ( Uniphy EF ) + UCHAR ucRefClkSource:2; //bit5:4: PPLL1 =0, PPLL2=1, EXT_CLK=2 + UCHAR ucEncoderSel:1; //bit3=0: Data/Clk path source from DIGA/C/E. =1: Data/clk path source from DIGB/D/F + UCHAR ucLinkSel:1; //bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E + // =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F + UCHAR fCoherentMode:1; //bit1=1: Coherent Mode ( for DVI/HDMI mode ) + UCHAR fDualLinkConnector:1; //bit0=1: Dual Link DVI connector +#else + UCHAR fDualLinkConnector:1; //bit0=1: Dual Link DVI connector + UCHAR fCoherentMode:1; //bit1=1: Coherent Mode ( for DVI/HDMI mode ) + UCHAR ucLinkSel:1; //bit2=0: Uniphy LINKA or C or E when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is A or C or E + // =1: Uniphy LINKB or D or F when fDualLinkConnector=0. when fDualLinkConnector=1, it means master link of dual link is B or D or F + UCHAR ucEncoderSel:1; //bit3=0: Data/Clk path source from DIGA/C/E. =1: Data/clk path source from DIGB/D/F + UCHAR ucRefClkSource:2; //bit5:4: PPLL1 =0, PPLL2=1, EXT_CLK=2 + UCHAR ucTransmitterSel:2; //bit7:6: =0 Dig Transmitter 1 ( Uniphy AB ) + // =1 Dig Transmitter 2 ( Uniphy CD ) + // =2 Dig Transmitter 3 ( Uniphy EF ) +#endif +}ATOM_DIG_TRANSMITTER_CONFIG_V3; - UCHAR aucPadding[3]; /* padding to DWORD aligned */ -} DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS; +typedef struct _DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 +{ + union + { + USHORT usPixelClock; // in 10KHz; for bios convenient + USHORT usInitInfo; // when init uniphy,lower 8bit is used for connector type defined in objectid.h + ATOM_DP_VS_MODE asMode; // DP Voltage swing mode + }; + ATOM_DIG_TRANSMITTER_CONFIG_V3 acConfig; + UCHAR ucAction; // define as ATOM_TRANSMITER_ACTION_XXX + UCHAR ucLaneNum; + UCHAR ucReserved[3]; +}DIG_TRANSMITTER_CONTROL_PARAMETERS_V3; + +//ucConfig +//Bit0 +#define ATOM_TRANSMITTER_CONFIG_V3_DUAL_LINK_CONNECTOR 0x01 + +//Bit1 +#define ATOM_TRANSMITTER_CONFIG_V3_COHERENT 0x02 + +//Bit2 +#define ATOM_TRANSMITTER_CONFIG_V3_LINK_SEL_MASK 0x04 +#define ATOM_TRANSMITTER_CONFIG_V3_LINKA 0x00 +#define ATOM_TRANSMITTER_CONFIG_V3_LINKB 0x04 + +// Bit3 +#define ATOM_TRANSMITTER_CONFIG_V3_ENCODER_SEL_MASK 0x08 +#define ATOM_TRANSMITTER_CONFIG_V3_DIG1_ENCODER 0x00 +#define ATOM_TRANSMITTER_CONFIG_V3_DIG2_ENCODER 0x08 + +// Bit5:4 +#define ATOM_TRASMITTER_CONFIG_V3_REFCLK_SEL_MASK 0x30 +#define ATOM_TRASMITTER_CONFIG_V3_P1PLL 0x00 +#define ATOM_TRASMITTER_CONFIG_V3_P2PLL 0x10 +#define ATOM_TRASMITTER_CONFIG_V3_REFCLK_SRC_EXT 0x20 + +// Bit7:6 +#define ATOM_TRANSMITTER_CONFIG_V3_TRANSMITTER_SEL_MASK 0xC0 +#define ATOM_TRANSMITTER_CONFIG_V3_TRANSMITTER1 0x00 //AB +#define ATOM_TRANSMITTER_CONFIG_V3_TRANSMITTER2 0x40 //CD +#define ATOM_TRANSMITTER_CONFIG_V3_TRANSMITTER3 0x80 //EF + +/****************************************************************************/ +// Structures used by DAC1OuputControlTable +// DAC2OuputControlTable +// LVTMAOutputControlTable (Before DEC30) +// TMDSAOutputControlTable (Before DEC30) +/****************************************************************************/ +typedef struct _DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS +{ + UCHAR ucAction; // Possible input:ATOM_ENABLE||ATOMDISABLE + // When the display is LCD, in addition to above: + // ATOM_LCD_BLOFF|| ATOM_LCD_BLON ||ATOM_LCD_BL_BRIGHTNESS_CONTROL||ATOM_LCD_SELFTEST_START|| + // ATOM_LCD_SELFTEST_STOP + + UCHAR aucPadding[3]; // padding to DWORD aligned +}DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS; #define DISPLAY_DEVICE_OUTPUT_CONTROL_PS_ALLOCATION DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS -#define CRT1_OUTPUT_CONTROL_PARAMETERS DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS + +#define CRT1_OUTPUT_CONTROL_PARAMETERS DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS #define CRT1_OUTPUT_CONTROL_PS_ALLOCATION DISPLAY_DEVICE_OUTPUT_CONTROL_PS_ALLOCATION -#define CRT2_OUTPUT_CONTROL_PARAMETERS DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS +#define CRT2_OUTPUT_CONTROL_PARAMETERS DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS #define CRT2_OUTPUT_CONTROL_PS_ALLOCATION DISPLAY_DEVICE_OUTPUT_CONTROL_PS_ALLOCATION #define CV1_OUTPUT_CONTROL_PARAMETERS DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS @@ -782,397 +980,550 @@ typedef struct _DISPLAY_DEVICE_OUTPUT_CONTROL_PARAMETERS { #define DVO_OUTPUT_CONTROL_PS_ALLOCATION DIG_TRANSMITTER_CONTROL_PS_ALLOCATION #define DVO_OUTPUT_CONTROL_PARAMETERS_V3 DIG_TRANSMITTER_CONTROL_PARAMETERS -/****************************************************************************/ -/* Structures used by BlankCRTCTable */ -/****************************************************************************/ -typedef struct _BLANK_CRTC_PARAMETERS { - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucBlanking; /* ATOM_BLANKING or ATOM_BLANKINGOFF */ - USHORT usBlackColorRCr; - USHORT usBlackColorGY; - USHORT usBlackColorBCb; -} BLANK_CRTC_PARAMETERS; +/****************************************************************************/ +// Structures used by BlankCRTCTable +/****************************************************************************/ +typedef struct _BLANK_CRTC_PARAMETERS +{ + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucBlanking; // ATOM_BLANKING or ATOM_BLANKINGOFF + USHORT usBlackColorRCr; + USHORT usBlackColorGY; + USHORT usBlackColorBCb; +}BLANK_CRTC_PARAMETERS; #define BLANK_CRTC_PS_ALLOCATION BLANK_CRTC_PARAMETERS -/****************************************************************************/ -/* Structures used by EnableCRTCTable */ -/* EnableCRTCMemReqTable */ -/* UpdateCRTC_DoubleBufferRegistersTable */ -/****************************************************************************/ -typedef struct _ENABLE_CRTC_PARAMETERS { - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucPadding[2]; -} ENABLE_CRTC_PARAMETERS; +/****************************************************************************/ +// Structures used by EnableCRTCTable +// EnableCRTCMemReqTable +// UpdateCRTC_DoubleBufferRegistersTable +/****************************************************************************/ +typedef struct _ENABLE_CRTC_PARAMETERS +{ + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + UCHAR ucPadding[2]; +}ENABLE_CRTC_PARAMETERS; #define ENABLE_CRTC_PS_ALLOCATION ENABLE_CRTC_PARAMETERS -/****************************************************************************/ -/* Structures used by SetCRTC_OverScanTable */ -/****************************************************************************/ -typedef struct _SET_CRTC_OVERSCAN_PARAMETERS { - USHORT usOverscanRight; /* right */ - USHORT usOverscanLeft; /* left */ - USHORT usOverscanBottom; /* bottom */ - USHORT usOverscanTop; /* top */ - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucPadding[3]; -} SET_CRTC_OVERSCAN_PARAMETERS; +/****************************************************************************/ +// Structures used by SetCRTC_OverScanTable +/****************************************************************************/ +typedef struct _SET_CRTC_OVERSCAN_PARAMETERS +{ + USHORT usOverscanRight; // right + USHORT usOverscanLeft; // left + USHORT usOverscanBottom; // bottom + USHORT usOverscanTop; // top + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucPadding[3]; +}SET_CRTC_OVERSCAN_PARAMETERS; #define SET_CRTC_OVERSCAN_PS_ALLOCATION SET_CRTC_OVERSCAN_PARAMETERS -/****************************************************************************/ -/* Structures used by SetCRTC_ReplicationTable */ -/****************************************************************************/ -typedef struct _SET_CRTC_REPLICATION_PARAMETERS { - UCHAR ucH_Replication; /* horizontal replication */ - UCHAR ucV_Replication; /* vertical replication */ - UCHAR usCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucPadding; -} SET_CRTC_REPLICATION_PARAMETERS; +/****************************************************************************/ +// Structures used by SetCRTC_ReplicationTable +/****************************************************************************/ +typedef struct _SET_CRTC_REPLICATION_PARAMETERS +{ + UCHAR ucH_Replication; // horizontal replication + UCHAR ucV_Replication; // vertical replication + UCHAR usCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucPadding; +}SET_CRTC_REPLICATION_PARAMETERS; #define SET_CRTC_REPLICATION_PS_ALLOCATION SET_CRTC_REPLICATION_PARAMETERS -/****************************************************************************/ -/* Structures used by SelectCRTC_SourceTable */ -/****************************************************************************/ -typedef struct _SELECT_CRTC_SOURCE_PARAMETERS { - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucDevice; /* ATOM_DEVICE_CRT1|ATOM_DEVICE_CRT2|.... */ - UCHAR ucPadding[2]; -} SELECT_CRTC_SOURCE_PARAMETERS; +/****************************************************************************/ +// Structures used by SelectCRTC_SourceTable +/****************************************************************************/ +typedef struct _SELECT_CRTC_SOURCE_PARAMETERS +{ + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucDevice; // ATOM_DEVICE_CRT1|ATOM_DEVICE_CRT2|.... + UCHAR ucPadding[2]; +}SELECT_CRTC_SOURCE_PARAMETERS; #define SELECT_CRTC_SOURCE_PS_ALLOCATION SELECT_CRTC_SOURCE_PARAMETERS -typedef struct _SELECT_CRTC_SOURCE_PARAMETERS_V2 { - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucEncoderID; /* DAC1/DAC2/TVOUT/DIG1/DIG2/DVO */ - UCHAR ucEncodeMode; /* Encoding mode, only valid when using DIG1/DIG2/DVO */ - UCHAR ucPadding; -} SELECT_CRTC_SOURCE_PARAMETERS_V2; - -/* ucEncoderID */ -/* #define ASIC_INT_DAC1_ENCODER_ID 0x00 */ -/* #define ASIC_INT_TV_ENCODER_ID 0x02 */ -/* #define ASIC_INT_DIG1_ENCODER_ID 0x03 */ -/* #define ASIC_INT_DAC2_ENCODER_ID 0x04 */ -/* #define ASIC_EXT_TV_ENCODER_ID 0x06 */ -/* #define ASIC_INT_DVO_ENCODER_ID 0x07 */ -/* #define ASIC_INT_DIG2_ENCODER_ID 0x09 */ -/* #define ASIC_EXT_DIG_ENCODER_ID 0x05 */ - -/* ucEncodeMode */ -/* #define ATOM_ENCODER_MODE_DP 0 */ -/* #define ATOM_ENCODER_MODE_LVDS 1 */ -/* #define ATOM_ENCODER_MODE_DVI 2 */ -/* #define ATOM_ENCODER_MODE_HDMI 3 */ -/* #define ATOM_ENCODER_MODE_SDVO 4 */ -/* #define ATOM_ENCODER_MODE_TV 13 */ -/* #define ATOM_ENCODER_MODE_CV 14 */ -/* #define ATOM_ENCODER_MODE_CRT 15 */ - -/****************************************************************************/ -/* Structures used by SetPixelClockTable */ -/* GetPixelClockTable */ -/****************************************************************************/ -/* Major revision=1., Minor revision=1 */ -typedef struct _PIXEL_CLOCK_PARAMETERS { - USHORT usPixelClock; /* in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) */ - /* 0 means disable PPLL */ - USHORT usRefDiv; /* Reference divider */ - USHORT usFbDiv; /* feedback divider */ - UCHAR ucPostDiv; /* post divider */ - UCHAR ucFracFbDiv; /* fractional feedback divider */ - UCHAR ucPpll; /* ATOM_PPLL1 or ATOM_PPL2 */ - UCHAR ucRefDivSrc; /* ATOM_PJITTER or ATO_NONPJITTER */ - UCHAR ucCRTC; /* Which CRTC uses this Ppll */ - UCHAR ucPadding; -} PIXEL_CLOCK_PARAMETERS; - -/* Major revision=1., Minor revision=2, add ucMiscIfno */ -/* ucMiscInfo: */ +typedef struct _SELECT_CRTC_SOURCE_PARAMETERS_V2 +{ + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucEncoderID; // DAC1/DAC2/TVOUT/DIG1/DIG2/DVO + UCHAR ucEncodeMode; // Encoding mode, only valid when using DIG1/DIG2/DVO + UCHAR ucPadding; +}SELECT_CRTC_SOURCE_PARAMETERS_V2; + +//ucEncoderID +//#define ASIC_INT_DAC1_ENCODER_ID 0x00 +//#define ASIC_INT_TV_ENCODER_ID 0x02 +//#define ASIC_INT_DIG1_ENCODER_ID 0x03 +//#define ASIC_INT_DAC2_ENCODER_ID 0x04 +//#define ASIC_EXT_TV_ENCODER_ID 0x06 +//#define ASIC_INT_DVO_ENCODER_ID 0x07 +//#define ASIC_INT_DIG2_ENCODER_ID 0x09 +//#define ASIC_EXT_DIG_ENCODER_ID 0x05 + +//ucEncodeMode +//#define ATOM_ENCODER_MODE_DP 0 +//#define ATOM_ENCODER_MODE_LVDS 1 +//#define ATOM_ENCODER_MODE_DVI 2 +//#define ATOM_ENCODER_MODE_HDMI 3 +//#define ATOM_ENCODER_MODE_SDVO 4 +//#define ATOM_ENCODER_MODE_TV 13 +//#define ATOM_ENCODER_MODE_CV 14 +//#define ATOM_ENCODER_MODE_CRT 15 + +/****************************************************************************/ +// Structures used by SetPixelClockTable +// GetPixelClockTable +/****************************************************************************/ +//Major revision=1., Minor revision=1 +typedef struct _PIXEL_CLOCK_PARAMETERS +{ + USHORT usPixelClock; // in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) + // 0 means disable PPLL + USHORT usRefDiv; // Reference divider + USHORT usFbDiv; // feedback divider + UCHAR ucPostDiv; // post divider + UCHAR ucFracFbDiv; // fractional feedback divider + UCHAR ucPpll; // ATOM_PPLL1 or ATOM_PPL2 + UCHAR ucRefDivSrc; // ATOM_PJITTER or ATO_NONPJITTER + UCHAR ucCRTC; // Which CRTC uses this Ppll + UCHAR ucPadding; +}PIXEL_CLOCK_PARAMETERS; + +//Major revision=1., Minor revision=2, add ucMiscIfno +//ucMiscInfo: #define MISC_FORCE_REPROG_PIXEL_CLOCK 0x1 #define MISC_DEVICE_INDEX_MASK 0xF0 #define MISC_DEVICE_INDEX_SHIFT 4 -typedef struct _PIXEL_CLOCK_PARAMETERS_V2 { - USHORT usPixelClock; /* in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) */ - /* 0 means disable PPLL */ - USHORT usRefDiv; /* Reference divider */ - USHORT usFbDiv; /* feedback divider */ - UCHAR ucPostDiv; /* post divider */ - UCHAR ucFracFbDiv; /* fractional feedback divider */ - UCHAR ucPpll; /* ATOM_PPLL1 or ATOM_PPL2 */ - UCHAR ucRefDivSrc; /* ATOM_PJITTER or ATO_NONPJITTER */ - UCHAR ucCRTC; /* Which CRTC uses this Ppll */ - UCHAR ucMiscInfo; /* Different bits for different purpose, bit [7:4] as device index, bit[0]=Force prog */ -} PIXEL_CLOCK_PARAMETERS_V2; - -/* Major revision=1., Minor revision=3, structure/definition change */ -/* ucEncoderMode: */ -/* ATOM_ENCODER_MODE_DP */ -/* ATOM_ENOCDER_MODE_LVDS */ -/* ATOM_ENOCDER_MODE_DVI */ -/* ATOM_ENOCDER_MODE_HDMI */ -/* ATOM_ENOCDER_MODE_SDVO */ -/* ATOM_ENCODER_MODE_TV 13 */ -/* ATOM_ENCODER_MODE_CV 14 */ -/* ATOM_ENCODER_MODE_CRT 15 */ - -/* ucDVOConfig */ -/* #define DVO_ENCODER_CONFIG_RATE_SEL 0x01 */ -/* #define DVO_ENCODER_CONFIG_DDR_SPEED 0x00 */ -/* #define DVO_ENCODER_CONFIG_SDR_SPEED 0x01 */ -/* #define DVO_ENCODER_CONFIG_OUTPUT_SEL 0x0c */ -/* #define DVO_ENCODER_CONFIG_LOW12BIT 0x00 */ -/* #define DVO_ENCODER_CONFIG_UPPER12BIT 0x04 */ -/* #define DVO_ENCODER_CONFIG_24BIT 0x08 */ - -/* ucMiscInfo: also changed, see below */ +typedef struct _PIXEL_CLOCK_PARAMETERS_V2 +{ + USHORT usPixelClock; // in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) + // 0 means disable PPLL + USHORT usRefDiv; // Reference divider + USHORT usFbDiv; // feedback divider + UCHAR ucPostDiv; // post divider + UCHAR ucFracFbDiv; // fractional feedback divider + UCHAR ucPpll; // ATOM_PPLL1 or ATOM_PPL2 + UCHAR ucRefDivSrc; // ATOM_PJITTER or ATO_NONPJITTER + UCHAR ucCRTC; // Which CRTC uses this Ppll + UCHAR ucMiscInfo; // Different bits for different purpose, bit [7:4] as device index, bit[0]=Force prog +}PIXEL_CLOCK_PARAMETERS_V2; + +//Major revision=1., Minor revision=3, structure/definition change +//ucEncoderMode: +//ATOM_ENCODER_MODE_DP +//ATOM_ENOCDER_MODE_LVDS +//ATOM_ENOCDER_MODE_DVI +//ATOM_ENOCDER_MODE_HDMI +//ATOM_ENOCDER_MODE_SDVO +//ATOM_ENCODER_MODE_TV 13 +//ATOM_ENCODER_MODE_CV 14 +//ATOM_ENCODER_MODE_CRT 15 + +//ucDVOConfig +//#define DVO_ENCODER_CONFIG_RATE_SEL 0x01 +//#define DVO_ENCODER_CONFIG_DDR_SPEED 0x00 +//#define DVO_ENCODER_CONFIG_SDR_SPEED 0x01 +//#define DVO_ENCODER_CONFIG_OUTPUT_SEL 0x0c +//#define DVO_ENCODER_CONFIG_LOW12BIT 0x00 +//#define DVO_ENCODER_CONFIG_UPPER12BIT 0x04 +//#define DVO_ENCODER_CONFIG_24BIT 0x08 + +//ucMiscInfo: also changed, see below #define PIXEL_CLOCK_MISC_FORCE_PROG_PPLL 0x01 #define PIXEL_CLOCK_MISC_VGA_MODE 0x02 #define PIXEL_CLOCK_MISC_CRTC_SEL_MASK 0x04 #define PIXEL_CLOCK_MISC_CRTC_SEL_CRTC1 0x00 #define PIXEL_CLOCK_MISC_CRTC_SEL_CRTC2 0x04 #define PIXEL_CLOCK_MISC_USE_ENGINE_FOR_DISPCLK 0x08 +#define PIXEL_CLOCK_MISC_REF_DIV_SRC 0x10 +// V1.4 for RoadRunner +#define PIXEL_CLOCK_V4_MISC_SS_ENABLE 0x10 +#define PIXEL_CLOCK_V4_MISC_COHERENT_MODE 0x20 -typedef struct _PIXEL_CLOCK_PARAMETERS_V3 { - USHORT usPixelClock; /* in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) */ - /* 0 means disable PPLL. For VGA PPLL,make sure this value is not 0. */ - USHORT usRefDiv; /* Reference divider */ - USHORT usFbDiv; /* feedback divider */ - UCHAR ucPostDiv; /* post divider */ - UCHAR ucFracFbDiv; /* fractional feedback divider */ - UCHAR ucPpll; /* ATOM_PPLL1 or ATOM_PPL2 */ - UCHAR ucTransmitterId; /* graphic encoder id defined in objectId.h */ - union { - UCHAR ucEncoderMode; /* encoder type defined as ATOM_ENCODER_MODE_DP/DVI/HDMI/ */ - UCHAR ucDVOConfig; /* when use DVO, need to know SDR/DDR, 12bit or 24bit */ +typedef struct _PIXEL_CLOCK_PARAMETERS_V3 +{ + USHORT usPixelClock; // in 10kHz unit; for bios convenient = (RefClk*FB_Div)/(Ref_Div*Post_Div) + // 0 means disable PPLL. For VGA PPLL,make sure this value is not 0. + USHORT usRefDiv; // Reference divider + USHORT usFbDiv; // feedback divider + UCHAR ucPostDiv; // post divider + UCHAR ucFracFbDiv; // fractional feedback divider + UCHAR ucPpll; // ATOM_PPLL1 or ATOM_PPL2 + UCHAR ucTransmitterId; // graphic encoder id defined in objectId.h + union + { + UCHAR ucEncoderMode; // encoder type defined as ATOM_ENCODER_MODE_DP/DVI/HDMI/ + UCHAR ucDVOConfig; // when use DVO, need to know SDR/DDR, 12bit or 24bit }; - UCHAR ucMiscInfo; /* bit[0]=Force program, bit[1]= set pclk for VGA, b[2]= CRTC sel */ - /* bit[3]=0:use PPLL for dispclk source, =1: use engine clock for dispclock source */ -} PIXEL_CLOCK_PARAMETERS_V3; + UCHAR ucMiscInfo; // bit[0]=Force program, bit[1]= set pclk for VGA, b[2]= CRTC sel + // bit[3]=0:use PPLL for dispclk source, =1: use engine clock for dispclock source + // bit[4]=0:use XTALIN as the source of reference divider,=1 use the pre-defined clock as the source of reference divider +}PIXEL_CLOCK_PARAMETERS_V3; #define PIXEL_CLOCK_PARAMETERS_LAST PIXEL_CLOCK_PARAMETERS_V2 #define GET_PIXEL_CLOCK_PS_ALLOCATION PIXEL_CLOCK_PARAMETERS_LAST -/****************************************************************************/ -/* Structures used by AdjustDisplayPllTable */ -/****************************************************************************/ -typedef struct _ADJUST_DISPLAY_PLL_PARAMETERS { +typedef struct _PIXEL_CLOCK_PARAMETERS_V5 +{ + UCHAR ucCRTC; // ATOM_CRTC1~6, indicate the CRTC controller to + // drive the pixel clock. not used for DCPLL case. + union{ + UCHAR ucReserved; + UCHAR ucFracFbDiv; // [gphan] temporary to prevent build problem. remove it after driver code is changed. + }; + USHORT usPixelClock; // target the pixel clock to drive the CRTC timing + // 0 means disable PPLL/DCPLL. + USHORT usFbDiv; // feedback divider integer part. + UCHAR ucPostDiv; // post divider. + UCHAR ucRefDiv; // Reference divider + UCHAR ucPpll; // ATOM_PPLL1/ATOM_PPLL2/ATOM_DCPLL + UCHAR ucTransmitterID; // ASIC encoder id defined in objectId.h, + // indicate which graphic encoder will be used. + UCHAR ucEncoderMode; // Encoder mode: + UCHAR ucMiscInfo; // bit[0]= Force program PPLL + // bit[1]= when VGA timing is used. + // bit[3:2]= HDMI panel bit depth: =0: 24bpp =1:30bpp, =2:32bpp + // bit[4]= RefClock source for PPLL. + // =0: XTLAIN( default mode ) + // =1: other external clock source, which is pre-defined + // by VBIOS depend on the feature required. + // bit[7:5]: reserved. + ULONG ulFbDivDecFrac; // 20 bit feedback divider decimal fraction part, range from 1~999999 ( 0.000001 to 0.999999 ) + +}PIXEL_CLOCK_PARAMETERS_V5; + +#define PIXEL_CLOCK_V5_MISC_FORCE_PROG_PPLL 0x01 +#define PIXEL_CLOCK_V5_MISC_VGA_MODE 0x02 +#define PIXEL_CLOCK_V5_MISC_HDMI_BPP_MASK 0x0c +#define PIXEL_CLOCK_V5_MISC_HDMI_24BPP 0x00 +#define PIXEL_CLOCK_V5_MISC_HDMI_30BPP 0x04 +#define PIXEL_CLOCK_V5_MISC_HDMI_32BPP 0x08 +#define PIXEL_CLOCK_V5_MISC_REF_DIV_SRC 0x10 + +typedef struct _GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V2 +{ + PIXEL_CLOCK_PARAMETERS_V3 sDispClkInput; +}GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V2; + +typedef struct _GET_DISP_PLL_STATUS_OUTPUT_PARAMETERS_V2 +{ + UCHAR ucStatus; + UCHAR ucRefDivSrc; // =1: reference clock source from XTALIN, =0: source from PCIE ref clock + UCHAR ucReserved[2]; +}GET_DISP_PLL_STATUS_OUTPUT_PARAMETERS_V2; + +typedef struct _GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V3 +{ + PIXEL_CLOCK_PARAMETERS_V5 sDispClkInput; +}GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V3; + +/****************************************************************************/ +// Structures used by AdjustDisplayPllTable +/****************************************************************************/ +typedef struct _ADJUST_DISPLAY_PLL_PARAMETERS +{ USHORT usPixelClock; UCHAR ucTransmitterID; UCHAR ucEncodeMode; - union { - UCHAR ucDVOConfig; /* if DVO, need passing link rate and output 12bitlow or 24bit */ - UCHAR ucConfig; /* if none DVO, not defined yet */ + union + { + UCHAR ucDVOConfig; //if DVO, need passing link rate and output 12bitlow or 24bit + UCHAR ucConfig; //if none DVO, not defined yet }; UCHAR ucReserved[3]; -} ADJUST_DISPLAY_PLL_PARAMETERS; +}ADJUST_DISPLAY_PLL_PARAMETERS; #define ADJUST_DISPLAY_CONFIG_SS_ENABLE 0x10 - #define ADJUST_DISPLAY_PLL_PS_ALLOCATION ADJUST_DISPLAY_PLL_PARAMETERS -/****************************************************************************/ -/* Structures used by EnableYUVTable */ -/****************************************************************************/ -typedef struct _ENABLE_YUV_PARAMETERS { - UCHAR ucEnable; /* ATOM_ENABLE:Enable YUV or ATOM_DISABLE:Disable YUV (RGB) */ - UCHAR ucCRTC; /* Which CRTC needs this YUV or RGB format */ - UCHAR ucPadding[2]; -} ENABLE_YUV_PARAMETERS; +typedef struct _ADJUST_DISPLAY_PLL_INPUT_PARAMETERS_V3 +{ + USHORT usPixelClock; // target pixel clock + UCHAR ucTransmitterID; // transmitter id defined in objectid.h + UCHAR ucEncodeMode; // encoder mode: CRT, LVDS, DP, TMDS or HDMI + UCHAR ucDispPllConfig; // display pll configure parameter defined as following DISPPLL_CONFIG_XXXX + UCHAR ucReserved[3]; +}ADJUST_DISPLAY_PLL_INPUT_PARAMETERS_V3; + +// usDispPllConfig v1.2 for RoadRunner +#define DISPPLL_CONFIG_DVO_RATE_SEL 0x0001 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_DDR_SPEED 0x0000 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_SDR_SPEED 0x0001 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_OUTPUT_SEL 0x000c // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_LOW12BIT 0x0000 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_UPPER12BIT 0x0004 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_DVO_24BIT 0x0008 // need only when ucTransmitterID = DVO +#define DISPPLL_CONFIG_SS_ENABLE 0x0010 // Only used when ucEncoderMode = DP or LVDS +#define DISPPLL_CONFIG_COHERENT_MODE 0x0020 // Only used when ucEncoderMode = TMDS or HDMI +#define DISPPLL_CONFIG_DUAL_LINK 0x0040 // Only used when ucEncoderMode = TMDS or LVDS + + +typedef struct _ADJUST_DISPLAY_PLL_OUTPUT_PARAMETERS_V3 +{ + ULONG ulDispPllFreq; // return display PPLL freq which is used to generate the pixclock, and related idclk, symclk etc + UCHAR ucRefDiv; // if it is none-zero, it is used to be calculated the other ppll parameter fb_divider and post_div ( if it is not given ) + UCHAR ucPostDiv; // if it is none-zero, it is used to be calculated the other ppll parameter fb_divider + UCHAR ucReserved[2]; +}ADJUST_DISPLAY_PLL_OUTPUT_PARAMETERS_V3; + +typedef struct _ADJUST_DISPLAY_PLL_PS_ALLOCATION_V3 +{ + union + { + ADJUST_DISPLAY_PLL_INPUT_PARAMETERS_V3 sInput; + ADJUST_DISPLAY_PLL_OUTPUT_PARAMETERS_V3 sOutput; + }; +} ADJUST_DISPLAY_PLL_PS_ALLOCATION_V3; + +/****************************************************************************/ +// Structures used by EnableYUVTable +/****************************************************************************/ +typedef struct _ENABLE_YUV_PARAMETERS +{ + UCHAR ucEnable; // ATOM_ENABLE:Enable YUV or ATOM_DISABLE:Disable YUV (RGB) + UCHAR ucCRTC; // Which CRTC needs this YUV or RGB format + UCHAR ucPadding[2]; +}ENABLE_YUV_PARAMETERS; #define ENABLE_YUV_PS_ALLOCATION ENABLE_YUV_PARAMETERS -/****************************************************************************/ -/* Structures used by GetMemoryClockTable */ -/****************************************************************************/ -typedef struct _GET_MEMORY_CLOCK_PARAMETERS { - ULONG ulReturnMemoryClock; /* current memory speed in 10KHz unit */ +/****************************************************************************/ +// Structures used by GetMemoryClockTable +/****************************************************************************/ +typedef struct _GET_MEMORY_CLOCK_PARAMETERS +{ + ULONG ulReturnMemoryClock; // current memory speed in 10KHz unit } GET_MEMORY_CLOCK_PARAMETERS; #define GET_MEMORY_CLOCK_PS_ALLOCATION GET_MEMORY_CLOCK_PARAMETERS -/****************************************************************************/ -/* Structures used by GetEngineClockTable */ -/****************************************************************************/ -typedef struct _GET_ENGINE_CLOCK_PARAMETERS { - ULONG ulReturnEngineClock; /* current engine speed in 10KHz unit */ +/****************************************************************************/ +// Structures used by GetEngineClockTable +/****************************************************************************/ +typedef struct _GET_ENGINE_CLOCK_PARAMETERS +{ + ULONG ulReturnEngineClock; // current engine speed in 10KHz unit } GET_ENGINE_CLOCK_PARAMETERS; #define GET_ENGINE_CLOCK_PS_ALLOCATION GET_ENGINE_CLOCK_PARAMETERS -/****************************************************************************/ -/* Following Structures and constant may be obsolete */ -/****************************************************************************/ -/* Maxium 8 bytes,the data read in will be placed in the parameter space. */ -/* Read operaion successeful when the paramter space is non-zero, otherwise read operation failed */ -typedef struct _READ_EDID_FROM_HW_I2C_DATA_PARAMETERS { - USHORT usPrescale; /* Ratio between Engine clock and I2C clock */ - USHORT usVRAMAddress; /* Adress in Frame Buffer where to pace raw EDID */ - USHORT usStatus; /* When use output: lower byte EDID checksum, high byte hardware status */ - /* WHen use input: lower byte as 'byte to read':currently limited to 128byte or 1byte */ - UCHAR ucSlaveAddr; /* Read from which slave */ - UCHAR ucLineNumber; /* Read from which HW assisted line */ -} READ_EDID_FROM_HW_I2C_DATA_PARAMETERS; +/****************************************************************************/ +// Following Structures and constant may be obsolete +/****************************************************************************/ +//Maxium 8 bytes,the data read in will be placed in the parameter space. +//Read operaion successeful when the paramter space is non-zero, otherwise read operation failed +typedef struct _READ_EDID_FROM_HW_I2C_DATA_PARAMETERS +{ + USHORT usPrescale; //Ratio between Engine clock and I2C clock + USHORT usVRAMAddress; //Adress in Frame Buffer where to pace raw EDID + USHORT usStatus; //When use output: lower byte EDID checksum, high byte hardware status + //WHen use input: lower byte as 'byte to read':currently limited to 128byte or 1byte + UCHAR ucSlaveAddr; //Read from which slave + UCHAR ucLineNumber; //Read from which HW assisted line +}READ_EDID_FROM_HW_I2C_DATA_PARAMETERS; #define READ_EDID_FROM_HW_I2C_DATA_PS_ALLOCATION READ_EDID_FROM_HW_I2C_DATA_PARAMETERS + #define ATOM_WRITE_I2C_FORMAT_PSOFFSET_PSDATABYTE 0 #define ATOM_WRITE_I2C_FORMAT_PSOFFSET_PSTWODATABYTES 1 #define ATOM_WRITE_I2C_FORMAT_PSCOUNTER_PSOFFSET_IDDATABLOCK 2 #define ATOM_WRITE_I2C_FORMAT_PSCOUNTER_IDOFFSET_PLUS_IDDATABLOCK 3 #define ATOM_WRITE_I2C_FORMAT_IDCOUNTER_IDOFFSET_IDDATABLOCK 4 -typedef struct _WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS { - USHORT usPrescale; /* Ratio between Engine clock and I2C clock */ - USHORT usByteOffset; /* Write to which byte */ - /* Upper portion of usByteOffset is Format of data */ - /* 1bytePS+offsetPS */ - /* 2bytesPS+offsetPS */ - /* blockID+offsetPS */ - /* blockID+offsetID */ - /* blockID+counterID+offsetID */ - UCHAR ucData; /* PS data1 */ - UCHAR ucStatus; /* Status byte 1=success, 2=failure, Also is used as PS data2 */ - UCHAR ucSlaveAddr; /* Write to which slave */ - UCHAR ucLineNumber; /* Write from which HW assisted line */ -} WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS; +typedef struct _WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS +{ + USHORT usPrescale; //Ratio between Engine clock and I2C clock + USHORT usByteOffset; //Write to which byte + //Upper portion of usByteOffset is Format of data + //1bytePS+offsetPS + //2bytesPS+offsetPS + //blockID+offsetPS + //blockID+offsetID + //blockID+counterID+offsetID + UCHAR ucData; //PS data1 + UCHAR ucStatus; //Status byte 1=success, 2=failure, Also is used as PS data2 + UCHAR ucSlaveAddr; //Write to which slave + UCHAR ucLineNumber; //Write from which HW assisted line +}WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS; #define WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS -typedef struct _SET_UP_HW_I2C_DATA_PARAMETERS { - USHORT usPrescale; /* Ratio between Engine clock and I2C clock */ - UCHAR ucSlaveAddr; /* Write to which slave */ - UCHAR ucLineNumber; /* Write from which HW assisted line */ -} SET_UP_HW_I2C_DATA_PARAMETERS; +typedef struct _SET_UP_HW_I2C_DATA_PARAMETERS +{ + USHORT usPrescale; //Ratio between Engine clock and I2C clock + UCHAR ucSlaveAddr; //Write to which slave + UCHAR ucLineNumber; //Write from which HW assisted line +}SET_UP_HW_I2C_DATA_PARAMETERS; + /**************************************************************************/ #define SPEED_FAN_CONTROL_PS_ALLOCATION WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS -/****************************************************************************/ -/* Structures used by PowerConnectorDetectionTable */ -/****************************************************************************/ -typedef struct _POWER_CONNECTOR_DETECTION_PARAMETERS { - UCHAR ucPowerConnectorStatus; /* Used for return value 0: detected, 1:not detected */ - UCHAR ucPwrBehaviorId; - USHORT usPwrBudget; /* how much power currently boot to in unit of watt */ -} POWER_CONNECTOR_DETECTION_PARAMETERS; - -typedef struct POWER_CONNECTOR_DETECTION_PS_ALLOCATION { - UCHAR ucPowerConnectorStatus; /* Used for return value 0: detected, 1:not detected */ - UCHAR ucReserved; - USHORT usPwrBudget; /* how much power currently boot to in unit of watt */ - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; -} POWER_CONNECTOR_DETECTION_PS_ALLOCATION; +/****************************************************************************/ +// Structures used by PowerConnectorDetectionTable +/****************************************************************************/ +typedef struct _POWER_CONNECTOR_DETECTION_PARAMETERS +{ + UCHAR ucPowerConnectorStatus; //Used for return value 0: detected, 1:not detected + UCHAR ucPwrBehaviorId; + USHORT usPwrBudget; //how much power currently boot to in unit of watt +}POWER_CONNECTOR_DETECTION_PARAMETERS; + +typedef struct POWER_CONNECTOR_DETECTION_PS_ALLOCATION +{ + UCHAR ucPowerConnectorStatus; //Used for return value 0: detected, 1:not detected + UCHAR ucReserved; + USHORT usPwrBudget; //how much power currently boot to in unit of watt + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; +}POWER_CONNECTOR_DETECTION_PS_ALLOCATION; /****************************LVDS SS Command Table Definitions**********************/ -/****************************************************************************/ -/* Structures used by EnableSpreadSpectrumOnPPLLTable */ -/****************************************************************************/ -typedef struct _ENABLE_LVDS_SS_PARAMETERS { - USHORT usSpreadSpectrumPercentage; - UCHAR ucSpreadSpectrumType; /* Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD */ - UCHAR ucSpreadSpectrumStepSize_Delay; /* bits3:2 SS_STEP_SIZE; bit 6:4 SS_DELAY */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucPadding[3]; -} ENABLE_LVDS_SS_PARAMETERS; - -/* ucTableFormatRevision=1,ucTableContentRevision=2 */ -typedef struct _ENABLE_LVDS_SS_PARAMETERS_V2 { - USHORT usSpreadSpectrumPercentage; - UCHAR ucSpreadSpectrumType; /* Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD */ - UCHAR ucSpreadSpectrumStep; /* */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucSpreadSpectrumDelay; - UCHAR ucSpreadSpectrumRange; - UCHAR ucPadding; -} ENABLE_LVDS_SS_PARAMETERS_V2; - -/* This new structure is based on ENABLE_LVDS_SS_PARAMETERS but expands to SS on PPLL, so other devices can use SS. */ -typedef struct _ENABLE_SPREAD_SPECTRUM_ON_PPLL { - USHORT usSpreadSpectrumPercentage; - UCHAR ucSpreadSpectrumType; /* Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD */ - UCHAR ucSpreadSpectrumStep; /* */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucSpreadSpectrumDelay; - UCHAR ucSpreadSpectrumRange; - UCHAR ucPpll; /* ATOM_PPLL1/ATOM_PPLL2 */ -} ENABLE_SPREAD_SPECTRUM_ON_PPLL; +/****************************************************************************/ +// Structures used by EnableSpreadSpectrumOnPPLLTable +/****************************************************************************/ +typedef struct _ENABLE_LVDS_SS_PARAMETERS +{ + USHORT usSpreadSpectrumPercentage; + UCHAR ucSpreadSpectrumType; //Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD + UCHAR ucSpreadSpectrumStepSize_Delay; //bits3:2 SS_STEP_SIZE; bit 6:4 SS_DELAY + UCHAR ucEnable; //ATOM_ENABLE or ATOM_DISABLE + UCHAR ucPadding[3]; +}ENABLE_LVDS_SS_PARAMETERS; + +//ucTableFormatRevision=1,ucTableContentRevision=2 +typedef struct _ENABLE_LVDS_SS_PARAMETERS_V2 +{ + USHORT usSpreadSpectrumPercentage; + UCHAR ucSpreadSpectrumType; //Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD + UCHAR ucSpreadSpectrumStep; // + UCHAR ucEnable; //ATOM_ENABLE or ATOM_DISABLE + UCHAR ucSpreadSpectrumDelay; + UCHAR ucSpreadSpectrumRange; + UCHAR ucPadding; +}ENABLE_LVDS_SS_PARAMETERS_V2; + +//This new structure is based on ENABLE_LVDS_SS_PARAMETERS but expands to SS on PPLL, so other devices can use SS. +typedef struct _ENABLE_SPREAD_SPECTRUM_ON_PPLL +{ + USHORT usSpreadSpectrumPercentage; + UCHAR ucSpreadSpectrumType; // Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD + UCHAR ucSpreadSpectrumStep; // + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + UCHAR ucSpreadSpectrumDelay; + UCHAR ucSpreadSpectrumRange; + UCHAR ucPpll; // ATOM_PPLL1/ATOM_PPLL2 +}ENABLE_SPREAD_SPECTRUM_ON_PPLL; + +typedef struct _ENABLE_SPREAD_SPECTRUM_ON_PPLL_V2 +{ + USHORT usSpreadSpectrumPercentage; + UCHAR ucSpreadSpectrumType; // Bit[0]: 0-Down Spread,1-Center Spread. + // Bit[1]: 1-Ext. 0-Int. + // Bit[3:2]: =0 P1PLL =1 P2PLL =2 DCPLL + // Bits[7:4] reserved + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + USHORT usSpreadSpectrumAmount; // Includes SS_AMOUNT_FBDIV[7:0] and SS_AMOUNT_NFRAC_SLIP[11:8] + USHORT usSpreadSpectrumStep; // SS_STEP_SIZE_DSFRAC +}ENABLE_SPREAD_SPECTRUM_ON_PPLL_V2; + +#define ATOM_PPLL_SS_TYPE_V2_DOWN_SPREAD 0x00 +#define ATOM_PPLL_SS_TYPE_V2_CENTRE_SPREAD 0x01 +#define ATOM_PPLL_SS_TYPE_V2_EXT_SPREAD 0x02 +#define ATOM_PPLL_SS_TYPE_V2_PPLL_SEL_MASK 0x0c +#define ATOM_PPLL_SS_TYPE_V2_P1PLL 0x00 +#define ATOM_PPLL_SS_TYPE_V2_P2PLL 0x04 +#define ATOM_PPLL_SS_TYPE_V2_DCPLL 0x08 +#define ATOM_PPLL_SS_AMOUNT_V2_FBDIV_MASK 0x00FF +#define ATOM_PPLL_SS_AMOUNT_V2_FBDIV_SHIFT 0 +#define ATOM_PPLL_SS_AMOUNT_V2_NFRAC_MASK 0x0F00 +#define ATOM_PPLL_SS_AMOUNT_V2_NFRAC_SHIFT 8 #define ENABLE_SPREAD_SPECTRUM_ON_PPLL_PS_ALLOCATION ENABLE_SPREAD_SPECTRUM_ON_PPLL /**************************************************************************/ -typedef struct _SET_PIXEL_CLOCK_PS_ALLOCATION { - PIXEL_CLOCK_PARAMETERS sPCLKInput; - ENABLE_SPREAD_SPECTRUM_ON_PPLL sReserved; /* Caller doesn't need to init this portion */ -} SET_PIXEL_CLOCK_PS_ALLOCATION; +typedef struct _SET_PIXEL_CLOCK_PS_ALLOCATION +{ + PIXEL_CLOCK_PARAMETERS sPCLKInput; + ENABLE_SPREAD_SPECTRUM_ON_PPLL sReserved;//Caller doesn't need to init this portion +}SET_PIXEL_CLOCK_PS_ALLOCATION; #define ENABLE_VGA_RENDER_PS_ALLOCATION SET_PIXEL_CLOCK_PS_ALLOCATION -/****************************************************************************/ -/* Structures used by ### */ -/****************************************************************************/ -typedef struct _MEMORY_TRAINING_PARAMETERS { - ULONG ulTargetMemoryClock; /* In 10Khz unit */ -} MEMORY_TRAINING_PARAMETERS; +/****************************************************************************/ +// Structures used by ### +/****************************************************************************/ +typedef struct _MEMORY_TRAINING_PARAMETERS +{ + ULONG ulTargetMemoryClock; //In 10Khz unit +}MEMORY_TRAINING_PARAMETERS; #define MEMORY_TRAINING_PS_ALLOCATION MEMORY_TRAINING_PARAMETERS + /****************************LVDS and other encoder command table definitions **********************/ -/****************************************************************************/ -/* Structures used by LVDSEncoderControlTable (Before DCE30) */ -/* LVTMAEncoderControlTable (Before DCE30) */ -/* TMDSAEncoderControlTable (Before DCE30) */ -/****************************************************************************/ -typedef struct _LVDS_ENCODER_CONTROL_PARAMETERS { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - UCHAR ucMisc; /* bit0=0: Enable single link */ - /* =1: Enable dual link */ - /* Bit1=0: 666RGB */ - /* =1: 888RGB */ - UCHAR ucAction; /* 0: turn off encoder */ - /* 1: setup and turn on encoder */ -} LVDS_ENCODER_CONTROL_PARAMETERS; -#define LVDS_ENCODER_CONTROL_PS_ALLOCATION LVDS_ENCODER_CONTROL_PARAMETERS +/****************************************************************************/ +// Structures used by LVDSEncoderControlTable (Before DCE30) +// LVTMAEncoderControlTable (Before DCE30) +// TMDSAEncoderControlTable (Before DCE30) +/****************************************************************************/ +typedef struct _LVDS_ENCODER_CONTROL_PARAMETERS +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + UCHAR ucMisc; // bit0=0: Enable single link + // =1: Enable dual link + // Bit1=0: 666RGB + // =1: 888RGB + UCHAR ucAction; // 0: turn off encoder + // 1: setup and turn on encoder +}LVDS_ENCODER_CONTROL_PARAMETERS; +#define LVDS_ENCODER_CONTROL_PS_ALLOCATION LVDS_ENCODER_CONTROL_PARAMETERS + #define TMDS1_ENCODER_CONTROL_PARAMETERS LVDS_ENCODER_CONTROL_PARAMETERS #define TMDS1_ENCODER_CONTROL_PS_ALLOCATION TMDS1_ENCODER_CONTROL_PARAMETERS #define TMDS2_ENCODER_CONTROL_PARAMETERS TMDS1_ENCODER_CONTROL_PARAMETERS #define TMDS2_ENCODER_CONTROL_PS_ALLOCATION TMDS2_ENCODER_CONTROL_PARAMETERS -/* ucTableFormatRevision=1,ucTableContentRevision=2 */ -typedef struct _LVDS_ENCODER_CONTROL_PARAMETERS_V2 { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - UCHAR ucMisc; /* see PANEL_ENCODER_MISC_xx definitions below */ - UCHAR ucAction; /* 0: turn off encoder */ - /* 1: setup and turn on encoder */ - UCHAR ucTruncate; /* bit0=0: Disable truncate */ - /* =1: Enable truncate */ - /* bit4=0: 666RGB */ - /* =1: 888RGB */ - UCHAR ucSpatial; /* bit0=0: Disable spatial dithering */ - /* =1: Enable spatial dithering */ - /* bit4=0: 666RGB */ - /* =1: 888RGB */ - UCHAR ucTemporal; /* bit0=0: Disable temporal dithering */ - /* =1: Enable temporal dithering */ - /* bit4=0: 666RGB */ - /* =1: 888RGB */ - /* bit5=0: Gray level 2 */ - /* =1: Gray level 4 */ - UCHAR ucFRC; /* bit4=0: 25FRC_SEL pattern E */ - /* =1: 25FRC_SEL pattern F */ - /* bit6:5=0: 50FRC_SEL pattern A */ - /* =1: 50FRC_SEL pattern B */ - /* =2: 50FRC_SEL pattern C */ - /* =3: 50FRC_SEL pattern D */ - /* bit7=0: 75FRC_SEL pattern E */ - /* =1: 75FRC_SEL pattern F */ -} LVDS_ENCODER_CONTROL_PARAMETERS_V2; -#define LVDS_ENCODER_CONTROL_PS_ALLOCATION_V2 LVDS_ENCODER_CONTROL_PARAMETERS_V2 +//ucTableFormatRevision=1,ucTableContentRevision=2 +typedef struct _LVDS_ENCODER_CONTROL_PARAMETERS_V2 +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + UCHAR ucMisc; // see PANEL_ENCODER_MISC_xx defintions below + UCHAR ucAction; // 0: turn off encoder + // 1: setup and turn on encoder + UCHAR ucTruncate; // bit0=0: Disable truncate + // =1: Enable truncate + // bit4=0: 666RGB + // =1: 888RGB + UCHAR ucSpatial; // bit0=0: Disable spatial dithering + // =1: Enable spatial dithering + // bit4=0: 666RGB + // =1: 888RGB + UCHAR ucTemporal; // bit0=0: Disable temporal dithering + // =1: Enable temporal dithering + // bit4=0: 666RGB + // =1: 888RGB + // bit5=0: Gray level 2 + // =1: Gray level 4 + UCHAR ucFRC; // bit4=0: 25FRC_SEL pattern E + // =1: 25FRC_SEL pattern F + // bit6:5=0: 50FRC_SEL pattern A + // =1: 50FRC_SEL pattern B + // =2: 50FRC_SEL pattern C + // =3: 50FRC_SEL pattern D + // bit7=0: 75FRC_SEL pattern E + // =1: 75FRC_SEL pattern F +}LVDS_ENCODER_CONTROL_PARAMETERS_V2; +#define LVDS_ENCODER_CONTROL_PS_ALLOCATION_V2 LVDS_ENCODER_CONTROL_PARAMETERS_V2 + #define TMDS1_ENCODER_CONTROL_PARAMETERS_V2 LVDS_ENCODER_CONTROL_PARAMETERS_V2 #define TMDS1_ENCODER_CONTROL_PS_ALLOCATION_V2 TMDS1_ENCODER_CONTROL_PARAMETERS_V2 - + #define TMDS2_ENCODER_CONTROL_PARAMETERS_V2 TMDS1_ENCODER_CONTROL_PARAMETERS_V2 #define TMDS2_ENCODER_CONTROL_PS_ALLOCATION_V2 TMDS2_ENCODER_CONTROL_PARAMETERS_V2 @@ -1185,38 +1536,42 @@ typedef struct _LVDS_ENCODER_CONTROL_PARAMETERS_V2 { #define TMDS2_ENCODER_CONTROL_PARAMETERS_V3 LVDS_ENCODER_CONTROL_PARAMETERS_V3 #define TMDS2_ENCODER_CONTROL_PS_ALLOCATION_V3 TMDS2_ENCODER_CONTROL_PARAMETERS_V3 -/****************************************************************************/ -/* Structures used by ### */ -/****************************************************************************/ -typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS { - UCHAR ucEnable; /* Enable or Disable External TMDS encoder */ - UCHAR ucMisc; /* Bit0=0:Enable Single link;=1:Enable Dual link;Bit1 {=0:666RGB, =1:888RGB} */ - UCHAR ucPadding[2]; -} ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS; - -typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION { - ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS sXTmdsEncoder; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; /* Caller doesn't need to init this portion */ -} ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION; +/****************************************************************************/ +// Structures used by ### +/****************************************************************************/ +typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS +{ + UCHAR ucEnable; // Enable or Disable External TMDS encoder + UCHAR ucMisc; // Bit0=0:Enable Single link;=1:Enable Dual link;Bit1 {=0:666RGB, =1:888RGB} + UCHAR ucPadding[2]; +}ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS; + +typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION +{ + ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS sXTmdsEncoder; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; //Caller doesn't need to init this portion +}ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION; #define ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS_V2 LVDS_ENCODER_CONTROL_PARAMETERS_V2 -typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION_V2 { - ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS_V2 sXTmdsEncoder; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; /* Caller doesn't need to init this portion */ -} ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION_V2; +typedef struct _ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION_V2 +{ + ENABLE_EXTERNAL_TMDS_ENCODER_PARAMETERS_V2 sXTmdsEncoder; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; //Caller doesn't need to init this portion +}ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION_V2; -typedef struct _EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION { - DIG_ENCODER_CONTROL_PARAMETERS sDigEncoder; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; -} EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION; +typedef struct _EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION +{ + DIG_ENCODER_CONTROL_PARAMETERS sDigEncoder; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; +}EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION; -/****************************************************************************/ -/* Structures used by DVOEncoderControlTable */ -/****************************************************************************/ -/* ucTableFormatRevision=1,ucTableContentRevision=3 */ +/****************************************************************************/ +// Structures used by DVOEncoderControlTable +/****************************************************************************/ +//ucTableFormatRevision=1,ucTableContentRevision=3 -/* ucDVOConfig: */ +//ucDVOConfig: #define DVO_ENCODER_CONFIG_RATE_SEL 0x01 #define DVO_ENCODER_CONFIG_DDR_SPEED 0x00 #define DVO_ENCODER_CONFIG_SDR_SPEED 0x01 @@ -1225,21 +1580,22 @@ typedef struct _EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION { #define DVO_ENCODER_CONFIG_UPPER12BIT 0x04 #define DVO_ENCODER_CONFIG_24BIT 0x08 -typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3 { - USHORT usPixelClock; - UCHAR ucDVOConfig; - UCHAR ucAction; /* ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT */ - UCHAR ucReseved[4]; -} DVO_ENCODER_CONTROL_PARAMETERS_V3; +typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3 +{ + USHORT usPixelClock; + UCHAR ucDVOConfig; + UCHAR ucAction; //ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT + UCHAR ucReseved[4]; +}DVO_ENCODER_CONTROL_PARAMETERS_V3; #define DVO_ENCODER_CONTROL_PS_ALLOCATION_V3 DVO_ENCODER_CONTROL_PARAMETERS_V3 -/* ucTableFormatRevision=1 */ -/* ucTableContentRevision=3 structure is not changed but usMisc add bit 1 as another input for */ -/* bit1=0: non-coherent mode */ -/* =1: coherent mode */ +//ucTableFormatRevision=1 +//ucTableContentRevision=3 structure is not changed but usMisc add bit 1 as another input for +// bit1=0: non-coherent mode +// =1: coherent mode -/* ========================================================================================== */ -/* Only change is here next time when changing encoder parameter definitions again! */ +//========================================================================================== +//Only change is here next time when changing encoder parameter definitions again! #define LVDS_ENCODER_CONTROL_PARAMETERS_LAST LVDS_ENCODER_CONTROL_PARAMETERS_V3 #define LVDS_ENCODER_CONTROL_PS_ALLOCATION_LAST LVDS_ENCODER_CONTROL_PARAMETERS_LAST @@ -1252,7 +1608,7 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3 { #define DVO_ENCODER_CONTROL_PARAMETERS_LAST DVO_ENCODER_CONTROL_PARAMETERS #define DVO_ENCODER_CONTROL_PS_ALLOCATION_LAST DVO_ENCODER_CONTROL_PS_ALLOCATION -/* ========================================================================================== */ +//========================================================================================== #define PANEL_ENCODER_MISC_DUAL 0x01 #define PANEL_ENCODER_MISC_COHERENT 0x02 #define PANEL_ENCODER_MISC_TMDS_LINKB 0x04 @@ -1281,159 +1637,159 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3 { #define PANEL_ENCODER_75FRC_E 0x00 #define PANEL_ENCODER_75FRC_F 0x80 -/****************************************************************************/ -/* Structures used by SetVoltageTable */ -/****************************************************************************/ +/****************************************************************************/ +// Structures used by SetVoltageTable +/****************************************************************************/ #define SET_VOLTAGE_TYPE_ASIC_VDDC 1 #define SET_VOLTAGE_TYPE_ASIC_MVDDC 2 #define SET_VOLTAGE_TYPE_ASIC_MVDDQ 3 #define SET_VOLTAGE_TYPE_ASIC_VDDCI 4 #define SET_VOLTAGE_INIT_MODE 5 -#define SET_VOLTAGE_GET_MAX_VOLTAGE 6 /* Gets the Max. voltage for the soldered Asic */ +#define SET_VOLTAGE_GET_MAX_VOLTAGE 6 //Gets the Max. voltage for the soldered Asic #define SET_ASIC_VOLTAGE_MODE_ALL_SOURCE 0x1 #define SET_ASIC_VOLTAGE_MODE_SOURCE_A 0x2 #define SET_ASIC_VOLTAGE_MODE_SOURCE_B 0x4 #define SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE 0x0 -#define SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL 0x1 +#define SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL 0x1 #define SET_ASIC_VOLTAGE_MODE_GET_GPIOMASK 0x2 -typedef struct _SET_VOLTAGE_PARAMETERS { - UCHAR ucVoltageType; /* To tell which voltage to set up, VDDC/MVDDC/MVDDQ */ - UCHAR ucVoltageMode; /* To set all, to set source A or source B or ... */ - UCHAR ucVoltageIndex; /* An index to tell which voltage level */ - UCHAR ucReserved; -} SET_VOLTAGE_PARAMETERS; - -typedef struct _SET_VOLTAGE_PARAMETERS_V2 { - UCHAR ucVoltageType; /* To tell which voltage to set up, VDDC/MVDDC/MVDDQ */ - UCHAR ucVoltageMode; /* Not used, maybe use for state machine for differen power mode */ - USHORT usVoltageLevel; /* real voltage level */ -} SET_VOLTAGE_PARAMETERS_V2; - -typedef struct _SET_VOLTAGE_PS_ALLOCATION { - SET_VOLTAGE_PARAMETERS sASICSetVoltage; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; -} SET_VOLTAGE_PS_ALLOCATION; - -/****************************************************************************/ -/* Structures used by TVEncoderControlTable */ -/****************************************************************************/ -typedef struct _TV_ENCODER_CONTROL_PARAMETERS { - USHORT usPixelClock; /* in 10KHz; for bios convenient */ - UCHAR ucTvStandard; /* See definition "ATOM_TV_NTSC ..." */ - UCHAR ucAction; /* 0: turn off encoder */ - /* 1: setup and turn on encoder */ -} TV_ENCODER_CONTROL_PARAMETERS; - -typedef struct _TV_ENCODER_CONTROL_PS_ALLOCATION { - TV_ENCODER_CONTROL_PARAMETERS sTVEncoder; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; /* Don't set this one */ -} TV_ENCODER_CONTROL_PS_ALLOCATION; - -/* ==============================Data Table Portion==================================== */ - -#ifdef UEFI_BUILD -#define UTEMP USHORT -#define USHORT void* -#endif - -/****************************************************************************/ -/* Structure used in Data.mtb */ -/****************************************************************************/ -typedef struct _ATOM_MASTER_LIST_OF_DATA_TABLES { - USHORT UtilityPipeLine; /* Offest for the utility to get parser info,Don't change this position! */ - USHORT MultimediaCapabilityInfo; /* Only used by MM Lib,latest version 1.1, not configuable from Bios, need to include the table to build Bios */ - USHORT MultimediaConfigInfo; /* Only used by MM Lib,latest version 2.1, not configuable from Bios, need to include the table to build Bios */ - USHORT StandardVESA_Timing; /* Only used by Bios */ - USHORT FirmwareInfo; /* Shared by various SW components,latest version 1.4 */ - USHORT DAC_Info; /* Will be obsolete from R600 */ - USHORT LVDS_Info; /* Shared by various SW components,latest version 1.1 */ - USHORT TMDS_Info; /* Will be obsolete from R600 */ - USHORT AnalogTV_Info; /* Shared by various SW components,latest version 1.1 */ - USHORT SupportedDevicesInfo; /* Will be obsolete from R600 */ - USHORT GPIO_I2C_Info; /* Shared by various SW components,latest version 1.2 will be used from R600 */ - USHORT VRAM_UsageByFirmware; /* Shared by various SW components,latest version 1.3 will be used from R600 */ - USHORT GPIO_Pin_LUT; /* Shared by various SW components,latest version 1.1 */ - USHORT VESA_ToInternalModeLUT; /* Only used by Bios */ - USHORT ComponentVideoInfo; /* Shared by various SW components,latest version 2.1 will be used from R600 */ - USHORT PowerPlayInfo; /* Shared by various SW components,latest version 2.1,new design from R600 */ - USHORT CompassionateData; /* Will be obsolete from R600 */ - USHORT SaveRestoreInfo; /* Only used by Bios */ - USHORT PPLL_SS_Info; /* Shared by various SW components,latest version 1.2, used to call SS_Info, change to new name because of int ASIC SS info */ - USHORT OemInfo; /* Defined and used by external SW, should be obsolete soon */ - USHORT XTMDS_Info; /* Will be obsolete from R600 */ - USHORT MclkSS_Info; /* Shared by various SW components,latest version 1.1, only enabled when ext SS chip is used */ - USHORT Object_Header; /* Shared by various SW components,latest version 1.1 */ - USHORT IndirectIOAccess; /* Only used by Bios,this table position can't change at all!! */ - USHORT MC_InitParameter; /* Only used by command table */ - USHORT ASIC_VDDC_Info; /* Will be obsolete from R600 */ - USHORT ASIC_InternalSS_Info; /* New tabel name from R600, used to be called "ASIC_MVDDC_Info" */ - USHORT TV_VideoMode; /* Only used by command table */ - USHORT VRAM_Info; /* Only used by command table, latest version 1.3 */ - USHORT MemoryTrainingInfo; /* Used for VBIOS and Diag utility for memory training purpose since R600. the new table rev start from 2.1 */ - USHORT IntegratedSystemInfo; /* Shared by various SW components */ - USHORT ASIC_ProfilingInfo; /* New table name from R600, used to be called "ASIC_VDDCI_Info" for pre-R600 */ - USHORT VoltageObjectInfo; /* Shared by various SW components, latest version 1.1 */ - USHORT PowerSourceInfo; /* Shared by various SW components, latest versoin 1.1 */ -} ATOM_MASTER_LIST_OF_DATA_TABLES; - -#ifdef UEFI_BUILD -#define USHORT UTEMP -#endif +typedef struct _SET_VOLTAGE_PARAMETERS +{ + UCHAR ucVoltageType; // To tell which voltage to set up, VDDC/MVDDC/MVDDQ + UCHAR ucVoltageMode; // To set all, to set source A or source B or ... + UCHAR ucVoltageIndex; // An index to tell which voltage level + UCHAR ucReserved; +}SET_VOLTAGE_PARAMETERS; -typedef struct _ATOM_MASTER_DATA_TABLE { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_MASTER_LIST_OF_DATA_TABLES ListOfDataTables; -} ATOM_MASTER_DATA_TABLE; +typedef struct _SET_VOLTAGE_PARAMETERS_V2 +{ + UCHAR ucVoltageType; // To tell which voltage to set up, VDDC/MVDDC/MVDDQ + UCHAR ucVoltageMode; // Not used, maybe use for state machine for differen power mode + USHORT usVoltageLevel; // real voltage level +}SET_VOLTAGE_PARAMETERS_V2; -/****************************************************************************/ -/* Structure used in MultimediaCapabilityInfoTable */ -/****************************************************************************/ -typedef struct _ATOM_MULTIMEDIA_CAPABILITY_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulSignature; /* HW info table signature string "$ATI" */ - UCHAR ucI2C_Type; /* I2C type (normal GP_IO, ImpactTV GP_IO, Dedicated I2C pin, etc) */ - UCHAR ucTV_OutInfo; /* Type of TV out supported (3:0) and video out crystal frequency (6:4) and TV data port (7) */ - UCHAR ucVideoPortInfo; /* Provides the video port capabilities */ - UCHAR ucHostPortInfo; /* Provides host port configuration information */ -} ATOM_MULTIMEDIA_CAPABILITY_INFO; +typedef struct _SET_VOLTAGE_PS_ALLOCATION +{ + SET_VOLTAGE_PARAMETERS sASICSetVoltage; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; +}SET_VOLTAGE_PS_ALLOCATION; + +/****************************************************************************/ +// Structures used by TVEncoderControlTable +/****************************************************************************/ +typedef struct _TV_ENCODER_CONTROL_PARAMETERS +{ + USHORT usPixelClock; // in 10KHz; for bios convenient + UCHAR ucTvStandard; // See definition "ATOM_TV_NTSC ..." + UCHAR ucAction; // 0: turn off encoder + // 1: setup and turn on encoder +}TV_ENCODER_CONTROL_PARAMETERS; -/****************************************************************************/ -/* Structure used in MultimediaConfigInfoTable */ -/****************************************************************************/ -typedef struct _ATOM_MULTIMEDIA_CONFIG_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulSignature; /* MM info table signature sting "$MMT" */ - UCHAR ucTunerInfo; /* Type of tuner installed on the adapter (4:0) and video input for tuner (7:5) */ - UCHAR ucAudioChipInfo; /* List the audio chip type (3:0) product type (4) and OEM revision (7:5) */ - UCHAR ucProductID; /* Defines as OEM ID or ATI board ID dependent on product type setting */ - UCHAR ucMiscInfo1; /* Tuner voltage (1:0) HW teletext support (3:2) FM audio decoder (5:4) reserved (6) audio scrambling (7) */ - UCHAR ucMiscInfo2; /* I2S input config (0) I2S output config (1) I2S Audio Chip (4:2) SPDIF Output Config (5) reserved (7:6) */ - UCHAR ucMiscInfo3; /* Video Decoder Type (3:0) Video In Standard/Crystal (7:4) */ - UCHAR ucMiscInfo4; /* Video Decoder Host Config (2:0) reserved (7:3) */ - UCHAR ucVideoInput0Info; /* Video Input 0 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) */ - UCHAR ucVideoInput1Info; /* Video Input 1 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) */ - UCHAR ucVideoInput2Info; /* Video Input 2 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) */ - UCHAR ucVideoInput3Info; /* Video Input 3 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) */ - UCHAR ucVideoInput4Info; /* Video Input 4 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) */ -} ATOM_MULTIMEDIA_CONFIG_INFO; +typedef struct _TV_ENCODER_CONTROL_PS_ALLOCATION +{ + TV_ENCODER_CONTROL_PARAMETERS sTVEncoder; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; // Don't set this one +}TV_ENCODER_CONTROL_PS_ALLOCATION; -/****************************************************************************/ -/* Structures used in FirmwareInfoTable */ -/****************************************************************************/ +//==============================Data Table Portion==================================== -/* usBIOSCapability Definition: */ -/* Bit 0 = 0: Bios image is not Posted, =1:Bios image is Posted; */ -/* Bit 1 = 0: Dual CRTC is not supported, =1: Dual CRTC is supported; */ -/* Bit 2 = 0: Extended Desktop is not supported, =1: Extended Desktop is supported; */ -/* Others: Reserved */ +/****************************************************************************/ +// Structure used in Data.mtb +/****************************************************************************/ +typedef struct _ATOM_MASTER_LIST_OF_DATA_TABLES +{ + USHORT UtilityPipeLine; // Offest for the utility to get parser info,Don't change this position! + USHORT MultimediaCapabilityInfo; // Only used by MM Lib,latest version 1.1, not configuable from Bios, need to include the table to build Bios + USHORT MultimediaConfigInfo; // Only used by MM Lib,latest version 2.1, not configuable from Bios, need to include the table to build Bios + USHORT StandardVESA_Timing; // Only used by Bios + USHORT FirmwareInfo; // Shared by various SW components,latest version 1.4 + USHORT DAC_Info; // Will be obsolete from R600 + USHORT LVDS_Info; // Shared by various SW components,latest version 1.1 + USHORT TMDS_Info; // Will be obsolete from R600 + USHORT AnalogTV_Info; // Shared by various SW components,latest version 1.1 + USHORT SupportedDevicesInfo; // Will be obsolete from R600 + USHORT GPIO_I2C_Info; // Shared by various SW components,latest version 1.2 will be used from R600 + USHORT VRAM_UsageByFirmware; // Shared by various SW components,latest version 1.3 will be used from R600 + USHORT GPIO_Pin_LUT; // Shared by various SW components,latest version 1.1 + USHORT VESA_ToInternalModeLUT; // Only used by Bios + USHORT ComponentVideoInfo; // Shared by various SW components,latest version 2.1 will be used from R600 + USHORT PowerPlayInfo; // Shared by various SW components,latest version 2.1,new design from R600 + USHORT CompassionateData; // Will be obsolete from R600 + USHORT SaveRestoreInfo; // Only used by Bios + USHORT PPLL_SS_Info; // Shared by various SW components,latest version 1.2, used to call SS_Info, change to new name because of int ASIC SS info + USHORT OemInfo; // Defined and used by external SW, should be obsolete soon + USHORT XTMDS_Info; // Will be obsolete from R600 + USHORT MclkSS_Info; // Shared by various SW components,latest version 1.1, only enabled when ext SS chip is used + USHORT Object_Header; // Shared by various SW components,latest version 1.1 + USHORT IndirectIOAccess; // Only used by Bios,this table position can't change at all!! + USHORT MC_InitParameter; // Only used by command table + USHORT ASIC_VDDC_Info; // Will be obsolete from R600 + USHORT ASIC_InternalSS_Info; // New tabel name from R600, used to be called "ASIC_MVDDC_Info" + USHORT TV_VideoMode; // Only used by command table + USHORT VRAM_Info; // Only used by command table, latest version 1.3 + USHORT MemoryTrainingInfo; // Used for VBIOS and Diag utility for memory training purpose since R600. the new table rev start from 2.1 + USHORT IntegratedSystemInfo; // Shared by various SW components + USHORT ASIC_ProfilingInfo; // New table name from R600, used to be called "ASIC_VDDCI_Info" for pre-R600 + USHORT VoltageObjectInfo; // Shared by various SW components, latest version 1.1 + USHORT PowerSourceInfo; // Shared by various SW components, latest versoin 1.1 +}ATOM_MASTER_LIST_OF_DATA_TABLES; + +typedef struct _ATOM_MASTER_DATA_TABLE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_MASTER_LIST_OF_DATA_TABLES ListOfDataTables; +}ATOM_MASTER_DATA_TABLE; + +/****************************************************************************/ +// Structure used in MultimediaCapabilityInfoTable +/****************************************************************************/ +typedef struct _ATOM_MULTIMEDIA_CAPABILITY_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulSignature; // HW info table signature string "$ATI" + UCHAR ucI2C_Type; // I2C type (normal GP_IO, ImpactTV GP_IO, Dedicated I2C pin, etc) + UCHAR ucTV_OutInfo; // Type of TV out supported (3:0) and video out crystal frequency (6:4) and TV data port (7) + UCHAR ucVideoPortInfo; // Provides the video port capabilities + UCHAR ucHostPortInfo; // Provides host port configuration information +}ATOM_MULTIMEDIA_CAPABILITY_INFO; + +/****************************************************************************/ +// Structure used in MultimediaConfigInfoTable +/****************************************************************************/ +typedef struct _ATOM_MULTIMEDIA_CONFIG_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulSignature; // MM info table signature sting "$MMT" + UCHAR ucTunerInfo; // Type of tuner installed on the adapter (4:0) and video input for tuner (7:5) + UCHAR ucAudioChipInfo; // List the audio chip type (3:0) product type (4) and OEM revision (7:5) + UCHAR ucProductID; // Defines as OEM ID or ATI board ID dependent on product type setting + UCHAR ucMiscInfo1; // Tuner voltage (1:0) HW teletext support (3:2) FM audio decoder (5:4) reserved (6) audio scrambling (7) + UCHAR ucMiscInfo2; // I2S input config (0) I2S output config (1) I2S Audio Chip (4:2) SPDIF Output Config (5) reserved (7:6) + UCHAR ucMiscInfo3; // Video Decoder Type (3:0) Video In Standard/Crystal (7:4) + UCHAR ucMiscInfo4; // Video Decoder Host Config (2:0) reserved (7:3) + UCHAR ucVideoInput0Info;// Video Input 0 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) + UCHAR ucVideoInput1Info;// Video Input 1 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) + UCHAR ucVideoInput2Info;// Video Input 2 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) + UCHAR ucVideoInput3Info;// Video Input 3 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) + UCHAR ucVideoInput4Info;// Video Input 4 Type (1:0) F/B setting (2) physical connector ID (5:3) reserved (7:6) +}ATOM_MULTIMEDIA_CONFIG_INFO; + +/****************************************************************************/ +// Structures used in FirmwareInfoTable +/****************************************************************************/ + +// usBIOSCapability Defintion: +// Bit 0 = 0: Bios image is not Posted, =1:Bios image is Posted; +// Bit 1 = 0: Dual CRTC is not supported, =1: Dual CRTC is supported; +// Bit 2 = 0: Extended Desktop is not supported, =1: Extended Desktop is supported; +// Others: Reserved #define ATOM_BIOS_INFO_ATOM_FIRMWARE_POSTED 0x0001 #define ATOM_BIOS_INFO_DUAL_CRTC_SUPPORT 0x0002 #define ATOM_BIOS_INFO_EXTENDED_DESKTOP_SUPPORT 0x0004 -#define ATOM_BIOS_INFO_MEMORY_CLOCK_SS_SUPPORT 0x0008 -#define ATOM_BIOS_INFO_ENGINE_CLOCK_SS_SUPPORT 0x0010 +#define ATOM_BIOS_INFO_MEMORY_CLOCK_SS_SUPPORT 0x0008 // (valid from v1.1 ~v1.4):=1: memclk SS enable, =0 memclk SS disable. +#define ATOM_BIOS_INFO_ENGINE_CLOCK_SS_SUPPORT 0x0010 // (valid from v1.1 ~v1.4):=1: engclk SS enable, =0 engclk SS disable. #define ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU 0x0020 #define ATOM_BIOS_INFO_WMI_SUPPORT 0x0040 #define ATOM_BIOS_INFO_PPMODE_ASSIGNGED_BY_SYSTEM 0x0080 @@ -1441,242 +1797,292 @@ typedef struct _ATOM_MULTIMEDIA_CONFIG_INFO { #define ATOM_BIOS_INFO_HYPERMEMORY_SIZE_MASK 0x1E00 #define ATOM_BIOS_INFO_VPOST_WITHOUT_FIRST_MODE_SET 0x2000 #define ATOM_BIOS_INFO_BIOS_SCRATCH6_SCL2_REDEFINE 0x4000 +#define ATOM_BIOS_INFO_MEMORY_CLOCK_EXT_SS_SUPPORT 0x0008 // (valid from v2.1 ): =1: memclk ss enable with external ss chip +#define ATOM_BIOS_INFO_ENGINE_CLOCK_EXT_SS_SUPPORT 0x0010 // (valid from v2.1 ): =1: engclk ss enable with external ss chip #ifndef _H2INC -/* Please don't add or expand this bitfield structure below, this one will retire soon.! */ -typedef struct _ATOM_FIRMWARE_CAPABILITY { +//Please don't add or expand this bitfield structure below, this one will retire soon.! +typedef struct _ATOM_FIRMWARE_CAPABILITY +{ #if ATOM_BIG_ENDIAN - USHORT Reserved:3; - USHORT HyperMemory_Size:4; - USHORT HyperMemory_Support:1; - USHORT PPMode_Assigned:1; - USHORT WMI_SUPPORT:1; - USHORT GPUControlsBL:1; - USHORT EngineClockSS_Support:1; - USHORT MemoryClockSS_Support:1; - USHORT ExtendedDesktopSupport:1; - USHORT DualCRTC_Support:1; - USHORT FirmwarePosted:1; + USHORT Reserved:3; + USHORT HyperMemory_Size:4; + USHORT HyperMemory_Support:1; + USHORT PPMode_Assigned:1; + USHORT WMI_SUPPORT:1; + USHORT GPUControlsBL:1; + USHORT EngineClockSS_Support:1; + USHORT MemoryClockSS_Support:1; + USHORT ExtendedDesktopSupport:1; + USHORT DualCRTC_Support:1; + USHORT FirmwarePosted:1; #else - USHORT FirmwarePosted:1; - USHORT DualCRTC_Support:1; - USHORT ExtendedDesktopSupport:1; - USHORT MemoryClockSS_Support:1; - USHORT EngineClockSS_Support:1; - USHORT GPUControlsBL:1; - USHORT WMI_SUPPORT:1; - USHORT PPMode_Assigned:1; - USHORT HyperMemory_Support:1; - USHORT HyperMemory_Size:4; - USHORT Reserved:3; + USHORT FirmwarePosted:1; + USHORT DualCRTC_Support:1; + USHORT ExtendedDesktopSupport:1; + USHORT MemoryClockSS_Support:1; + USHORT EngineClockSS_Support:1; + USHORT GPUControlsBL:1; + USHORT WMI_SUPPORT:1; + USHORT PPMode_Assigned:1; + USHORT HyperMemory_Support:1; + USHORT HyperMemory_Size:4; + USHORT Reserved:3; #endif -} ATOM_FIRMWARE_CAPABILITY; +}ATOM_FIRMWARE_CAPABILITY; -typedef union _ATOM_FIRMWARE_CAPABILITY_ACCESS { - ATOM_FIRMWARE_CAPABILITY sbfAccess; - USHORT susAccess; -} ATOM_FIRMWARE_CAPABILITY_ACCESS; +typedef union _ATOM_FIRMWARE_CAPABILITY_ACCESS +{ + ATOM_FIRMWARE_CAPABILITY sbfAccess; + USHORT susAccess; +}ATOM_FIRMWARE_CAPABILITY_ACCESS; #else -typedef union _ATOM_FIRMWARE_CAPABILITY_ACCESS { - USHORT susAccess; -} ATOM_FIRMWARE_CAPABILITY_ACCESS; +typedef union _ATOM_FIRMWARE_CAPABILITY_ACCESS +{ + USHORT susAccess; +}ATOM_FIRMWARE_CAPABILITY_ACCESS; #endif -typedef struct _ATOM_FIRMWARE_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulFirmwareRevision; - ULONG ulDefaultEngineClock; /* In 10Khz unit */ - ULONG ulDefaultMemoryClock; /* In 10Khz unit */ - ULONG ulDriverTargetEngineClock; /* In 10Khz unit */ - ULONG ulDriverTargetMemoryClock; /* In 10Khz unit */ - ULONG ulMaxEngineClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxMemoryClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxPixelClockPLL_Output; /* In 10Khz unit */ - ULONG ulASICMaxEngineClock; /* In 10Khz unit */ - ULONG ulASICMaxMemoryClock; /* In 10Khz unit */ - UCHAR ucASICMaxTemperature; - UCHAR ucPadding[3]; /* Don't use them */ - ULONG aulReservedForBIOS[3]; /* Don't use them */ - USHORT usMinEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Output; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Output; /* In 10Khz unit */ - USHORT usMaxPixelClock; /* In 10Khz unit, Max. Pclk */ - USHORT usMinPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMinPixelClockPLL_Output; /* In 10Khz unit, the definitions above can't change!!! */ - ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; - USHORT usReferenceClock; /* In 10Khz unit */ - USHORT usPM_RTS_Location; /* RTS PM4 starting location in ROM in 1Kb unit */ - UCHAR ucPM_RTS_StreamSize; /* RTS PM4 packets in Kb unit */ - UCHAR ucDesign_ID; /* Indicate what is the board design */ - UCHAR ucMemoryModule_ID; /* Indicate what is the board design */ -} ATOM_FIRMWARE_INFO; - -typedef struct _ATOM_FIRMWARE_INFO_V1_2 { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulFirmwareRevision; - ULONG ulDefaultEngineClock; /* In 10Khz unit */ - ULONG ulDefaultMemoryClock; /* In 10Khz unit */ - ULONG ulDriverTargetEngineClock; /* In 10Khz unit */ - ULONG ulDriverTargetMemoryClock; /* In 10Khz unit */ - ULONG ulMaxEngineClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxMemoryClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxPixelClockPLL_Output; /* In 10Khz unit */ - ULONG ulASICMaxEngineClock; /* In 10Khz unit */ - ULONG ulASICMaxMemoryClock; /* In 10Khz unit */ - UCHAR ucASICMaxTemperature; - UCHAR ucMinAllowedBL_Level; - UCHAR ucPadding[2]; /* Don't use them */ - ULONG aulReservedForBIOS[2]; /* Don't use them */ - ULONG ulMinPixelClockPLL_Output; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Output; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Output; /* In 10Khz unit */ - USHORT usMaxPixelClock; /* In 10Khz unit, Max. Pclk */ - USHORT usMinPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMinPixelClockPLL_Output; /* In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output */ - ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; - USHORT usReferenceClock; /* In 10Khz unit */ - USHORT usPM_RTS_Location; /* RTS PM4 starting location in ROM in 1Kb unit */ - UCHAR ucPM_RTS_StreamSize; /* RTS PM4 packets in Kb unit */ - UCHAR ucDesign_ID; /* Indicate what is the board design */ - UCHAR ucMemoryModule_ID; /* Indicate what is the board design */ -} ATOM_FIRMWARE_INFO_V1_2; - -typedef struct _ATOM_FIRMWARE_INFO_V1_3 { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulFirmwareRevision; - ULONG ulDefaultEngineClock; /* In 10Khz unit */ - ULONG ulDefaultMemoryClock; /* In 10Khz unit */ - ULONG ulDriverTargetEngineClock; /* In 10Khz unit */ - ULONG ulDriverTargetMemoryClock; /* In 10Khz unit */ - ULONG ulMaxEngineClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxMemoryClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxPixelClockPLL_Output; /* In 10Khz unit */ - ULONG ulASICMaxEngineClock; /* In 10Khz unit */ - ULONG ulASICMaxMemoryClock; /* In 10Khz unit */ - UCHAR ucASICMaxTemperature; - UCHAR ucMinAllowedBL_Level; - UCHAR ucPadding[2]; /* Don't use them */ - ULONG aulReservedForBIOS; /* Don't use them */ - ULONG ul3DAccelerationEngineClock; /* In 10Khz unit */ - ULONG ulMinPixelClockPLL_Output; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Output; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Output; /* In 10Khz unit */ - USHORT usMaxPixelClock; /* In 10Khz unit, Max. Pclk */ - USHORT usMinPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMinPixelClockPLL_Output; /* In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output */ - ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; - USHORT usReferenceClock; /* In 10Khz unit */ - USHORT usPM_RTS_Location; /* RTS PM4 starting location in ROM in 1Kb unit */ - UCHAR ucPM_RTS_StreamSize; /* RTS PM4 packets in Kb unit */ - UCHAR ucDesign_ID; /* Indicate what is the board design */ - UCHAR ucMemoryModule_ID; /* Indicate what is the board design */ -} ATOM_FIRMWARE_INFO_V1_3; - -typedef struct _ATOM_FIRMWARE_INFO_V1_4 { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulFirmwareRevision; - ULONG ulDefaultEngineClock; /* In 10Khz unit */ - ULONG ulDefaultMemoryClock; /* In 10Khz unit */ - ULONG ulDriverTargetEngineClock; /* In 10Khz unit */ - ULONG ulDriverTargetMemoryClock; /* In 10Khz unit */ - ULONG ulMaxEngineClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxMemoryClockPLL_Output; /* In 10Khz unit */ - ULONG ulMaxPixelClockPLL_Output; /* In 10Khz unit */ - ULONG ulASICMaxEngineClock; /* In 10Khz unit */ - ULONG ulASICMaxMemoryClock; /* In 10Khz unit */ - UCHAR ucASICMaxTemperature; - UCHAR ucMinAllowedBL_Level; - USHORT usBootUpVDDCVoltage; /* In MV unit */ - USHORT usLcdMinPixelClockPLL_Output; /* In MHz unit */ - USHORT usLcdMaxPixelClockPLL_Output; /* In MHz unit */ - ULONG ul3DAccelerationEngineClock; /* In 10Khz unit */ - ULONG ulMinPixelClockPLL_Output; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxEngineClockPLL_Input; /* In 10Khz unit */ - USHORT usMinEngineClockPLL_Output; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxMemoryClockPLL_Input; /* In 10Khz unit */ - USHORT usMinMemoryClockPLL_Output; /* In 10Khz unit */ - USHORT usMaxPixelClock; /* In 10Khz unit, Max. Pclk */ - USHORT usMinPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMaxPixelClockPLL_Input; /* In 10Khz unit */ - USHORT usMinPixelClockPLL_Output; /* In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output */ - ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; - USHORT usReferenceClock; /* In 10Khz unit */ - USHORT usPM_RTS_Location; /* RTS PM4 starting location in ROM in 1Kb unit */ - UCHAR ucPM_RTS_StreamSize; /* RTS PM4 packets in Kb unit */ - UCHAR ucDesign_ID; /* Indicate what is the board design */ - UCHAR ucMemoryModule_ID; /* Indicate what is the board design */ -} ATOM_FIRMWARE_INFO_V1_4; - -#define ATOM_FIRMWARE_INFO_LAST ATOM_FIRMWARE_INFO_V1_4 - -/****************************************************************************/ -/* Structures used in IntegratedSystemInfoTable */ -/****************************************************************************/ +typedef struct _ATOM_FIRMWARE_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulFirmwareRevision; + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit + ULONG ulDriverTargetEngineClock; //In 10Khz unit + ULONG ulDriverTargetMemoryClock; //In 10Khz unit + ULONG ulMaxEngineClockPLL_Output; //In 10Khz unit + ULONG ulMaxMemoryClockPLL_Output; //In 10Khz unit + ULONG ulMaxPixelClockPLL_Output; //In 10Khz unit + ULONG ulASICMaxEngineClock; //In 10Khz unit + ULONG ulASICMaxMemoryClock; //In 10Khz unit + UCHAR ucASICMaxTemperature; + UCHAR ucPadding[3]; //Don't use them + ULONG aulReservedForBIOS[3]; //Don't use them + USHORT usMinEngineClockPLL_Input; //In 10Khz unit + USHORT usMaxEngineClockPLL_Input; //In 10Khz unit + USHORT usMinEngineClockPLL_Output; //In 10Khz unit + USHORT usMinMemoryClockPLL_Input; //In 10Khz unit + USHORT usMaxMemoryClockPLL_Input; //In 10Khz unit + USHORT usMinMemoryClockPLL_Output; //In 10Khz unit + USHORT usMaxPixelClock; //In 10Khz unit, Max. Pclk + USHORT usMinPixelClockPLL_Input; //In 10Khz unit + USHORT usMaxPixelClockPLL_Input; //In 10Khz unit + USHORT usMinPixelClockPLL_Output; //In 10Khz unit, the definitions above can't change!!! + ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; + USHORT usReferenceClock; //In 10Khz unit + USHORT usPM_RTS_Location; //RTS PM4 starting location in ROM in 1Kb unit + UCHAR ucPM_RTS_StreamSize; //RTS PM4 packets in Kb unit + UCHAR ucDesign_ID; //Indicate what is the board design + UCHAR ucMemoryModule_ID; //Indicate what is the board design +}ATOM_FIRMWARE_INFO; + +typedef struct _ATOM_FIRMWARE_INFO_V1_2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulFirmwareRevision; + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit + ULONG ulDriverTargetEngineClock; //In 10Khz unit + ULONG ulDriverTargetMemoryClock; //In 10Khz unit + ULONG ulMaxEngineClockPLL_Output; //In 10Khz unit + ULONG ulMaxMemoryClockPLL_Output; //In 10Khz unit + ULONG ulMaxPixelClockPLL_Output; //In 10Khz unit + ULONG ulASICMaxEngineClock; //In 10Khz unit + ULONG ulASICMaxMemoryClock; //In 10Khz unit + UCHAR ucASICMaxTemperature; + UCHAR ucMinAllowedBL_Level; + UCHAR ucPadding[2]; //Don't use them + ULONG aulReservedForBIOS[2]; //Don't use them + ULONG ulMinPixelClockPLL_Output; //In 10Khz unit + USHORT usMinEngineClockPLL_Input; //In 10Khz unit + USHORT usMaxEngineClockPLL_Input; //In 10Khz unit + USHORT usMinEngineClockPLL_Output; //In 10Khz unit + USHORT usMinMemoryClockPLL_Input; //In 10Khz unit + USHORT usMaxMemoryClockPLL_Input; //In 10Khz unit + USHORT usMinMemoryClockPLL_Output; //In 10Khz unit + USHORT usMaxPixelClock; //In 10Khz unit, Max. Pclk + USHORT usMinPixelClockPLL_Input; //In 10Khz unit + USHORT usMaxPixelClockPLL_Input; //In 10Khz unit + USHORT usMinPixelClockPLL_Output; //In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output + ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; + USHORT usReferenceClock; //In 10Khz unit + USHORT usPM_RTS_Location; //RTS PM4 starting location in ROM in 1Kb unit + UCHAR ucPM_RTS_StreamSize; //RTS PM4 packets in Kb unit + UCHAR ucDesign_ID; //Indicate what is the board design + UCHAR ucMemoryModule_ID; //Indicate what is the board design +}ATOM_FIRMWARE_INFO_V1_2; + +typedef struct _ATOM_FIRMWARE_INFO_V1_3 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulFirmwareRevision; + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit + ULONG ulDriverTargetEngineClock; //In 10Khz unit + ULONG ulDriverTargetMemoryClock; //In 10Khz unit + ULONG ulMaxEngineClockPLL_Output; //In 10Khz unit + ULONG ulMaxMemoryClockPLL_Output; //In 10Khz unit + ULONG ulMaxPixelClockPLL_Output; //In 10Khz unit + ULONG ulASICMaxEngineClock; //In 10Khz unit + ULONG ulASICMaxMemoryClock; //In 10Khz unit + UCHAR ucASICMaxTemperature; + UCHAR ucMinAllowedBL_Level; + UCHAR ucPadding[2]; //Don't use them + ULONG aulReservedForBIOS; //Don't use them + ULONG ul3DAccelerationEngineClock;//In 10Khz unit + ULONG ulMinPixelClockPLL_Output; //In 10Khz unit + USHORT usMinEngineClockPLL_Input; //In 10Khz unit + USHORT usMaxEngineClockPLL_Input; //In 10Khz unit + USHORT usMinEngineClockPLL_Output; //In 10Khz unit + USHORT usMinMemoryClockPLL_Input; //In 10Khz unit + USHORT usMaxMemoryClockPLL_Input; //In 10Khz unit + USHORT usMinMemoryClockPLL_Output; //In 10Khz unit + USHORT usMaxPixelClock; //In 10Khz unit, Max. Pclk + USHORT usMinPixelClockPLL_Input; //In 10Khz unit + USHORT usMaxPixelClockPLL_Input; //In 10Khz unit + USHORT usMinPixelClockPLL_Output; //In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output + ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; + USHORT usReferenceClock; //In 10Khz unit + USHORT usPM_RTS_Location; //RTS PM4 starting location in ROM in 1Kb unit + UCHAR ucPM_RTS_StreamSize; //RTS PM4 packets in Kb unit + UCHAR ucDesign_ID; //Indicate what is the board design + UCHAR ucMemoryModule_ID; //Indicate what is the board design +}ATOM_FIRMWARE_INFO_V1_3; + +typedef struct _ATOM_FIRMWARE_INFO_V1_4 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulFirmwareRevision; + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit + ULONG ulDriverTargetEngineClock; //In 10Khz unit + ULONG ulDriverTargetMemoryClock; //In 10Khz unit + ULONG ulMaxEngineClockPLL_Output; //In 10Khz unit + ULONG ulMaxMemoryClockPLL_Output; //In 10Khz unit + ULONG ulMaxPixelClockPLL_Output; //In 10Khz unit + ULONG ulASICMaxEngineClock; //In 10Khz unit + ULONG ulASICMaxMemoryClock; //In 10Khz unit + UCHAR ucASICMaxTemperature; + UCHAR ucMinAllowedBL_Level; + USHORT usBootUpVDDCVoltage; //In MV unit + USHORT usLcdMinPixelClockPLL_Output; // In MHz unit + USHORT usLcdMaxPixelClockPLL_Output; // In MHz unit + ULONG ul3DAccelerationEngineClock;//In 10Khz unit + ULONG ulMinPixelClockPLL_Output; //In 10Khz unit + USHORT usMinEngineClockPLL_Input; //In 10Khz unit + USHORT usMaxEngineClockPLL_Input; //In 10Khz unit + USHORT usMinEngineClockPLL_Output; //In 10Khz unit + USHORT usMinMemoryClockPLL_Input; //In 10Khz unit + USHORT usMaxMemoryClockPLL_Input; //In 10Khz unit + USHORT usMinMemoryClockPLL_Output; //In 10Khz unit + USHORT usMaxPixelClock; //In 10Khz unit, Max. Pclk + USHORT usMinPixelClockPLL_Input; //In 10Khz unit + USHORT usMaxPixelClockPLL_Input; //In 10Khz unit + USHORT usMinPixelClockPLL_Output; //In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output + ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; + USHORT usReferenceClock; //In 10Khz unit + USHORT usPM_RTS_Location; //RTS PM4 starting location in ROM in 1Kb unit + UCHAR ucPM_RTS_StreamSize; //RTS PM4 packets in Kb unit + UCHAR ucDesign_ID; //Indicate what is the board design + UCHAR ucMemoryModule_ID; //Indicate what is the board design +}ATOM_FIRMWARE_INFO_V1_4; + +//the structure below to be used from Cypress +typedef struct _ATOM_FIRMWARE_INFO_V2_1 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulFirmwareRevision; + ULONG ulDefaultEngineClock; //In 10Khz unit + ULONG ulDefaultMemoryClock; //In 10Khz unit + ULONG ulReserved1; + ULONG ulReserved2; + ULONG ulMaxEngineClockPLL_Output; //In 10Khz unit + ULONG ulMaxMemoryClockPLL_Output; //In 10Khz unit + ULONG ulMaxPixelClockPLL_Output; //In 10Khz unit + ULONG ulBinaryAlteredInfo; //Was ulASICMaxEngineClock + ULONG ulDefaultDispEngineClkFreq; //In 10Khz unit + UCHAR ucReserved1; //Was ucASICMaxTemperature; + UCHAR ucMinAllowedBL_Level; + USHORT usBootUpVDDCVoltage; //In MV unit + USHORT usLcdMinPixelClockPLL_Output; // In MHz unit + USHORT usLcdMaxPixelClockPLL_Output; // In MHz unit + ULONG ulReserved4; //Was ulAsicMaximumVoltage + ULONG ulMinPixelClockPLL_Output; //In 10Khz unit + USHORT usMinEngineClockPLL_Input; //In 10Khz unit + USHORT usMaxEngineClockPLL_Input; //In 10Khz unit + USHORT usMinEngineClockPLL_Output; //In 10Khz unit + USHORT usMinMemoryClockPLL_Input; //In 10Khz unit + USHORT usMaxMemoryClockPLL_Input; //In 10Khz unit + USHORT usMinMemoryClockPLL_Output; //In 10Khz unit + USHORT usMaxPixelClock; //In 10Khz unit, Max. Pclk + USHORT usMinPixelClockPLL_Input; //In 10Khz unit + USHORT usMaxPixelClockPLL_Input; //In 10Khz unit + USHORT usMinPixelClockPLL_Output; //In 10Khz unit - lower 16bit of ulMinPixelClockPLL_Output + ATOM_FIRMWARE_CAPABILITY_ACCESS usFirmwareCapability; + USHORT usCoreReferenceClock; //In 10Khz unit + USHORT usMemoryReferenceClock; //In 10Khz unit + USHORT usUniphyDPModeExtClkFreq; //In 10Khz unit, if it is 0, In DP Mode Uniphy Input clock from internal PPLL, otherwise Input clock from external Spread clock + UCHAR ucMemoryModule_ID; //Indicate what is the board design + UCHAR ucReserved4[3]; +}ATOM_FIRMWARE_INFO_V2_1; + + +#define ATOM_FIRMWARE_INFO_LAST ATOM_FIRMWARE_INFO_V2_1 + +/****************************************************************************/ +// Structures used in IntegratedSystemInfoTable +/****************************************************************************/ #define IGP_CAP_FLAG_DYNAMIC_CLOCK_EN 0x2 #define IGP_CAP_FLAG_AC_CARD 0x4 #define IGP_CAP_FLAG_SDVO_CARD 0x8 #define IGP_CAP_FLAG_POSTDIV_BY_2_MODE 0x10 -typedef struct _ATOM_INTEGRATED_SYSTEM_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulBootUpEngineClock; /* in 10kHz unit */ - ULONG ulBootUpMemoryClock; /* in 10kHz unit */ - ULONG ulMaxSystemMemoryClock; /* in 10kHz unit */ - ULONG ulMinSystemMemoryClock; /* in 10kHz unit */ - UCHAR ucNumberOfCyclesInPeriodHi; - UCHAR ucLCDTimingSel; /* =0:not valid.!=0 sel this timing descriptor from LCD EDID. */ - USHORT usReserved1; - USHORT usInterNBVoltageLow; /* An intermidiate PMW value to set the voltage */ - USHORT usInterNBVoltageHigh; /* Another intermidiate PMW value to set the voltage */ - ULONG ulReserved[2]; - - USHORT usFSBClock; /* In MHz unit */ - USHORT usCapabilityFlag; /* Bit0=1 indicates the fake HDMI support,Bit1=0/1 for Dynamic clocking dis/enable */ - /* Bit[3:2]== 0:No PCIE card, 1:AC card, 2:SDVO card */ - /* Bit[4]==1: P/2 mode, ==0: P/1 mode */ - USHORT usPCIENBCfgReg7; /* bit[7:0]=MUX_Sel, bit[9:8]=MUX_SEL_LEVEL2, bit[10]=Lane_Reversal */ - USHORT usK8MemoryClock; /* in MHz unit */ - USHORT usK8SyncStartDelay; /* in 0.01 us unit */ - USHORT usK8DataReturnTime; /* in 0.01 us unit */ - UCHAR ucMaxNBVoltage; - UCHAR ucMinNBVoltage; - UCHAR ucMemoryType; /* [7:4]=1:DDR1;=2:DDR2;=3:DDR3.[3:0] is reserved */ - UCHAR ucNumberOfCyclesInPeriod; /* CG.FVTHROT_PWM_CTRL_REG0.NumberOfCyclesInPeriod */ - UCHAR ucStartingPWM_HighTime; /* CG.FVTHROT_PWM_CTRL_REG0.StartingPWM_HighTime */ - UCHAR ucHTLinkWidth; /* 16 bit vs. 8 bit */ - UCHAR ucMaxNBVoltageHigh; - UCHAR ucMinNBVoltageHigh; -} ATOM_INTEGRATED_SYSTEM_INFO; +typedef struct _ATOM_INTEGRATED_SYSTEM_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulBootUpEngineClock; //in 10kHz unit + ULONG ulBootUpMemoryClock; //in 10kHz unit + ULONG ulMaxSystemMemoryClock; //in 10kHz unit + ULONG ulMinSystemMemoryClock; //in 10kHz unit + UCHAR ucNumberOfCyclesInPeriodHi; + UCHAR ucLCDTimingSel; //=0:not valid.!=0 sel this timing descriptor from LCD EDID. + USHORT usReserved1; + USHORT usInterNBVoltageLow; //An intermidiate PMW value to set the voltage + USHORT usInterNBVoltageHigh; //Another intermidiate PMW value to set the voltage + ULONG ulReserved[2]; + + USHORT usFSBClock; //In MHz unit + USHORT usCapabilityFlag; //Bit0=1 indicates the fake HDMI support,Bit1=0/1 for Dynamic clocking dis/enable + //Bit[3:2]== 0:No PCIE card, 1:AC card, 2:SDVO card + //Bit[4]==1: P/2 mode, ==0: P/1 mode + USHORT usPCIENBCfgReg7; //bit[7:0]=MUX_Sel, bit[9:8]=MUX_SEL_LEVEL2, bit[10]=Lane_Reversal + USHORT usK8MemoryClock; //in MHz unit + USHORT usK8SyncStartDelay; //in 0.01 us unit + USHORT usK8DataReturnTime; //in 0.01 us unit + UCHAR ucMaxNBVoltage; + UCHAR ucMinNBVoltage; + UCHAR ucMemoryType; //[7:4]=1:DDR1;=2:DDR2;=3:DDR3.[3:0] is reserved + UCHAR ucNumberOfCyclesInPeriod; //CG.FVTHROT_PWM_CTRL_REG0.NumberOfCyclesInPeriod + UCHAR ucStartingPWM_HighTime; //CG.FVTHROT_PWM_CTRL_REG0.StartingPWM_HighTime + UCHAR ucHTLinkWidth; //16 bit vs. 8 bit + UCHAR ucMaxNBVoltageHigh; + UCHAR ucMinNBVoltageHigh; +}ATOM_INTEGRATED_SYSTEM_INFO; /* Explanation on entries in ATOM_INTEGRATED_SYSTEM_INFO -ulBootUpMemoryClock: For Intel IGP,it's the UMA system memory clock +ulBootUpMemoryClock: For Intel IGP,it's the UMA system memory clock For AMD IGP,it's 0 if no SidePort memory installed or it's the boot-up SidePort memory clock ulMaxSystemMemoryClock: For Intel IGP,it's the Max freq from memory SPD if memory runs in ASYNC mode or otherwise (SYNC mode) it's 0 For AMD IGP,for now this can be 0 -ulMinSystemMemoryClock: For Intel IGP,it's 133MHz if memory runs in ASYNC mode or otherwise (SYNC mode) it's 0 +ulMinSystemMemoryClock: For Intel IGP,it's 133MHz if memory runs in ASYNC mode or otherwise (SYNC mode) it's 0 For AMD IGP,for now this can be 0 -usFSBClock: For Intel IGP,it's FSB Freq +usFSBClock: For Intel IGP,it's FSB Freq For AMD IGP,it's HT Link Speed usK8MemoryClock: For AMD IGP only. For RevF CPU, set it to 200 @@ -1687,98 +2093,113 @@ VC:Voltage Control ucMaxNBVoltage: Voltage regulator dependent PWM value. Low 8 bits of the value for the max voltage.Set this one to 0xFF if VC without PWM. Set this to 0x0 if no VC at all. ucMinNBVoltage: Voltage regulator dependent PWM value. Low 8 bits of the value for the min voltage.Set this one to 0x00 if VC without PWM or no VC at all. -ucNumberOfCyclesInPeriod: Indicate how many cycles when PWM duty is 100%. low 8 bits of the value. -ucNumberOfCyclesInPeriodHi: Indicate how many cycles when PWM duty is 100%. high 8 bits of the value.If the PWM has an inverter,set bit [7]==1,otherwise set it 0 +ucNumberOfCyclesInPeriod: Indicate how many cycles when PWM duty is 100%. low 8 bits of the value. +ucNumberOfCyclesInPeriodHi: Indicate how many cycles when PWM duty is 100%. high 8 bits of the value.If the PWM has an inverter,set bit [7]==1,otherwise set it 0 ucMaxNBVoltageHigh: Voltage regulator dependent PWM value. High 8 bits of the value for the max voltage.Set this one to 0xFF if VC without PWM. Set this to 0x0 if no VC at all. ucMinNBVoltageHigh: Voltage regulator dependent PWM value. High 8 bits of the value for the min voltage.Set this one to 0x00 if VC without PWM or no VC at all. + usInterNBVoltageLow: Voltage regulator dependent PWM value. The value makes the the voltage >=Min NB voltage but <=InterNBVoltageHigh. Set this to 0x0000 if VC without PWM or no VC at all. usInterNBVoltageHigh: Voltage regulator dependent PWM value. The value makes the the voltage >=InterNBVoltageLow but <=Max NB voltage.Set this to 0x0000 if VC without PWM or no VC at all. */ + /* The following IGP table is introduced from RS780, which is supposed to be put by SBIOS in FB before IGP VBIOS starts VPOST; -Then VBIOS will copy the whole structure to its image so all GPU SW components can access this data structure to get whatever they need. +Then VBIOS will copy the whole structure to its image so all GPU SW components can access this data structure to get whatever they need. The enough reservation should allow us to never change table revisions. Whenever needed, a GPU SW component can use reserved portion for new data entries. SW components can access the IGP system infor structure in the same way as before */ -typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 { - ATOM_COMMON_TABLE_HEADER sHeader; - ULONG ulBootUpEngineClock; /* in 10kHz unit */ - ULONG ulReserved1[2]; /* must be 0x0 for the reserved */ - ULONG ulBootUpUMAClock; /* in 10kHz unit */ - ULONG ulBootUpSidePortClock; /* in 10kHz unit */ - ULONG ulMinSidePortClock; /* in 10kHz unit */ - ULONG ulReserved2[6]; /* must be 0x0 for the reserved */ - ULONG ulSystemConfig; /* see explanation below */ - ULONG ulBootUpReqDisplayVector; - ULONG ulOtherDisplayMisc; - ULONG ulDDISlot1Config; - ULONG ulDDISlot2Config; - UCHAR ucMemoryType; /* [3:0]=1:DDR1;=2:DDR2;=3:DDR3.[7:4] is reserved */ - UCHAR ucUMAChannelNumber; - UCHAR ucDockingPinBit; - UCHAR ucDockingPinPolarity; - ULONG ulDockingPinCFGInfo; - ULONG ulCPUCapInfo; - USHORT usNumberOfCyclesInPeriod; - USHORT usMaxNBVoltage; - USHORT usMinNBVoltage; - USHORT usBootUpNBVoltage; - ULONG ulHTLinkFreq; /* in 10Khz */ - USHORT usMinHTLinkWidth; - USHORT usMaxHTLinkWidth; - USHORT usUMASyncStartDelay; - USHORT usUMADataReturnTime; - USHORT usLinkStatusZeroTime; - USHORT usReserved; - ULONG ulHighVoltageHTLinkFreq; /* in 10Khz */ - ULONG ulLowVoltageHTLinkFreq; /* in 10Khz */ - USHORT usMaxUpStreamHTLinkWidth; - USHORT usMaxDownStreamHTLinkWidth; - USHORT usMinUpStreamHTLinkWidth; - USHORT usMinDownStreamHTLinkWidth; - ULONG ulReserved3[97]; /* must be 0x0 */ -} ATOM_INTEGRATED_SYSTEM_INFO_V2; + +typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulBootUpEngineClock; //in 10kHz unit + ULONG ulReserved1[2]; //must be 0x0 for the reserved + ULONG ulBootUpUMAClock; //in 10kHz unit + ULONG ulBootUpSidePortClock; //in 10kHz unit + ULONG ulMinSidePortClock; //in 10kHz unit + ULONG ulReserved2[6]; //must be 0x0 for the reserved + ULONG ulSystemConfig; //see explanation below + ULONG ulBootUpReqDisplayVector; + ULONG ulOtherDisplayMisc; + ULONG ulDDISlot1Config; + ULONG ulDDISlot2Config; + UCHAR ucMemoryType; //[3:0]=1:DDR1;=2:DDR2;=3:DDR3.[7:4] is reserved + UCHAR ucUMAChannelNumber; + UCHAR ucDockingPinBit; + UCHAR ucDockingPinPolarity; + ULONG ulDockingPinCFGInfo; + ULONG ulCPUCapInfo; + USHORT usNumberOfCyclesInPeriod; + USHORT usMaxNBVoltage; + USHORT usMinNBVoltage; + USHORT usBootUpNBVoltage; + ULONG ulHTLinkFreq; //in 10Khz + USHORT usMinHTLinkWidth; + USHORT usMaxHTLinkWidth; + USHORT usUMASyncStartDelay; + USHORT usUMADataReturnTime; + USHORT usLinkStatusZeroTime; + USHORT usDACEfuse; //for storing badgap value (for RS880 only) + ULONG ulHighVoltageHTLinkFreq; // in 10Khz + ULONG ulLowVoltageHTLinkFreq; // in 10Khz + USHORT usMaxUpStreamHTLinkWidth; + USHORT usMaxDownStreamHTLinkWidth; + USHORT usMinUpStreamHTLinkWidth; + USHORT usMinDownStreamHTLinkWidth; + USHORT usFirmwareVersion; //0 means FW is not supported. Otherwise it's the FW version loaded by SBIOS and driver should enable FW. + USHORT usFullT0Time; // Input to calculate minimum HT link change time required by NB P-State. Unit is 0.01us. + ULONG ulReserved3[96]; //must be 0x0 +}ATOM_INTEGRATED_SYSTEM_INFO_V2; /* ulBootUpEngineClock: Boot-up Engine Clock in 10Khz; ulBootUpUMAClock: Boot-up UMA Clock in 10Khz; it must be 0x0 when UMA is not present ulBootUpSidePortClock: Boot-up SidePort Clock in 10Khz; it must be 0x0 when SidePort Memory is not present,this could be equal to or less than maximum supported Sideport memory clock -ulSystemConfig: -Bit[0]=1: PowerExpress mode =0 Non-PowerExpress mode; +ulSystemConfig: +Bit[0]=1: PowerExpress mode =0 Non-PowerExpress mode; Bit[1]=1: system boots up at AMD overdrived state or user customized mode. In this case, driver will just stick to this boot-up mode. No other PowerPlay state =0: system boots up at driver control state. Power state depends on PowerPlay table. Bit[2]=1: PWM method is used on NB voltage control. =0: GPIO method is used. Bit[3]=1: Only one power state(Performance) will be supported. =0: Multiple power states supported from PowerPlay table. -Bit[4]=1: CLMC is supported and enabled on current system. - =0: CLMC is not supported or enabled on current system. SBIOS need to support HT link/freq change through ATIF interface. -Bit[5]=1: Enable CDLW for all driver control power states. Max HT width is from SBIOS, while Min HT width is determined by display requirement. +Bit[4]=1: CLMC is supported and enabled on current system. + =0: CLMC is not supported or enabled on current system. SBIOS need to support HT link/freq change through ATIF interface. +Bit[5]=1: Enable CDLW for all driver control power states. Max HT width is from SBIOS, while Min HT width is determined by display requirement. =0: CDLW is disabled. If CLMC is enabled case, Min HT width will be set equal to Max HT width. If CLMC disabled case, Max HT width will be applied. Bit[6]=1: High Voltage requested for all power states. In this case, voltage will be forced at 1.1v and powerplay table voltage drop/throttling request will be ignored. =0: Voltage settings is determined by powerplay table. Bit[7]=1: Enable CLMC as hybrid Mode. CDLD and CILR will be disabled in this case and we're using legacy C1E. This is workaround for CPU(Griffin) performance issue. =0: Enable CLMC as regular mode, CDLD and CILR will be enabled. +Bit[8]=1: CDLF is supported and enabled on current system. + =0: CDLF is not supported or enabled on current system. +Bit[9]=1: DLL Shut Down feature is enabled on current system. + =0: DLL Shut Down feature is not enabled or supported on current system. ulBootUpReqDisplayVector: This dword is a bit vector indicates what display devices are requested during boot-up. Refer to ATOM_DEVICE_xxx_SUPPORT for the bit vector definitions. ulOtherDisplayMisc: [15:8]- Bootup LCD Expansion selection; 0-center, 1-full panel size expansion; - [7:0] - BootupTV standard selection; This is a bit vector to indicate what TV standards are supported by the system. Refer to ucTVSuppportedStd definition; + [7:0] - BootupTV standard selection; This is a bit vector to indicate what TV standards are supported by the system. Refer to ucTVSupportedStd definition; ulDDISlot1Config: Describes the PCIE lane configuration on this DDI PCIE slot (ADD2 card) or connector (Mobile design). [3:0] - Bit vector to indicate PCIE lane config of the DDI slot/connector on chassis (bit 0=1 lane 3:0; bit 1=1 lane 7:4; bit 2=1 lane 11:8; bit 3=1 lane 15:12) - [7:4] - Bit vector to indicate PCIE lane config of the same DDI slot/connector on docking station (bit 0=1 lane 3:0; bit 1=1 lane 7:4; bit 2=1 lane 11:8; bit 3=1 lane 15:12) - [15:8] - Lane configuration attribute; + [7:4] - Bit vector to indicate PCIE lane config of the same DDI slot/connector on docking station (bit 4=1 lane 3:0; bit 5=1 lane 7:4; bit 6=1 lane 11:8; bit 7=1 lane 15:12) + When a DDI connector is not "paired" (meaming two connections mutualexclusive on chassis or docking, only one of them can be connected at one time. + in both chassis and docking, SBIOS has to duplicate the same PCIE lane info from chassis to docking or vice versa. For example: + one DDI connector is only populated in docking with PCIE lane 8-11, but there is no paired connection on chassis, SBIOS has to copy bit 6 to bit 2. + + [15:8] - Lane configuration attribute; [23:16]- Connector type, possible value: CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D CONNECTOR_OBJECT_ID_HDMI_TYPE_A CONNECTOR_OBJECT_ID_DISPLAYPORT + CONNECTOR_OBJECT_ID_eDP [31:24]- Reserved ulDDISlot2Config: Same as Slot1. @@ -1787,29 +2208,31 @@ For IGP, Hypermemory is the only memory type showed in CCC. ucUMAChannelNumber: how many channels for the UMA; -ulDockingPinCFGInfo: [15:0]-Bus/Device/Function # to CFG to read this Docking Pin; [31:16]-reg offset in CFG to read this pin +ulDockingPinCFGInfo: [15:0]-Bus/Device/Function # to CFG to read this Docking Pin; [31:16]-reg offset in CFG to read this pin ucDockingPinBit: which bit in this register to read the pin status; ucDockingPinPolarity:Polarity of the pin when docked; ulCPUCapInfo: [7:0]=1:Griffin;[7:0]=2:Greyhound;[7:0]=3:K8, other bits reserved for now and must be 0x0 usNumberOfCyclesInPeriod:Indicate how many cycles when PWM duty is 100%. -usMaxNBVoltage:Max. voltage control value in either PWM or GPIO mode. + +usMaxNBVoltage:Max. voltage control value in either PWM or GPIO mode. usMinNBVoltage:Min. voltage control value in either PWM or GPIO mode. GPIO mode: both usMaxNBVoltage & usMinNBVoltage have a valid value ulSystemConfig.SYSTEM_CONFIG_USE_PWM_ON_VOLTAGE=0 PWM mode: both usMaxNBVoltage & usMinNBVoltage have a valid value ulSystemConfig.SYSTEM_CONFIG_USE_PWM_ON_VOLTAGE=1 GPU SW don't control mode: usMaxNBVoltage & usMinNBVoltage=0 and no care about ulSystemConfig.SYSTEM_CONFIG_USE_PWM_ON_VOLTAGE + usBootUpNBVoltage:Boot-up voltage regulator dependent PWM value. ulHTLinkFreq: Bootup HT link Frequency in 10Khz. -usMinHTLinkWidth: Bootup minimum HT link width. If CDLW disabled, this is equal to usMaxHTLinkWidth. - If CDLW enabled, both upstream and downstream width should be the same during bootup. -usMaxHTLinkWidth: Bootup maximum HT link width. If CDLW disabled, this is equal to usMinHTLinkWidth. +usMinHTLinkWidth: Bootup minimum HT link width. If CDLW disabled, this is equal to usMaxHTLinkWidth. If CDLW enabled, both upstream and downstream width should be the same during bootup. +usMaxHTLinkWidth: Bootup maximum HT link width. If CDLW disabled, this is equal to usMinHTLinkWidth. + If CDLW enabled, both upstream and downstream width should be the same during bootup. -usUMASyncStartDelay: Memory access latency, required for watermark calculation +usUMASyncStartDelay: Memory access latency, required for watermark calculation usUMADataReturnTime: Memory access latency, required for watermark calculation -usLinkStatusZeroTime:Memory access latency required for watermark calculation, set this to 0x0 for K8 CPU, set a proper value in 0.01 the unit of us +usLinkStatusZeroTime:Memory access latency required for watermark calculation, set this to 0x0 for K8 CPU, set a proper value in 0.01 the unit of us for Griffin or Greyhound. SBIOS needs to convert to actual time by: if T0Ttime [5:4]=00b, then usLinkStatusZeroTime=T0Ttime [3:0]*0.1us (0.0 to 1.5us) if T0Ttime [5:4]=01b, then usLinkStatusZeroTime=T0Ttime [3:0]*0.5us (0.0 to 7.5us) @@ -1817,7 +2240,7 @@ for Griffin or Greyhound. SBIOS needs to convert to actual time by: if T0Ttime [5:4]=11b, and T0Ttime [3:0]=0x0 to 0xa, then usLinkStatusZeroTime=T0Ttime [3:0]*20us (0.0 to 200us) ulHighVoltageHTLinkFreq: HT link frequency for power state with low voltage. If boot up runs in HT1, this must be 0. - This must be less than or equal to ulHTLinkFreq(bootup frequency). + This must be less than or equal to ulHTLinkFreq(bootup frequency). ulLowVoltageHTLinkFreq: HT link frequency for power state with low voltage or voltage scaling 1.0v~1.1v. If boot up runs in HT1, this must be 0. This must be less than or equal to ulHighVoltageHTLinkFreq. @@ -1827,14 +2250,17 @@ usMinUpStreamHTLinkWidth: Asymmetric link width support in the future, to rep usMinDownStreamHTLinkWidth: same as above. */ + #define SYSTEM_CONFIG_POWEREXPRESS_ENABLE 0x00000001 #define SYSTEM_CONFIG_RUN_AT_OVERDRIVE_ENGINE 0x00000002 -#define SYSTEM_CONFIG_USE_PWM_ON_VOLTAGE 0x00000004 +#define SYSTEM_CONFIG_USE_PWM_ON_VOLTAGE 0x00000004 #define SYSTEM_CONFIG_PERFORMANCE_POWERSTATE_ONLY 0x00000008 #define SYSTEM_CONFIG_CLMC_ENABLED 0x00000010 #define SYSTEM_CONFIG_CDLW_ENABLED 0x00000020 #define SYSTEM_CONFIG_HIGH_VOLTAGE_REQUESTED 0x00000040 #define SYSTEM_CONFIG_CLMC_HYBRID_MODE_ENABLED 0x00000080 +#define SYSTEM_CONFIG_CDLF_ENABLED 0x00000100 +#define SYSTEM_CONFIG_DLL_SHUTDOWN_ENABLED 0x00000200 #define IGP_DDI_SLOT_LANE_CONFIG_MASK 0x000000FF @@ -1851,6 +2277,41 @@ usMinDownStreamHTLinkWidth: same as above. #define IGP_DDI_SLOT_CONNECTOR_TYPE_MASK 0x00FF0000 +// IntegratedSystemInfoTable new Rev is V5 after V2, because of the real rev of V2 is v1.4. This rev is used for RR +typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulBootUpEngineClock; //in 10kHz unit + ULONG ulDentistVCOFreq; //Dentist VCO clock in 10kHz unit, the source of GPU SCLK, LCLK, UCLK and VCLK. + ULONG ulLClockFreq; //GPU Lclk freq in 10kHz unit, have relationship with NCLK in NorthBridge + ULONG ulBootUpUMAClock; //in 10kHz unit + ULONG ulReserved1[8]; //must be 0x0 for the reserved + ULONG ulBootUpReqDisplayVector; + ULONG ulOtherDisplayMisc; + ULONG ulReserved2[4]; //must be 0x0 for the reserved + ULONG ulSystemConfig; //TBD + ULONG ulCPUCapInfo; //TBD + USHORT usMaxNBVoltage; //high NB voltage, calculated using current VDDNB (D24F2xDC) and VDDNB offset fuse; + USHORT usMinNBVoltage; //low NB voltage, calculated using current VDDNB (D24F2xDC) and VDDNB offset fuse; + USHORT usBootUpNBVoltage; //boot up NB voltage + UCHAR ucHtcTmpLmt; //bit [22:16] of D24F3x64 Hardware Thermal Control (HTC) Register, may not be needed, TBD + UCHAR ucTjOffset; //bit [28:22] of D24F3xE4 Thermtrip Status Register,may not be needed, TBD + ULONG ulReserved3[4]; //must be 0x0 for the reserved + ULONG ulDDISlot1Config; //see above ulDDISlot1Config definition + ULONG ulDDISlot2Config; + ULONG ulDDISlot3Config; + ULONG ulDDISlot4Config; + ULONG ulReserved4[4]; //must be 0x0 for the reserved + UCHAR ucMemoryType; //[3:0]=1:DDR1;=2:DDR2;=3:DDR3.[7:4] is reserved + UCHAR ucUMAChannelNumber; + USHORT usReserved; + ULONG ulReserved5[4]; //must be 0x0 for the reserved + ULONG ulCSR_M3_ARB_CNTL_DEFAULT[10];//arrays with values for CSR M3 arbiter for default + ULONG ulCSR_M3_ARB_CNTL_UVD[10]; //arrays with values for CSR M3 arbiter for UVD playback + ULONG ulCSR_M3_ARB_CNTL_FS3D[10];//arrays with values for CSR M3 arbiter for Full Screen 3D applications + ULONG ulReserved6[61]; //must be 0x0 +}ATOM_INTEGRATED_SYSTEM_INFO_V5; + #define ATOM_CRT_INT_ENCODER1_INDEX 0x00000000 #define ATOM_LCD_INT_ENCODER1_INDEX 0x00000001 #define ATOM_TV_INT_ENCODER1_INDEX 0x00000002 @@ -1866,8 +2327,8 @@ usMinDownStreamHTLinkWidth: same as above. #define ATOM_DFP_INT_ENCODER3_INDEX 0x0000000C #define ATOM_DFP_INT_ENCODER4_INDEX 0x0000000D -/* define ASIC internal encoder id ( bit vector ) */ -#define ASIC_INT_DAC1_ENCODER_ID 0x00 +// define ASIC internal encoder id ( bit vector ), used for CRTC_SourceSelTable +#define ASIC_INT_DAC1_ENCODER_ID 0x00 #define ASIC_INT_TV_ENCODER_ID 0x02 #define ASIC_INT_DIG1_ENCODER_ID 0x03 #define ASIC_INT_DAC2_ENCODER_ID 0x04 @@ -1875,10 +2336,24 @@ usMinDownStreamHTLinkWidth: same as above. #define ASIC_INT_DVO_ENCODER_ID 0x07 #define ASIC_INT_DIG2_ENCODER_ID 0x09 #define ASIC_EXT_DIG_ENCODER_ID 0x05 +#define ASIC_EXT_DIG2_ENCODER_ID 0x08 +#define ASIC_INT_DIG3_ENCODER_ID 0x0a +#define ASIC_INT_DIG4_ENCODER_ID 0x0b +#define ASIC_INT_DIG5_ENCODER_ID 0x0c +#define ASIC_INT_DIG6_ENCODER_ID 0x0d -/* define Encoder attribute */ +//define Encoder attribute #define ATOM_ANALOG_ENCODER 0 -#define ATOM_DIGITAL_ENCODER 1 +#define ATOM_DIGITAL_ENCODER 1 +#define ATOM_DP_ENCODER 2 + +#define ATOM_ENCODER_ENUM_MASK 0x70 +#define ATOM_ENCODER_ENUM_ID1 0x00 +#define ATOM_ENCODER_ENUM_ID2 0x10 +#define ATOM_ENCODER_ENUM_ID3 0x20 +#define ATOM_ENCODER_ENUM_ID4 0x30 +#define ATOM_ENCODER_ENUM_ID5 0x40 +#define ATOM_ENCODER_ENUM_ID6 0x50 #define ATOM_DEVICE_CRT1_INDEX 0x00000000 #define ATOM_DEVICE_LCD1_INDEX 0x00000001 @@ -1886,45 +2361,40 @@ usMinDownStreamHTLinkWidth: same as above. #define ATOM_DEVICE_DFP1_INDEX 0x00000003 #define ATOM_DEVICE_CRT2_INDEX 0x00000004 #define ATOM_DEVICE_LCD2_INDEX 0x00000005 -#define ATOM_DEVICE_TV2_INDEX 0x00000006 +#define ATOM_DEVICE_DFP6_INDEX 0x00000006 #define ATOM_DEVICE_DFP2_INDEX 0x00000007 #define ATOM_DEVICE_CV_INDEX 0x00000008 -#define ATOM_DEVICE_DFP3_INDEX 0x00000009 -#define ATOM_DEVICE_DFP4_INDEX 0x0000000A -#define ATOM_DEVICE_DFP5_INDEX 0x0000000B +#define ATOM_DEVICE_DFP3_INDEX 0x00000009 +#define ATOM_DEVICE_DFP4_INDEX 0x0000000A +#define ATOM_DEVICE_DFP5_INDEX 0x0000000B + #define ATOM_DEVICE_RESERVEDC_INDEX 0x0000000C #define ATOM_DEVICE_RESERVEDD_INDEX 0x0000000D #define ATOM_DEVICE_RESERVEDE_INDEX 0x0000000E #define ATOM_DEVICE_RESERVEDF_INDEX 0x0000000F #define ATOM_MAX_SUPPORTED_DEVICE_INFO (ATOM_DEVICE_DFP3_INDEX+1) #define ATOM_MAX_SUPPORTED_DEVICE_INFO_2 ATOM_MAX_SUPPORTED_DEVICE_INFO -#define ATOM_MAX_SUPPORTED_DEVICE_INFO_3 (ATOM_DEVICE_DFP5_INDEX + 1) +#define ATOM_MAX_SUPPORTED_DEVICE_INFO_3 (ATOM_DEVICE_DFP5_INDEX + 1 ) #define ATOM_MAX_SUPPORTED_DEVICE (ATOM_DEVICE_RESERVEDF_INDEX+1) -#define ATOM_DEVICE_CRT1_SUPPORT (0x1L << ATOM_DEVICE_CRT1_INDEX) -#define ATOM_DEVICE_LCD1_SUPPORT (0x1L << ATOM_DEVICE_LCD1_INDEX) -#define ATOM_DEVICE_TV1_SUPPORT (0x1L << ATOM_DEVICE_TV1_INDEX) -#define ATOM_DEVICE_DFP1_SUPPORT (0x1L << ATOM_DEVICE_DFP1_INDEX) -#define ATOM_DEVICE_CRT2_SUPPORT (0x1L << ATOM_DEVICE_CRT2_INDEX) -#define ATOM_DEVICE_LCD2_SUPPORT (0x1L << ATOM_DEVICE_LCD2_INDEX) -#define ATOM_DEVICE_TV2_SUPPORT (0x1L << ATOM_DEVICE_TV2_INDEX) -#define ATOM_DEVICE_DFP2_SUPPORT (0x1L << ATOM_DEVICE_DFP2_INDEX) -#define ATOM_DEVICE_CV_SUPPORT (0x1L << ATOM_DEVICE_CV_INDEX) -#define ATOM_DEVICE_DFP3_SUPPORT (0x1L << ATOM_DEVICE_DFP3_INDEX) -#define ATOM_DEVICE_DFP4_SUPPORT (0x1L << ATOM_DEVICE_DFP4_INDEX ) -#define ATOM_DEVICE_DFP5_SUPPORT (0x1L << ATOM_DEVICE_DFP5_INDEX) - -#define ATOM_DEVICE_CRT_SUPPORT \ - (ATOM_DEVICE_CRT1_SUPPORT | ATOM_DEVICE_CRT2_SUPPORT) -#define ATOM_DEVICE_DFP_SUPPORT \ - (ATOM_DEVICE_DFP1_SUPPORT | ATOM_DEVICE_DFP2_SUPPORT | \ - ATOM_DEVICE_DFP3_SUPPORT | ATOM_DEVICE_DFP4_SUPPORT | \ - ATOM_DEVICE_DFP5_SUPPORT) -#define ATOM_DEVICE_TV_SUPPORT \ - (ATOM_DEVICE_TV1_SUPPORT | ATOM_DEVICE_TV2_SUPPORT) -#define ATOM_DEVICE_LCD_SUPPORT \ - (ATOM_DEVICE_LCD1_SUPPORT | ATOM_DEVICE_LCD2_SUPPORT) +#define ATOM_DEVICE_CRT1_SUPPORT (0x1L << ATOM_DEVICE_CRT1_INDEX ) +#define ATOM_DEVICE_LCD1_SUPPORT (0x1L << ATOM_DEVICE_LCD1_INDEX ) +#define ATOM_DEVICE_TV1_SUPPORT (0x1L << ATOM_DEVICE_TV1_INDEX ) +#define ATOM_DEVICE_DFP1_SUPPORT (0x1L << ATOM_DEVICE_DFP1_INDEX ) +#define ATOM_DEVICE_CRT2_SUPPORT (0x1L << ATOM_DEVICE_CRT2_INDEX ) +#define ATOM_DEVICE_LCD2_SUPPORT (0x1L << ATOM_DEVICE_LCD2_INDEX ) +#define ATOM_DEVICE_DFP6_SUPPORT (0x1L << ATOM_DEVICE_DFP6_INDEX ) +#define ATOM_DEVICE_DFP2_SUPPORT (0x1L << ATOM_DEVICE_DFP2_INDEX ) +#define ATOM_DEVICE_CV_SUPPORT (0x1L << ATOM_DEVICE_CV_INDEX ) +#define ATOM_DEVICE_DFP3_SUPPORT (0x1L << ATOM_DEVICE_DFP3_INDEX ) +#define ATOM_DEVICE_DFP4_SUPPORT (0x1L << ATOM_DEVICE_DFP4_INDEX ) +#define ATOM_DEVICE_DFP5_SUPPORT (0x1L << ATOM_DEVICE_DFP5_INDEX ) + +#define ATOM_DEVICE_CRT_SUPPORT (ATOM_DEVICE_CRT1_SUPPORT | ATOM_DEVICE_CRT2_SUPPORT) +#define ATOM_DEVICE_DFP_SUPPORT (ATOM_DEVICE_DFP1_SUPPORT | ATOM_DEVICE_DFP2_SUPPORT | ATOM_DEVICE_DFP3_SUPPORT | ATOM_DEVICE_DFP4_SUPPORT | ATOM_DEVICE_DFP5_SUPPORT | ATOM_DEVICE_DFP6_SUPPORT) +#define ATOM_DEVICE_TV_SUPPORT (ATOM_DEVICE_TV1_SUPPORT) +#define ATOM_DEVICE_LCD_SUPPORT (ATOM_DEVICE_LCD1_SUPPORT | ATOM_DEVICE_LCD2_SUPPORT) #define ATOM_DEVICE_CONNECTOR_TYPE_MASK 0x000000F0 #define ATOM_DEVICE_CONNECTOR_TYPE_SHIFT 0x00000004 @@ -1942,6 +2412,7 @@ usMinDownStreamHTLinkWidth: same as above. #define ATOM_DEVICE_CONNECTOR_CASE_1 0x0000000E #define ATOM_DEVICE_CONNECTOR_DISPLAYPORT 0x0000000F + #define ATOM_DEVICE_DAC_INFO_MASK 0x0000000F #define ATOM_DEVICE_DAC_INFO_SHIFT 0x00000000 #define ATOM_DEVICE_DAC_INFO_NODAC 0x00000000 @@ -1958,139 +2429,150 @@ usMinDownStreamHTLinkWidth: same as above. #define ATOM_DEVICE_I2C_ID_SHIFT 0x00000004 #define ATOM_DEVICE_I2C_ID_IS_FOR_NON_MM_USE 0x00000001 #define ATOM_DEVICE_I2C_ID_IS_FOR_MM_USE 0x00000002 -#define ATOM_DEVICE_I2C_ID_IS_FOR_SDVO_USE 0x00000003 /* For IGP RS600 */ -#define ATOM_DEVICE_I2C_ID_IS_FOR_DAC_SCL 0x00000004 /* For IGP RS690 */ +#define ATOM_DEVICE_I2C_ID_IS_FOR_SDVO_USE 0x00000003 //For IGP RS600 +#define ATOM_DEVICE_I2C_ID_IS_FOR_DAC_SCL 0x00000004 //For IGP RS690 #define ATOM_DEVICE_I2C_HARDWARE_CAP_MASK 0x00000080 #define ATOM_DEVICE_I2C_HARDWARE_CAP_SHIFT 0x00000007 #define ATOM_DEVICE_USES_SOFTWARE_ASSISTED_I2C 0x00000000 #define ATOM_DEVICE_USES_HARDWARE_ASSISTED_I2C 0x00000001 -/* usDeviceSupport: */ -/* Bits0 = 0 - no CRT1 support= 1- CRT1 is supported */ -/* Bit 1 = 0 - no LCD1 support= 1- LCD1 is supported */ -/* Bit 2 = 0 - no TV1 support= 1- TV1 is supported */ -/* Bit 3 = 0 - no DFP1 support= 1- DFP1 is supported */ -/* Bit 4 = 0 - no CRT2 support= 1- CRT2 is supported */ -/* Bit 5 = 0 - no LCD2 support= 1- LCD2 is supported */ -/* Bit 6 = 0 - no TV2 support= 1- TV2 is supported */ -/* Bit 7 = 0 - no DFP2 support= 1- DFP2 is supported */ -/* Bit 8 = 0 - no CV support= 1- CV is supported */ -/* Bit 9 = 0 - no DFP3 support= 1- DFP3 is supported */ -/* Byte1 (Supported Device Info) */ -/* Bit 0 = = 0 - no CV support= 1- CV is supported */ -/* */ -/* */ - -/* ucI2C_ConfigID */ -/* [7:0] - I2C LINE Associate ID */ -/* = 0 - no I2C */ -/* [7] - HW_Cap = 1, [6:0]=HW assisted I2C ID(HW line selection) */ -/* = 0, [6:0]=SW assisted I2C ID */ -/* [6-4] - HW_ENGINE_ID = 1, HW engine for NON multimedia use */ -/* = 2, HW engine for Multimedia use */ -/* = 3-7 Reserved for future I2C engines */ -/* [3-0] - I2C_LINE_MUX = A Mux number when it's HW assisted I2C or GPIO ID when it's SW I2C */ - -typedef struct _ATOM_I2C_ID_CONFIG { -#if ATOM_BIG_ENDIAN - UCHAR bfHW_Capable:1; - UCHAR bfHW_EngineID:3; - UCHAR bfI2C_LineMux:4; -#else - UCHAR bfI2C_LineMux:4; - UCHAR bfHW_EngineID:3; - UCHAR bfHW_Capable:1; -#endif -} ATOM_I2C_ID_CONFIG; - -typedef union _ATOM_I2C_ID_CONFIG_ACCESS { - ATOM_I2C_ID_CONFIG sbfAccess; - UCHAR ucAccess; -} ATOM_I2C_ID_CONFIG_ACCESS; +// usDeviceSupport: +// Bits0 = 0 - no CRT1 support= 1- CRT1 is supported +// Bit 1 = 0 - no LCD1 support= 1- LCD1 is supported +// Bit 2 = 0 - no TV1 support= 1- TV1 is supported +// Bit 3 = 0 - no DFP1 support= 1- DFP1 is supported +// Bit 4 = 0 - no CRT2 support= 1- CRT2 is supported +// Bit 5 = 0 - no LCD2 support= 1- LCD2 is supported +// Bit 6 = 0 - no DFP6 support= 1- DFP6 is supported +// Bit 7 = 0 - no DFP2 support= 1- DFP2 is supported +// Bit 8 = 0 - no CV support= 1- CV is supported +// Bit 9 = 0 - no DFP3 support= 1- DFP3 is supported +// Bit 10 = 0 - no DFP4 support= 1- DFP4 is supported +// Bit 11 = 0 - no DFP5 support= 1- DFP5 is supported +// +// /****************************************************************************/ -/* Structure used in GPIO_I2C_InfoTable */ +/* Structure used in MclkSS_InfoTable */ /****************************************************************************/ -typedef struct _ATOM_GPIO_I2C_ASSIGMENT { - USHORT usClkMaskRegisterIndex; - USHORT usClkEnRegisterIndex; - USHORT usClkY_RegisterIndex; - USHORT usClkA_RegisterIndex; - USHORT usDataMaskRegisterIndex; - USHORT usDataEnRegisterIndex; - USHORT usDataY_RegisterIndex; - USHORT usDataA_RegisterIndex; - ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; - UCHAR ucClkMaskShift; - UCHAR ucClkEnShift; - UCHAR ucClkY_Shift; - UCHAR ucClkA_Shift; - UCHAR ucDataMaskShift; - UCHAR ucDataEnShift; - UCHAR ucDataY_Shift; - UCHAR ucDataA_Shift; - UCHAR ucReserved1; - UCHAR ucReserved2; -} ATOM_GPIO_I2C_ASSIGMENT; - -typedef struct _ATOM_GPIO_I2C_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_GPIO_I2C_ASSIGMENT asGPIO_Info[ATOM_MAX_SUPPORTED_DEVICE]; -} ATOM_GPIO_I2C_INFO; +// ucI2C_ConfigID +// [7:0] - I2C LINE Associate ID +// = 0 - no I2C +// [7] - HW_Cap = 1, [6:0]=HW assisted I2C ID(HW line selection) +// = 0, [6:0]=SW assisted I2C ID +// [6-4] - HW_ENGINE_ID = 1, HW engine for NON multimedia use +// = 2, HW engine for Multimedia use +// = 3-7 Reserved for future I2C engines +// [3-0] - I2C_LINE_MUX = A Mux number when it's HW assisted I2C or GPIO ID when it's SW I2C + +typedef struct _ATOM_I2C_ID_CONFIG +{ +#if ATOM_BIG_ENDIAN + UCHAR bfHW_Capable:1; + UCHAR bfHW_EngineID:3; + UCHAR bfI2C_LineMux:4; +#else + UCHAR bfI2C_LineMux:4; + UCHAR bfHW_EngineID:3; + UCHAR bfHW_Capable:1; +#endif +}ATOM_I2C_ID_CONFIG; -/****************************************************************************/ -/* Common Structure used in other structures */ -/****************************************************************************/ +typedef union _ATOM_I2C_ID_CONFIG_ACCESS +{ + ATOM_I2C_ID_CONFIG sbfAccess; + UCHAR ucAccess; +}ATOM_I2C_ID_CONFIG_ACCESS; + + +/****************************************************************************/ +// Structure used in GPIO_I2C_InfoTable +/****************************************************************************/ +typedef struct _ATOM_GPIO_I2C_ASSIGMENT +{ + USHORT usClkMaskRegisterIndex; + USHORT usClkEnRegisterIndex; + USHORT usClkY_RegisterIndex; + USHORT usClkA_RegisterIndex; + USHORT usDataMaskRegisterIndex; + USHORT usDataEnRegisterIndex; + USHORT usDataY_RegisterIndex; + USHORT usDataA_RegisterIndex; + ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; + UCHAR ucClkMaskShift; + UCHAR ucClkEnShift; + UCHAR ucClkY_Shift; + UCHAR ucClkA_Shift; + UCHAR ucDataMaskShift; + UCHAR ucDataEnShift; + UCHAR ucDataY_Shift; + UCHAR ucDataA_Shift; + UCHAR ucReserved1; + UCHAR ucReserved2; +}ATOM_GPIO_I2C_ASSIGMENT; + +typedef struct _ATOM_GPIO_I2C_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_GPIO_I2C_ASSIGMENT asGPIO_Info[ATOM_MAX_SUPPORTED_DEVICE]; +}ATOM_GPIO_I2C_INFO; + +/****************************************************************************/ +// Common Structure used in other structures +/****************************************************************************/ #ifndef _H2INC - -/* Please don't add or expand this bitfield structure below, this one will retire soon.! */ -typedef struct _ATOM_MODE_MISC_INFO { + +//Please don't add or expand this bitfield structure below, this one will retire soon.! +typedef struct _ATOM_MODE_MISC_INFO +{ #if ATOM_BIG_ENDIAN - USHORT Reserved:6; - USHORT RGB888:1; - USHORT DoubleClock:1; - USHORT Interlace:1; - USHORT CompositeSync:1; - USHORT V_ReplicationBy2:1; - USHORT H_ReplicationBy2:1; - USHORT VerticalCutOff:1; - USHORT VSyncPolarity:1; /* 0=Active High, 1=Active Low */ - USHORT HSyncPolarity:1; /* 0=Active High, 1=Active Low */ - USHORT HorizontalCutOff:1; + USHORT Reserved:6; + USHORT RGB888:1; + USHORT DoubleClock:1; + USHORT Interlace:1; + USHORT CompositeSync:1; + USHORT V_ReplicationBy2:1; + USHORT H_ReplicationBy2:1; + USHORT VerticalCutOff:1; + USHORT VSyncPolarity:1; //0=Active High, 1=Active Low + USHORT HSyncPolarity:1; //0=Active High, 1=Active Low + USHORT HorizontalCutOff:1; #else - USHORT HorizontalCutOff:1; - USHORT HSyncPolarity:1; /* 0=Active High, 1=Active Low */ - USHORT VSyncPolarity:1; /* 0=Active High, 1=Active Low */ - USHORT VerticalCutOff:1; - USHORT H_ReplicationBy2:1; - USHORT V_ReplicationBy2:1; - USHORT CompositeSync:1; - USHORT Interlace:1; - USHORT DoubleClock:1; - USHORT RGB888:1; - USHORT Reserved:6; + USHORT HorizontalCutOff:1; + USHORT HSyncPolarity:1; //0=Active High, 1=Active Low + USHORT VSyncPolarity:1; //0=Active High, 1=Active Low + USHORT VerticalCutOff:1; + USHORT H_ReplicationBy2:1; + USHORT V_ReplicationBy2:1; + USHORT CompositeSync:1; + USHORT Interlace:1; + USHORT DoubleClock:1; + USHORT RGB888:1; + USHORT Reserved:6; #endif -} ATOM_MODE_MISC_INFO; - -typedef union _ATOM_MODE_MISC_INFO_ACCESS { - ATOM_MODE_MISC_INFO sbfAccess; - USHORT usAccess; -} ATOM_MODE_MISC_INFO_ACCESS; - +}ATOM_MODE_MISC_INFO; + +typedef union _ATOM_MODE_MISC_INFO_ACCESS +{ + ATOM_MODE_MISC_INFO sbfAccess; + USHORT usAccess; +}ATOM_MODE_MISC_INFO_ACCESS; + #else - -typedef union _ATOM_MODE_MISC_INFO_ACCESS { - USHORT usAccess; -} ATOM_MODE_MISC_INFO_ACCESS; - + +typedef union _ATOM_MODE_MISC_INFO_ACCESS +{ + USHORT usAccess; +}ATOM_MODE_MISC_INFO_ACCESS; + #endif -/* usModeMiscInfo- */ +// usModeMiscInfo- #define ATOM_H_CUTOFF 0x01 -#define ATOM_HSYNC_POLARITY 0x02 /* 0=Active High, 1=Active Low */ -#define ATOM_VSYNC_POLARITY 0x04 /* 0=Active High, 1=Active Low */ +#define ATOM_HSYNC_POLARITY 0x02 //0=Active High, 1=Active Low +#define ATOM_VSYNC_POLARITY 0x04 //0=Active High, 1=Active Low #define ATOM_V_CUTOFF 0x08 #define ATOM_H_REPLICATIONBY2 0x10 #define ATOM_V_REPLICATIONBY2 0x20 @@ -2099,10 +2581,10 @@ typedef union _ATOM_MODE_MISC_INFO_ACCESS { #define ATOM_DOUBLE_CLOCK_MODE 0x100 #define ATOM_RGB888_MODE 0x200 -/* usRefreshRate- */ +//usRefreshRate- #define ATOM_REFRESH_43 43 #define ATOM_REFRESH_47 47 -#define ATOM_REFRESH_56 56 +#define ATOM_REFRESH_56 56 #define ATOM_REFRESH_60 60 #define ATOM_REFRESH_65 65 #define ATOM_REFRESH_70 70 @@ -2110,192 +2592,233 @@ typedef union _ATOM_MODE_MISC_INFO_ACCESS { #define ATOM_REFRESH_75 75 #define ATOM_REFRESH_85 85 -/* ATOM_MODE_TIMING data are exactly the same as VESA timing data. */ -/* Translation from EDID to ATOM_MODE_TIMING, use the following formula. */ -/* */ -/* VESA_HTOTAL = VESA_ACTIVE + 2* VESA_BORDER + VESA_BLANK */ -/* = EDID_HA + EDID_HBL */ -/* VESA_HDISP = VESA_ACTIVE = EDID_HA */ -/* VESA_HSYNC_START = VESA_ACTIVE + VESA_BORDER + VESA_FRONT_PORCH */ -/* = EDID_HA + EDID_HSO */ -/* VESA_HSYNC_WIDTH = VESA_HSYNC_TIME = EDID_HSPW */ -/* VESA_BORDER = EDID_BORDER */ - -/****************************************************************************/ -/* Structure used in SetCRTC_UsingDTDTimingTable */ -/****************************************************************************/ -typedef struct _SET_CRTC_USING_DTD_TIMING_PARAMETERS { - USHORT usH_Size; - USHORT usH_Blanking_Time; - USHORT usV_Size; - USHORT usV_Blanking_Time; - USHORT usH_SyncOffset; - USHORT usH_SyncWidth; - USHORT usV_SyncOffset; - USHORT usV_SyncWidth; - ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; - UCHAR ucH_Border; /* From DFP EDID */ - UCHAR ucV_Border; - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucPadding[3]; -} SET_CRTC_USING_DTD_TIMING_PARAMETERS; - -/****************************************************************************/ -/* Structure used in SetCRTC_TimingTable */ -/****************************************************************************/ -typedef struct _SET_CRTC_TIMING_PARAMETERS { - USHORT usH_Total; /* horizontal total */ - USHORT usH_Disp; /* horizontal display */ - USHORT usH_SyncStart; /* horozontal Sync start */ - USHORT usH_SyncWidth; /* horizontal Sync width */ - USHORT usV_Total; /* vertical total */ - USHORT usV_Disp; /* vertical display */ - USHORT usV_SyncStart; /* vertical Sync start */ - USHORT usV_SyncWidth; /* vertical Sync width */ - ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; - UCHAR ucCRTC; /* ATOM_CRTC1 or ATOM_CRTC2 */ - UCHAR ucOverscanRight; /* right */ - UCHAR ucOverscanLeft; /* left */ - UCHAR ucOverscanBottom; /* bottom */ - UCHAR ucOverscanTop; /* top */ - UCHAR ucReserved; -} SET_CRTC_TIMING_PARAMETERS; +// ATOM_MODE_TIMING data are exactly the same as VESA timing data. +// Translation from EDID to ATOM_MODE_TIMING, use the following formula. +// +// VESA_HTOTAL = VESA_ACTIVE + 2* VESA_BORDER + VESA_BLANK +// = EDID_HA + EDID_HBL +// VESA_HDISP = VESA_ACTIVE = EDID_HA +// VESA_HSYNC_START = VESA_ACTIVE + VESA_BORDER + VESA_FRONT_PORCH +// = EDID_HA + EDID_HSO +// VESA_HSYNC_WIDTH = VESA_HSYNC_TIME = EDID_HSPW +// VESA_BORDER = EDID_BORDER + +/****************************************************************************/ +// Structure used in SetCRTC_UsingDTDTimingTable +/****************************************************************************/ +typedef struct _SET_CRTC_USING_DTD_TIMING_PARAMETERS +{ + USHORT usH_Size; + USHORT usH_Blanking_Time; + USHORT usV_Size; + USHORT usV_Blanking_Time; + USHORT usH_SyncOffset; + USHORT usH_SyncWidth; + USHORT usV_SyncOffset; + USHORT usV_SyncWidth; + ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; + UCHAR ucH_Border; // From DFP EDID + UCHAR ucV_Border; + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucPadding[3]; +}SET_CRTC_USING_DTD_TIMING_PARAMETERS; + +/****************************************************************************/ +// Structure used in SetCRTC_TimingTable +/****************************************************************************/ +typedef struct _SET_CRTC_TIMING_PARAMETERS +{ + USHORT usH_Total; // horizontal total + USHORT usH_Disp; // horizontal display + USHORT usH_SyncStart; // horozontal Sync start + USHORT usH_SyncWidth; // horizontal Sync width + USHORT usV_Total; // vertical total + USHORT usV_Disp; // vertical display + USHORT usV_SyncStart; // vertical Sync start + USHORT usV_SyncWidth; // vertical Sync width + ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; + UCHAR ucCRTC; // ATOM_CRTC1 or ATOM_CRTC2 + UCHAR ucOverscanRight; // right + UCHAR ucOverscanLeft; // left + UCHAR ucOverscanBottom; // bottom + UCHAR ucOverscanTop; // top + UCHAR ucReserved; +}SET_CRTC_TIMING_PARAMETERS; #define SET_CRTC_TIMING_PARAMETERS_PS_ALLOCATION SET_CRTC_TIMING_PARAMETERS -/****************************************************************************/ -/* Structure used in StandardVESA_TimingTable */ -/* AnalogTV_InfoTable */ -/* ComponentVideoInfoTable */ -/****************************************************************************/ -typedef struct _ATOM_MODE_TIMING { - USHORT usCRTC_H_Total; - USHORT usCRTC_H_Disp; - USHORT usCRTC_H_SyncStart; - USHORT usCRTC_H_SyncWidth; - USHORT usCRTC_V_Total; - USHORT usCRTC_V_Disp; - USHORT usCRTC_V_SyncStart; - USHORT usCRTC_V_SyncWidth; - USHORT usPixelClock; /* in 10Khz unit */ - ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; - USHORT usCRTC_OverscanRight; - USHORT usCRTC_OverscanLeft; - USHORT usCRTC_OverscanBottom; - USHORT usCRTC_OverscanTop; - USHORT usReserve; - UCHAR ucInternalModeNumber; - UCHAR ucRefreshRate; -} ATOM_MODE_TIMING; - -typedef struct _ATOM_DTD_FORMAT { - USHORT usPixClk; - USHORT usHActive; - USHORT usHBlanking_Time; - USHORT usVActive; - USHORT usVBlanking_Time; - USHORT usHSyncOffset; - USHORT usHSyncWidth; - USHORT usVSyncOffset; - USHORT usVSyncWidth; - USHORT usImageHSize; - USHORT usImageVSize; - UCHAR ucHBorder; - UCHAR ucVBorder; - ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; - UCHAR ucInternalModeNumber; - UCHAR ucRefreshRate; -} ATOM_DTD_FORMAT; - -/****************************************************************************/ -/* Structure used in LVDS_InfoTable */ -/* * Need a document to describe this table */ -/****************************************************************************/ +/****************************************************************************/ +// Structure used in StandardVESA_TimingTable +// AnalogTV_InfoTable +// ComponentVideoInfoTable +/****************************************************************************/ +typedef struct _ATOM_MODE_TIMING +{ + USHORT usCRTC_H_Total; + USHORT usCRTC_H_Disp; + USHORT usCRTC_H_SyncStart; + USHORT usCRTC_H_SyncWidth; + USHORT usCRTC_V_Total; + USHORT usCRTC_V_Disp; + USHORT usCRTC_V_SyncStart; + USHORT usCRTC_V_SyncWidth; + USHORT usPixelClock; //in 10Khz unit + ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; + USHORT usCRTC_OverscanRight; + USHORT usCRTC_OverscanLeft; + USHORT usCRTC_OverscanBottom; + USHORT usCRTC_OverscanTop; + USHORT usReserve; + UCHAR ucInternalModeNumber; + UCHAR ucRefreshRate; +}ATOM_MODE_TIMING; + +typedef struct _ATOM_DTD_FORMAT +{ + USHORT usPixClk; + USHORT usHActive; + USHORT usHBlanking_Time; + USHORT usVActive; + USHORT usVBlanking_Time; + USHORT usHSyncOffset; + USHORT usHSyncWidth; + USHORT usVSyncOffset; + USHORT usVSyncWidth; + USHORT usImageHSize; + USHORT usImageVSize; + UCHAR ucHBorder; + UCHAR ucVBorder; + ATOM_MODE_MISC_INFO_ACCESS susModeMiscInfo; + UCHAR ucInternalModeNumber; + UCHAR ucRefreshRate; +}ATOM_DTD_FORMAT; + +/****************************************************************************/ +// Structure used in LVDS_InfoTable +// * Need a document to describe this table +/****************************************************************************/ #define SUPPORTED_LCD_REFRESHRATE_30Hz 0x0004 #define SUPPORTED_LCD_REFRESHRATE_40Hz 0x0008 #define SUPPORTED_LCD_REFRESHRATE_50Hz 0x0010 #define SUPPORTED_LCD_REFRESHRATE_60Hz 0x0020 -/* Once DAL sees this CAP is set, it will read EDID from LCD on its own instead of using sLCDTiming in ATOM_LVDS_INFO_V12. */ -/* Other entries in ATOM_LVDS_INFO_V12 are still valid/useful to DAL */ -#define LCDPANEL_CAP_READ_EDID 0x1 - -/* ucTableFormatRevision=1 */ -/* ucTableContentRevision=1 */ -typedef struct _ATOM_LVDS_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_DTD_FORMAT sLCDTiming; - USHORT usModePatchTableOffset; - USHORT usSupportedRefreshRate; /* Refer to panel info table in ATOMBIOS extension Spec. */ - USHORT usOffDelayInMs; - UCHAR ucPowerSequenceDigOntoDEin10Ms; - UCHAR ucPowerSequenceDEtoBLOnin10Ms; - UCHAR ucLVDS_Misc; /* Bit0:{=0:single, =1:dual},Bit1 {=0:666RGB, =1:888RGB},Bit2:3:{Grey level} */ - /* Bit4:{=0:LDI format for RGB888, =1 FPDI format for RGB888} */ - /* Bit5:{=0:Spatial Dithering disabled;1 Spatial Dithering enabled} */ - /* Bit6:{=0:Temporal Dithering disabled;1 Temporal Dithering enabled} */ - UCHAR ucPanelDefaultRefreshRate; - UCHAR ucPanelIdentification; - UCHAR ucSS_Id; -} ATOM_LVDS_INFO; - -/* ucTableFormatRevision=1 */ -/* ucTableContentRevision=2 */ -typedef struct _ATOM_LVDS_INFO_V12 { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_DTD_FORMAT sLCDTiming; - USHORT usExtInfoTableOffset; - USHORT usSupportedRefreshRate; /* Refer to panel info table in ATOMBIOS extension Spec. */ - USHORT usOffDelayInMs; - UCHAR ucPowerSequenceDigOntoDEin10Ms; - UCHAR ucPowerSequenceDEtoBLOnin10Ms; - UCHAR ucLVDS_Misc; /* Bit0:{=0:single, =1:dual},Bit1 {=0:666RGB, =1:888RGB},Bit2:3:{Grey level} */ - /* Bit4:{=0:LDI format for RGB888, =1 FPDI format for RGB888} */ - /* Bit5:{=0:Spatial Dithering disabled;1 Spatial Dithering enabled} */ - /* Bit6:{=0:Temporal Dithering disabled;1 Temporal Dithering enabled} */ - UCHAR ucPanelDefaultRefreshRate; - UCHAR ucPanelIdentification; - UCHAR ucSS_Id; - USHORT usLCDVenderID; - USHORT usLCDProductID; - UCHAR ucLCDPanel_SpecialHandlingCap; - UCHAR ucPanelInfoSize; /* start from ATOM_DTD_FORMAT to end of panel info, include ExtInfoTable */ - UCHAR ucReserved[2]; -} ATOM_LVDS_INFO_V12; +//ucTableFormatRevision=1 +//ucTableContentRevision=1 +typedef struct _ATOM_LVDS_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_DTD_FORMAT sLCDTiming; + USHORT usModePatchTableOffset; + USHORT usSupportedRefreshRate; //Refer to panel info table in ATOMBIOS extension Spec. + USHORT usOffDelayInMs; + UCHAR ucPowerSequenceDigOntoDEin10Ms; + UCHAR ucPowerSequenceDEtoBLOnin10Ms; + UCHAR ucLVDS_Misc; // Bit0:{=0:single, =1:dual},Bit1 {=0:666RGB, =1:888RGB},Bit2:3:{Grey level} + // Bit4:{=0:LDI format for RGB888, =1 FPDI format for RGB888} + // Bit5:{=0:Spatial Dithering disabled;1 Spatial Dithering enabled} + // Bit6:{=0:Temporal Dithering disabled;1 Temporal Dithering enabled} + UCHAR ucPanelDefaultRefreshRate; + UCHAR ucPanelIdentification; + UCHAR ucSS_Id; +}ATOM_LVDS_INFO; + +//ucTableFormatRevision=1 +//ucTableContentRevision=2 +typedef struct _ATOM_LVDS_INFO_V12 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_DTD_FORMAT sLCDTiming; + USHORT usExtInfoTableOffset; + USHORT usSupportedRefreshRate; //Refer to panel info table in ATOMBIOS extension Spec. + USHORT usOffDelayInMs; + UCHAR ucPowerSequenceDigOntoDEin10Ms; + UCHAR ucPowerSequenceDEtoBLOnin10Ms; + UCHAR ucLVDS_Misc; // Bit0:{=0:single, =1:dual},Bit1 {=0:666RGB, =1:888RGB},Bit2:3:{Grey level} + // Bit4:{=0:LDI format for RGB888, =1 FPDI format for RGB888} + // Bit5:{=0:Spatial Dithering disabled;1 Spatial Dithering enabled} + // Bit6:{=0:Temporal Dithering disabled;1 Temporal Dithering enabled} + UCHAR ucPanelDefaultRefreshRate; + UCHAR ucPanelIdentification; + UCHAR ucSS_Id; + USHORT usLCDVenderID; + USHORT usLCDProductID; + UCHAR ucLCDPanel_SpecialHandlingCap; + UCHAR ucPanelInfoSize; // start from ATOM_DTD_FORMAT to end of panel info, include ExtInfoTable + UCHAR ucReserved[2]; +}ATOM_LVDS_INFO_V12; + +//Definitions for ucLCDPanel_SpecialHandlingCap: + +//Once DAL sees this CAP is set, it will read EDID from LCD on its own instead of using sLCDTiming in ATOM_LVDS_INFO_V12. +//Other entries in ATOM_LVDS_INFO_V12 are still valid/useful to DAL +#define LCDPANEL_CAP_READ_EDID 0x1 + +//If a design supports DRR (dynamic refresh rate) on internal panels (LVDS or EDP), this cap is set in ucLCDPanel_SpecialHandlingCap together +//with multiple supported refresh rates@usSupportedRefreshRate. This cap should not be set when only slow refresh rate is supported (static +//refresh rate switch by SW. This is only valid from ATOM_LVDS_INFO_V12 +#define LCDPANEL_CAP_DRR_SUPPORTED 0x2 + +//Use this cap bit for a quick reference whether an embadded panel (LCD1 ) is LVDS or eDP. +#define LCDPANEL_CAP_eDP 0x4 + + +//Color Bit Depth definition in EDID V1.4 @BYTE 14h +//Bit 6 5 4 + // 0 0 0 - Color bit depth is undefined + // 0 0 1 - 6 Bits per Primary Color + // 0 1 0 - 8 Bits per Primary Color + // 0 1 1 - 10 Bits per Primary Color + // 1 0 0 - 12 Bits per Primary Color + // 1 0 1 - 14 Bits per Primary Color + // 1 1 0 - 16 Bits per Primary Color + // 1 1 1 - Reserved + +#define PANEL_COLOR_BIT_DEPTH_MASK 0x70 + +// Bit7:{=0:Random Dithering disabled;1 Random Dithering enabled} +#define PANEL_RANDOM_DITHER 0x80 +#define PANEL_RANDOM_DITHER_MASK 0x80 + #define ATOM_LVDS_INFO_LAST ATOM_LVDS_INFO_V12 -typedef struct _ATOM_PATCH_RECORD_MODE { - UCHAR ucRecordType; - USHORT usHDisp; - USHORT usVDisp; -} ATOM_PATCH_RECORD_MODE; +typedef struct _ATOM_PATCH_RECORD_MODE +{ + UCHAR ucRecordType; + USHORT usHDisp; + USHORT usVDisp; +}ATOM_PATCH_RECORD_MODE; -typedef struct _ATOM_LCD_RTS_RECORD { - UCHAR ucRecordType; - UCHAR ucRTSValue; -} ATOM_LCD_RTS_RECORD; +typedef struct _ATOM_LCD_RTS_RECORD +{ + UCHAR ucRecordType; + UCHAR ucRTSValue; +}ATOM_LCD_RTS_RECORD; -/* !! If the record below exits, it shoud always be the first record for easy use in command table!!! */ -typedef struct _ATOM_LCD_MODE_CONTROL_CAP { - UCHAR ucRecordType; - USHORT usLCDCap; -} ATOM_LCD_MODE_CONTROL_CAP; +//!! If the record below exits, it shoud always be the first record for easy use in command table!!! +// The record below is only used when LVDS_Info is present. From ATOM_LVDS_INFO_V12, use ucLCDPanel_SpecialHandlingCap instead. +typedef struct _ATOM_LCD_MODE_CONTROL_CAP +{ + UCHAR ucRecordType; + USHORT usLCDCap; +}ATOM_LCD_MODE_CONTROL_CAP; #define LCD_MODE_CAP_BL_OFF 1 #define LCD_MODE_CAP_CRTC_OFF 2 #define LCD_MODE_CAP_PANEL_OFF 4 -typedef struct _ATOM_FAKE_EDID_PATCH_RECORD { - UCHAR ucRecordType; - UCHAR ucFakeEDIDLength; - UCHAR ucFakeEDIDString[1]; /* This actually has ucFakeEdidLength elements. */ +typedef struct _ATOM_FAKE_EDID_PATCH_RECORD +{ + UCHAR ucRecordType; + UCHAR ucFakeEDIDLength; + UCHAR ucFakeEDIDString[1]; // This actually has ucFakeEdidLength elements. } ATOM_FAKE_EDID_PATCH_RECORD; -typedef struct _ATOM_PANEL_RESOLUTION_PATCH_RECORD { - UCHAR ucRecordType; - USHORT usHSize; - USHORT usVSize; -} ATOM_PANEL_RESOLUTION_PATCH_RECORD; +typedef struct _ATOM_PANEL_RESOLUTION_PATCH_RECORD +{ + UCHAR ucRecordType; + USHORT usHSize; + USHORT usVSize; +}ATOM_PANEL_RESOLUTION_PATCH_RECORD; #define LCD_MODE_PATCH_RECORD_MODE_TYPE 1 #define LCD_RTS_RECORD_TYPE 2 @@ -2306,21 +2829,25 @@ typedef struct _ATOM_PANEL_RESOLUTION_PATCH_RECORD { /****************************Spread Spectrum Info Table Definitions **********************/ -/* ucTableFormatRevision=1 */ -/* ucTableContentRevision=2 */ -typedef struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT { - USHORT usSpreadSpectrumPercentage; - UCHAR ucSpreadSpectrumType; /* Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Others:TBD */ - UCHAR ucSS_Step; - UCHAR ucSS_Delay; - UCHAR ucSS_Id; - UCHAR ucRecommendedRef_Div; - UCHAR ucSS_Range; /* it was reserved for V11 */ -} ATOM_SPREAD_SPECTRUM_ASSIGNMENT; +//ucTableFormatRevision=1 +//ucTableContentRevision=2 +typedef struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT +{ + USHORT usSpreadSpectrumPercentage; + UCHAR ucSpreadSpectrumType; //Bit1=0 Down Spread,=1 Center Spread. Bit1=1 Ext. =0 Int. Bit2=1: PCIE REFCLK SS =0 iternal PPLL SS Others:TBD + UCHAR ucSS_Step; + UCHAR ucSS_Delay; + UCHAR ucSS_Id; + UCHAR ucRecommendedRef_Div; + UCHAR ucSS_Range; //it was reserved for V11 +}ATOM_SPREAD_SPECTRUM_ASSIGNMENT; #define ATOM_MAX_SS_ENTRY 16 -#define ATOM_DP_SS_ID1 0x0f1 /* SS modulation freq=30k */ -#define ATOM_DP_SS_ID2 0x0f2 /* SS modulation freq=33k */ +#define ATOM_DP_SS_ID1 0x0f1 // SS ID for internal DP stream at 2.7Ghz. if ATOM_DP_SS_ID2 does not exist in SS_InfoTable, it is used for internal DP stream at 1.62Ghz as well. +#define ATOM_DP_SS_ID2 0x0f2 // SS ID for internal DP stream at 1.62Ghz, if it exists in SS_InfoTable. +#define ATOM_LVLINK_2700MHz_SS_ID 0x0f3 // SS ID for LV link translator chip at 2.7Ghz +#define ATOM_LVLINK_1620MHz_SS_ID 0x0f4 // SS ID for LV link translator chip at 1.62Ghz + #define ATOM_SS_DOWN_SPREAD_MODE_MASK 0x00000000 #define ATOM_SS_DOWN_SPREAD_MODE 0x00000000 @@ -2329,29 +2856,30 @@ typedef struct _ATOM_SPREAD_SPECTRUM_ASSIGNMENT { #define ATOM_INTERNAL_SS_MASK 0x00000000 #define ATOM_EXTERNAL_SS_MASK 0x00000002 #define EXEC_SS_STEP_SIZE_SHIFT 2 -#define EXEC_SS_DELAY_SHIFT 4 +#define EXEC_SS_DELAY_SHIFT 4 #define ACTIVEDATA_TO_BLON_DELAY_SHIFT 4 -typedef struct _ATOM_SPREAD_SPECTRUM_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_SPREAD_SPECTRUM_ASSIGNMENT asSS_Info[ATOM_MAX_SS_ENTRY]; -} ATOM_SPREAD_SPECTRUM_INFO; - -/****************************************************************************/ -/* Structure used in AnalogTV_InfoTable (Top level) */ -/****************************************************************************/ -/* ucTVBootUpDefaultStd definiton: */ - -/* ATOM_TV_NTSC 1 */ -/* ATOM_TV_NTSCJ 2 */ -/* ATOM_TV_PAL 3 */ -/* ATOM_TV_PALM 4 */ -/* ATOM_TV_PALCN 5 */ -/* ATOM_TV_PALN 6 */ -/* ATOM_TV_PAL60 7 */ -/* ATOM_TV_SECAM 8 */ - -/* ucTVSuppportedStd definition: */ +typedef struct _ATOM_SPREAD_SPECTRUM_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_SPREAD_SPECTRUM_ASSIGNMENT asSS_Info[ATOM_MAX_SS_ENTRY]; +}ATOM_SPREAD_SPECTRUM_INFO; + +/****************************************************************************/ +// Structure used in AnalogTV_InfoTable (Top level) +/****************************************************************************/ +//ucTVBootUpDefaultStd definiton: + +//ATOM_TV_NTSC 1 +//ATOM_TV_NTSCJ 2 +//ATOM_TV_PAL 3 +//ATOM_TV_PALM 4 +//ATOM_TV_PALCN 5 +//ATOM_TV_PALN 6 +//ATOM_TV_PAL60 7 +//ATOM_TV_SECAM 8 + +//ucTVSupportedStd definition: #define NTSC_SUPPORT 0x1 #define NTSCJ_SUPPORT 0x2 @@ -2364,46 +2892,58 @@ typedef struct _ATOM_SPREAD_SPECTRUM_INFO { #define MAX_SUPPORTED_TV_TIMING 2 -typedef struct _ATOM_ANALOG_TV_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucTV_SupportedStandard; - UCHAR ucTV_BootUpDefaultStandard; - UCHAR ucExt_TV_ASIC_ID; - UCHAR ucExt_TV_ASIC_SlaveAddr; - /*ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_TV_TIMING]; */ - ATOM_MODE_TIMING aModeTimings[MAX_SUPPORTED_TV_TIMING]; -} ATOM_ANALOG_TV_INFO; +typedef struct _ATOM_ANALOG_TV_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucTV_SupportedStandard; + UCHAR ucTV_BootUpDefaultStandard; + UCHAR ucExt_TV_ASIC_ID; + UCHAR ucExt_TV_ASIC_SlaveAddr; + /*ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_TV_TIMING];*/ + ATOM_MODE_TIMING aModeTimings[MAX_SUPPORTED_TV_TIMING]; +}ATOM_ANALOG_TV_INFO; #define MAX_SUPPORTED_TV_TIMING_V1_2 3 -typedef struct _ATOM_ANALOG_TV_INFO_V1_2 { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucTV_SupportedStandard; - UCHAR ucTV_BootUpDefaultStandard; - UCHAR ucExt_TV_ASIC_ID; - UCHAR ucExt_TV_ASIC_SlaveAddr; - ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_TV_TIMING]; -} ATOM_ANALOG_TV_INFO_V1_2; +typedef struct _ATOM_ANALOG_TV_INFO_V1_2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucTV_SupportedStandard; + UCHAR ucTV_BootUpDefaultStandard; + UCHAR ucExt_TV_ASIC_ID; + UCHAR ucExt_TV_ASIC_SlaveAddr; + ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_TV_TIMING]; +}ATOM_ANALOG_TV_INFO_V1_2; + +typedef struct _ATOM_DPCD_INFO +{ + UCHAR ucRevisionNumber; //10h : Revision 1.0; 11h : Revision 1.1 + UCHAR ucMaxLinkRate; //06h : 1.62Gbps per lane; 0Ah = 2.7Gbps per lane + UCHAR ucMaxLane; //Bits 4:0 = MAX_LANE_COUNT (1/2/4). Bit 7 = ENHANCED_FRAME_CAP + UCHAR ucMaxDownSpread; //Bit0 = 0: No Down spread; Bit0 = 1: 0.5% (Subject to change according to DP spec) +}ATOM_DPCD_INFO; + +#define ATOM_DPCD_MAX_LANE_MASK 0x1F /**************************************************************************/ -/* VRAM usage and their definitions */ +// VRAM usage and their defintions -/* One chunk of VRAM used by Bios are for HWICON surfaces,EDID data. */ -/* Current Mode timing and Dail Timing and/or STD timing data EACH device. They can be broken down as below. */ -/* All the addresses below are the offsets from the frame buffer start.They all MUST be Dword aligned! */ -/* To driver: The physical address of this memory portion=mmFB_START(4K aligned)+ATOMBIOS_VRAM_USAGE_START_ADDR+ATOM_x_ADDR */ -/* To Bios: ATOMBIOS_VRAM_USAGE_START_ADDR+ATOM_x_ADDR->MM_INDEX */ +// One chunk of VRAM used by Bios are for HWICON surfaces,EDID data. +// Current Mode timing and Dail Timing and/or STD timing data EACH device. They can be broken down as below. +// All the addresses below are the offsets from the frame buffer start.They all MUST be Dword aligned! +// To driver: The physical address of this memory portion=mmFB_START(4K aligned)+ATOMBIOS_VRAM_USAGE_START_ADDR+ATOM_x_ADDR +// To Bios: ATOMBIOS_VRAM_USAGE_START_ADDR+ATOM_x_ADDR->MM_INDEX #ifndef VESA_MEMORY_IN_64K_BLOCK -#define VESA_MEMORY_IN_64K_BLOCK 0x100 /* 256*64K=16Mb (Max. VESA memory is 16Mb!) */ +#define VESA_MEMORY_IN_64K_BLOCK 0x100 //256*64K=16Mb (Max. VESA memory is 16Mb!) #endif -#define ATOM_EDID_RAW_DATASIZE 256 /* In Bytes */ -#define ATOM_HWICON_SURFACE_SIZE 4096 /* In Bytes */ +#define ATOM_EDID_RAW_DATASIZE 256 //In Bytes +#define ATOM_HWICON_SURFACE_SIZE 4096 //In Bytes #define ATOM_HWICON_INFOTABLE_SIZE 32 #define MAX_DTD_MODE_IN_VRAM 6 -#define ATOM_DTD_MODE_SUPPORT_TBL_SIZE (MAX_DTD_MODE_IN_VRAM*28) /* 28= (SIZEOF ATOM_DTD_FORMAT) */ -#define ATOM_STD_MODE_SUPPORT_TBL_SIZE (32*8) /* 32 is a predefined number,8= (SIZEOF ATOM_STD_FORMAT) */ +#define ATOM_DTD_MODE_SUPPORT_TBL_SIZE (MAX_DTD_MODE_IN_VRAM*28) //28= (SIZEOF ATOM_DTD_FORMAT) +#define ATOM_STD_MODE_SUPPORT_TBL_SIZE 32*8 //32 is a predefined number,8= (SIZEOF ATOM_STD_FORMAT) #define DFP_ENCODER_TYPE_OFFSET 0x80 #define DP_ENCODER_LANE_NUM_OFFSET 0x84 #define DP_ENCODER_LINK_RATE_OFFSET 0x88 @@ -2417,7 +2957,7 @@ typedef struct _ATOM_ANALOG_TV_INFO_V1_2 { #define ATOM_LCD1_EDID_ADDR (ATOM_CRT1_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) #define ATOM_LCD1_DTD_MODE_TBL_ADDR (ATOM_LCD1_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) -#define ATOM_LCD1_STD_MODE_TBL_ADDR (ATOM_LCD1_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_LCD1_STD_MODE_TBL_ADDR (ATOM_LCD1_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) #define ATOM_TV1_DTD_MODE_TBL_ADDR (ATOM_LCD1_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) @@ -2431,13 +2971,13 @@ typedef struct _ATOM_ANALOG_TV_INFO_V1_2 { #define ATOM_LCD2_EDID_ADDR (ATOM_CRT2_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) #define ATOM_LCD2_DTD_MODE_TBL_ADDR (ATOM_LCD2_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) -#define ATOM_LCD2_STD_MODE_TBL_ADDR (ATOM_LCD2_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_LCD2_STD_MODE_TBL_ADDR (ATOM_LCD2_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) -#define ATOM_TV2_EDID_ADDR (ATOM_LCD2_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) -#define ATOM_TV2_DTD_MODE_TBL_ADDR (ATOM_TV2_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) -#define ATOM_TV2_STD_MODE_TBL_ADDR (ATOM_TV2_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_DFP6_EDID_ADDR (ATOM_LCD2_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_DFP6_DTD_MODE_TBL_ADDR (ATOM_DFP6_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) +#define ATOM_DFP6_STD_MODE_TBL_ADDR (ATOM_DFP6_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) -#define ATOM_DFP2_EDID_ADDR (ATOM_TV2_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_DFP2_EDID_ADDR (ATOM_DFP6_STD_MODE_TBL_ADDR + ATOM_STD_MODE_SUPPORT_TBL_SIZE) #define ATOM_DFP2_DTD_MODE_TBL_ADDR (ATOM_DFP2_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) #define ATOM_DFP2_STD_MODE_TBL_ADDR (ATOM_DFP2_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) @@ -2457,533 +2997,850 @@ typedef struct _ATOM_ANALOG_TV_INFO_V1_2 { #define ATOM_DFP5_DTD_MODE_TBL_ADDR (ATOM_DFP5_EDID_ADDR + ATOM_EDID_RAW_DATASIZE) #define ATOM_DFP5_STD_MODE_TBL_ADDR (ATOM_DFP5_DTD_MODE_TBL_ADDR + ATOM_DTD_MODE_SUPPORT_TBL_SIZE) -#define ATOM_DP_TRAINING_TBL_ADDR (ATOM_DFP5_STD_MODE_TBL_ADDR+ATOM_STD_MODE_SUPPORT_TBL_SIZE) +#define ATOM_DP_TRAINING_TBL_ADDR (ATOM_DFP5_STD_MODE_TBL_ADDR+ATOM_STD_MODE_SUPPORT_TBL_SIZE) -#define ATOM_STACK_STORAGE_START (ATOM_DP_TRAINING_TBL_ADDR + 256) -#define ATOM_STACK_STORAGE_END (ATOM_STACK_STORAGE_START + 512) +#define ATOM_STACK_STORAGE_START (ATOM_DP_TRAINING_TBL_ADDR+256) +#define ATOM_STACK_STORAGE_END ATOM_STACK_STORAGE_START+512 -/* The size below is in Kb! */ +//The size below is in Kb! #define ATOM_VRAM_RESERVE_SIZE ((((ATOM_STACK_STORAGE_END - ATOM_HWICON1_SURFACE_ADDR)>>10)+4)&0xFFFC) - + #define ATOM_VRAM_OPERATION_FLAGS_MASK 0xC0000000L #define ATOM_VRAM_OPERATION_FLAGS_SHIFT 30 #define ATOM_VRAM_BLOCK_NEEDS_NO_RESERVATION 0x1 #define ATOM_VRAM_BLOCK_NEEDS_RESERVATION 0x0 -/***********************************************************************************/ -/* Structure used in VRAM_UsageByFirmwareTable */ -/* Note1: This table is filled by SetBiosReservationStartInFB in CoreCommSubs.asm */ -/* at running time. */ -/* note2: From RV770, the memory is more than 32bit addressable, so we will change */ -/* ucTableFormatRevision=1,ucTableContentRevision=4, the strcuture remains */ -/* exactly same as 1.1 and 1.2 (1.3 is never in use), but ulStartAddrUsedByFirmware */ -/* (in offset to start of memory address) is KB aligned instead of byte aligend. */ -/***********************************************************************************/ +/***********************************************************************************/ +// Structure used in VRAM_UsageByFirmwareTable +// Note1: This table is filled by SetBiosReservationStartInFB in CoreCommSubs.asm +// at running time. +// note2: From RV770, the memory is more than 32bit addressable, so we will change +// ucTableFormatRevision=1,ucTableContentRevision=4, the strcuture remains +// exactly same as 1.1 and 1.2 (1.3 is never in use), but ulStartAddrUsedByFirmware +// (in offset to start of memory address) is KB aligned instead of byte aligend. +/***********************************************************************************/ +// Note3: +/* If we change usReserved to "usFBUsedbyDrvInKB", then to VBIOS this usFBUsedbyDrvInKB is a predefined, unchanged constant across VGA or non VGA adapter, +for CAIL, The size of FB access area is known, only thing missing is the Offset of FB Access area, so we can have: + +If (ulStartAddrUsedByFirmware!=0) +FBAccessAreaOffset= ulStartAddrUsedByFirmware - usFBUsedbyDrvInKB; +Reserved area has been claimed by VBIOS including this FB access area; CAIL doesn't need to reserve any extra area for this purpose +else //Non VGA case + if (FB_Size<=2Gb) + FBAccessAreaOffset= FB_Size - usFBUsedbyDrvInKB; + else + FBAccessAreaOffset= Aper_Size - usFBUsedbyDrvInKB + +CAIL needs to claim an reserved area defined by FBAccessAreaOffset and usFBUsedbyDrvInKB in non VGA case.*/ + #define ATOM_MAX_FIRMWARE_VRAM_USAGE_INFO 1 -typedef struct _ATOM_FIRMWARE_VRAM_RESERVE_INFO { - ULONG ulStartAddrUsedByFirmware; - USHORT usFirmwareUseInKb; - USHORT usReserved; -} ATOM_FIRMWARE_VRAM_RESERVE_INFO; +typedef struct _ATOM_FIRMWARE_VRAM_RESERVE_INFO +{ + ULONG ulStartAddrUsedByFirmware; + USHORT usFirmwareUseInKb; + USHORT usReserved; +}ATOM_FIRMWARE_VRAM_RESERVE_INFO; -typedef struct _ATOM_VRAM_USAGE_BY_FIRMWARE { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_FIRMWARE_VRAM_RESERVE_INFO - asFirmwareVramReserveInfo[ATOM_MAX_FIRMWARE_VRAM_USAGE_INFO]; -} ATOM_VRAM_USAGE_BY_FIRMWARE; +typedef struct _ATOM_VRAM_USAGE_BY_FIRMWARE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_FIRMWARE_VRAM_RESERVE_INFO asFirmwareVramReserveInfo[ATOM_MAX_FIRMWARE_VRAM_USAGE_INFO]; +}ATOM_VRAM_USAGE_BY_FIRMWARE; -/****************************************************************************/ -/* Structure used in GPIO_Pin_LUTTable */ -/****************************************************************************/ -typedef struct _ATOM_GPIO_PIN_ASSIGNMENT { - USHORT usGpioPin_AIndex; - UCHAR ucGpioPinBitShift; - UCHAR ucGPIO_ID; -} ATOM_GPIO_PIN_ASSIGNMENT; +// change verion to 1.5, when allow driver to allocate the vram area for command table access. +typedef struct _ATOM_FIRMWARE_VRAM_RESERVE_INFO_V1_5 +{ + ULONG ulStartAddrUsedByFirmware; + USHORT usFirmwareUseInKb; + USHORT usFBUsedByDrvInKb; +}ATOM_FIRMWARE_VRAM_RESERVE_INFO_V1_5; -typedef struct _ATOM_GPIO_PIN_LUT { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_GPIO_PIN_ASSIGNMENT asGPIO_Pin[1]; -} ATOM_GPIO_PIN_LUT; +typedef struct _ATOM_VRAM_USAGE_BY_FIRMWARE_V1_5 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_FIRMWARE_VRAM_RESERVE_INFO_V1_5 asFirmwareVramReserveInfo[ATOM_MAX_FIRMWARE_VRAM_USAGE_INFO]; +}ATOM_VRAM_USAGE_BY_FIRMWARE_V1_5; + +/****************************************************************************/ +// Structure used in GPIO_Pin_LUTTable +/****************************************************************************/ +typedef struct _ATOM_GPIO_PIN_ASSIGNMENT +{ + USHORT usGpioPin_AIndex; + UCHAR ucGpioPinBitShift; + UCHAR ucGPIO_ID; +}ATOM_GPIO_PIN_ASSIGNMENT; -/****************************************************************************/ -/* Structure used in ComponentVideoInfoTable */ -/****************************************************************************/ +typedef struct _ATOM_GPIO_PIN_LUT +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_GPIO_PIN_ASSIGNMENT asGPIO_Pin[1]; +}ATOM_GPIO_PIN_LUT; + +/****************************************************************************/ +// Structure used in ComponentVideoInfoTable +/****************************************************************************/ #define GPIO_PIN_ACTIVE_HIGH 0x1 #define MAX_SUPPORTED_CV_STANDARDS 5 -/* definitions for ATOM_D_INFO.ucSettings */ -#define ATOM_GPIO_SETTINGS_BITSHIFT_MASK 0x1F /* [4:0] */ -#define ATOM_GPIO_SETTINGS_RESERVED_MASK 0x60 /* [6:5] = must be zeroed out */ -#define ATOM_GPIO_SETTINGS_ACTIVE_MASK 0x80 /* [7] */ +// definitions for ATOM_D_INFO.ucSettings +#define ATOM_GPIO_SETTINGS_BITSHIFT_MASK 0x1F // [4:0] +#define ATOM_GPIO_SETTINGS_RESERVED_MASK 0x60 // [6:5] = must be zeroed out +#define ATOM_GPIO_SETTINGS_ACTIVE_MASK 0x80 // [7] -typedef struct _ATOM_GPIO_INFO { - USHORT usAOffset; - UCHAR ucSettings; - UCHAR ucReserved; -} ATOM_GPIO_INFO; +typedef struct _ATOM_GPIO_INFO +{ + USHORT usAOffset; + UCHAR ucSettings; + UCHAR ucReserved; +}ATOM_GPIO_INFO; -/* definitions for ATOM_COMPONENT_VIDEO_INFO.ucMiscInfo (bit vector) */ +// definitions for ATOM_COMPONENT_VIDEO_INFO.ucMiscInfo (bit vector) #define ATOM_CV_RESTRICT_FORMAT_SELECTION 0x2 -/* definitions for ATOM_COMPONENT_VIDEO_INFO.uc480i/uc480p/uc720p/uc1080i */ -#define ATOM_GPIO_DEFAULT_MODE_EN 0x80 /* [7]; */ -#define ATOM_GPIO_SETTING_PERMODE_MASK 0x7F /* [6:0] */ - -/* definitions for ATOM_COMPONENT_VIDEO_INFO.ucLetterBoxMode */ -/* Line 3 out put 5V. */ -#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_A 0x01 /* represent gpio 3 state for 16:9 */ -#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_B 0x02 /* represent gpio 4 state for 16:9 */ -#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_SHIFT 0x0 - -/* Line 3 out put 2.2V */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_A 0x04 /* represent gpio 3 state for 4:3 Letter box */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_B 0x08 /* represent gpio 4 state for 4:3 Letter box */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_SHIFT 0x2 - -/* Line 3 out put 0V */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_A 0x10 /* represent gpio 3 state for 4:3 */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_B 0x20 /* represent gpio 4 state for 4:3 */ -#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_SHIFT 0x4 - -#define ATOM_CV_LINE3_ASPECTRATIO_MASK 0x3F /* bit [5:0] */ - -#define ATOM_CV_LINE3_ASPECTRATIO_EXIST 0x80 /* bit 7 */ - -/* GPIO bit index in gpio setting per mode value, also represend the block no. in gpio blocks. */ -#define ATOM_GPIO_INDEX_LINE3_ASPECRATIO_GPIO_A 3 /* bit 3 in uc480i/uc480p/uc720p/uc1080i, which represend the default gpio bit setting for the mode. */ -#define ATOM_GPIO_INDEX_LINE3_ASPECRATIO_GPIO_B 4 /* bit 4 in uc480i/uc480p/uc720p/uc1080i, which represend the default gpio bit setting for the mode. */ - -typedef struct _ATOM_COMPONENT_VIDEO_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usMask_PinRegisterIndex; - USHORT usEN_PinRegisterIndex; - USHORT usY_PinRegisterIndex; - USHORT usA_PinRegisterIndex; - UCHAR ucBitShift; - UCHAR ucPinActiveState; /* ucPinActiveState: Bit0=1 active high, =0 active low */ - ATOM_DTD_FORMAT sReserved; /* must be zeroed out */ - UCHAR ucMiscInfo; - UCHAR uc480i; - UCHAR uc480p; - UCHAR uc720p; - UCHAR uc1080i; - UCHAR ucLetterBoxMode; - UCHAR ucReserved[3]; - UCHAR ucNumOfWbGpioBlocks; /* For Component video D-Connector support. If zere, NTSC type connector */ - ATOM_GPIO_INFO aWbGpioStateBlock[MAX_SUPPORTED_CV_STANDARDS]; - ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_CV_STANDARDS]; -} ATOM_COMPONENT_VIDEO_INFO; - -/* ucTableFormatRevision=2 */ -/* ucTableContentRevision=1 */ -typedef struct _ATOM_COMPONENT_VIDEO_INFO_V21 { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucMiscInfo; - UCHAR uc480i; - UCHAR uc480p; - UCHAR uc720p; - UCHAR uc1080i; - UCHAR ucReserved; - UCHAR ucLetterBoxMode; - UCHAR ucNumOfWbGpioBlocks; /* For Component video D-Connector support. If zere, NTSC type connector */ - ATOM_GPIO_INFO aWbGpioStateBlock[MAX_SUPPORTED_CV_STANDARDS]; - ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_CV_STANDARDS]; -} ATOM_COMPONENT_VIDEO_INFO_V21; +// definitions for ATOM_COMPONENT_VIDEO_INFO.uc480i/uc480p/uc720p/uc1080i +#define ATOM_GPIO_DEFAULT_MODE_EN 0x80 //[7]; +#define ATOM_GPIO_SETTING_PERMODE_MASK 0x7F //[6:0] + +// definitions for ATOM_COMPONENT_VIDEO_INFO.ucLetterBoxMode +//Line 3 out put 5V. +#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_A 0x01 //represent gpio 3 state for 16:9 +#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_B 0x02 //represent gpio 4 state for 16:9 +#define ATOM_CV_LINE3_ASPECTRATIO_16_9_GPIO_SHIFT 0x0 + +//Line 3 out put 2.2V +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_A 0x04 //represent gpio 3 state for 4:3 Letter box +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_B 0x08 //represent gpio 4 state for 4:3 Letter box +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_LETBOX_GPIO_SHIFT 0x2 + +//Line 3 out put 0V +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_A 0x10 //represent gpio 3 state for 4:3 +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_B 0x20 //represent gpio 4 state for 4:3 +#define ATOM_CV_LINE3_ASPECTRATIO_4_3_GPIO_SHIFT 0x4 + +#define ATOM_CV_LINE3_ASPECTRATIO_MASK 0x3F // bit [5:0] + +#define ATOM_CV_LINE3_ASPECTRATIO_EXIST 0x80 //bit 7 + +//GPIO bit index in gpio setting per mode value, also represend the block no. in gpio blocks. +#define ATOM_GPIO_INDEX_LINE3_ASPECRATIO_GPIO_A 3 //bit 3 in uc480i/uc480p/uc720p/uc1080i, which represend the default gpio bit setting for the mode. +#define ATOM_GPIO_INDEX_LINE3_ASPECRATIO_GPIO_B 4 //bit 4 in uc480i/uc480p/uc720p/uc1080i, which represend the default gpio bit setting for the mode. + + +typedef struct _ATOM_COMPONENT_VIDEO_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usMask_PinRegisterIndex; + USHORT usEN_PinRegisterIndex; + USHORT usY_PinRegisterIndex; + USHORT usA_PinRegisterIndex; + UCHAR ucBitShift; + UCHAR ucPinActiveState; //ucPinActiveState: Bit0=1 active high, =0 active low + ATOM_DTD_FORMAT sReserved; // must be zeroed out + UCHAR ucMiscInfo; + UCHAR uc480i; + UCHAR uc480p; + UCHAR uc720p; + UCHAR uc1080i; + UCHAR ucLetterBoxMode; + UCHAR ucReserved[3]; + UCHAR ucNumOfWbGpioBlocks; //For Component video D-Connector support. If zere, NTSC type connector + ATOM_GPIO_INFO aWbGpioStateBlock[MAX_SUPPORTED_CV_STANDARDS]; + ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_CV_STANDARDS]; +}ATOM_COMPONENT_VIDEO_INFO; + +//ucTableFormatRevision=2 +//ucTableContentRevision=1 +typedef struct _ATOM_COMPONENT_VIDEO_INFO_V21 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucMiscInfo; + UCHAR uc480i; + UCHAR uc480p; + UCHAR uc720p; + UCHAR uc1080i; + UCHAR ucReserved; + UCHAR ucLetterBoxMode; + UCHAR ucNumOfWbGpioBlocks; //For Component video D-Connector support. If zere, NTSC type connector + ATOM_GPIO_INFO aWbGpioStateBlock[MAX_SUPPORTED_CV_STANDARDS]; + ATOM_DTD_FORMAT aModeTimings[MAX_SUPPORTED_CV_STANDARDS]; +}ATOM_COMPONENT_VIDEO_INFO_V21; #define ATOM_COMPONENT_VIDEO_INFO_LAST ATOM_COMPONENT_VIDEO_INFO_V21 -/****************************************************************************/ -/* Structure used in object_InfoTable */ -/****************************************************************************/ -typedef struct _ATOM_OBJECT_HEADER { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usDeviceSupport; - USHORT usConnectorObjectTableOffset; - USHORT usRouterObjectTableOffset; - USHORT usEncoderObjectTableOffset; - USHORT usProtectionObjectTableOffset; /* only available when Protection block is independent. */ - USHORT usDisplayPathTableOffset; -} ATOM_OBJECT_HEADER; - -typedef struct _ATOM_DISPLAY_OBJECT_PATH { - USHORT usDeviceTag; /* supported device */ - USHORT usSize; /* the size of ATOM_DISPLAY_OBJECT_PATH */ - USHORT usConnObjectId; /* Connector Object ID */ - USHORT usGPUObjectId; /* GPU ID */ - USHORT usGraphicObjIds[1]; /* 1st Encoder Obj source from GPU to last Graphic Obj destinate to connector. */ -} ATOM_DISPLAY_OBJECT_PATH; - -typedef struct _ATOM_DISPLAY_OBJECT_PATH_TABLE { - UCHAR ucNumOfDispPath; - UCHAR ucVersion; - UCHAR ucPadding[2]; - ATOM_DISPLAY_OBJECT_PATH asDispPath[1]; -} ATOM_DISPLAY_OBJECT_PATH_TABLE; - -typedef struct _ATOM_OBJECT /* each object has this structure */ -{ - USHORT usObjectID; - USHORT usSrcDstTableOffset; - USHORT usRecordOffset; /* this pointing to a bunch of records defined below */ - USHORT usReserved; -} ATOM_OBJECT; - -typedef struct _ATOM_OBJECT_TABLE /* Above 4 object table offset pointing to a bunch of objects all have this structure */ -{ - UCHAR ucNumberOfObjects; - UCHAR ucPadding[3]; - ATOM_OBJECT asObjects[1]; -} ATOM_OBJECT_TABLE; - -typedef struct _ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT /* usSrcDstTableOffset pointing to this structure */ -{ - UCHAR ucNumberOfSrc; - USHORT usSrcObjectID[1]; - UCHAR ucNumberOfDst; - USHORT usDstObjectID[1]; -} ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT; - -/* Related definitions, all records are differnt but they have a commond header */ -typedef struct _ATOM_COMMON_RECORD_HEADER { - UCHAR ucRecordType; /* An emun to indicate the record type */ - UCHAR ucRecordSize; /* The size of the whole record in byte */ -} ATOM_COMMON_RECORD_HEADER; - -#define ATOM_I2C_RECORD_TYPE 1 +/****************************************************************************/ +// Structure used in object_InfoTable +/****************************************************************************/ +typedef struct _ATOM_OBJECT_HEADER +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usDeviceSupport; + USHORT usConnectorObjectTableOffset; + USHORT usRouterObjectTableOffset; + USHORT usEncoderObjectTableOffset; + USHORT usProtectionObjectTableOffset; //only available when Protection block is independent. + USHORT usDisplayPathTableOffset; +}ATOM_OBJECT_HEADER; + +typedef struct _ATOM_OBJECT_HEADER_V3 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usDeviceSupport; + USHORT usConnectorObjectTableOffset; + USHORT usRouterObjectTableOffset; + USHORT usEncoderObjectTableOffset; + USHORT usProtectionObjectTableOffset; //only available when Protection block is independent. + USHORT usDisplayPathTableOffset; + USHORT usMiscObjectTableOffset; +}ATOM_OBJECT_HEADER_V3; + +typedef struct _ATOM_DISPLAY_OBJECT_PATH +{ + USHORT usDeviceTag; //supported device + USHORT usSize; //the size of ATOM_DISPLAY_OBJECT_PATH + USHORT usConnObjectId; //Connector Object ID + USHORT usGPUObjectId; //GPU ID + USHORT usGraphicObjIds[1]; //1st Encoder Obj source from GPU to last Graphic Obj destinate to connector. +}ATOM_DISPLAY_OBJECT_PATH; + +typedef struct _ATOM_DISPLAY_OBJECT_PATH_TABLE +{ + UCHAR ucNumOfDispPath; + UCHAR ucVersion; + UCHAR ucPadding[2]; + ATOM_DISPLAY_OBJECT_PATH asDispPath[1]; +}ATOM_DISPLAY_OBJECT_PATH_TABLE; + + +typedef struct _ATOM_OBJECT //each object has this structure +{ + USHORT usObjectID; + USHORT usSrcDstTableOffset; + USHORT usRecordOffset; //this pointing to a bunch of records defined below + USHORT usReserved; +}ATOM_OBJECT; + +typedef struct _ATOM_OBJECT_TABLE //Above 4 object table offset pointing to a bunch of objects all have this structure +{ + UCHAR ucNumberOfObjects; + UCHAR ucPadding[3]; + ATOM_OBJECT asObjects[1]; +}ATOM_OBJECT_TABLE; + +typedef struct _ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT //usSrcDstTableOffset pointing to this structure +{ + UCHAR ucNumberOfSrc; + USHORT usSrcObjectID[1]; + UCHAR ucNumberOfDst; + USHORT usDstObjectID[1]; +}ATOM_SRC_DST_TABLE_FOR_ONE_OBJECT; + + +//Two definitions below are for OPM on MXM module designs + +#define EXT_HPDPIN_LUTINDEX_0 0 +#define EXT_HPDPIN_LUTINDEX_1 1 +#define EXT_HPDPIN_LUTINDEX_2 2 +#define EXT_HPDPIN_LUTINDEX_3 3 +#define EXT_HPDPIN_LUTINDEX_4 4 +#define EXT_HPDPIN_LUTINDEX_5 5 +#define EXT_HPDPIN_LUTINDEX_6 6 +#define EXT_HPDPIN_LUTINDEX_7 7 +#define MAX_NUMBER_OF_EXT_HPDPIN_LUT_ENTRIES (EXT_HPDPIN_LUTINDEX_7+1) + +#define EXT_AUXDDC_LUTINDEX_0 0 +#define EXT_AUXDDC_LUTINDEX_1 1 +#define EXT_AUXDDC_LUTINDEX_2 2 +#define EXT_AUXDDC_LUTINDEX_3 3 +#define EXT_AUXDDC_LUTINDEX_4 4 +#define EXT_AUXDDC_LUTINDEX_5 5 +#define EXT_AUXDDC_LUTINDEX_6 6 +#define EXT_AUXDDC_LUTINDEX_7 7 +#define MAX_NUMBER_OF_EXT_AUXDDC_LUT_ENTRIES (EXT_AUXDDC_LUTINDEX_7+1) + +typedef struct _EXT_DISPLAY_PATH +{ + USHORT usDeviceTag; //A bit vector to show what devices are supported + USHORT usDeviceACPIEnum; //16bit device ACPI id. + USHORT usDeviceConnector; //A physical connector for displays to plug in, using object connector definitions + UCHAR ucExtAUXDDCLutIndex; //An index into external AUX/DDC channel LUT + UCHAR ucExtHPDPINLutIndex; //An index into external HPD pin LUT + USHORT usExtEncoderObjId; //external encoder object id + USHORT usReserved[3]; +}EXT_DISPLAY_PATH; + +#define NUMBER_OF_UCHAR_FOR_GUID 16 +#define MAX_NUMBER_OF_EXT_DISPLAY_PATH 7 + +typedef struct _ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucGuid [NUMBER_OF_UCHAR_FOR_GUID]; // a GUID is a 16 byte long string + EXT_DISPLAY_PATH sPath[MAX_NUMBER_OF_EXT_DISPLAY_PATH]; // total of fixed 7 entries. + UCHAR ucChecksum; // a simple Checksum of the sum of whole structure equal to 0x0. + UCHAR Reserved [7]; // for potential expansion +}ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO; + +//Related definitions, all records are differnt but they have a commond header +typedef struct _ATOM_COMMON_RECORD_HEADER +{ + UCHAR ucRecordType; //An emun to indicate the record type + UCHAR ucRecordSize; //The size of the whole record in byte +}ATOM_COMMON_RECORD_HEADER; + + +#define ATOM_I2C_RECORD_TYPE 1 #define ATOM_HPD_INT_RECORD_TYPE 2 #define ATOM_OUTPUT_PROTECTION_RECORD_TYPE 3 #define ATOM_CONNECTOR_DEVICE_TAG_RECORD_TYPE 4 -#define ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD_TYPE 5 /* Obsolete, switch to use GPIO_CNTL_RECORD_TYPE */ -#define ATOM_ENCODER_FPGA_CONTROL_RECORD_TYPE 6 /* Obsolete, switch to use GPIO_CNTL_RECORD_TYPE */ +#define ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD_TYPE 5 //Obsolete, switch to use GPIO_CNTL_RECORD_TYPE +#define ATOM_ENCODER_FPGA_CONTROL_RECORD_TYPE 6 //Obsolete, switch to use GPIO_CNTL_RECORD_TYPE #define ATOM_CONNECTOR_CVTV_SHARE_DIN_RECORD_TYPE 7 -#define ATOM_JTAG_RECORD_TYPE 8 /* Obsolete, switch to use GPIO_CNTL_RECORD_TYPE */ +#define ATOM_JTAG_RECORD_TYPE 8 //Obsolete, switch to use GPIO_CNTL_RECORD_TYPE #define ATOM_OBJECT_GPIO_CNTL_RECORD_TYPE 9 #define ATOM_ENCODER_DVO_CF_RECORD_TYPE 10 #define ATOM_CONNECTOR_CF_RECORD_TYPE 11 #define ATOM_CONNECTOR_HARDCODE_DTD_RECORD_TYPE 12 #define ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD_TYPE 13 -#define ATOM_ROUTER_DDC_PATH_SELECT_RECORD_TYPE 14 -#define ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD_TYPE 15 - -/* Must be updated when new record type is added,equal to that record definition! */ -#define ATOM_MAX_OBJECT_RECORD_NUMBER ATOM_CONNECTOR_CF_RECORD_TYPE - -typedef struct _ATOM_I2C_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - ATOM_I2C_ID_CONFIG sucI2cId; - UCHAR ucI2CAddr; /* The slave address, it's 0 when the record is attached to connector for DDC */ -} ATOM_I2C_RECORD; - -typedef struct _ATOM_HPD_INT_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucHPDIntGPIOID; /* Corresponding block in GPIO_PIN_INFO table gives the pin info */ - UCHAR ucPlugged_PinState; -} ATOM_HPD_INT_RECORD; - -typedef struct _ATOM_OUTPUT_PROTECTION_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucProtectionFlag; - UCHAR ucReserved; -} ATOM_OUTPUT_PROTECTION_RECORD; - -typedef struct _ATOM_CONNECTOR_DEVICE_TAG { - ULONG ulACPIDeviceEnum; /* Reserved for now */ - USHORT usDeviceID; /* This Id is same as "ATOM_DEVICE_XXX_SUPPORT" */ - USHORT usPadding; -} ATOM_CONNECTOR_DEVICE_TAG; - -typedef struct _ATOM_CONNECTOR_DEVICE_TAG_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucNumberOfDevice; - UCHAR ucReserved; - ATOM_CONNECTOR_DEVICE_TAG asDeviceTag[1]; /* This Id is same as "ATOM_DEVICE_XXX_SUPPORT", 1 is only for allocation */ -} ATOM_CONNECTOR_DEVICE_TAG_RECORD; - -typedef struct _ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucConfigGPIOID; - UCHAR ucConfigGPIOState; /* Set to 1 when it's active high to enable external flow in */ - UCHAR ucFlowinGPIPID; - UCHAR ucExtInGPIPID; -} ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD; - -typedef struct _ATOM_ENCODER_FPGA_CONTROL_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucCTL1GPIO_ID; - UCHAR ucCTL1GPIOState; /* Set to 1 when it's active high */ - UCHAR ucCTL2GPIO_ID; - UCHAR ucCTL2GPIOState; /* Set to 1 when it's active high */ - UCHAR ucCTL3GPIO_ID; - UCHAR ucCTL3GPIOState; /* Set to 1 when it's active high */ - UCHAR ucCTLFPGA_IN_ID; - UCHAR ucPadding[3]; -} ATOM_ENCODER_FPGA_CONTROL_RECORD; - -typedef struct _ATOM_CONNECTOR_CVTV_SHARE_DIN_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucGPIOID; /* Corresponding block in GPIO_PIN_INFO table gives the pin info */ - UCHAR ucTVActiveState; /* Indicating when the pin==0 or 1 when TV is connected */ -} ATOM_CONNECTOR_CVTV_SHARE_DIN_RECORD; - -typedef struct _ATOM_JTAG_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucTMSGPIO_ID; - UCHAR ucTMSGPIOState; /* Set to 1 when it's active high */ - UCHAR ucTCKGPIO_ID; - UCHAR ucTCKGPIOState; /* Set to 1 when it's active high */ - UCHAR ucTDOGPIO_ID; - UCHAR ucTDOGPIOState; /* Set to 1 when it's active high */ - UCHAR ucTDIGPIO_ID; - UCHAR ucTDIGPIOState; /* Set to 1 when it's active high */ - UCHAR ucPadding[2]; -} ATOM_JTAG_RECORD; - -/* The following generic object gpio pin control record type will replace JTAG_RECORD/FPGA_CONTROL_RECORD/DVI_EXT_INPUT_RECORD above gradually */ -typedef struct _ATOM_GPIO_PIN_CONTROL_PAIR { - UCHAR ucGPIOID; /* GPIO_ID, find the corresponding ID in GPIO_LUT table */ - UCHAR ucGPIO_PinState; /* Pin state showing how to set-up the pin */ -} ATOM_GPIO_PIN_CONTROL_PAIR; - -typedef struct _ATOM_OBJECT_GPIO_CNTL_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucFlags; /* Future expnadibility */ - UCHAR ucNumberOfPins; /* Number of GPIO pins used to control the object */ - ATOM_GPIO_PIN_CONTROL_PAIR asGpio[1]; /* the real gpio pin pair determined by number of pins ucNumberOfPins */ -} ATOM_OBJECT_GPIO_CNTL_RECORD; - -/* Definitions for GPIO pin state */ +#define ATOM_ROUTER_DDC_PATH_SELECT_RECORD_TYPE 14 +#define ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD_TYPE 15 +#define ATOM_CONNECTOR_HPDPIN_LUT_RECORD_TYPE 16 //This is for the case when connectors are not known to object table +#define ATOM_CONNECTOR_AUXDDC_LUT_RECORD_TYPE 17 //This is for the case when connectors are not known to object table +#define ATOM_OBJECT_LINK_RECORD_TYPE 18 //Once this record is present under one object, it indicats the oobject is linked to another obj described by the record +#define ATOM_CONNECTOR_REMOTE_CAP_RECORD_TYPE 19 + + +//Must be updated when new record type is added,equal to that record definition! +#define ATOM_MAX_OBJECT_RECORD_NUMBER ATOM_CONNECTOR_REMOTE_CAP_RECORD_TYPE + +typedef struct _ATOM_I2C_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + ATOM_I2C_ID_CONFIG sucI2cId; + UCHAR ucI2CAddr; //The slave address, it's 0 when the record is attached to connector for DDC +}ATOM_I2C_RECORD; + +typedef struct _ATOM_HPD_INT_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucHPDIntGPIOID; //Corresponding block in GPIO_PIN_INFO table gives the pin info + UCHAR ucPlugged_PinState; +}ATOM_HPD_INT_RECORD; + + +typedef struct _ATOM_OUTPUT_PROTECTION_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucProtectionFlag; + UCHAR ucReserved; +}ATOM_OUTPUT_PROTECTION_RECORD; + +typedef struct _ATOM_CONNECTOR_DEVICE_TAG +{ + ULONG ulACPIDeviceEnum; //Reserved for now + USHORT usDeviceID; //This Id is same as "ATOM_DEVICE_XXX_SUPPORT" + USHORT usPadding; +}ATOM_CONNECTOR_DEVICE_TAG; + +typedef struct _ATOM_CONNECTOR_DEVICE_TAG_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucNumberOfDevice; + UCHAR ucReserved; + ATOM_CONNECTOR_DEVICE_TAG asDeviceTag[1]; //This Id is same as "ATOM_DEVICE_XXX_SUPPORT", 1 is only for allocation +}ATOM_CONNECTOR_DEVICE_TAG_RECORD; + + +typedef struct _ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucConfigGPIOID; + UCHAR ucConfigGPIOState; //Set to 1 when it's active high to enable external flow in + UCHAR ucFlowinGPIPID; + UCHAR ucExtInGPIPID; +}ATOM_CONNECTOR_DVI_EXT_INPUT_RECORD; + +typedef struct _ATOM_ENCODER_FPGA_CONTROL_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucCTL1GPIO_ID; + UCHAR ucCTL1GPIOState; //Set to 1 when it's active high + UCHAR ucCTL2GPIO_ID; + UCHAR ucCTL2GPIOState; //Set to 1 when it's active high + UCHAR ucCTL3GPIO_ID; + UCHAR ucCTL3GPIOState; //Set to 1 when it's active high + UCHAR ucCTLFPGA_IN_ID; + UCHAR ucPadding[3]; +}ATOM_ENCODER_FPGA_CONTROL_RECORD; + +typedef struct _ATOM_CONNECTOR_CVTV_SHARE_DIN_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucGPIOID; //Corresponding block in GPIO_PIN_INFO table gives the pin info + UCHAR ucTVActiveState; //Indicating when the pin==0 or 1 when TV is connected +}ATOM_CONNECTOR_CVTV_SHARE_DIN_RECORD; + +typedef struct _ATOM_JTAG_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucTMSGPIO_ID; + UCHAR ucTMSGPIOState; //Set to 1 when it's active high + UCHAR ucTCKGPIO_ID; + UCHAR ucTCKGPIOState; //Set to 1 when it's active high + UCHAR ucTDOGPIO_ID; + UCHAR ucTDOGPIOState; //Set to 1 when it's active high + UCHAR ucTDIGPIO_ID; + UCHAR ucTDIGPIOState; //Set to 1 when it's active high + UCHAR ucPadding[2]; +}ATOM_JTAG_RECORD; + + +//The following generic object gpio pin control record type will replace JTAG_RECORD/FPGA_CONTROL_RECORD/DVI_EXT_INPUT_RECORD above gradually +typedef struct _ATOM_GPIO_PIN_CONTROL_PAIR +{ + UCHAR ucGPIOID; // GPIO_ID, find the corresponding ID in GPIO_LUT table + UCHAR ucGPIO_PinState; // Pin state showing how to set-up the pin +}ATOM_GPIO_PIN_CONTROL_PAIR; + +typedef struct _ATOM_OBJECT_GPIO_CNTL_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucFlags; // Future expnadibility + UCHAR ucNumberOfPins; // Number of GPIO pins used to control the object + ATOM_GPIO_PIN_CONTROL_PAIR asGpio[1]; // the real gpio pin pair determined by number of pins ucNumberOfPins +}ATOM_OBJECT_GPIO_CNTL_RECORD; + +//Definitions for GPIO pin state #define GPIO_PIN_TYPE_INPUT 0x00 #define GPIO_PIN_TYPE_OUTPUT 0x10 #define GPIO_PIN_TYPE_HW_CONTROL 0x20 -/* For GPIO_PIN_TYPE_OUTPUT the following is defined */ +//For GPIO_PIN_TYPE_OUTPUT the following is defined #define GPIO_PIN_OUTPUT_STATE_MASK 0x01 #define GPIO_PIN_OUTPUT_STATE_SHIFT 0 #define GPIO_PIN_STATE_ACTIVE_LOW 0x0 #define GPIO_PIN_STATE_ACTIVE_HIGH 0x1 -typedef struct _ATOM_ENCODER_DVO_CF_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - ULONG ulStrengthControl; /* DVOA strength control for CF */ - UCHAR ucPadding[2]; -} ATOM_ENCODER_DVO_CF_RECORD; +// Indexes to GPIO array in GLSync record +#define ATOM_GPIO_INDEX_GLSYNC_REFCLK 0 +#define ATOM_GPIO_INDEX_GLSYNC_HSYNC 1 +#define ATOM_GPIO_INDEX_GLSYNC_VSYNC 2 +#define ATOM_GPIO_INDEX_GLSYNC_SWAP_REQ 3 +#define ATOM_GPIO_INDEX_GLSYNC_SWAP_GNT 4 +#define ATOM_GPIO_INDEX_GLSYNC_INTERRUPT 5 +#define ATOM_GPIO_INDEX_GLSYNC_V_RESET 6 +#define ATOM_GPIO_INDEX_GLSYNC_MAX 7 + +typedef struct _ATOM_ENCODER_DVO_CF_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + ULONG ulStrengthControl; // DVOA strength control for CF + UCHAR ucPadding[2]; +}ATOM_ENCODER_DVO_CF_RECORD; -/* value for ATOM_CONNECTOR_CF_RECORD.ucConnectedDvoBundle */ +// value for ATOM_CONNECTOR_CF_RECORD.ucConnectedDvoBundle #define ATOM_CONNECTOR_CF_RECORD_CONNECTED_UPPER12BITBUNDLEA 1 #define ATOM_CONNECTOR_CF_RECORD_CONNECTED_LOWER12BITBUNDLEB 2 -typedef struct _ATOM_CONNECTOR_CF_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - USHORT usMaxPixClk; - UCHAR ucFlowCntlGpioId; - UCHAR ucSwapCntlGpioId; - UCHAR ucConnectedDvoBundle; - UCHAR ucPadding; -} ATOM_CONNECTOR_CF_RECORD; - -typedef struct _ATOM_CONNECTOR_HARDCODE_DTD_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - ATOM_DTD_FORMAT asTiming; -} ATOM_CONNECTOR_HARDCODE_DTD_RECORD; - -typedef struct _ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; /* ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD_TYPE */ - UCHAR ucSubConnectorType; /* CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D|X_ID_DUAL_LINK_DVI_D|HDMI_TYPE_A */ - UCHAR ucReserved; -} ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD; - -typedef struct _ATOM_ROUTER_DDC_PATH_SELECT_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucMuxType; /* decide the number of ucMuxState, =0, no pin state, =1: single state with complement, >1: multiple state */ - UCHAR ucMuxControlPin; - UCHAR ucMuxState[2]; /* for alligment purpose */ -} ATOM_ROUTER_DDC_PATH_SELECT_RECORD; - -typedef struct _ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD { - ATOM_COMMON_RECORD_HEADER sheader; - UCHAR ucMuxType; - UCHAR ucMuxControlPin; - UCHAR ucMuxState[2]; /* for alligment purpose */ -} ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD; - -/* define ucMuxType */ +typedef struct _ATOM_CONNECTOR_CF_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + USHORT usMaxPixClk; + UCHAR ucFlowCntlGpioId; + UCHAR ucSwapCntlGpioId; + UCHAR ucConnectedDvoBundle; + UCHAR ucPadding; +}ATOM_CONNECTOR_CF_RECORD; + +typedef struct _ATOM_CONNECTOR_HARDCODE_DTD_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + ATOM_DTD_FORMAT asTiming; +}ATOM_CONNECTOR_HARDCODE_DTD_RECORD; + +typedef struct _ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; //ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD_TYPE + UCHAR ucSubConnectorType; //CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D|X_ID_DUAL_LINK_DVI_D|HDMI_TYPE_A + UCHAR ucReserved; +}ATOM_CONNECTOR_PCIE_SUBCONNECTOR_RECORD; + + +typedef struct _ATOM_ROUTER_DDC_PATH_SELECT_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucMuxType; //decide the number of ucMuxState, =0, no pin state, =1: single state with complement, >1: multiple state + UCHAR ucMuxControlPin; + UCHAR ucMuxState[2]; //for alligment purpose +}ATOM_ROUTER_DDC_PATH_SELECT_RECORD; + +typedef struct _ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucMuxType; + UCHAR ucMuxControlPin; + UCHAR ucMuxState[2]; //for alligment purpose +}ATOM_ROUTER_DATA_CLOCK_PATH_SELECT_RECORD; + +// define ucMuxType #define ATOM_ROUTER_MUX_PIN_STATE_MASK 0x0f #define ATOM_ROUTER_MUX_PIN_SINGLE_STATE_COMPLEMENT 0x01 -/****************************************************************************/ -/* ASIC voltage data table */ -/****************************************************************************/ -typedef struct _ATOM_VOLTAGE_INFO_HEADER { - USHORT usVDDCBaseLevel; /* In number of 50mv unit */ - USHORT usReserved; /* For possible extension table offset */ - UCHAR ucNumOfVoltageEntries; - UCHAR ucBytesPerVoltageEntry; - UCHAR ucVoltageStep; /* Indicating in how many mv increament is one step, 0.5mv unit */ - UCHAR ucDefaultVoltageEntry; - UCHAR ucVoltageControlI2cLine; - UCHAR ucVoltageControlAddress; - UCHAR ucVoltageControlOffset; -} ATOM_VOLTAGE_INFO_HEADER; - -typedef struct _ATOM_VOLTAGE_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_VOLTAGE_INFO_HEADER viHeader; - UCHAR ucVoltageEntries[64]; /* 64 is for allocation, the actual number of entry is present at ucNumOfVoltageEntries*ucBytesPerVoltageEntry */ -} ATOM_VOLTAGE_INFO; - -typedef struct _ATOM_VOLTAGE_FORMULA { - USHORT usVoltageBaseLevel; /* In number of 1mv unit */ - USHORT usVoltageStep; /* Indicating in how many mv increament is one step, 1mv unit */ - UCHAR ucNumOfVoltageEntries; /* Number of Voltage Entry, which indicate max Voltage */ - UCHAR ucFlag; /* bit0=0 :step is 1mv =1 0.5mv */ - UCHAR ucBaseVID; /* if there is no lookup table, VID= BaseVID + ( Vol - BaseLevle ) /VoltageStep */ - UCHAR ucReserved; - UCHAR ucVIDAdjustEntries[32]; /* 32 is for allocation, the actual number of entry is present at ucNumOfVoltageEntries */ -} ATOM_VOLTAGE_FORMULA; - -typedef struct _ATOM_VOLTAGE_CONTROL { - UCHAR ucVoltageControlId; /* Indicate it is controlled by I2C or GPIO or HW state machine */ - UCHAR ucVoltageControlI2cLine; - UCHAR ucVoltageControlAddress; - UCHAR ucVoltageControlOffset; - USHORT usGpioPin_AIndex; /* GPIO_PAD register index */ - UCHAR ucGpioPinBitShift[9]; /* at most 8 pin support 255 VIDs, termintate with 0xff */ - UCHAR ucReserved; -} ATOM_VOLTAGE_CONTROL; - -/* Define ucVoltageControlId */ +typedef struct _ATOM_CONNECTOR_HPDPIN_LUT_RECORD //record for ATOM_CONNECTOR_HPDPIN_LUT_RECORD_TYPE +{ + ATOM_COMMON_RECORD_HEADER sheader; + UCHAR ucHPDPINMap[MAX_NUMBER_OF_EXT_HPDPIN_LUT_ENTRIES]; //An fixed size array which maps external pins to internal GPIO_PIN_INFO table +}ATOM_CONNECTOR_HPDPIN_LUT_RECORD; + +typedef struct _ATOM_CONNECTOR_AUXDDC_LUT_RECORD //record for ATOM_CONNECTOR_AUXDDC_LUT_RECORD_TYPE +{ + ATOM_COMMON_RECORD_HEADER sheader; + ATOM_I2C_ID_CONFIG ucAUXDDCMap[MAX_NUMBER_OF_EXT_AUXDDC_LUT_ENTRIES]; //An fixed size array which maps external pins to internal DDC ID +}ATOM_CONNECTOR_AUXDDC_LUT_RECORD; + +typedef struct _ATOM_OBJECT_LINK_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + USHORT usObjectID; //could be connector, encorder or other object in object.h +}ATOM_OBJECT_LINK_RECORD; + +typedef struct _ATOM_CONNECTOR_REMOTE_CAP_RECORD +{ + ATOM_COMMON_RECORD_HEADER sheader; + USHORT usReserved; +}ATOM_CONNECTOR_REMOTE_CAP_RECORD; + +/****************************************************************************/ +// ASIC voltage data table +/****************************************************************************/ +typedef struct _ATOM_VOLTAGE_INFO_HEADER +{ + USHORT usVDDCBaseLevel; //In number of 50mv unit + USHORT usReserved; //For possible extension table offset + UCHAR ucNumOfVoltageEntries; + UCHAR ucBytesPerVoltageEntry; + UCHAR ucVoltageStep; //Indicating in how many mv increament is one step, 0.5mv unit + UCHAR ucDefaultVoltageEntry; + UCHAR ucVoltageControlI2cLine; + UCHAR ucVoltageControlAddress; + UCHAR ucVoltageControlOffset; +}ATOM_VOLTAGE_INFO_HEADER; + +typedef struct _ATOM_VOLTAGE_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_VOLTAGE_INFO_HEADER viHeader; + UCHAR ucVoltageEntries[64]; //64 is for allocation, the actual number of entry is present at ucNumOfVoltageEntries*ucBytesPerVoltageEntry +}ATOM_VOLTAGE_INFO; + + +typedef struct _ATOM_VOLTAGE_FORMULA +{ + USHORT usVoltageBaseLevel; // In number of 1mv unit + USHORT usVoltageStep; // Indicating in how many mv increament is one step, 1mv unit + UCHAR ucNumOfVoltageEntries; // Number of Voltage Entry, which indicate max Voltage + UCHAR ucFlag; // bit0=0 :step is 1mv =1 0.5mv + UCHAR ucBaseVID; // if there is no lookup table, VID= BaseVID + ( Vol - BaseLevle ) /VoltageStep + UCHAR ucReserved; + UCHAR ucVIDAdjustEntries[32]; // 32 is for allocation, the actual number of entry is present at ucNumOfVoltageEntries +}ATOM_VOLTAGE_FORMULA; + +typedef struct _VOLTAGE_LUT_ENTRY +{ + USHORT usVoltageCode; // The Voltage ID, either GPIO or I2C code + USHORT usVoltageValue; // The corresponding Voltage Value, in mV +}VOLTAGE_LUT_ENTRY; + +typedef struct _ATOM_VOLTAGE_FORMULA_V2 +{ + UCHAR ucNumOfVoltageEntries; // Number of Voltage Entry, which indicate max Voltage + UCHAR ucReserved[3]; + VOLTAGE_LUT_ENTRY asVIDAdjustEntries[32];// 32 is for allocation, the actual number of entries is in ucNumOfVoltageEntries +}ATOM_VOLTAGE_FORMULA_V2; + +typedef struct _ATOM_VOLTAGE_CONTROL +{ + UCHAR ucVoltageControlId; //Indicate it is controlled by I2C or GPIO or HW state machine + UCHAR ucVoltageControlI2cLine; + UCHAR ucVoltageControlAddress; + UCHAR ucVoltageControlOffset; + USHORT usGpioPin_AIndex; //GPIO_PAD register index + UCHAR ucGpioPinBitShift[9]; //at most 8 pin support 255 VIDs, termintate with 0xff + UCHAR ucReserved; +}ATOM_VOLTAGE_CONTROL; + +// Define ucVoltageControlId #define VOLTAGE_CONTROLLED_BY_HW 0x00 #define VOLTAGE_CONTROLLED_BY_I2C_MASK 0x7F #define VOLTAGE_CONTROLLED_BY_GPIO 0x80 -#define VOLTAGE_CONTROL_ID_LM64 0x01 /* I2C control, used for R5xx Core Voltage */ -#define VOLTAGE_CONTROL_ID_DAC 0x02 /* I2C control, used for R5xx/R6xx MVDDC,MVDDQ or VDDCI */ -#define VOLTAGE_CONTROL_ID_VT116xM 0x03 /* I2C control, used for R6xx Core Voltage */ -#define VOLTAGE_CONTROL_ID_DS4402 0x04 - -typedef struct _ATOM_VOLTAGE_OBJECT { - UCHAR ucVoltageType; /* Indicate Voltage Source: VDDC, MVDDC, MVDDQ or MVDDCI */ - UCHAR ucSize; /* Size of Object */ - ATOM_VOLTAGE_CONTROL asControl; /* describ how to control */ - ATOM_VOLTAGE_FORMULA asFormula; /* Indicate How to convert real Voltage to VID */ -} ATOM_VOLTAGE_OBJECT; - -typedef struct _ATOM_VOLTAGE_OBJECT_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_VOLTAGE_OBJECT asVoltageObj[3]; /* Info for Voltage control */ -} ATOM_VOLTAGE_OBJECT_INFO; - -typedef struct _ATOM_LEAKID_VOLTAGE { - UCHAR ucLeakageId; - UCHAR ucReserved; - USHORT usVoltage; -} ATOM_LEAKID_VOLTAGE; - -typedef struct _ATOM_ASIC_PROFILE_VOLTAGE { - UCHAR ucProfileId; - UCHAR ucReserved; - USHORT usSize; - USHORT usEfuseSpareStartAddr; - USHORT usFuseIndex[8]; /* from LSB to MSB, Max 8bit,end of 0xffff if less than 8 efuse id, */ - ATOM_LEAKID_VOLTAGE asLeakVol[2]; /* Leakid and relatd voltage */ -} ATOM_ASIC_PROFILE_VOLTAGE; - -/* ucProfileId */ -#define ATOM_ASIC_PROFILE_ID_EFUSE_VOLTAGE 1 +#define VOLTAGE_CONTROL_ID_LM64 0x01 //I2C control, used for R5xx Core Voltage +#define VOLTAGE_CONTROL_ID_DAC 0x02 //I2C control, used for R5xx/R6xx MVDDC,MVDDQ or VDDCI +#define VOLTAGE_CONTROL_ID_VT116xM 0x03 //I2C control, used for R6xx Core Voltage +#define VOLTAGE_CONTROL_ID_DS4402 0x04 + +typedef struct _ATOM_VOLTAGE_OBJECT +{ + UCHAR ucVoltageType; //Indicate Voltage Source: VDDC, MVDDC, MVDDQ or MVDDCI + UCHAR ucSize; //Size of Object + ATOM_VOLTAGE_CONTROL asControl; //describ how to control + ATOM_VOLTAGE_FORMULA asFormula; //Indicate How to convert real Voltage to VID +}ATOM_VOLTAGE_OBJECT; + +typedef struct _ATOM_VOLTAGE_OBJECT_V2 +{ + UCHAR ucVoltageType; //Indicate Voltage Source: VDDC, MVDDC, MVDDQ or MVDDCI + UCHAR ucSize; //Size of Object + ATOM_VOLTAGE_CONTROL asControl; //describ how to control + ATOM_VOLTAGE_FORMULA_V2 asFormula; //Indicate How to convert real Voltage to VID +}ATOM_VOLTAGE_OBJECT_V2; + +typedef struct _ATOM_VOLTAGE_OBJECT_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_VOLTAGE_OBJECT asVoltageObj[3]; //Info for Voltage control +}ATOM_VOLTAGE_OBJECT_INFO; + +typedef struct _ATOM_VOLTAGE_OBJECT_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_VOLTAGE_OBJECT_V2 asVoltageObj[3]; //Info for Voltage control +}ATOM_VOLTAGE_OBJECT_INFO_V2; + +typedef struct _ATOM_LEAKID_VOLTAGE +{ + UCHAR ucLeakageId; + UCHAR ucReserved; + USHORT usVoltage; +}ATOM_LEAKID_VOLTAGE; + +typedef struct _ATOM_ASIC_PROFILE_VOLTAGE +{ + UCHAR ucProfileId; + UCHAR ucReserved; + USHORT usSize; + USHORT usEfuseSpareStartAddr; + USHORT usFuseIndex[8]; //from LSB to MSB, Max 8bit,end of 0xffff if less than 8 efuse id, + ATOM_LEAKID_VOLTAGE asLeakVol[2]; //Leakid and relatd voltage +}ATOM_ASIC_PROFILE_VOLTAGE; + +//ucProfileId +#define ATOM_ASIC_PROFILE_ID_EFUSE_VOLTAGE 1 #define ATOM_ASIC_PROFILE_ID_EFUSE_PERFORMANCE_VOLTAGE 1 #define ATOM_ASIC_PROFILE_ID_EFUSE_THERMAL_VOLTAGE 2 -typedef struct _ATOM_ASIC_PROFILING_INFO { - ATOM_COMMON_TABLE_HEADER asHeader; - ATOM_ASIC_PROFILE_VOLTAGE asVoltage; -} ATOM_ASIC_PROFILING_INFO; - -typedef struct _ATOM_POWER_SOURCE_OBJECT { - UCHAR ucPwrSrcId; /* Power source */ - UCHAR ucPwrSensorType; /* GPIO, I2C or none */ - UCHAR ucPwrSensId; /* if GPIO detect, it is GPIO id, if I2C detect, it is I2C id */ - UCHAR ucPwrSensSlaveAddr; /* Slave address if I2C detect */ - UCHAR ucPwrSensRegIndex; /* I2C register Index if I2C detect */ - UCHAR ucPwrSensRegBitMask; /* detect which bit is used if I2C detect */ - UCHAR ucPwrSensActiveState; /* high active or low active */ - UCHAR ucReserve[3]; /* reserve */ - USHORT usSensPwr; /* in unit of watt */ -} ATOM_POWER_SOURCE_OBJECT; - -typedef struct _ATOM_POWER_SOURCE_INFO { - ATOM_COMMON_TABLE_HEADER asHeader; - UCHAR asPwrbehave[16]; - ATOM_POWER_SOURCE_OBJECT asPwrObj[1]; -} ATOM_POWER_SOURCE_INFO; - -/* Define ucPwrSrcId */ +typedef struct _ATOM_ASIC_PROFILING_INFO +{ + ATOM_COMMON_TABLE_HEADER asHeader; + ATOM_ASIC_PROFILE_VOLTAGE asVoltage; +}ATOM_ASIC_PROFILING_INFO; + +typedef struct _ATOM_POWER_SOURCE_OBJECT +{ + UCHAR ucPwrSrcId; // Power source + UCHAR ucPwrSensorType; // GPIO, I2C or none + UCHAR ucPwrSensId; // if GPIO detect, it is GPIO id, if I2C detect, it is I2C id + UCHAR ucPwrSensSlaveAddr; // Slave address if I2C detect + UCHAR ucPwrSensRegIndex; // I2C register Index if I2C detect + UCHAR ucPwrSensRegBitMask; // detect which bit is used if I2C detect + UCHAR ucPwrSensActiveState; // high active or low active + UCHAR ucReserve[3]; // reserve + USHORT usSensPwr; // in unit of watt +}ATOM_POWER_SOURCE_OBJECT; + +typedef struct _ATOM_POWER_SOURCE_INFO +{ + ATOM_COMMON_TABLE_HEADER asHeader; + UCHAR asPwrbehave[16]; + ATOM_POWER_SOURCE_OBJECT asPwrObj[1]; +}ATOM_POWER_SOURCE_INFO; + + +//Define ucPwrSrcId #define POWERSOURCE_PCIE_ID1 0x00 #define POWERSOURCE_6PIN_CONNECTOR_ID1 0x01 #define POWERSOURCE_8PIN_CONNECTOR_ID1 0x02 #define POWERSOURCE_6PIN_CONNECTOR_ID2 0x04 #define POWERSOURCE_8PIN_CONNECTOR_ID2 0x08 -/* define ucPwrSensorId */ +//define ucPwrSensorId #define POWER_SENSOR_ALWAYS 0x00 #define POWER_SENSOR_GPIO 0x01 #define POWER_SENSOR_I2C 0x02 +typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ULONG ulBootUpEngineClock; + ULONG ulDentistVCOFreq; + ULONG ulBootUpUMAClock; + ULONG ulReserved1[8]; + ULONG ulBootUpReqDisplayVector; + ULONG ulOtherDisplayMisc; + ULONG ulGPUCapInfo; + ULONG ulReserved2[3]; + ULONG ulSystemConfig; + ULONG ulCPUCapInfo; + USHORT usMaxNBVoltage; + USHORT usMinNBVoltage; + USHORT usBootUpNBVoltage; + USHORT usExtDispConnInfoOffset; + UCHAR ucHtcTmpLmt; + UCHAR ucTjOffset; + UCHAR ucMemoryType; + UCHAR ucUMAChannelNumber; + ULONG ulCSR_M3_ARB_CNTL_DEFAULT[10]; + ULONG ulCSR_M3_ARB_CNTL_UVD[10]; + ULONG ulCSR_M3_ARB_CNTL_FS3D[10]; + ULONG ulReserved3[42]; + ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO sExtDispConnInfo; +}ATOM_INTEGRATED_SYSTEM_INFO_V6; + +/********************************************************************************************************************** +// ATOM_INTEGRATED_SYSTEM_INFO_V6 Description +//ulBootUpEngineClock: VBIOS bootup Engine clock frequency, in 10kHz unit. +//ulDentistVCOFreq: Dentist VCO clock in 10kHz unit. +//ulBootUpUMAClock: System memory boot up clock frequency in 10Khz unit. +//ulReserved1[8] Reserved by now, must be 0x0. +//ulBootUpReqDisplayVector VBIOS boot up display IDs +// ATOM_DEVICE_CRT1_SUPPORT 0x0001 +// ATOM_DEVICE_CRT2_SUPPORT 0x0010 +// ATOM_DEVICE_DFP1_SUPPORT 0x0008 +// ATOM_DEVICE_DFP6_SUPPORT 0x0040 +// ATOM_DEVICE_DFP2_SUPPORT 0x0080 +// ATOM_DEVICE_DFP3_SUPPORT 0x0200 +// ATOM_DEVICE_DFP4_SUPPORT 0x0400 +// ATOM_DEVICE_DFP5_SUPPORT 0x0800 +// ATOM_DEVICE_LCD1_SUPPORT 0x0002 +//ulOtherDisplayMisc Other display related flags, not defined yet. +//ulGPUCapInfo TBD +//ulReserved2[3] must be 0x0 for the reserved. +//ulSystemConfig TBD +//ulCPUCapInfo TBD +//usMaxNBVoltage High NB voltage in unit of mv, calculated using current VDDNB (D24F2xDC) and VDDNB offset fuse. +//usMinNBVoltage Low NB voltage in unit of mv, calculated using current VDDNB (D24F2xDC) and VDDNB offset fuse. +//usBootUpNBVoltage Boot up NB voltage in unit of mv. +//ucHtcTmpLmt Bit [22:16] of D24F3x64 Thermal Control (HTC) Register. +//ucTjOffset Bit [28:22] of D24F3xE4 Thermtrip Status Register,may not be needed. +//ucMemoryType [3:0]=1:DDR1;=2:DDR2;=3:DDR3.[7:4] is reserved. +//ucUMAChannelNumber System memory channel numbers. +//usExtDispConnectionInfoOffset ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO offset relative to beginning of this table. +//ulCSR_M3_ARB_CNTL_DEFAULT[10] Arrays with values for CSR M3 arbiter for default +//ulCSR_M3_ARB_CNTL_UVD[10] Arrays with values for CSR M3 arbiter for UVD playback. +//ulCSR_M3_ARB_CNTL_FS3D[10] Arrays with values for CSR M3 arbiter for Full Screen 3D applications. +**********************************************************************************************************************/ + /**************************************************************************/ -/* This portion is only used when ext thermal chip or engine/memory clock SS chip is populated on a design */ -/* Memory SS Info Table */ -/* Define Memory Clock SS chip ID */ +// This portion is only used when ext thermal chip or engine/memory clock SS chip is populated on a design +//Memory SS Info Table +//Define Memory Clock SS chip ID #define ICS91719 1 #define ICS91720 2 -/* Define one structure to inform SW a "block of data" writing to external SS chip via I2C protocol */ -typedef struct _ATOM_I2C_DATA_RECORD { - UCHAR ucNunberOfBytes; /* Indicates how many bytes SW needs to write to the external ASIC for one block, besides to "Start" and "Stop" */ - UCHAR ucI2CData[1]; /* I2C data in bytes, should be less than 16 bytes usually */ -} ATOM_I2C_DATA_RECORD; - -/* Define one structure to inform SW how many blocks of data writing to external SS chip via I2C protocol, in addition to other information */ -typedef struct _ATOM_I2C_DEVICE_SETUP_INFO { - ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; /* I2C line and HW/SW assisted cap. */ - UCHAR ucSSChipID; /* SS chip being used */ - UCHAR ucSSChipSlaveAddr; /* Slave Address to set up this SS chip */ - UCHAR ucNumOfI2CDataRecords; /* number of data block */ - ATOM_I2C_DATA_RECORD asI2CData[1]; -} ATOM_I2C_DEVICE_SETUP_INFO; - -/* ========================================================================================== */ -typedef struct _ATOM_ASIC_MVDD_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_I2C_DEVICE_SETUP_INFO asI2CSetup[1]; -} ATOM_ASIC_MVDD_INFO; - -/* ========================================================================================== */ +//Define one structure to inform SW a "block of data" writing to external SS chip via I2C protocol +typedef struct _ATOM_I2C_DATA_RECORD +{ + UCHAR ucNunberOfBytes; //Indicates how many bytes SW needs to write to the external ASIC for one block, besides to "Start" and "Stop" + UCHAR ucI2CData[1]; //I2C data in bytes, should be less than 16 bytes usually +}ATOM_I2C_DATA_RECORD; + + +//Define one structure to inform SW how many blocks of data writing to external SS chip via I2C protocol, in addition to other information +typedef struct _ATOM_I2C_DEVICE_SETUP_INFO +{ + ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; //I2C line and HW/SW assisted cap. + UCHAR ucSSChipID; //SS chip being used + UCHAR ucSSChipSlaveAddr; //Slave Address to set up this SS chip + UCHAR ucNumOfI2CDataRecords; //number of data block + ATOM_I2C_DATA_RECORD asI2CData[1]; +}ATOM_I2C_DEVICE_SETUP_INFO; + +//========================================================================================== +typedef struct _ATOM_ASIC_MVDD_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_I2C_DEVICE_SETUP_INFO asI2CSetup[1]; +}ATOM_ASIC_MVDD_INFO; + +//========================================================================================== #define ATOM_MCLK_SS_INFO ATOM_ASIC_MVDD_INFO -/* ========================================================================================== */ +//========================================================================================== /**************************************************************************/ -typedef struct _ATOM_ASIC_SS_ASSIGNMENT { - ULONG ulTargetClockRange; /* Clock Out frequence (VCO ), in unit of 10Khz */ - USHORT usSpreadSpectrumPercentage; /* in unit of 0.01% */ - USHORT usSpreadRateInKhz; /* in unit of kHz, modulation freq */ - UCHAR ucClockIndication; /* Indicate which clock source needs SS */ - UCHAR ucSpreadSpectrumMode; /* Bit1=0 Down Spread,=1 Center Spread. */ - UCHAR ucReserved[2]; -} ATOM_ASIC_SS_ASSIGNMENT; - -/* Define ucSpreadSpectrumType */ +typedef struct _ATOM_ASIC_SS_ASSIGNMENT +{ + ULONG ulTargetClockRange; //Clock Out frequence (VCO ), in unit of 10Khz + USHORT usSpreadSpectrumPercentage; //in unit of 0.01% + USHORT usSpreadRateInKhz; //in unit of kHz, modulation freq + UCHAR ucClockIndication; //Indicate which clock source needs SS + UCHAR ucSpreadSpectrumMode; //Bit1=0 Down Spread,=1 Center Spread. + UCHAR ucReserved[2]; +}ATOM_ASIC_SS_ASSIGNMENT; + +//Define ucClockIndication, SW uses the IDs below to search if the SS is requried/enabled on a clock branch/signal type. +//SS is not required or enabled if a match is not found. #define ASIC_INTERNAL_MEMORY_SS 1 #define ASIC_INTERNAL_ENGINE_SS 2 -#define ASIC_INTERNAL_UVD_SS 3 +#define ASIC_INTERNAL_UVD_SS 3 +#define ASIC_INTERNAL_SS_ON_TMDS 4 +#define ASIC_INTERNAL_SS_ON_HDMI 5 +#define ASIC_INTERNAL_SS_ON_LVDS 6 +#define ASIC_INTERNAL_SS_ON_DP 7 +#define ASIC_INTERNAL_SS_ON_DCPLL 8 + +typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V2 +{ + ULONG ulTargetClockRange; //For mem/engine/uvd, Clock Out frequence (VCO ), in unit of 10Khz + //For TMDS/HDMI/LVDS, it is pixel clock , for DP, it is link clock ( 27000 or 16200 ) + USHORT usSpreadSpectrumPercentage; //in unit of 0.01% + USHORT usSpreadRateIn10Hz; //in unit of 10Hz, modulation freq + UCHAR ucClockIndication; //Indicate which clock source needs SS + UCHAR ucSpreadSpectrumMode; //Bit0=0 Down Spread,=1 Center Spread, bit1=0: internal SS bit1=1: external SS + UCHAR ucReserved[2]; +}ATOM_ASIC_SS_ASSIGNMENT_V2; + +//ucSpreadSpectrumMode +//#define ATOM_SS_DOWN_SPREAD_MODE_MASK 0x00000000 +//#define ATOM_SS_DOWN_SPREAD_MODE 0x00000000 +//#define ATOM_SS_CENTRE_SPREAD_MODE_MASK 0x00000001 +//#define ATOM_SS_CENTRE_SPREAD_MODE 0x00000001 +//#define ATOM_INTERNAL_SS_MASK 0x00000000 +//#define ATOM_EXTERNAL_SS_MASK 0x00000002 + +typedef struct _ATOM_ASIC_INTERNAL_SS_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_ASIC_SS_ASSIGNMENT asSpreadSpectrum[4]; +}ATOM_ASIC_INTERNAL_SS_INFO; -typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_ASIC_SS_ASSIGNMENT asSpreadSpectrum[4]; -} ATOM_ASIC_INTERNAL_SS_INFO; +typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_ASIC_SS_ASSIGNMENT_V2 asSpreadSpectrum[1]; //this is point only. +}ATOM_ASIC_INTERNAL_SS_INFO_V2; -/* ==============================Scratch Pad Definition Portion=============================== */ +typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V3 +{ + ULONG ulTargetClockRange; //For mem/engine/uvd, Clock Out frequence (VCO ), in unit of 10Khz + //For TMDS/HDMI/LVDS, it is pixel clock , for DP, it is link clock ( 27000 or 16200 ) + USHORT usSpreadSpectrumPercentage; //in unit of 0.01% + USHORT usSpreadRateIn10Hz; //in unit of 10Hz, modulation freq + UCHAR ucClockIndication; //Indicate which clock source needs SS + UCHAR ucSpreadSpectrumMode; //Bit0=0 Down Spread,=1 Center Spread, bit1=0: internal SS bit1=1: external SS + UCHAR ucReserved[2]; +}ATOM_ASIC_SS_ASSIGNMENT_V3; + +typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_ASIC_SS_ASSIGNMENT_V3 asSpreadSpectrum[1]; //this is pointer only. +}ATOM_ASIC_INTERNAL_SS_INFO_V3; + + +//==============================Scratch Pad Definition Portion=============================== #define ATOM_DEVICE_CONNECT_INFO_DEF 0 #define ATOM_ROM_LOCATION_DEF 1 #define ATOM_TV_STANDARD_DEF 2 @@ -2995,7 +3852,8 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_I2C_CHANNEL_STATUS_DEF 8 #define ATOM_I2C_CHANNEL_STATUS1_DEF 9 -/* BIOS_0_SCRATCH Definition */ + +// BIOS_0_SCRATCH Definition #define ATOM_S0_CRT1_MONO 0x00000001L #define ATOM_S0_CRT1_COLOR 0x00000002L #define ATOM_S0_CRT1_MASK (ATOM_S0_CRT1_MONO+ATOM_S0_CRT1_COLOR) @@ -3008,6 +3866,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S0_CV_DIN_A 0x00000020L #define ATOM_S0_CV_MASK_A (ATOM_S0_CV_A+ATOM_S0_CV_DIN_A) + #define ATOM_S0_CRT2_MONO 0x00000100L #define ATOM_S0_CRT2_COLOR 0x00000200L #define ATOM_S0_CRT2_MASK (ATOM_S0_CRT2_MONO+ATOM_S0_CRT2_COLOR) @@ -3025,28 +3884,27 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S0_DFP2 0x00020000L #define ATOM_S0_LCD1 0x00040000L #define ATOM_S0_LCD2 0x00080000L -#define ATOM_S0_TV2 0x00100000L -#define ATOM_S0_DFP3 0x00200000L -#define ATOM_S0_DFP4 0x00400000L -#define ATOM_S0_DFP5 0x00800000L +#define ATOM_S0_DFP6 0x00100000L +#define ATOM_S0_DFP3 0x00200000L +#define ATOM_S0_DFP4 0x00400000L +#define ATOM_S0_DFP5 0x00800000L -#define ATOM_S0_DFP_MASK \ - (ATOM_S0_DFP1 | ATOM_S0_DFP2 | ATOM_S0_DFP3 | ATOM_S0_DFP4 | ATOM_S0_DFP5) +#define ATOM_S0_DFP_MASK ATOM_S0_DFP1 | ATOM_S0_DFP2 | ATOM_S0_DFP3 | ATOM_S0_DFP4 | ATOM_S0_DFP5 | ATOM_S0_DFP6 -#define ATOM_S0_FAD_REGISTER_BUG 0x02000000L /* If set, indicates we are running a PCIE asic with */ - /* the FAD/HDP reg access bug. Bit is read by DAL */ +#define ATOM_S0_FAD_REGISTER_BUG 0x02000000L // If set, indicates we are running a PCIE asic with + // the FAD/HDP reg access bug. Bit is read by DAL, this is obsolete from RV5xx #define ATOM_S0_THERMAL_STATE_MASK 0x1C000000L #define ATOM_S0_THERMAL_STATE_SHIFT 26 #define ATOM_S0_SYSTEM_POWER_STATE_MASK 0xE0000000L -#define ATOM_S0_SYSTEM_POWER_STATE_SHIFT 29 +#define ATOM_S0_SYSTEM_POWER_STATE_SHIFT 29 #define ATOM_S0_SYSTEM_POWER_STATE_VALUE_AC 1 #define ATOM_S0_SYSTEM_POWER_STATE_VALUE_DC 2 #define ATOM_S0_SYSTEM_POWER_STATE_VALUE_LITEAC 3 -/* Byte aligned definition for BIOS usage */ +//Byte aligned defintion for BIOS usage #define ATOM_S0_CRT1_MONOb0 0x01 #define ATOM_S0_CRT1_COLORb0 0x02 #define ATOM_S0_CRT1_MASKb0 (ATOM_S0_CRT1_MONOb0+ATOM_S0_CRT1_COLORb0) @@ -3076,8 +3934,11 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S0_DFP2b2 0x02 #define ATOM_S0_LCD1b2 0x04 #define ATOM_S0_LCD2b2 0x08 -#define ATOM_S0_TV2b2 0x10 -#define ATOM_S0_DFP3b2 0x20 +#define ATOM_S0_DFP6b2 0x10 +#define ATOM_S0_DFP3b2 0x20 +#define ATOM_S0_DFP4b2 0x40 +#define ATOM_S0_DFP5b2 0x80 + #define ATOM_S0_THERMAL_STATE_MASKb3 0x1C #define ATOM_S0_THERMAL_STATE_SHIFTb3 2 @@ -3085,43 +3946,20 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S0_SYSTEM_POWER_STATE_MASKb3 0xE0 #define ATOM_S0_LCD1_SHIFT 18 -/* BIOS_1_SCRATCH Definition */ +// BIOS_1_SCRATCH Definition #define ATOM_S1_ROM_LOCATION_MASK 0x0000FFFFL #define ATOM_S1_PCI_BUS_DEV_MASK 0xFFFF0000L -/* BIOS_2_SCRATCH Definition */ +// BIOS_2_SCRATCH Definition #define ATOM_S2_TV1_STANDARD_MASK 0x0000000FL #define ATOM_S2_CURRENT_BL_LEVEL_MASK 0x0000FF00L #define ATOM_S2_CURRENT_BL_LEVEL_SHIFT 8 -#define ATOM_S2_CRT1_DPMS_STATE 0x00010000L -#define ATOM_S2_LCD1_DPMS_STATE 0x00020000L -#define ATOM_S2_TV1_DPMS_STATE 0x00040000L -#define ATOM_S2_DFP1_DPMS_STATE 0x00080000L -#define ATOM_S2_CRT2_DPMS_STATE 0x00100000L -#define ATOM_S2_LCD2_DPMS_STATE 0x00200000L -#define ATOM_S2_TV2_DPMS_STATE 0x00400000L -#define ATOM_S2_DFP2_DPMS_STATE 0x00800000L -#define ATOM_S2_CV_DPMS_STATE 0x01000000L -#define ATOM_S2_DFP3_DPMS_STATE 0x02000000L -#define ATOM_S2_DFP4_DPMS_STATE 0x04000000L -#define ATOM_S2_DFP5_DPMS_STATE 0x08000000L - -#define ATOM_S2_DFP_DPM_STATE \ - (ATOM_S2_DFP1_DPMS_STATE | ATOM_S2_DFP2_DPMS_STATE | \ - ATOM_S2_DFP3_DPMS_STATE | ATOM_S2_DFP4_DPMS_STATE | \ - ATOM_S2_DFP5_DPMS_STATE) - -#define ATOM_S2_DEVICE_DPMS_STATE \ - (ATOM_S2_CRT1_DPMS_STATE + ATOM_S2_LCD1_DPMS_STATE + \ - ATOM_S2_TV1_DPMS_STATE + ATOM_S2_DFP_DPMS_STATE + \ - ATOM_S2_CRT2_DPMS_STATE + ATOM_S2_LCD2_DPMS_STATE + \ - ATOM_S2_TV2_DPMS_STATE + ATOM_S2_CV_DPMS_STATE) - #define ATOM_S2_FORCEDLOWPWRMODE_STATE_MASK 0x0C000000L #define ATOM_S2_FORCEDLOWPWRMODE_STATE_MASK_SHIFT 26 #define ATOM_S2_FORCEDLOWPWRMODE_STATE_CHANGE 0x10000000L +#define ATOM_S2_DEVICE_DPMS_STATE 0x00010000L #define ATOM_S2_VRI_BRIGHT_ENABLE 0x20000000L #define ATOM_S2_DISPLAY_ROTATION_0_DEGREE 0x0 @@ -3131,21 +3969,11 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S2_DISPLAY_ROTATION_DEGREE_SHIFT 30 #define ATOM_S2_DISPLAY_ROTATION_ANGLE_MASK 0xC0000000L -/* Byte aligned definition for BIOS usage */ + +//Byte aligned defintion for BIOS usage #define ATOM_S2_TV1_STANDARD_MASKb0 0x0F #define ATOM_S2_CURRENT_BL_LEVEL_MASKb1 0xFF -#define ATOM_S2_CRT1_DPMS_STATEb2 0x01 -#define ATOM_S2_LCD1_DPMS_STATEb2 0x02 -#define ATOM_S2_TV1_DPMS_STATEb2 0x04 -#define ATOM_S2_DFP1_DPMS_STATEb2 0x08 -#define ATOM_S2_CRT2_DPMS_STATEb2 0x10 -#define ATOM_S2_LCD2_DPMS_STATEb2 0x20 -#define ATOM_S2_TV2_DPMS_STATEb2 0x40 -#define ATOM_S2_DFP2_DPMS_STATEb2 0x80 -#define ATOM_S2_CV_DPMS_STATEb3 0x01 -#define ATOM_S2_DFP3_DPMS_STATEb3 0x02 -#define ATOM_S2_DFP4_DPMS_STATEb3 0x04 -#define ATOM_S2_DFP5_DPMS_STATEb3 0x08 +#define ATOM_S2_DEVICE_DPMS_STATEb2 0x01 #define ATOM_S2_DEVICE_DPMS_MASKw1 0x3FF #define ATOM_S2_FORCEDLOWPWRMODE_STATE_MASKb3 0x0C @@ -3153,21 +3981,22 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S2_VRI_BRIGHT_ENABLEb3 0x20 #define ATOM_S2_ROTATION_STATE_MASKb3 0xC0 -/* BIOS_3_SCRATCH Definition */ + +// BIOS_3_SCRATCH Definition #define ATOM_S3_CRT1_ACTIVE 0x00000001L #define ATOM_S3_LCD1_ACTIVE 0x00000002L #define ATOM_S3_TV1_ACTIVE 0x00000004L #define ATOM_S3_DFP1_ACTIVE 0x00000008L #define ATOM_S3_CRT2_ACTIVE 0x00000010L #define ATOM_S3_LCD2_ACTIVE 0x00000020L -#define ATOM_S3_TV2_ACTIVE 0x00000040L +#define ATOM_S3_DFP6_ACTIVE 0x00000040L #define ATOM_S3_DFP2_ACTIVE 0x00000080L #define ATOM_S3_CV_ACTIVE 0x00000100L #define ATOM_S3_DFP3_ACTIVE 0x00000200L #define ATOM_S3_DFP4_ACTIVE 0x00000400L #define ATOM_S3_DFP5_ACTIVE 0x00000800L -#define ATOM_S3_DEVICE_ACTIVE_MASK 0x000003FFL +#define ATOM_S3_DEVICE_ACTIVE_MASK 0x00000FFFL #define ATOM_S3_LCD_FULLEXPANSION_ACTIVE 0x00001000L #define ATOM_S3_LCD_EXPANSION_ASPEC_RATIO_ACTIVE 0x00002000L @@ -3178,7 +4007,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S3_DFP1_CRTC_ACTIVE 0x00080000L #define ATOM_S3_CRT2_CRTC_ACTIVE 0x00100000L #define ATOM_S3_LCD2_CRTC_ACTIVE 0x00200000L -#define ATOM_S3_TV2_CRTC_ACTIVE 0x00400000L +#define ATOM_S3_DFP6_CRTC_ACTIVE 0x00400000L #define ATOM_S3_DFP2_CRTC_ACTIVE 0x00800000L #define ATOM_S3_CV_CRTC_ACTIVE 0x01000000L #define ATOM_S3_DFP3_CRTC_ACTIVE 0x02000000L @@ -3187,17 +4016,18 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S3_DEVICE_CRTC_ACTIVE_MASK 0x0FFF0000L #define ATOM_S3_ASIC_GUI_ENGINE_HUNG 0x20000000L +//Below two definitions are not supported in pplib, but in the old powerplay in DAL #define ATOM_S3_ALLOW_FAST_PWR_SWITCH 0x40000000L #define ATOM_S3_RQST_GPU_USE_MIN_PWR 0x80000000L -/* Byte aligned definition for BIOS usage */ +//Byte aligned defintion for BIOS usage #define ATOM_S3_CRT1_ACTIVEb0 0x01 #define ATOM_S3_LCD1_ACTIVEb0 0x02 #define ATOM_S3_TV1_ACTIVEb0 0x04 #define ATOM_S3_DFP1_ACTIVEb0 0x08 #define ATOM_S3_CRT2_ACTIVEb0 0x10 #define ATOM_S3_LCD2_ACTIVEb0 0x20 -#define ATOM_S3_TV2_ACTIVEb0 0x40 +#define ATOM_S3_DFP6_ACTIVEb0 0x40 #define ATOM_S3_DFP2_ACTIVEb0 0x80 #define ATOM_S3_CV_ACTIVEb1 0x01 #define ATOM_S3_DFP3_ACTIVEb1 0x02 @@ -3212,7 +4042,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S3_DFP1_CRTC_ACTIVEb2 0x08 #define ATOM_S3_CRT2_CRTC_ACTIVEb2 0x10 #define ATOM_S3_LCD2_CRTC_ACTIVEb2 0x20 -#define ATOM_S3_TV2_CRTC_ACTIVEb2 0x40 +#define ATOM_S3_DFP6_CRTC_ACTIVEb2 0x40 #define ATOM_S3_DFP2_CRTC_ACTIVEb2 0x80 #define ATOM_S3_CV_CRTC_ACTIVEb3 0x01 #define ATOM_S3_DFP3_CRTC_ACTIVEb3 0x02 @@ -3221,35 +4051,31 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S3_ACTIVE_CRTC2w1 0xFFF -#define ATOM_S3_ASIC_GUI_ENGINE_HUNGb3 0x20 -#define ATOM_S3_ALLOW_FAST_PWR_SWITCHb3 0x40 -#define ATOM_S3_RQST_GPU_USE_MIN_PWRb3 0x80 - -/* BIOS_4_SCRATCH Definition */ +// BIOS_4_SCRATCH Definition #define ATOM_S4_LCD1_PANEL_ID_MASK 0x000000FFL #define ATOM_S4_LCD1_REFRESH_MASK 0x0000FF00L #define ATOM_S4_LCD1_REFRESH_SHIFT 8 -/* Byte aligned definition for BIOS usage */ +//Byte aligned defintion for BIOS usage #define ATOM_S4_LCD1_PANEL_ID_MASKb0 0x0FF #define ATOM_S4_LCD1_REFRESH_MASKb1 ATOM_S4_LCD1_PANEL_ID_MASKb0 #define ATOM_S4_VRAM_INFO_MASKb2 ATOM_S4_LCD1_PANEL_ID_MASKb0 -/* BIOS_5_SCRATCH Definition, BIOS_5_SCRATCH is used by Firmware only !!!! */ +// BIOS_5_SCRATCH Definition, BIOS_5_SCRATCH is used by Firmware only !!!! #define ATOM_S5_DOS_REQ_CRT1b0 0x01 #define ATOM_S5_DOS_REQ_LCD1b0 0x02 #define ATOM_S5_DOS_REQ_TV1b0 0x04 #define ATOM_S5_DOS_REQ_DFP1b0 0x08 #define ATOM_S5_DOS_REQ_CRT2b0 0x10 #define ATOM_S5_DOS_REQ_LCD2b0 0x20 -#define ATOM_S5_DOS_REQ_TV2b0 0x40 +#define ATOM_S5_DOS_REQ_DFP6b0 0x40 #define ATOM_S5_DOS_REQ_DFP2b0 0x80 #define ATOM_S5_DOS_REQ_CVb1 0x01 #define ATOM_S5_DOS_REQ_DFP3b1 0x02 #define ATOM_S5_DOS_REQ_DFP4b1 0x04 #define ATOM_S5_DOS_REQ_DFP5b1 0x08 -#define ATOM_S5_DOS_REQ_DEVICEw0 0x03FF +#define ATOM_S5_DOS_REQ_DEVICEw0 0x0FFF #define ATOM_S5_DOS_REQ_CRT1 0x0001 #define ATOM_S5_DOS_REQ_LCD1 0x0002 @@ -3257,22 +4083,21 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S5_DOS_REQ_DFP1 0x0008 #define ATOM_S5_DOS_REQ_CRT2 0x0010 #define ATOM_S5_DOS_REQ_LCD2 0x0020 -#define ATOM_S5_DOS_REQ_TV2 0x0040 +#define ATOM_S5_DOS_REQ_DFP6 0x0040 #define ATOM_S5_DOS_REQ_DFP2 0x0080 #define ATOM_S5_DOS_REQ_CV 0x0100 -#define ATOM_S5_DOS_REQ_DFP3 0x0200 -#define ATOM_S5_DOS_REQ_DFP4 0x0400 -#define ATOM_S5_DOS_REQ_DFP5 0x0800 +#define ATOM_S5_DOS_REQ_DFP3 0x0200 +#define ATOM_S5_DOS_REQ_DFP4 0x0400 +#define ATOM_S5_DOS_REQ_DFP5 0x0800 #define ATOM_S5_DOS_FORCE_CRT1b2 ATOM_S5_DOS_REQ_CRT1b0 #define ATOM_S5_DOS_FORCE_TV1b2 ATOM_S5_DOS_REQ_TV1b0 #define ATOM_S5_DOS_FORCE_CRT2b2 ATOM_S5_DOS_REQ_CRT2b0 #define ATOM_S5_DOS_FORCE_CVb3 ATOM_S5_DOS_REQ_CVb1 -#define ATOM_S5_DOS_FORCE_DEVICEw1 \ - (ATOM_S5_DOS_FORCE_CRT1b2 + ATOM_S5_DOS_FORCE_TV1b2 + \ - ATOM_S5_DOS_FORCE_CRT2b2 + (ATOM_S5_DOS_FORCE_CVb3 << 8)) +#define ATOM_S5_DOS_FORCE_DEVICEw1 (ATOM_S5_DOS_FORCE_CRT1b2+ATOM_S5_DOS_FORCE_TV1b2+ATOM_S5_DOS_FORCE_CRT2b2+\ + (ATOM_S5_DOS_FORCE_CVb3<<8)) -/* BIOS_6_SCRATCH Definition */ +// BIOS_6_SCRATCH Definition #define ATOM_S6_DEVICE_CHANGE 0x00000001L #define ATOM_S6_SCALER_CHANGE 0x00000002L #define ATOM_S6_LID_CHANGE 0x00000004L @@ -3285,11 +4110,11 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_HW_I2C_BUSY_STATE 0x00000200L #define ATOM_S6_THERMAL_STATE_CHANGE 0x00000400L #define ATOM_S6_INTERRUPT_SET_BY_BIOS 0x00000800L -#define ATOM_S6_REQ_LCD_EXPANSION_FULL 0x00001000L /* Normal expansion Request bit for LCD */ -#define ATOM_S6_REQ_LCD_EXPANSION_ASPEC_RATIO 0x00002000L /* Aspect ratio expansion Request bit for LCD */ +#define ATOM_S6_REQ_LCD_EXPANSION_FULL 0x00001000L //Normal expansion Request bit for LCD +#define ATOM_S6_REQ_LCD_EXPANSION_ASPEC_RATIO 0x00002000L //Aspect ratio expansion Request bit for LCD -#define ATOM_S6_DISPLAY_STATE_CHANGE 0x00004000L /* This bit is recycled when ATOM_BIOS_INFO_BIOS_SCRATCH6_SCL2_REDEFINE is set,previously it's SCL2_H_expansion */ -#define ATOM_S6_I2C_STATE_CHANGE 0x00008000L /* This bit is recycled,when ATOM_BIOS_INFO_BIOS_SCRATCH6_SCL2_REDEFINE is set,previously it's SCL2_V_expansion */ +#define ATOM_S6_DISPLAY_STATE_CHANGE 0x00004000L //This bit is recycled when ATOM_BIOS_INFO_BIOS_SCRATCH6_SCL2_REDEFINE is set,previously it's SCL2_H_expansion +#define ATOM_S6_I2C_STATE_CHANGE 0x00008000L //This bit is recycled,when ATOM_BIOS_INFO_BIOS_SCRATCH6_SCL2_REDEFINE is set,previously it's SCL2_V_expansion #define ATOM_S6_ACC_REQ_CRT1 0x00010000L #define ATOM_S6_ACC_REQ_LCD1 0x00020000L @@ -3297,7 +4122,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_ACC_REQ_DFP1 0x00080000L #define ATOM_S6_ACC_REQ_CRT2 0x00100000L #define ATOM_S6_ACC_REQ_LCD2 0x00200000L -#define ATOM_S6_ACC_REQ_TV2 0x00400000L +#define ATOM_S6_ACC_REQ_DFP6 0x00400000L #define ATOM_S6_ACC_REQ_DFP2 0x00800000L #define ATOM_S6_ACC_REQ_CV 0x01000000L #define ATOM_S6_ACC_REQ_DFP3 0x02000000L @@ -3310,7 +4135,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_VRI_BRIGHTNESS_CHANGE 0x40000000L #define ATOM_S6_CONFIG_DISPLAY_CHANGE_MASK 0x80000000L -/* Byte aligned definition for BIOS usage */ +//Byte aligned defintion for BIOS usage #define ATOM_S6_DEVICE_CHANGEb0 0x01 #define ATOM_S6_SCALER_CHANGEb0 0x02 #define ATOM_S6_LID_CHANGEb0 0x04 @@ -3320,11 +4145,11 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_LID_STATEb0 0x40 #define ATOM_S6_DOCK_STATEb0 0x80 #define ATOM_S6_CRITICAL_STATEb1 0x01 -#define ATOM_S6_HW_I2C_BUSY_STATEb1 0x02 +#define ATOM_S6_HW_I2C_BUSY_STATEb1 0x02 #define ATOM_S6_THERMAL_STATE_CHANGEb1 0x04 #define ATOM_S6_INTERRUPT_SET_BY_BIOSb1 0x08 -#define ATOM_S6_REQ_LCD_EXPANSION_FULLb1 0x10 -#define ATOM_S6_REQ_LCD_EXPANSION_ASPEC_RATIOb1 0x20 +#define ATOM_S6_REQ_LCD_EXPANSION_FULLb1 0x10 +#define ATOM_S6_REQ_LCD_EXPANSION_ASPEC_RATIOb1 0x20 #define ATOM_S6_ACC_REQ_CRT1b2 0x01 #define ATOM_S6_ACC_REQ_LCD1b2 0x02 @@ -3332,12 +4157,12 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_ACC_REQ_DFP1b2 0x08 #define ATOM_S6_ACC_REQ_CRT2b2 0x10 #define ATOM_S6_ACC_REQ_LCD2b2 0x20 -#define ATOM_S6_ACC_REQ_TV2b2 0x40 +#define ATOM_S6_ACC_REQ_DFP6b2 0x40 #define ATOM_S6_ACC_REQ_DFP2b2 0x80 #define ATOM_S6_ACC_REQ_CVb3 0x01 -#define ATOM_S6_ACC_REQ_DFP3b3 0x02 -#define ATOM_S6_ACC_REQ_DFP4b3 0x04 -#define ATOM_S6_ACC_REQ_DFP5b3 0x08 +#define ATOM_S6_ACC_REQ_DFP3b3 0x02 +#define ATOM_S6_ACC_REQ_DFP4b3 0x04 +#define ATOM_S6_ACC_REQ_DFP5b3 0x08 #define ATOM_S6_ACC_REQ_DEVICEw1 ATOM_S5_DOS_REQ_DEVICEw0 #define ATOM_S6_SYSTEM_POWER_MODE_CHANGEb3 0x10 @@ -3366,7 +4191,7 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S6_VRI_BRIGHTNESS_CHANGE_SHIFT 30 #define ATOM_S6_CONFIG_DISPLAY_CHANGE_SHIFT 31 -/* BIOS_7_SCRATCH Definition, BIOS_7_SCRATCH is used by Firmware only !!!! */ +// BIOS_7_SCRATCH Definition, BIOS_7_SCRATCH is used by Firmware only !!!! #define ATOM_S7_DOS_MODE_TYPEb0 0x03 #define ATOM_S7_DOS_MODE_VGAb0 0x00 #define ATOM_S7_DOS_MODE_VESAb0 0x01 @@ -3378,220 +4203,194 @@ typedef struct _ATOM_ASIC_INTERNAL_SS_INFO { #define ATOM_S7_DOS_8BIT_DAC_EN_SHIFT 8 -/* BIOS_8_SCRATCH Definition */ +// BIOS_8_SCRATCH Definition #define ATOM_S8_I2C_CHANNEL_BUSY_MASK 0x00000FFFF -#define ATOM_S8_I2C_HW_ENGINE_BUSY_MASK 0x0FFFF0000 +#define ATOM_S8_I2C_HW_ENGINE_BUSY_MASK 0x0FFFF0000 #define ATOM_S8_I2C_CHANNEL_BUSY_SHIFT 0 #define ATOM_S8_I2C_ENGINE_BUSY_SHIFT 16 -/* BIOS_9_SCRATCH Definition */ -#ifndef ATOM_S9_I2C_CHANNEL_COMPLETED_MASK +// BIOS_9_SCRATCH Definition +#ifndef ATOM_S9_I2C_CHANNEL_COMPLETED_MASK #define ATOM_S9_I2C_CHANNEL_COMPLETED_MASK 0x0000FFFF #endif -#ifndef ATOM_S9_I2C_CHANNEL_ABORTED_MASK +#ifndef ATOM_S9_I2C_CHANNEL_ABORTED_MASK #define ATOM_S9_I2C_CHANNEL_ABORTED_MASK 0xFFFF0000 #endif -#ifndef ATOM_S9_I2C_CHANNEL_COMPLETED_SHIFT +#ifndef ATOM_S9_I2C_CHANNEL_COMPLETED_SHIFT #define ATOM_S9_I2C_CHANNEL_COMPLETED_SHIFT 0 #endif -#ifndef ATOM_S9_I2C_CHANNEL_ABORTED_SHIFT +#ifndef ATOM_S9_I2C_CHANNEL_ABORTED_SHIFT #define ATOM_S9_I2C_CHANNEL_ABORTED_SHIFT 16 #endif + #define ATOM_FLAG_SET 0x20 #define ATOM_FLAG_CLEAR 0 -#define CLEAR_ATOM_S6_ACC_MODE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_ACC_MODE_SHIFT | ATOM_FLAG_CLEAR) -#define SET_ATOM_S6_DEVICE_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_DEVICE_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_VRI_BRIGHTNESS_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_VRI_BRIGHTNESS_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_SCALER_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_SCALER_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_LID_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_LID_CHANGE_SHIFT | ATOM_FLAG_SET) - -#define SET_ATOM_S6_LID_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) |\ - ATOM_S6_LID_STATE_SHIFT | ATOM_FLAG_SET) -#define CLEAR_ATOM_S6_LID_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_LID_STATE_SHIFT | ATOM_FLAG_CLEAR) - -#define SET_ATOM_S6_DOCK_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8)| \ - ATOM_S6_DOCKING_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_DOCK_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_DOCK_STATE_SHIFT | ATOM_FLAG_SET) -#define CLEAR_ATOM_S6_DOCK_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_DOCK_STATE_SHIFT | ATOM_FLAG_CLEAR) - -#define SET_ATOM_S6_THERMAL_STATE_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_THERMAL_STATE_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_SYSTEM_POWER_MODE_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_SYSTEM_POWER_MODE_CHANGE_SHIFT | ATOM_FLAG_SET) -#define SET_ATOM_S6_INTERRUPT_SET_BY_BIOS \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_INTERRUPT_SET_BY_BIOS_SHIFT | ATOM_FLAG_SET) - -#define SET_ATOM_S6_CRITICAL_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_CRITICAL_STATE_SHIFT | ATOM_FLAG_SET) -#define CLEAR_ATOM_S6_CRITICAL_STATE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_CRITICAL_STATE_SHIFT | ATOM_FLAG_CLEAR) - -#define SET_ATOM_S6_REQ_SCALER \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_REQ_SCALER_SHIFT | ATOM_FLAG_SET) -#define CLEAR_ATOM_S6_REQ_SCALER \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_REQ_SCALER_SHIFT | ATOM_FLAG_CLEAR ) - -#define SET_ATOM_S6_REQ_SCALER_ARATIO \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_REQ_SCALER_ARATIO_SHIFT | ATOM_FLAG_SET ) -#define CLEAR_ATOM_S6_REQ_SCALER_ARATIO \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_REQ_SCALER_ARATIO_SHIFT | ATOM_FLAG_CLEAR ) - -#define SET_ATOM_S6_I2C_STATE_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_I2C_STATE_CHANGE_SHIFT | ATOM_FLAG_SET ) - -#define SET_ATOM_S6_DISPLAY_STATE_CHANGE \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_DISPLAY_STATE_CHANGE_SHIFT | ATOM_FLAG_SET ) - -#define SET_ATOM_S6_DEVICE_RECONFIG \ - ((ATOM_ACC_CHANGE_INFO_DEF << 8) | \ - ATOM_S6_CONFIG_DISPLAY_CHANGE_SHIFT | ATOM_FLAG_SET) -#define CLEAR_ATOM_S0_LCD1 \ - ((ATOM_DEVICE_CONNECT_INFO_DEF << 8 ) | \ - ATOM_S0_LCD1_SHIFT | ATOM_FLAG_CLEAR ) -#define SET_ATOM_S7_DOS_8BIT_DAC_EN \ - ((ATOM_DOS_MODE_INFO_DEF << 8) | \ - ATOM_S7_DOS_8BIT_DAC_EN_SHIFT | ATOM_FLAG_SET ) -#define CLEAR_ATOM_S7_DOS_8BIT_DAC_EN \ - ((ATOM_DOS_MODE_INFO_DEF << 8) | \ - ATOM_S7_DOS_8BIT_DAC_EN_SHIFT | ATOM_FLAG_CLEAR ) +#define CLEAR_ATOM_S6_ACC_MODE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_ACC_MODE_SHIFT | ATOM_FLAG_CLEAR) +#define SET_ATOM_S6_DEVICE_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_DEVICE_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_VRI_BRIGHTNESS_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_VRI_BRIGHTNESS_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_SCALER_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_SCALER_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_LID_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_LID_CHANGE_SHIFT | ATOM_FLAG_SET) -/****************************************************************************/ -/* Portion II: Definitinos only used in Driver */ +#define SET_ATOM_S6_LID_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_LID_STATE_SHIFT | ATOM_FLAG_SET) +#define CLEAR_ATOM_S6_LID_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_LID_STATE_SHIFT | ATOM_FLAG_CLEAR) + +#define SET_ATOM_S6_DOCK_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_DOCKING_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_DOCK_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_DOCK_STATE_SHIFT | ATOM_FLAG_SET) +#define CLEAR_ATOM_S6_DOCK_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_DOCK_STATE_SHIFT | ATOM_FLAG_CLEAR) + +#define SET_ATOM_S6_THERMAL_STATE_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_THERMAL_STATE_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_SYSTEM_POWER_MODE_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_SYSTEM_POWER_MODE_CHANGE_SHIFT | ATOM_FLAG_SET) +#define SET_ATOM_S6_INTERRUPT_SET_BY_BIOS ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_INTERRUPT_SET_BY_BIOS_SHIFT | ATOM_FLAG_SET) + +#define SET_ATOM_S6_CRITICAL_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_CRITICAL_STATE_SHIFT | ATOM_FLAG_SET) +#define CLEAR_ATOM_S6_CRITICAL_STATE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_CRITICAL_STATE_SHIFT | ATOM_FLAG_CLEAR) + +#define SET_ATOM_S6_REQ_SCALER ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_REQ_SCALER_SHIFT | ATOM_FLAG_SET) +#define CLEAR_ATOM_S6_REQ_SCALER ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_REQ_SCALER_SHIFT | ATOM_FLAG_CLEAR ) + +#define SET_ATOM_S6_REQ_SCALER_ARATIO ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_REQ_SCALER_ARATIO_SHIFT | ATOM_FLAG_SET ) +#define CLEAR_ATOM_S6_REQ_SCALER_ARATIO ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_REQ_SCALER_ARATIO_SHIFT | ATOM_FLAG_CLEAR ) + +#define SET_ATOM_S6_I2C_STATE_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_I2C_STATE_CHANGE_SHIFT | ATOM_FLAG_SET ) + +#define SET_ATOM_S6_DISPLAY_STATE_CHANGE ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_DISPLAY_STATE_CHANGE_SHIFT | ATOM_FLAG_SET ) + +#define SET_ATOM_S6_DEVICE_RECONFIG ((ATOM_ACC_CHANGE_INFO_DEF << 8 )|ATOM_S6_CONFIG_DISPLAY_CHANGE_SHIFT | ATOM_FLAG_SET) +#define CLEAR_ATOM_S0_LCD1 ((ATOM_DEVICE_CONNECT_INFO_DEF << 8 )| ATOM_S0_LCD1_SHIFT | ATOM_FLAG_CLEAR ) +#define SET_ATOM_S7_DOS_8BIT_DAC_EN ((ATOM_DOS_MODE_INFO_DEF << 8 )|ATOM_S7_DOS_8BIT_DAC_EN_SHIFT | ATOM_FLAG_SET ) +#define CLEAR_ATOM_S7_DOS_8BIT_DAC_EN ((ATOM_DOS_MODE_INFO_DEF << 8 )|ATOM_S7_DOS_8BIT_DAC_EN_SHIFT | ATOM_FLAG_CLEAR ) + +/****************************************************************************/ +//Portion II: Definitinos only used in Driver /****************************************************************************/ -/* Macros used by driver */ +// Macros used by driver +#ifdef __cplusplus +#define GetIndexIntoMasterTable(MasterOrData, FieldName) ((reinterpret_cast<char*>(&(static_cast<ATOM_MASTER_LIST_OF_##MasterOrData##_TABLES*>(0))->FieldName)-static_cast<char*>(0))/sizeof(USHORT)) -#define GetIndexIntoMasterTable(MasterOrData, FieldName) (((char *)(&((ATOM_MASTER_LIST_OF_##MasterOrData##_TABLES *)0)->FieldName)-(char *)0)/sizeof(USHORT)) +#define GET_COMMAND_TABLE_COMMANDSET_REVISION(TABLE_HEADER_OFFSET) (((static_cast<ATOM_COMMON_TABLE_HEADER*>(TABLE_HEADER_OFFSET))->ucTableFormatRevision )&0x3F) +#define GET_COMMAND_TABLE_PARAMETER_REVISION(TABLE_HEADER_OFFSET) (((static_cast<ATOM_COMMON_TABLE_HEADER*>(TABLE_HEADER_OFFSET))->ucTableContentRevision)&0x3F) +#else // not __cplusplus +#define GetIndexIntoMasterTable(MasterOrData, FieldName) (((char*)(&((ATOM_MASTER_LIST_OF_##MasterOrData##_TABLES*)0)->FieldName)-(char*)0)/sizeof(USHORT)) #define GET_COMMAND_TABLE_COMMANDSET_REVISION(TABLE_HEADER_OFFSET) ((((ATOM_COMMON_TABLE_HEADER*)TABLE_HEADER_OFFSET)->ucTableFormatRevision)&0x3F) #define GET_COMMAND_TABLE_PARAMETER_REVISION(TABLE_HEADER_OFFSET) ((((ATOM_COMMON_TABLE_HEADER*)TABLE_HEADER_OFFSET)->ucTableContentRevision)&0x3F) +#endif // __cplusplus #define GET_DATA_TABLE_MAJOR_REVISION GET_COMMAND_TABLE_COMMANDSET_REVISION #define GET_DATA_TABLE_MINOR_REVISION GET_COMMAND_TABLE_PARAMETER_REVISION -/****************************************************************************/ -/* Portion III: Definitinos only used in VBIOS */ +/****************************************************************************/ +//Portion III: Definitinos only used in VBIOS /****************************************************************************/ #define ATOM_DAC_SRC 0x80 #define ATOM_SRC_DAC1 0 #define ATOM_SRC_DAC2 0x80 -#ifdef UEFI_BUILD -#define USHORT UTEMP -#endif - -typedef struct _MEMORY_PLLINIT_PARAMETERS { - ULONG ulTargetMemoryClock; /* In 10Khz unit */ - UCHAR ucAction; /* not define yet */ - UCHAR ucFbDiv_Hi; /* Fbdiv Hi byte */ - UCHAR ucFbDiv; /* FB value */ - UCHAR ucPostDiv; /* Post div */ -} MEMORY_PLLINIT_PARAMETERS; +typedef struct _MEMORY_PLLINIT_PARAMETERS +{ + ULONG ulTargetMemoryClock; //In 10Khz unit + UCHAR ucAction; //not define yet + UCHAR ucFbDiv_Hi; //Fbdiv Hi byte + UCHAR ucFbDiv; //FB value + UCHAR ucPostDiv; //Post div +}MEMORY_PLLINIT_PARAMETERS; #define MEMORY_PLLINIT_PS_ALLOCATION MEMORY_PLLINIT_PARAMETERS -#define GPIO_PIN_WRITE 0x01 + +#define GPIO_PIN_WRITE 0x01 #define GPIO_PIN_READ 0x00 -typedef struct _GPIO_PIN_CONTROL_PARAMETERS { - UCHAR ucGPIO_ID; /* return value, read from GPIO pins */ - UCHAR ucGPIOBitShift; /* define which bit in uGPIOBitVal need to be update */ - UCHAR ucGPIOBitVal; /* Set/Reset corresponding bit defined in ucGPIOBitMask */ - UCHAR ucAction; /* =GPIO_PIN_WRITE: Read; =GPIO_PIN_READ: Write */ -} GPIO_PIN_CONTROL_PARAMETERS; - -typedef struct _ENABLE_SCALER_PARAMETERS { - UCHAR ucScaler; /* ATOM_SCALER1, ATOM_SCALER2 */ - UCHAR ucEnable; /* ATOM_SCALER_DISABLE or ATOM_SCALER_CENTER or ATOM_SCALER_EXPANSION */ - UCHAR ucTVStandard; /* */ - UCHAR ucPadding[1]; -} ENABLE_SCALER_PARAMETERS; -#define ENABLE_SCALER_PS_ALLOCATION ENABLE_SCALER_PARAMETERS - -/* ucEnable: */ +typedef struct _GPIO_PIN_CONTROL_PARAMETERS +{ + UCHAR ucGPIO_ID; //return value, read from GPIO pins + UCHAR ucGPIOBitShift; //define which bit in uGPIOBitVal need to be update + UCHAR ucGPIOBitVal; //Set/Reset corresponding bit defined in ucGPIOBitMask + UCHAR ucAction; //=GPIO_PIN_WRITE: Read; =GPIO_PIN_READ: Write +}GPIO_PIN_CONTROL_PARAMETERS; + +typedef struct _ENABLE_SCALER_PARAMETERS +{ + UCHAR ucScaler; // ATOM_SCALER1, ATOM_SCALER2 + UCHAR ucEnable; // ATOM_SCALER_DISABLE or ATOM_SCALER_CENTER or ATOM_SCALER_EXPANSION + UCHAR ucTVStandard; // + UCHAR ucPadding[1]; +}ENABLE_SCALER_PARAMETERS; +#define ENABLE_SCALER_PS_ALLOCATION ENABLE_SCALER_PARAMETERS + +//ucEnable: #define SCALER_BYPASS_AUTO_CENTER_NO_REPLICATION 0 #define SCALER_BYPASS_AUTO_CENTER_AUTO_REPLICATION 1 #define SCALER_ENABLE_2TAP_ALPHA_MODE 2 #define SCALER_ENABLE_MULTITAP_MODE 3 -typedef struct _ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS { - ULONG usHWIconHorzVertPosn; /* Hardware Icon Vertical position */ - UCHAR ucHWIconVertOffset; /* Hardware Icon Vertical offset */ - UCHAR ucHWIconHorzOffset; /* Hardware Icon Horizontal offset */ - UCHAR ucSelection; /* ATOM_CURSOR1 or ATOM_ICON1 or ATOM_CURSOR2 or ATOM_ICON2 */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ -} ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS; - -typedef struct _ENABLE_HARDWARE_ICON_CURSOR_PS_ALLOCATION { - ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS sEnableIcon; - ENABLE_CRTC_PARAMETERS sReserved; -} ENABLE_HARDWARE_ICON_CURSOR_PS_ALLOCATION; - -typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS { - USHORT usHight; /* Image Hight */ - USHORT usWidth; /* Image Width */ - UCHAR ucSurface; /* Surface 1 or 2 */ - UCHAR ucPadding[3]; -} ENABLE_GRAPH_SURFACE_PARAMETERS; - -typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS_V1_2 { - USHORT usHight; /* Image Hight */ - USHORT usWidth; /* Image Width */ - UCHAR ucSurface; /* Surface 1 or 2 */ - UCHAR ucEnable; /* ATOM_ENABLE or ATOM_DISABLE */ - UCHAR ucPadding[2]; -} ENABLE_GRAPH_SURFACE_PARAMETERS_V1_2; - -typedef struct _ENABLE_GRAPH_SURFACE_PS_ALLOCATION { - ENABLE_GRAPH_SURFACE_PARAMETERS sSetSurface; - ENABLE_YUV_PS_ALLOCATION sReserved; /* Don't set this one */ -} ENABLE_GRAPH_SURFACE_PS_ALLOCATION; - -typedef struct _MEMORY_CLEAN_UP_PARAMETERS { - USHORT usMemoryStart; /* in 8Kb boundry, offset from memory base address */ - USHORT usMemorySize; /* 8Kb blocks aligned */ -} MEMORY_CLEAN_UP_PARAMETERS; +typedef struct _ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS +{ + ULONG usHWIconHorzVertPosn; // Hardware Icon Vertical position + UCHAR ucHWIconVertOffset; // Hardware Icon Vertical offset + UCHAR ucHWIconHorzOffset; // Hardware Icon Horizontal offset + UCHAR ucSelection; // ATOM_CURSOR1 or ATOM_ICON1 or ATOM_CURSOR2 or ATOM_ICON2 + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE +}ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS; + +typedef struct _ENABLE_HARDWARE_ICON_CURSOR_PS_ALLOCATION +{ + ENABLE_HARDWARE_ICON_CURSOR_PARAMETERS sEnableIcon; + ENABLE_CRTC_PARAMETERS sReserved; +}ENABLE_HARDWARE_ICON_CURSOR_PS_ALLOCATION; + +typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS +{ + USHORT usHight; // Image Hight + USHORT usWidth; // Image Width + UCHAR ucSurface; // Surface 1 or 2 + UCHAR ucPadding[3]; +}ENABLE_GRAPH_SURFACE_PARAMETERS; + +typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS_V1_2 +{ + USHORT usHight; // Image Hight + USHORT usWidth; // Image Width + UCHAR ucSurface; // Surface 1 or 2 + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + UCHAR ucPadding[2]; +}ENABLE_GRAPH_SURFACE_PARAMETERS_V1_2; + +typedef struct _ENABLE_GRAPH_SURFACE_PARAMETERS_V1_3 +{ + USHORT usHight; // Image Hight + USHORT usWidth; // Image Width + UCHAR ucSurface; // Surface 1 or 2 + UCHAR ucEnable; // ATOM_ENABLE or ATOM_DISABLE + USHORT usDeviceId; // Active Device Id for this surface. If no device, set to 0. +}ENABLE_GRAPH_SURFACE_PARAMETERS_V1_3; + +typedef struct _ENABLE_GRAPH_SURFACE_PS_ALLOCATION +{ + ENABLE_GRAPH_SURFACE_PARAMETERS sSetSurface; + ENABLE_YUV_PS_ALLOCATION sReserved; // Don't set this one +}ENABLE_GRAPH_SURFACE_PS_ALLOCATION; + +typedef struct _MEMORY_CLEAN_UP_PARAMETERS +{ + USHORT usMemoryStart; //in 8Kb boundry, offset from memory base address + USHORT usMemorySize; //8Kb blocks aligned +}MEMORY_CLEAN_UP_PARAMETERS; #define MEMORY_CLEAN_UP_PS_ALLOCATION MEMORY_CLEAN_UP_PARAMETERS -typedef struct _GET_DISPLAY_SURFACE_SIZE_PARAMETERS { - USHORT usX_Size; /* When use as input parameter, usX_Size indicates which CRTC */ - USHORT usY_Size; -} GET_DISPLAY_SURFACE_SIZE_PARAMETERS; +typedef struct _GET_DISPLAY_SURFACE_SIZE_PARAMETERS +{ + USHORT usX_Size; //When use as input parameter, usX_Size indicates which CRTC + USHORT usY_Size; +}GET_DISPLAY_SURFACE_SIZE_PARAMETERS; -typedef struct _INDIRECT_IO_ACCESS { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR IOAccessSequence[256]; +typedef struct _INDIRECT_IO_ACCESS +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR IOAccessSequence[256]; } INDIRECT_IO_ACCESS; #define INDIRECT_READ 0x00 @@ -3615,93 +4414,108 @@ typedef struct _INDIRECT_IO_ACCESS { #define INDIRECT_IO_NBMISC_READ INDIRECT_IO_NBMISC | INDIRECT_READ #define INDIRECT_IO_NBMISC_WRITE INDIRECT_IO_NBMISC | INDIRECT_WRITE -typedef struct _ATOM_OEM_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; -} ATOM_OEM_INFO; - -typedef struct _ATOM_TV_MODE { - UCHAR ucVMode_Num; /* Video mode number */ - UCHAR ucTV_Mode_Num; /* Internal TV mode number */ -} ATOM_TV_MODE; - -typedef struct _ATOM_BIOS_INT_TVSTD_MODE { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usTV_Mode_LUT_Offset; /* Pointer to standard to internal number conversion table */ - USHORT usTV_FIFO_Offset; /* Pointer to FIFO entry table */ - USHORT usNTSC_Tbl_Offset; /* Pointer to SDTV_Mode_NTSC table */ - USHORT usPAL_Tbl_Offset; /* Pointer to SDTV_Mode_PAL table */ - USHORT usCV_Tbl_Offset; /* Pointer to SDTV_Mode_PAL table */ -} ATOM_BIOS_INT_TVSTD_MODE; - -typedef struct _ATOM_TV_MODE_SCALER_PTR { - USHORT ucFilter0_Offset; /* Pointer to filter format 0 coefficients */ - USHORT usFilter1_Offset; /* Pointer to filter format 0 coefficients */ - UCHAR ucTV_Mode_Num; -} ATOM_TV_MODE_SCALER_PTR; - -typedef struct _ATOM_STANDARD_VESA_TIMING { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_DTD_FORMAT aModeTimings[16]; /* 16 is not the real array number, just for initial allocation */ -} ATOM_STANDARD_VESA_TIMING; - -typedef struct _ATOM_STD_FORMAT { - USHORT usSTD_HDisp; - USHORT usSTD_VDisp; - USHORT usSTD_RefreshRate; - USHORT usReserved; -} ATOM_STD_FORMAT; - -typedef struct _ATOM_VESA_TO_EXTENDED_MODE { - USHORT usVESA_ModeNumber; - USHORT usExtendedModeNumber; -} ATOM_VESA_TO_EXTENDED_MODE; - -typedef struct _ATOM_VESA_TO_INTENAL_MODE_LUT { - ATOM_COMMON_TABLE_HEADER sHeader; - ATOM_VESA_TO_EXTENDED_MODE asVESA_ToExtendedModeInfo[76]; -} ATOM_VESA_TO_INTENAL_MODE_LUT; +typedef struct _ATOM_OEM_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; +}ATOM_OEM_INFO; + +typedef struct _ATOM_TV_MODE +{ + UCHAR ucVMode_Num; //Video mode number + UCHAR ucTV_Mode_Num; //Internal TV mode number +}ATOM_TV_MODE; + +typedef struct _ATOM_BIOS_INT_TVSTD_MODE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usTV_Mode_LUT_Offset; // Pointer to standard to internal number conversion table + USHORT usTV_FIFO_Offset; // Pointer to FIFO entry table + USHORT usNTSC_Tbl_Offset; // Pointer to SDTV_Mode_NTSC table + USHORT usPAL_Tbl_Offset; // Pointer to SDTV_Mode_PAL table + USHORT usCV_Tbl_Offset; // Pointer to SDTV_Mode_PAL table +}ATOM_BIOS_INT_TVSTD_MODE; + + +typedef struct _ATOM_TV_MODE_SCALER_PTR +{ + USHORT ucFilter0_Offset; //Pointer to filter format 0 coefficients + USHORT usFilter1_Offset; //Pointer to filter format 0 coefficients + UCHAR ucTV_Mode_Num; +}ATOM_TV_MODE_SCALER_PTR; + +typedef struct _ATOM_STANDARD_VESA_TIMING +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_DTD_FORMAT aModeTimings[16]; // 16 is not the real array number, just for initial allocation +}ATOM_STANDARD_VESA_TIMING; + + +typedef struct _ATOM_STD_FORMAT +{ + USHORT usSTD_HDisp; + USHORT usSTD_VDisp; + USHORT usSTD_RefreshRate; + USHORT usReserved; +}ATOM_STD_FORMAT; + +typedef struct _ATOM_VESA_TO_EXTENDED_MODE +{ + USHORT usVESA_ModeNumber; + USHORT usExtendedModeNumber; +}ATOM_VESA_TO_EXTENDED_MODE; + +typedef struct _ATOM_VESA_TO_INTENAL_MODE_LUT +{ + ATOM_COMMON_TABLE_HEADER sHeader; + ATOM_VESA_TO_EXTENDED_MODE asVESA_ToExtendedModeInfo[76]; +}ATOM_VESA_TO_INTENAL_MODE_LUT; /*************** ATOM Memory Related Data Structure ***********************/ -typedef struct _ATOM_MEMORY_VENDOR_BLOCK { - UCHAR ucMemoryType; - UCHAR ucMemoryVendor; - UCHAR ucAdjMCId; - UCHAR ucDynClkId; - ULONG ulDllResetClkRange; -} ATOM_MEMORY_VENDOR_BLOCK; - -typedef struct _ATOM_MEMORY_SETTING_ID_CONFIG { +typedef struct _ATOM_MEMORY_VENDOR_BLOCK{ + UCHAR ucMemoryType; + UCHAR ucMemoryVendor; + UCHAR ucAdjMCId; + UCHAR ucDynClkId; + ULONG ulDllResetClkRange; +}ATOM_MEMORY_VENDOR_BLOCK; + + +typedef struct _ATOM_MEMORY_SETTING_ID_CONFIG{ #if ATOM_BIG_ENDIAN - ULONG ucMemBlkId:8; - ULONG ulMemClockRange:24; + ULONG ucMemBlkId:8; + ULONG ulMemClockRange:24; #else - ULONG ulMemClockRange:24; - ULONG ucMemBlkId:8; + ULONG ulMemClockRange:24; + ULONG ucMemBlkId:8; #endif -} ATOM_MEMORY_SETTING_ID_CONFIG; - -typedef union _ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS { - ATOM_MEMORY_SETTING_ID_CONFIG slAccess; - ULONG ulAccess; -} ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS; - -typedef struct _ATOM_MEMORY_SETTING_DATA_BLOCK { - ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS ulMemoryID; - ULONG aulMemData[1]; -} ATOM_MEMORY_SETTING_DATA_BLOCK; - -typedef struct _ATOM_INIT_REG_INDEX_FORMAT { - USHORT usRegIndex; /* MC register index */ - UCHAR ucPreRegDataLength; /* offset in ATOM_INIT_REG_DATA_BLOCK.saRegDataBuf */ -} ATOM_INIT_REG_INDEX_FORMAT; - -typedef struct _ATOM_INIT_REG_BLOCK { - USHORT usRegIndexTblSize; /* size of asRegIndexBuf */ - USHORT usRegDataBlkSize; /* size of ATOM_MEMORY_SETTING_DATA_BLOCK */ - ATOM_INIT_REG_INDEX_FORMAT asRegIndexBuf[1]; - ATOM_MEMORY_SETTING_DATA_BLOCK asRegDataBuf[1]; -} ATOM_INIT_REG_BLOCK; +}ATOM_MEMORY_SETTING_ID_CONFIG; + +typedef union _ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS +{ + ATOM_MEMORY_SETTING_ID_CONFIG slAccess; + ULONG ulAccess; +}ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS; + + +typedef struct _ATOM_MEMORY_SETTING_DATA_BLOCK{ + ATOM_MEMORY_SETTING_ID_CONFIG_ACCESS ulMemoryID; + ULONG aulMemData[1]; +}ATOM_MEMORY_SETTING_DATA_BLOCK; + + +typedef struct _ATOM_INIT_REG_INDEX_FORMAT{ + USHORT usRegIndex; // MC register index + UCHAR ucPreRegDataLength; // offset in ATOM_INIT_REG_DATA_BLOCK.saRegDataBuf +}ATOM_INIT_REG_INDEX_FORMAT; + + +typedef struct _ATOM_INIT_REG_BLOCK{ + USHORT usRegIndexTblSize; //size of asRegIndexBuf + USHORT usRegDataBlkSize; //size of ATOM_MEMORY_SETTING_DATA_BLOCK + ATOM_INIT_REG_INDEX_FORMAT asRegIndexBuf[1]; + ATOM_MEMORY_SETTING_DATA_BLOCK asRegDataBuf[1]; +}ATOM_INIT_REG_BLOCK; #define END_OF_REG_INDEX_BLOCK 0x0ffff #define END_OF_REG_DATA_BLOCK 0x00000000 @@ -3716,16 +4530,19 @@ typedef struct _ATOM_INIT_REG_BLOCK { #define INDEX_ACCESS_RANGE_END (INDEX_ACCESS_RANGE_BEGIN + 1) #define VALUE_INDEX_ACCESS_SINGLE (INDEX_ACCESS_RANGE_END + 1) -typedef struct _ATOM_MC_INIT_PARAM_TABLE { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usAdjustARB_SEQDataOffset; - USHORT usMCInitMemTypeTblOffset; - USHORT usMCInitCommonTblOffset; - USHORT usMCInitPowerDownTblOffset; - ULONG ulARB_SEQDataBuf[32]; - ATOM_INIT_REG_BLOCK asMCInitMemType; - ATOM_INIT_REG_BLOCK asMCInitCommon; -} ATOM_MC_INIT_PARAM_TABLE; + +typedef struct _ATOM_MC_INIT_PARAM_TABLE +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usAdjustARB_SEQDataOffset; + USHORT usMCInitMemTypeTblOffset; + USHORT usMCInitCommonTblOffset; + USHORT usMCInitPowerDownTblOffset; + ULONG ulARB_SEQDataBuf[32]; + ATOM_INIT_REG_BLOCK asMCInitMemType; + ATOM_INIT_REG_BLOCK asMCInitCommon; +}ATOM_MC_INIT_PARAM_TABLE; + #define _4Mx16 0x2 #define _4Mx32 0x3 @@ -3751,221 +4568,272 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE { #define QIMONDA INFINEON #define PROMOS MOSEL +#define KRETON INFINEON -/* ///////////Support for GDDR5 MC uCode to reside in upper 64K of ROM///////////// */ +/////////////Support for GDDR5 MC uCode to reside in upper 64K of ROM///////////// #define UCODE_ROM_START_ADDRESS 0x1c000 -#define UCODE_SIGNATURE 0x4375434d /* 'MCuC' - MC uCode */ - -/* uCode block header for reference */ - -typedef struct _MCuCodeHeader { - ULONG ulSignature; - UCHAR ucRevision; - UCHAR ucChecksum; - UCHAR ucReserved1; - UCHAR ucReserved2; - USHORT usParametersLength; - USHORT usUCodeLength; - USHORT usReserved1; - USHORT usReserved2; +#define UCODE_SIGNATURE 0x4375434d // 'MCuC' - MC uCode + +//uCode block header for reference + +typedef struct _MCuCodeHeader +{ + ULONG ulSignature; + UCHAR ucRevision; + UCHAR ucChecksum; + UCHAR ucReserved1; + UCHAR ucReserved2; + USHORT usParametersLength; + USHORT usUCodeLength; + USHORT usReserved1; + USHORT usReserved2; } MCuCodeHeader; -/* //////////////////////////////////////////////////////////////////////////////// */ +////////////////////////////////////////////////////////////////////////////////// #define ATOM_MAX_NUMBER_OF_VRAM_MODULE 16 #define ATOM_VRAM_MODULE_MEMORY_VENDOR_ID_MASK 0xF -typedef struct _ATOM_VRAM_MODULE_V1 { - ULONG ulReserved; - USHORT usEMRSValue; - USHORT usMRSValue; - USHORT usReserved; - UCHAR ucExtMemoryID; /* An external indicator (by hardcode, callback or pin) to tell what is the current memory module */ - UCHAR ucMemoryType; /* [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] reserved; */ - UCHAR ucMemoryVenderID; /* Predefined,never change across designs or memory type/vender */ - UCHAR ucMemoryDeviceCfg; /* [7:4]=0x0:4M;=0x1:8M;=0x2:16M;0x3:32M....[3:0]=0x0:x4;=0x1:x8;=0x2:x16;=0x3:x32... */ - UCHAR ucRow; /* Number of Row,in power of 2; */ - UCHAR ucColumn; /* Number of Column,in power of 2; */ - UCHAR ucBank; /* Nunber of Bank; */ - UCHAR ucRank; /* Number of Rank, in power of 2 */ - UCHAR ucChannelNum; /* Number of channel; */ - UCHAR ucChannelConfig; /* [3:0]=Indication of what channel combination;[4:7]=Channel bit width, in number of 2 */ - UCHAR ucDefaultMVDDQ_ID; /* Default MVDDQ setting for this memory block, ID linking to MVDDQ info table to find real set-up data; */ - UCHAR ucDefaultMVDDC_ID; /* Default MVDDC setting for this memory block, ID linking to MVDDC info table to find real set-up data; */ - UCHAR ucReserved[2]; -} ATOM_VRAM_MODULE_V1; - -typedef struct _ATOM_VRAM_MODULE_V2 { - ULONG ulReserved; - ULONG ulFlags; /* To enable/disable functionalities based on memory type */ - ULONG ulEngineClock; /* Override of default engine clock for particular memory type */ - ULONG ulMemoryClock; /* Override of default memory clock for particular memory type */ - USHORT usEMRS2Value; /* EMRS2 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usEMRS3Value; /* EMRS3 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usEMRSValue; - USHORT usMRSValue; - USHORT usReserved; - UCHAR ucExtMemoryID; /* An external indicator (by hardcode, callback or pin) to tell what is the current memory module */ - UCHAR ucMemoryType; /* [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] - must not be used for now; */ - UCHAR ucMemoryVenderID; /* Predefined,never change across designs or memory type/vender. If not predefined, vendor detection table gets executed */ - UCHAR ucMemoryDeviceCfg; /* [7:4]=0x0:4M;=0x1:8M;=0x2:16M;0x3:32M....[3:0]=0x0:x4;=0x1:x8;=0x2:x16;=0x3:x32... */ - UCHAR ucRow; /* Number of Row,in power of 2; */ - UCHAR ucColumn; /* Number of Column,in power of 2; */ - UCHAR ucBank; /* Nunber of Bank; */ - UCHAR ucRank; /* Number of Rank, in power of 2 */ - UCHAR ucChannelNum; /* Number of channel; */ - UCHAR ucChannelConfig; /* [3:0]=Indication of what channel combination;[4:7]=Channel bit width, in number of 2 */ - UCHAR ucDefaultMVDDQ_ID; /* Default MVDDQ setting for this memory block, ID linking to MVDDQ info table to find real set-up data; */ - UCHAR ucDefaultMVDDC_ID; /* Default MVDDC setting for this memory block, ID linking to MVDDC info table to find real set-up data; */ - UCHAR ucRefreshRateFactor; - UCHAR ucReserved[3]; -} ATOM_VRAM_MODULE_V2; - -typedef struct _ATOM_MEMORY_TIMING_FORMAT { - ULONG ulClkRange; /* memory clock in 10kHz unit, when target memory clock is below this clock, use this memory timing */ - union { - USHORT usMRS; /* mode register */ - USHORT usDDR3_MR0; - }; - union { - USHORT usEMRS; /* extended mode register */ - USHORT usDDR3_MR1; - }; - UCHAR ucCL; /* CAS latency */ - UCHAR ucWL; /* WRITE Latency */ - UCHAR uctRAS; /* tRAS */ - UCHAR uctRC; /* tRC */ - UCHAR uctRFC; /* tRFC */ - UCHAR uctRCDR; /* tRCDR */ - UCHAR uctRCDW; /* tRCDW */ - UCHAR uctRP; /* tRP */ - UCHAR uctRRD; /* tRRD */ - UCHAR uctWR; /* tWR */ - UCHAR uctWTR; /* tWTR */ - UCHAR uctPDIX; /* tPDIX */ - UCHAR uctFAW; /* tFAW */ - UCHAR uctAOND; /* tAOND */ - union { - struct { - UCHAR ucflag; /* flag to control memory timing calculation. bit0= control EMRS2 Infineon */ - UCHAR ucReserved; - }; - USHORT usDDR3_MR2; - }; -} ATOM_MEMORY_TIMING_FORMAT; - -typedef struct _ATOM_MEMORY_TIMING_FORMAT_V1 { - ULONG ulClkRange; /* memory clock in 10kHz unit, when target memory clock is below this clock, use this memory timing */ - USHORT usMRS; /* mode register */ - USHORT usEMRS; /* extended mode register */ - UCHAR ucCL; /* CAS latency */ - UCHAR ucWL; /* WRITE Latency */ - UCHAR uctRAS; /* tRAS */ - UCHAR uctRC; /* tRC */ - UCHAR uctRFC; /* tRFC */ - UCHAR uctRCDR; /* tRCDR */ - UCHAR uctRCDW; /* tRCDW */ - UCHAR uctRP; /* tRP */ - UCHAR uctRRD; /* tRRD */ - UCHAR uctWR; /* tWR */ - UCHAR uctWTR; /* tWTR */ - UCHAR uctPDIX; /* tPDIX */ - UCHAR uctFAW; /* tFAW */ - UCHAR uctAOND; /* tAOND */ - UCHAR ucflag; /* flag to control memory timing calculation. bit0= control EMRS2 Infineon */ -/* ///////////////////////GDDR parameters/////////////////////////////////// */ - UCHAR uctCCDL; /* */ - UCHAR uctCRCRL; /* */ - UCHAR uctCRCWL; /* */ - UCHAR uctCKE; /* */ - UCHAR uctCKRSE; /* */ - UCHAR uctCKRSX; /* */ - UCHAR uctFAW32; /* */ - UCHAR ucReserved1; /* */ - UCHAR ucReserved2; /* */ - UCHAR ucTerminator; -} ATOM_MEMORY_TIMING_FORMAT_V1; - -typedef struct _ATOM_MEMORY_FORMAT { - ULONG ulDllDisClock; /* memory DLL will be disable when target memory clock is below this clock */ - union { - USHORT usEMRS2Value; /* EMRS2 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usDDR3_Reserved; /* Not used for DDR3 memory */ - }; - union { - USHORT usEMRS3Value; /* EMRS3 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usDDR3_MR3; /* Used for DDR3 memory */ - }; - UCHAR ucMemoryType; /* [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] - must not be used for now; */ - UCHAR ucMemoryVenderID; /* Predefined,never change across designs or memory type/vender. If not predefined, vendor detection table gets executed */ - UCHAR ucRow; /* Number of Row,in power of 2; */ - UCHAR ucColumn; /* Number of Column,in power of 2; */ - UCHAR ucBank; /* Nunber of Bank; */ - UCHAR ucRank; /* Number of Rank, in power of 2 */ - UCHAR ucBurstSize; /* burst size, 0= burst size=4 1= burst size=8 */ - UCHAR ucDllDisBit; /* position of DLL Enable/Disable bit in EMRS ( Extended Mode Register ) */ - UCHAR ucRefreshRateFactor; /* memory refresh rate in unit of ms */ - UCHAR ucDensity; /* _8Mx32, _16Mx32, _16Mx16, _32Mx16 */ - UCHAR ucPreamble; /* [7:4] Write Preamble, [3:0] Read Preamble */ - UCHAR ucMemAttrib; /* Memory Device Addribute, like RDBI/WDBI etc */ - ATOM_MEMORY_TIMING_FORMAT asMemTiming[5]; /* Memory Timing block sort from lower clock to higher clock */ -} ATOM_MEMORY_FORMAT; - -typedef struct _ATOM_VRAM_MODULE_V3 { - ULONG ulChannelMapCfg; /* board dependent paramenter:Channel combination */ - USHORT usSize; /* size of ATOM_VRAM_MODULE_V3 */ - USHORT usDefaultMVDDQ; /* board dependent parameter:Default Memory Core Voltage */ - USHORT usDefaultMVDDC; /* board dependent parameter:Default Memory IO Voltage */ - UCHAR ucExtMemoryID; /* An external indicator (by hardcode, callback or pin) to tell what is the current memory module */ - UCHAR ucChannelNum; /* board dependent parameter:Number of channel; */ - UCHAR ucChannelSize; /* board dependent parameter:32bit or 64bit */ - UCHAR ucVREFI; /* board dependnt parameter: EXT or INT +160mv to -140mv */ - UCHAR ucNPL_RT; /* board dependent parameter:NPL round trip delay, used for calculate memory timing parameters */ - UCHAR ucFlag; /* To enable/disable functionalities based on memory type */ - ATOM_MEMORY_FORMAT asMemory; /* describ all of video memory parameters from memory spec */ -} ATOM_VRAM_MODULE_V3; - -/* ATOM_VRAM_MODULE_V3.ucNPL_RT */ +typedef struct _ATOM_VRAM_MODULE_V1 +{ + ULONG ulReserved; + USHORT usEMRSValue; + USHORT usMRSValue; + USHORT usReserved; + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] reserved; + UCHAR ucMemoryVenderID; // Predefined,never change across designs or memory type/vender + UCHAR ucMemoryDeviceCfg; // [7:4]=0x0:4M;=0x1:8M;=0x2:16M;0x3:32M....[3:0]=0x0:x4;=0x1:x8;=0x2:x16;=0x3:x32... + UCHAR ucRow; // Number of Row,in power of 2; + UCHAR ucColumn; // Number of Column,in power of 2; + UCHAR ucBank; // Nunber of Bank; + UCHAR ucRank; // Number of Rank, in power of 2 + UCHAR ucChannelNum; // Number of channel; + UCHAR ucChannelConfig; // [3:0]=Indication of what channel combination;[4:7]=Channel bit width, in number of 2 + UCHAR ucDefaultMVDDQ_ID; // Default MVDDQ setting for this memory block, ID linking to MVDDQ info table to find real set-up data; + UCHAR ucDefaultMVDDC_ID; // Default MVDDC setting for this memory block, ID linking to MVDDC info table to find real set-up data; + UCHAR ucReserved[2]; +}ATOM_VRAM_MODULE_V1; + + +typedef struct _ATOM_VRAM_MODULE_V2 +{ + ULONG ulReserved; + ULONG ulFlags; // To enable/disable functionalities based on memory type + ULONG ulEngineClock; // Override of default engine clock for particular memory type + ULONG ulMemoryClock; // Override of default memory clock for particular memory type + USHORT usEMRS2Value; // EMRS2 Value is used for GDDR2 and GDDR4 memory type + USHORT usEMRS3Value; // EMRS3 Value is used for GDDR2 and GDDR4 memory type + USHORT usEMRSValue; + USHORT usMRSValue; + USHORT usReserved; + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] - must not be used for now; + UCHAR ucMemoryVenderID; // Predefined,never change across designs or memory type/vender. If not predefined, vendor detection table gets executed + UCHAR ucMemoryDeviceCfg; // [7:4]=0x0:4M;=0x1:8M;=0x2:16M;0x3:32M....[3:0]=0x0:x4;=0x1:x8;=0x2:x16;=0x3:x32... + UCHAR ucRow; // Number of Row,in power of 2; + UCHAR ucColumn; // Number of Column,in power of 2; + UCHAR ucBank; // Nunber of Bank; + UCHAR ucRank; // Number of Rank, in power of 2 + UCHAR ucChannelNum; // Number of channel; + UCHAR ucChannelConfig; // [3:0]=Indication of what channel combination;[4:7]=Channel bit width, in number of 2 + UCHAR ucDefaultMVDDQ_ID; // Default MVDDQ setting for this memory block, ID linking to MVDDQ info table to find real set-up data; + UCHAR ucDefaultMVDDC_ID; // Default MVDDC setting for this memory block, ID linking to MVDDC info table to find real set-up data; + UCHAR ucRefreshRateFactor; + UCHAR ucReserved[3]; +}ATOM_VRAM_MODULE_V2; + + +typedef struct _ATOM_MEMORY_TIMING_FORMAT +{ + ULONG ulClkRange; // memory clock in 10kHz unit, when target memory clock is below this clock, use this memory timing + union{ + USHORT usMRS; // mode register + USHORT usDDR3_MR0; + }; + union{ + USHORT usEMRS; // extended mode register + USHORT usDDR3_MR1; + }; + UCHAR ucCL; // CAS latency + UCHAR ucWL; // WRITE Latency + UCHAR uctRAS; // tRAS + UCHAR uctRC; // tRC + UCHAR uctRFC; // tRFC + UCHAR uctRCDR; // tRCDR + UCHAR uctRCDW; // tRCDW + UCHAR uctRP; // tRP + UCHAR uctRRD; // tRRD + UCHAR uctWR; // tWR + UCHAR uctWTR; // tWTR + UCHAR uctPDIX; // tPDIX + UCHAR uctFAW; // tFAW + UCHAR uctAOND; // tAOND + union + { + struct { + UCHAR ucflag; // flag to control memory timing calculation. bit0= control EMRS2 Infineon + UCHAR ucReserved; + }; + USHORT usDDR3_MR2; + }; +}ATOM_MEMORY_TIMING_FORMAT; + + +typedef struct _ATOM_MEMORY_TIMING_FORMAT_V1 +{ + ULONG ulClkRange; // memory clock in 10kHz unit, when target memory clock is below this clock, use this memory timing + USHORT usMRS; // mode register + USHORT usEMRS; // extended mode register + UCHAR ucCL; // CAS latency + UCHAR ucWL; // WRITE Latency + UCHAR uctRAS; // tRAS + UCHAR uctRC; // tRC + UCHAR uctRFC; // tRFC + UCHAR uctRCDR; // tRCDR + UCHAR uctRCDW; // tRCDW + UCHAR uctRP; // tRP + UCHAR uctRRD; // tRRD + UCHAR uctWR; // tWR + UCHAR uctWTR; // tWTR + UCHAR uctPDIX; // tPDIX + UCHAR uctFAW; // tFAW + UCHAR uctAOND; // tAOND + UCHAR ucflag; // flag to control memory timing calculation. bit0= control EMRS2 Infineon +////////////////////////////////////GDDR parameters/////////////////////////////////// + UCHAR uctCCDL; // + UCHAR uctCRCRL; // + UCHAR uctCRCWL; // + UCHAR uctCKE; // + UCHAR uctCKRSE; // + UCHAR uctCKRSX; // + UCHAR uctFAW32; // + UCHAR ucMR5lo; // + UCHAR ucMR5hi; // + UCHAR ucTerminator; +}ATOM_MEMORY_TIMING_FORMAT_V1; + +typedef struct _ATOM_MEMORY_TIMING_FORMAT_V2 +{ + ULONG ulClkRange; // memory clock in 10kHz unit, when target memory clock is below this clock, use this memory timing + USHORT usMRS; // mode register + USHORT usEMRS; // extended mode register + UCHAR ucCL; // CAS latency + UCHAR ucWL; // WRITE Latency + UCHAR uctRAS; // tRAS + UCHAR uctRC; // tRC + UCHAR uctRFC; // tRFC + UCHAR uctRCDR; // tRCDR + UCHAR uctRCDW; // tRCDW + UCHAR uctRP; // tRP + UCHAR uctRRD; // tRRD + UCHAR uctWR; // tWR + UCHAR uctWTR; // tWTR + UCHAR uctPDIX; // tPDIX + UCHAR uctFAW; // tFAW + UCHAR uctAOND; // tAOND + UCHAR ucflag; // flag to control memory timing calculation. bit0= control EMRS2 Infineon +////////////////////////////////////GDDR parameters/////////////////////////////////// + UCHAR uctCCDL; // + UCHAR uctCRCRL; // + UCHAR uctCRCWL; // + UCHAR uctCKE; // + UCHAR uctCKRSE; // + UCHAR uctCKRSX; // + UCHAR uctFAW32; // + UCHAR ucMR4lo; // + UCHAR ucMR4hi; // + UCHAR ucMR5lo; // + UCHAR ucMR5hi; // + UCHAR ucTerminator; + UCHAR ucReserved; +}ATOM_MEMORY_TIMING_FORMAT_V2; + +typedef struct _ATOM_MEMORY_FORMAT +{ + ULONG ulDllDisClock; // memory DLL will be disable when target memory clock is below this clock + union{ + USHORT usEMRS2Value; // EMRS2 Value is used for GDDR2 and GDDR4 memory type + USHORT usDDR3_Reserved; // Not used for DDR3 memory + }; + union{ + USHORT usEMRS3Value; // EMRS3 Value is used for GDDR2 and GDDR4 memory type + USHORT usDDR3_MR3; // Used for DDR3 memory + }; + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4;[3:0] - must not be used for now; + UCHAR ucMemoryVenderID; // Predefined,never change across designs or memory type/vender. If not predefined, vendor detection table gets executed + UCHAR ucRow; // Number of Row,in power of 2; + UCHAR ucColumn; // Number of Column,in power of 2; + UCHAR ucBank; // Nunber of Bank; + UCHAR ucRank; // Number of Rank, in power of 2 + UCHAR ucBurstSize; // burst size, 0= burst size=4 1= burst size=8 + UCHAR ucDllDisBit; // position of DLL Enable/Disable bit in EMRS ( Extended Mode Register ) + UCHAR ucRefreshRateFactor; // memory refresh rate in unit of ms + UCHAR ucDensity; // _8Mx32, _16Mx32, _16Mx16, _32Mx16 + UCHAR ucPreamble; //[7:4] Write Preamble, [3:0] Read Preamble + UCHAR ucMemAttrib; // Memory Device Addribute, like RDBI/WDBI etc + ATOM_MEMORY_TIMING_FORMAT asMemTiming[5]; //Memory Timing block sort from lower clock to higher clock +}ATOM_MEMORY_FORMAT; + + +typedef struct _ATOM_VRAM_MODULE_V3 +{ + ULONG ulChannelMapCfg; // board dependent paramenter:Channel combination + USHORT usSize; // size of ATOM_VRAM_MODULE_V3 + USHORT usDefaultMVDDQ; // board dependent parameter:Default Memory Core Voltage + USHORT usDefaultMVDDC; // board dependent parameter:Default Memory IO Voltage + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucChannelNum; // board dependent parameter:Number of channel; + UCHAR ucChannelSize; // board dependent parameter:32bit or 64bit + UCHAR ucVREFI; // board dependnt parameter: EXT or INT +160mv to -140mv + UCHAR ucNPL_RT; // board dependent parameter:NPL round trip delay, used for calculate memory timing parameters + UCHAR ucFlag; // To enable/disable functionalities based on memory type + ATOM_MEMORY_FORMAT asMemory; // describ all of video memory parameters from memory spec +}ATOM_VRAM_MODULE_V3; + + +//ATOM_VRAM_MODULE_V3.ucNPL_RT #define NPL_RT_MASK 0x0f #define BATTERY_ODT_MASK 0xc0 #define ATOM_VRAM_MODULE ATOM_VRAM_MODULE_V3 -typedef struct _ATOM_VRAM_MODULE_V4 { - ULONG ulChannelMapCfg; /* board dependent parameter: Channel combination */ - USHORT usModuleSize; /* size of ATOM_VRAM_MODULE_V4, make it easy for VBIOS to look for next entry of VRAM_MODULE */ - USHORT usPrivateReserved; /* BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! */ - /* MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS) */ - USHORT usReserved; - UCHAR ucExtMemoryID; /* An external indicator (by hardcode, callback or pin) to tell what is the current memory module */ - UCHAR ucMemoryType; /* [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4; 0x5:DDR5 [3:0] - Must be 0x0 for now; */ - UCHAR ucChannelNum; /* Number of channels present in this module config */ - UCHAR ucChannelWidth; /* 0 - 32 bits; 1 - 64 bits */ - UCHAR ucDensity; /* _8Mx32, _16Mx32, _16Mx16, _32Mx16 */ - UCHAR ucFlag; /* To enable/disable functionalities based on memory type */ - UCHAR ucMisc; /* bit0: 0 - single rank; 1 - dual rank; bit2: 0 - burstlength 4, 1 - burstlength 8 */ - UCHAR ucVREFI; /* board dependent parameter */ - UCHAR ucNPL_RT; /* board dependent parameter:NPL round trip delay, used for calculate memory timing parameters */ - UCHAR ucPreamble; /* [7:4] Write Preamble, [3:0] Read Preamble */ - UCHAR ucMemorySize; /* BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! */ - /* Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros */ - UCHAR ucReserved[3]; - -/* compare with V3, we flat the struct by merging ATOM_MEMORY_FORMAT (as is) into V4 as the same level */ - union { - USHORT usEMRS2Value; /* EMRS2 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usDDR3_Reserved; - }; - union { - USHORT usEMRS3Value; /* EMRS3 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usDDR3_MR3; /* Used for DDR3 memory */ - }; - UCHAR ucMemoryVenderID; /* Predefined, If not predefined, vendor detection table gets executed */ - UCHAR ucRefreshRateFactor; /* [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms) */ - UCHAR ucReserved2[2]; - ATOM_MEMORY_TIMING_FORMAT asMemTiming[5]; /* Memory Timing block sort from lower clock to higher clock */ -} ATOM_VRAM_MODULE_V4; +typedef struct _ATOM_VRAM_MODULE_V4 +{ + ULONG ulChannelMapCfg; // board dependent parameter: Channel combination + USHORT usModuleSize; // size of ATOM_VRAM_MODULE_V4, make it easy for VBIOS to look for next entry of VRAM_MODULE + USHORT usPrivateReserved; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS) + USHORT usReserved; + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4; 0x5:DDR5 [3:0] - Must be 0x0 for now; + UCHAR ucChannelNum; // Number of channels present in this module config + UCHAR ucChannelWidth; // 0 - 32 bits; 1 - 64 bits + UCHAR ucDensity; // _8Mx32, _16Mx32, _16Mx16, _32Mx16 + UCHAR ucFlag; // To enable/disable functionalities based on memory type + UCHAR ucMisc; // bit0: 0 - single rank; 1 - dual rank; bit2: 0 - burstlength 4, 1 - burstlength 8 + UCHAR ucVREFI; // board dependent parameter + UCHAR ucNPL_RT; // board dependent parameter:NPL round trip delay, used for calculate memory timing parameters + UCHAR ucPreamble; // [7:4] Write Preamble, [3:0] Read Preamble + UCHAR ucMemorySize; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros + UCHAR ucReserved[3]; + +//compare with V3, we flat the struct by merging ATOM_MEMORY_FORMAT (as is) into V4 as the same level + union{ + USHORT usEMRS2Value; // EMRS2 Value is used for GDDR2 and GDDR4 memory type + USHORT usDDR3_Reserved; + }; + union{ + USHORT usEMRS3Value; // EMRS3 Value is used for GDDR2 and GDDR4 memory type + USHORT usDDR3_MR3; // Used for DDR3 memory + }; + UCHAR ucMemoryVenderID; // Predefined, If not predefined, vendor detection table gets executed + UCHAR ucRefreshRateFactor; // [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms) + UCHAR ucReserved2[2]; + ATOM_MEMORY_TIMING_FORMAT asMemTiming[5];//Memory Timing block sort from lower clock to higher clock +}ATOM_VRAM_MODULE_V4; #define VRAM_MODULE_V4_MISC_RANK_MASK 0x3 #define VRAM_MODULE_V4_MISC_DUAL_RANK 0x1 @@ -3973,96 +4841,139 @@ typedef struct _ATOM_VRAM_MODULE_V4 { #define VRAM_MODULE_V4_MISC_BL8 0x4 #define VRAM_MODULE_V4_MISC_DUAL_CS 0x10 -typedef struct _ATOM_VRAM_MODULE_V5 { - ULONG ulChannelMapCfg; /* board dependent parameter: Channel combination */ - USHORT usModuleSize; /* size of ATOM_VRAM_MODULE_V4, make it easy for VBIOS to look for next entry of VRAM_MODULE */ - USHORT usPrivateReserved; /* BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! */ - /* MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS) */ - USHORT usReserved; - UCHAR ucExtMemoryID; /* An external indicator (by hardcode, callback or pin) to tell what is the current memory module */ - UCHAR ucMemoryType; /* [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4; 0x5:DDR5 [3:0] - Must be 0x0 for now; */ - UCHAR ucChannelNum; /* Number of channels present in this module config */ - UCHAR ucChannelWidth; /* 0 - 32 bits; 1 - 64 bits */ - UCHAR ucDensity; /* _8Mx32, _16Mx32, _16Mx16, _32Mx16 */ - UCHAR ucFlag; /* To enable/disable functionalities based on memory type */ - UCHAR ucMisc; /* bit0: 0 - single rank; 1 - dual rank; bit2: 0 - burstlength 4, 1 - burstlength 8 */ - UCHAR ucVREFI; /* board dependent parameter */ - UCHAR ucNPL_RT; /* board dependent parameter:NPL round trip delay, used for calculate memory timing parameters */ - UCHAR ucPreamble; /* [7:4] Write Preamble, [3:0] Read Preamble */ - UCHAR ucMemorySize; /* BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! */ - /* Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros */ - UCHAR ucReserved[3]; +typedef struct _ATOM_VRAM_MODULE_V5 +{ + ULONG ulChannelMapCfg; // board dependent parameter: Channel combination + USHORT usModuleSize; // size of ATOM_VRAM_MODULE_V4, make it easy for VBIOS to look for next entry of VRAM_MODULE + USHORT usPrivateReserved; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS) + USHORT usReserved; + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4; 0x5:DDR5 [3:0] - Must be 0x0 for now; + UCHAR ucChannelNum; // Number of channels present in this module config + UCHAR ucChannelWidth; // 0 - 32 bits; 1 - 64 bits + UCHAR ucDensity; // _8Mx32, _16Mx32, _16Mx16, _32Mx16 + UCHAR ucFlag; // To enable/disable functionalities based on memory type + UCHAR ucMisc; // bit0: 0 - single rank; 1 - dual rank; bit2: 0 - burstlength 4, 1 - burstlength 8 + UCHAR ucVREFI; // board dependent parameter + UCHAR ucNPL_RT; // board dependent parameter:NPL round trip delay, used for calculate memory timing parameters + UCHAR ucPreamble; // [7:4] Write Preamble, [3:0] Read Preamble + UCHAR ucMemorySize; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros + UCHAR ucReserved[3]; + +//compare with V3, we flat the struct by merging ATOM_MEMORY_FORMAT (as is) into V4 as the same level + USHORT usEMRS2Value; // EMRS2 Value is used for GDDR2 and GDDR4 memory type + USHORT usEMRS3Value; // EMRS3 Value is used for GDDR2 and GDDR4 memory type + UCHAR ucMemoryVenderID; // Predefined, If not predefined, vendor detection table gets executed + UCHAR ucRefreshRateFactor; // [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms) + UCHAR ucFIFODepth; // FIFO depth supposes to be detected during vendor detection, but if we dont do vendor detection we have to hardcode FIFO Depth + UCHAR ucCDR_Bandwidth; // [0:3]=Read CDR bandwidth, [4:7] - Write CDR Bandwidth + ATOM_MEMORY_TIMING_FORMAT_V1 asMemTiming[5];//Memory Timing block sort from lower clock to higher clock +}ATOM_VRAM_MODULE_V5; + +typedef struct _ATOM_VRAM_MODULE_V6 +{ + ULONG ulChannelMapCfg; // board dependent parameter: Channel combination + USHORT usModuleSize; // size of ATOM_VRAM_MODULE_V4, make it easy for VBIOS to look for next entry of VRAM_MODULE + USHORT usPrivateReserved; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // MC_ARB_RAMCFG (includes NOOFBANK,NOOFRANKS,NOOFROWS,NOOFCOLS) + USHORT usReserved; + UCHAR ucExtMemoryID; // An external indicator (by hardcode, callback or pin) to tell what is the current memory module + UCHAR ucMemoryType; // [7:4]=0x1:DDR1;=0x2:DDR2;=0x3:DDR3;=0x4:DDR4; 0x5:DDR5 [3:0] - Must be 0x0 for now; + UCHAR ucChannelNum; // Number of channels present in this module config + UCHAR ucChannelWidth; // 0 - 32 bits; 1 - 64 bits + UCHAR ucDensity; // _8Mx32, _16Mx32, _16Mx16, _32Mx16 + UCHAR ucFlag; // To enable/disable functionalities based on memory type + UCHAR ucMisc; // bit0: 0 - single rank; 1 - dual rank; bit2: 0 - burstlength 4, 1 - burstlength 8 + UCHAR ucVREFI; // board dependent parameter + UCHAR ucNPL_RT; // board dependent parameter:NPL round trip delay, used for calculate memory timing parameters + UCHAR ucPreamble; // [7:4] Write Preamble, [3:0] Read Preamble + UCHAR ucMemorySize; // BIOS internal reserved space to optimize code size, updated by the compiler, shouldn't be modified manually!! + // Total memory size in unit of 16MB for CONFIG_MEMSIZE - bit[23:0] zeros + UCHAR ucReserved[3]; + +//compare with V3, we flat the struct by merging ATOM_MEMORY_FORMAT (as is) into V4 as the same level + USHORT usEMRS2Value; // EMRS2 Value is used for GDDR2 and GDDR4 memory type + USHORT usEMRS3Value; // EMRS3 Value is used for GDDR2 and GDDR4 memory type + UCHAR ucMemoryVenderID; // Predefined, If not predefined, vendor detection table gets executed + UCHAR ucRefreshRateFactor; // [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms) + UCHAR ucFIFODepth; // FIFO depth supposes to be detected during vendor detection, but if we dont do vendor detection we have to hardcode FIFO Depth + UCHAR ucCDR_Bandwidth; // [0:3]=Read CDR bandwidth, [4:7] - Write CDR Bandwidth + ATOM_MEMORY_TIMING_FORMAT_V2 asMemTiming[5];//Memory Timing block sort from lower clock to higher clock +}ATOM_VRAM_MODULE_V6; + + + +typedef struct _ATOM_VRAM_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucNumOfVRAMModule; + ATOM_VRAM_MODULE aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; // just for allocation, real number of blocks is in ucNumOfVRAMModule; +}ATOM_VRAM_INFO_V2; -/* compare with V3, we flat the struct by merging ATOM_MEMORY_FORMAT (as is) into V4 as the same level */ - USHORT usEMRS2Value; /* EMRS2 Value is used for GDDR2 and GDDR4 memory type */ - USHORT usEMRS3Value; /* EMRS3 Value is used for GDDR2 and GDDR4 memory type */ - UCHAR ucMemoryVenderID; /* Predefined, If not predefined, vendor detection table gets executed */ - UCHAR ucRefreshRateFactor; /* [1:0]=RefreshFactor (00=8ms, 01=16ms, 10=32ms,11=64ms) */ - UCHAR ucFIFODepth; /* FIFO depth supposes to be detected during vendor detection, but if we dont do vendor detection we have to hardcode FIFO Depth */ - UCHAR ucCDR_Bandwidth; /* [0:3]=Read CDR bandwidth, [4:7] - Write CDR Bandwidth */ - ATOM_MEMORY_TIMING_FORMAT_V1 asMemTiming[5]; /* Memory Timing block sort from lower clock to higher clock */ -} ATOM_VRAM_MODULE_V5; - -typedef struct _ATOM_VRAM_INFO_V2 { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucNumOfVRAMModule; - ATOM_VRAM_MODULE aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; /* just for allocation, real number of blocks is in ucNumOfVRAMModule; */ -} ATOM_VRAM_INFO_V2; - -typedef struct _ATOM_VRAM_INFO_V3 { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usMemAdjustTblOffset; /* offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting */ - USHORT usMemClkPatchTblOffset; /* offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting */ - USHORT usRerseved; - UCHAR aVID_PinsShift[9]; /* 8 bit strap maximum+terminator */ - UCHAR ucNumOfVRAMModule; - ATOM_VRAM_MODULE aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; /* just for allocation, real number of blocks is in ucNumOfVRAMModule; */ - ATOM_INIT_REG_BLOCK asMemPatch; /* for allocation */ - /* ATOM_INIT_REG_BLOCK aMemAdjust; */ -} ATOM_VRAM_INFO_V3; +typedef struct _ATOM_VRAM_INFO_V3 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usMemAdjustTblOffset; // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting + USHORT usMemClkPatchTblOffset; // offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting + USHORT usRerseved; + UCHAR aVID_PinsShift[9]; // 8 bit strap maximum+terminator + UCHAR ucNumOfVRAMModule; + ATOM_VRAM_MODULE aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; // just for allocation, real number of blocks is in ucNumOfVRAMModule; + ATOM_INIT_REG_BLOCK asMemPatch; // for allocation + // ATOM_INIT_REG_BLOCK aMemAdjust; +}ATOM_VRAM_INFO_V3; #define ATOM_VRAM_INFO_LAST ATOM_VRAM_INFO_V3 -typedef struct _ATOM_VRAM_INFO_V4 { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usMemAdjustTblOffset; /* offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting */ - USHORT usMemClkPatchTblOffset; /* offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting */ - USHORT usRerseved; - UCHAR ucMemDQ7_0ByteRemap; /* DQ line byte remap, =0: Memory Data line BYTE0, =1: BYTE1, =2: BYTE2, =3: BYTE3 */ - ULONG ulMemDQ7_0BitRemap; /* each DQ line ( 7~0) use 3bits, like: DQ0=Bit[2:0], DQ1:[5:3], ... DQ7:[23:21] */ - UCHAR ucReservde[4]; - UCHAR ucNumOfVRAMModule; - ATOM_VRAM_MODULE_V4 aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; /* just for allocation, real number of blocks is in ucNumOfVRAMModule; */ - ATOM_INIT_REG_BLOCK asMemPatch; /* for allocation */ - /* ATOM_INIT_REG_BLOCK aMemAdjust; */ -} ATOM_VRAM_INFO_V4; - -typedef struct _ATOM_VRAM_GPIO_DETECTION_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR aVID_PinsShift[9]; /* 8 bit strap maximum+terminator */ -} ATOM_VRAM_GPIO_DETECTION_INFO; - -typedef struct _ATOM_MEMORY_TRAINING_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucTrainingLoop; - UCHAR ucReserved[3]; - ATOM_INIT_REG_BLOCK asMemTrainingSetting; -} ATOM_MEMORY_TRAINING_INFO; - -typedef struct SW_I2C_CNTL_DATA_PARAMETERS { - UCHAR ucControl; - UCHAR ucData; - UCHAR ucSatus; - UCHAR ucTemp; +typedef struct _ATOM_VRAM_INFO_V4 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usMemAdjustTblOffset; // offset of ATOM_INIT_REG_BLOCK structure for memory vendor specific MC adjust setting + USHORT usMemClkPatchTblOffset; // offset of ATOM_INIT_REG_BLOCK structure for memory clock specific MC setting + USHORT usRerseved; + UCHAR ucMemDQ7_0ByteRemap; // DQ line byte remap, =0: Memory Data line BYTE0, =1: BYTE1, =2: BYTE2, =3: BYTE3 + ULONG ulMemDQ7_0BitRemap; // each DQ line ( 7~0) use 3bits, like: DQ0=Bit[2:0], DQ1:[5:3], ... DQ7:[23:21] + UCHAR ucReservde[4]; + UCHAR ucNumOfVRAMModule; + ATOM_VRAM_MODULE_V4 aVramInfo[ATOM_MAX_NUMBER_OF_VRAM_MODULE]; // just for allocation, real number of blocks is in ucNumOfVRAMModule; + ATOM_INIT_REG_BLOCK asMemPatch; // for allocation + // ATOM_INIT_REG_BLOCK aMemAdjust; +}ATOM_VRAM_INFO_V4; + +typedef struct _ATOM_VRAM_GPIO_DETECTION_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR aVID_PinsShift[9]; //8 bit strap maximum+terminator +}ATOM_VRAM_GPIO_DETECTION_INFO; + + +typedef struct _ATOM_MEMORY_TRAINING_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucTrainingLoop; + UCHAR ucReserved[3]; + ATOM_INIT_REG_BLOCK asMemTrainingSetting; +}ATOM_MEMORY_TRAINING_INFO; + + +typedef struct SW_I2C_CNTL_DATA_PARAMETERS +{ + UCHAR ucControl; + UCHAR ucData; + UCHAR ucSatus; + UCHAR ucTemp; } SW_I2C_CNTL_DATA_PARAMETERS; #define SW_I2C_CNTL_DATA_PS_ALLOCATION SW_I2C_CNTL_DATA_PARAMETERS -typedef struct _SW_I2C_IO_DATA_PARAMETERS { - USHORT GPIO_Info; - UCHAR ucAct; - UCHAR ucData; -} SW_I2C_IO_DATA_PARAMETERS; +typedef struct _SW_I2C_IO_DATA_PARAMETERS +{ + USHORT GPIO_Info; + UCHAR ucAct; + UCHAR ucData; + } SW_I2C_IO_DATA_PARAMETERS; #define SW_I2C_IO_DATA_PS_ALLOCATION SW_I2C_IO_DATA_PARAMETERS @@ -4087,127 +4998,136 @@ typedef struct _SW_I2C_IO_DATA_PARAMETERS { #define SW_I2C_CNTL_CLOSE 5 #define SW_I2C_CNTL_WRITE1BIT 6 -/* ==============================VESA definition Portion=============================== */ +//==============================VESA definition Portion=============================== #define VESA_OEM_PRODUCT_REV '01.00' -#define VESA_MODE_ATTRIBUTE_MODE_SUPPORT 0xBB /* refer to VBE spec p.32, no TTY support */ +#define VESA_MODE_ATTRIBUTE_MODE_SUPPORT 0xBB //refer to VBE spec p.32, no TTY support #define VESA_MODE_WIN_ATTRIBUTE 7 #define VESA_WIN_SIZE 64 -typedef struct _PTR_32_BIT_STRUCTURE { - USHORT Offset16; - USHORT Segment16; +typedef struct _PTR_32_BIT_STRUCTURE +{ + USHORT Offset16; + USHORT Segment16; } PTR_32_BIT_STRUCTURE; -typedef union _PTR_32_BIT_UNION { - PTR_32_BIT_STRUCTURE SegmentOffset; - ULONG Ptr32_Bit; +typedef union _PTR_32_BIT_UNION +{ + PTR_32_BIT_STRUCTURE SegmentOffset; + ULONG Ptr32_Bit; } PTR_32_BIT_UNION; -typedef struct _VBE_1_2_INFO_BLOCK_UPDATABLE { - UCHAR VbeSignature[4]; - USHORT VbeVersion; - PTR_32_BIT_UNION OemStringPtr; - UCHAR Capabilities[4]; - PTR_32_BIT_UNION VideoModePtr; - USHORT TotalMemory; +typedef struct _VBE_1_2_INFO_BLOCK_UPDATABLE +{ + UCHAR VbeSignature[4]; + USHORT VbeVersion; + PTR_32_BIT_UNION OemStringPtr; + UCHAR Capabilities[4]; + PTR_32_BIT_UNION VideoModePtr; + USHORT TotalMemory; } VBE_1_2_INFO_BLOCK_UPDATABLE; -typedef struct _VBE_2_0_INFO_BLOCK_UPDATABLE { - VBE_1_2_INFO_BLOCK_UPDATABLE CommonBlock; - USHORT OemSoftRev; - PTR_32_BIT_UNION OemVendorNamePtr; - PTR_32_BIT_UNION OemProductNamePtr; - PTR_32_BIT_UNION OemProductRevPtr; + +typedef struct _VBE_2_0_INFO_BLOCK_UPDATABLE +{ + VBE_1_2_INFO_BLOCK_UPDATABLE CommonBlock; + USHORT OemSoftRev; + PTR_32_BIT_UNION OemVendorNamePtr; + PTR_32_BIT_UNION OemProductNamePtr; + PTR_32_BIT_UNION OemProductRevPtr; } VBE_2_0_INFO_BLOCK_UPDATABLE; -typedef union _VBE_VERSION_UNION { - VBE_2_0_INFO_BLOCK_UPDATABLE VBE_2_0_InfoBlock; - VBE_1_2_INFO_BLOCK_UPDATABLE VBE_1_2_InfoBlock; +typedef union _VBE_VERSION_UNION +{ + VBE_2_0_INFO_BLOCK_UPDATABLE VBE_2_0_InfoBlock; + VBE_1_2_INFO_BLOCK_UPDATABLE VBE_1_2_InfoBlock; } VBE_VERSION_UNION; -typedef struct _VBE_INFO_BLOCK { - VBE_VERSION_UNION UpdatableVBE_Info; - UCHAR Reserved[222]; - UCHAR OemData[256]; +typedef struct _VBE_INFO_BLOCK +{ + VBE_VERSION_UNION UpdatableVBE_Info; + UCHAR Reserved[222]; + UCHAR OemData[256]; } VBE_INFO_BLOCK; -typedef struct _VBE_FP_INFO { - USHORT HSize; - USHORT VSize; - USHORT FPType; - UCHAR RedBPP; - UCHAR GreenBPP; - UCHAR BlueBPP; - UCHAR ReservedBPP; - ULONG RsvdOffScrnMemSize; - ULONG RsvdOffScrnMEmPtr; - UCHAR Reserved[14]; +typedef struct _VBE_FP_INFO +{ + USHORT HSize; + USHORT VSize; + USHORT FPType; + UCHAR RedBPP; + UCHAR GreenBPP; + UCHAR BlueBPP; + UCHAR ReservedBPP; + ULONG RsvdOffScrnMemSize; + ULONG RsvdOffScrnMEmPtr; + UCHAR Reserved[14]; } VBE_FP_INFO; -typedef struct _VESA_MODE_INFO_BLOCK { -/* Mandatory information for all VBE revisions */ - USHORT ModeAttributes; /* dw ? ; mode attributes */ - UCHAR WinAAttributes; /* db ? ; window A attributes */ - UCHAR WinBAttributes; /* db ? ; window B attributes */ - USHORT WinGranularity; /* dw ? ; window granularity */ - USHORT WinSize; /* dw ? ; window size */ - USHORT WinASegment; /* dw ? ; window A start segment */ - USHORT WinBSegment; /* dw ? ; window B start segment */ - ULONG WinFuncPtr; /* dd ? ; real mode pointer to window function */ - USHORT BytesPerScanLine; /* dw ? ; bytes per scan line */ - -/* ; Mandatory information for VBE 1.2 and above */ - USHORT XResolution; /* dw ? ; horizontal resolution in pixels or characters */ - USHORT YResolution; /* dw ? ; vertical resolution in pixels or characters */ - UCHAR XCharSize; /* db ? ; character cell width in pixels */ - UCHAR YCharSize; /* db ? ; character cell height in pixels */ - UCHAR NumberOfPlanes; /* db ? ; number of memory planes */ - UCHAR BitsPerPixel; /* db ? ; bits per pixel */ - UCHAR NumberOfBanks; /* db ? ; number of banks */ - UCHAR MemoryModel; /* db ? ; memory model type */ - UCHAR BankSize; /* db ? ; bank size in KB */ - UCHAR NumberOfImagePages; /* db ? ; number of images */ - UCHAR ReservedForPageFunction; /* db 1 ; reserved for page function */ - -/* ; Direct Color fields(required for direct/6 and YUV/7 memory models) */ - UCHAR RedMaskSize; /* db ? ; size of direct color red mask in bits */ - UCHAR RedFieldPosition; /* db ? ; bit position of lsb of red mask */ - UCHAR GreenMaskSize; /* db ? ; size of direct color green mask in bits */ - UCHAR GreenFieldPosition; /* db ? ; bit position of lsb of green mask */ - UCHAR BlueMaskSize; /* db ? ; size of direct color blue mask in bits */ - UCHAR BlueFieldPosition; /* db ? ; bit position of lsb of blue mask */ - UCHAR RsvdMaskSize; /* db ? ; size of direct color reserved mask in bits */ - UCHAR RsvdFieldPosition; /* db ? ; bit position of lsb of reserved mask */ - UCHAR DirectColorModeInfo; /* db ? ; direct color mode attributes */ - -/* ; Mandatory information for VBE 2.0 and above */ - ULONG PhysBasePtr; /* dd ? ; physical address for flat memory frame buffer */ - ULONG Reserved_1; /* dd 0 ; reserved - always set to 0 */ - USHORT Reserved_2; /* dw 0 ; reserved - always set to 0 */ - -/* ; Mandatory information for VBE 3.0 and above */ - USHORT LinBytesPerScanLine; /* dw ? ; bytes per scan line for linear modes */ - UCHAR BnkNumberOfImagePages; /* db ? ; number of images for banked modes */ - UCHAR LinNumberOfImagPages; /* db ? ; number of images for linear modes */ - UCHAR LinRedMaskSize; /* db ? ; size of direct color red mask(linear modes) */ - UCHAR LinRedFieldPosition; /* db ? ; bit position of lsb of red mask(linear modes) */ - UCHAR LinGreenMaskSize; /* db ? ; size of direct color green mask(linear modes) */ - UCHAR LinGreenFieldPosition; /* db ? ; bit position of lsb of green mask(linear modes) */ - UCHAR LinBlueMaskSize; /* db ? ; size of direct color blue mask(linear modes) */ - UCHAR LinBlueFieldPosition; /* db ? ; bit position of lsb of blue mask(linear modes) */ - UCHAR LinRsvdMaskSize; /* db ? ; size of direct color reserved mask(linear modes) */ - UCHAR LinRsvdFieldPosition; /* db ? ; bit position of lsb of reserved mask(linear modes) */ - ULONG MaxPixelClock; /* dd ? ; maximum pixel clock(in Hz) for graphics mode */ - UCHAR Reserved; /* db 190 dup (0) */ +typedef struct _VESA_MODE_INFO_BLOCK +{ +// Mandatory information for all VBE revisions + USHORT ModeAttributes; // dw ? ; mode attributes + UCHAR WinAAttributes; // db ? ; window A attributes + UCHAR WinBAttributes; // db ? ; window B attributes + USHORT WinGranularity; // dw ? ; window granularity + USHORT WinSize; // dw ? ; window size + USHORT WinASegment; // dw ? ; window A start segment + USHORT WinBSegment; // dw ? ; window B start segment + ULONG WinFuncPtr; // dd ? ; real mode pointer to window function + USHORT BytesPerScanLine;// dw ? ; bytes per scan line + +//; Mandatory information for VBE 1.2 and above + USHORT XResolution; // dw ? ; horizontal resolution in pixels or characters + USHORT YResolution; // dw ? ; vertical resolution in pixels or characters + UCHAR XCharSize; // db ? ; character cell width in pixels + UCHAR YCharSize; // db ? ; character cell height in pixels + UCHAR NumberOfPlanes; // db ? ; number of memory planes + UCHAR BitsPerPixel; // db ? ; bits per pixel + UCHAR NumberOfBanks; // db ? ; number of banks + UCHAR MemoryModel; // db ? ; memory model type + UCHAR BankSize; // db ? ; bank size in KB + UCHAR NumberOfImagePages;// db ? ; number of images + UCHAR ReservedForPageFunction;//db 1 ; reserved for page function + +//; Direct Color fields(required for direct/6 and YUV/7 memory models) + UCHAR RedMaskSize; // db ? ; size of direct color red mask in bits + UCHAR RedFieldPosition; // db ? ; bit position of lsb of red mask + UCHAR GreenMaskSize; // db ? ; size of direct color green mask in bits + UCHAR GreenFieldPosition; // db ? ; bit position of lsb of green mask + UCHAR BlueMaskSize; // db ? ; size of direct color blue mask in bits + UCHAR BlueFieldPosition; // db ? ; bit position of lsb of blue mask + UCHAR RsvdMaskSize; // db ? ; size of direct color reserved mask in bits + UCHAR RsvdFieldPosition; // db ? ; bit position of lsb of reserved mask + UCHAR DirectColorModeInfo;// db ? ; direct color mode attributes + +//; Mandatory information for VBE 2.0 and above + ULONG PhysBasePtr; // dd ? ; physical address for flat memory frame buffer + ULONG Reserved_1; // dd 0 ; reserved - always set to 0 + USHORT Reserved_2; // dw 0 ; reserved - always set to 0 + +//; Mandatory information for VBE 3.0 and above + USHORT LinBytesPerScanLine; // dw ? ; bytes per scan line for linear modes + UCHAR BnkNumberOfImagePages;// db ? ; number of images for banked modes + UCHAR LinNumberOfImagPages; // db ? ; number of images for linear modes + UCHAR LinRedMaskSize; // db ? ; size of direct color red mask(linear modes) + UCHAR LinRedFieldPosition; // db ? ; bit position of lsb of red mask(linear modes) + UCHAR LinGreenMaskSize; // db ? ; size of direct color green mask(linear modes) + UCHAR LinGreenFieldPosition;// db ? ; bit position of lsb of green mask(linear modes) + UCHAR LinBlueMaskSize; // db ? ; size of direct color blue mask(linear modes) + UCHAR LinBlueFieldPosition; // db ? ; bit position of lsb of blue mask(linear modes) + UCHAR LinRsvdMaskSize; // db ? ; size of direct color reserved mask(linear modes) + UCHAR LinRsvdFieldPosition; // db ? ; bit position of lsb of reserved mask(linear modes) + ULONG MaxPixelClock; // dd ? ; maximum pixel clock(in Hz) for graphics mode + UCHAR Reserved; // db 190 dup (0) } VESA_MODE_INFO_BLOCK; -/* BIOS function CALLS */ -#define ATOM_BIOS_EXTENDED_FUNCTION_CODE 0xA0 /* ATI Extended Function code */ +// BIOS function CALLS +#define ATOM_BIOS_EXTENDED_FUNCTION_CODE 0xA0 // ATI Extended Function code #define ATOM_BIOS_FUNCTION_COP_MODE 0x00 #define ATOM_BIOS_FUNCTION_SHORT_QUERY1 0x04 #define ATOM_BIOS_FUNCTION_SHORT_QUERY2 0x05 #define ATOM_BIOS_FUNCTION_SHORT_QUERY3 0x06 -#define ATOM_BIOS_FUNCTION_GET_DDC 0x0B +#define ATOM_BIOS_FUNCTION_GET_DDC 0x0B #define ATOM_BIOS_FUNCTION_ASIC_DSTATE 0x0E #define ATOM_BIOS_FUNCTION_DEBUG_PLAY 0x0F #define ATOM_BIOS_FUNCTION_STV_STD 0x16 @@ -4217,100 +5137,135 @@ typedef struct _VESA_MODE_INFO_BLOCK { #define ATOM_BIOS_FUNCTION_PANEL_CONTROL 0x82 #define ATOM_BIOS_FUNCTION_OLD_DEVICE_DET 0x83 #define ATOM_BIOS_FUNCTION_OLD_DEVICE_SWITCH 0x84 -#define ATOM_BIOS_FUNCTION_HW_ICON 0x8A +#define ATOM_BIOS_FUNCTION_HW_ICON 0x8A #define ATOM_BIOS_FUNCTION_SET_CMOS 0x8B -#define SUB_FUNCTION_UPDATE_DISPLAY_INFO 0x8000 /* Sub function 80 */ -#define SUB_FUNCTION_UPDATE_EXPANSION_INFO 0x8100 /* Sub function 80 */ +#define SUB_FUNCTION_UPDATE_DISPLAY_INFO 0x8000 // Sub function 80 +#define SUB_FUNCTION_UPDATE_EXPANSION_INFO 0x8100 // Sub function 80 #define ATOM_BIOS_FUNCTION_DISPLAY_INFO 0x8D #define ATOM_BIOS_FUNCTION_DEVICE_ON_OFF 0x8E -#define ATOM_BIOS_FUNCTION_VIDEO_STATE 0x8F -#define ATOM_SUB_FUNCTION_GET_CRITICAL_STATE 0x0300 /* Sub function 03 */ -#define ATOM_SUB_FUNCTION_GET_LIDSTATE 0x0700 /* Sub function 7 */ -#define ATOM_SUB_FUNCTION_THERMAL_STATE_NOTICE 0x1400 /* Notify caller the current thermal state */ -#define ATOM_SUB_FUNCTION_CRITICAL_STATE_NOTICE 0x8300 /* Notify caller the current critical state */ -#define ATOM_SUB_FUNCTION_SET_LIDSTATE 0x8500 /* Sub function 85 */ -#define ATOM_SUB_FUNCTION_GET_REQ_DISPLAY_FROM_SBIOS_MODE 0x8900 /* Sub function 89 */ -#define ATOM_SUB_FUNCTION_INFORM_ADC_SUPPORT 0x9400 /* Notify caller that ADC is supported */ - -#define ATOM_BIOS_FUNCTION_VESA_DPMS 0x4F10 /* Set DPMS */ -#define ATOM_SUB_FUNCTION_SET_DPMS 0x0001 /* BL: Sub function 01 */ -#define ATOM_SUB_FUNCTION_GET_DPMS 0x0002 /* BL: Sub function 02 */ -#define ATOM_PARAMETER_VESA_DPMS_ON 0x0000 /* BH Parameter for DPMS ON. */ -#define ATOM_PARAMETER_VESA_DPMS_STANDBY 0x0100 /* BH Parameter for DPMS STANDBY */ -#define ATOM_PARAMETER_VESA_DPMS_SUSPEND 0x0200 /* BH Parameter for DPMS SUSPEND */ -#define ATOM_PARAMETER_VESA_DPMS_OFF 0x0400 /* BH Parameter for DPMS OFF */ -#define ATOM_PARAMETER_VESA_DPMS_REDUCE_ON 0x0800 /* BH Parameter for DPMS REDUCE ON (NOT SUPPORTED) */ +#define ATOM_BIOS_FUNCTION_VIDEO_STATE 0x8F +#define ATOM_SUB_FUNCTION_GET_CRITICAL_STATE 0x0300 // Sub function 03 +#define ATOM_SUB_FUNCTION_GET_LIDSTATE 0x0700 // Sub function 7 +#define ATOM_SUB_FUNCTION_THERMAL_STATE_NOTICE 0x1400 // Notify caller the current thermal state +#define ATOM_SUB_FUNCTION_CRITICAL_STATE_NOTICE 0x8300 // Notify caller the current critical state +#define ATOM_SUB_FUNCTION_SET_LIDSTATE 0x8500 // Sub function 85 +#define ATOM_SUB_FUNCTION_GET_REQ_DISPLAY_FROM_SBIOS_MODE 0x8900// Sub function 89 +#define ATOM_SUB_FUNCTION_INFORM_ADC_SUPPORT 0x9400 // Notify caller that ADC is supported + + +#define ATOM_BIOS_FUNCTION_VESA_DPMS 0x4F10 // Set DPMS +#define ATOM_SUB_FUNCTION_SET_DPMS 0x0001 // BL: Sub function 01 +#define ATOM_SUB_FUNCTION_GET_DPMS 0x0002 // BL: Sub function 02 +#define ATOM_PARAMETER_VESA_DPMS_ON 0x0000 // BH Parameter for DPMS ON. +#define ATOM_PARAMETER_VESA_DPMS_STANDBY 0x0100 // BH Parameter for DPMS STANDBY +#define ATOM_PARAMETER_VESA_DPMS_SUSPEND 0x0200 // BH Parameter for DPMS SUSPEND +#define ATOM_PARAMETER_VESA_DPMS_OFF 0x0400 // BH Parameter for DPMS OFF +#define ATOM_PARAMETER_VESA_DPMS_REDUCE_ON 0x0800 // BH Parameter for DPMS REDUCE ON (NOT SUPPORTED) #define ATOM_BIOS_RETURN_CODE_MASK 0x0000FF00L #define ATOM_BIOS_REG_HIGH_MASK 0x0000FF00L #define ATOM_BIOS_REG_LOW_MASK 0x000000FFL -/* structure used for VBIOS only */ +// structure used for VBIOS only -/* DispOutInfoTable */ -typedef struct _ASIC_TRANSMITTER_INFO { +//DispOutInfoTable +typedef struct _ASIC_TRANSMITTER_INFO +{ USHORT usTransmitterObjId; USHORT usSupportDevice; - UCHAR ucTransmitterCmdTblId; - UCHAR ucConfig; - UCHAR ucEncoderID; /* available 1st encoder ( default ) */ - UCHAR ucOptionEncoderID; /* available 2nd encoder ( optional ) */ - UCHAR uc2ndEncoderID; - UCHAR ucReserved; -} ASIC_TRANSMITTER_INFO; - -typedef struct _ASIC_ENCODER_INFO { + UCHAR ucTransmitterCmdTblId; + UCHAR ucConfig; + UCHAR ucEncoderID; //available 1st encoder ( default ) + UCHAR ucOptionEncoderID; //available 2nd encoder ( optional ) + UCHAR uc2ndEncoderID; + UCHAR ucReserved; +}ASIC_TRANSMITTER_INFO; + +typedef struct _ASIC_ENCODER_INFO +{ UCHAR ucEncoderID; UCHAR ucEncoderConfig; - USHORT usEncoderCmdTblId; -} ASIC_ENCODER_INFO; + USHORT usEncoderCmdTblId; +}ASIC_ENCODER_INFO; + +typedef struct _ATOM_DISP_OUT_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT ptrTransmitterInfo; + USHORT ptrEncoderInfo; + ASIC_TRANSMITTER_INFO asTransmitterInfo[1]; + ASIC_ENCODER_INFO asEncoderInfo[1]; +}ATOM_DISP_OUT_INFO; -typedef struct _ATOM_DISP_OUT_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; +typedef struct _ATOM_DISP_OUT_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; USHORT ptrTransmitterInfo; USHORT ptrEncoderInfo; - ASIC_TRANSMITTER_INFO asTransmitterInfo[1]; - ASIC_ENCODER_INFO asEncoderInfo[1]; -} ATOM_DISP_OUT_INFO; + USHORT ptrMainCallParserFar; // direct address of main parser call in VBIOS binary. + ASIC_TRANSMITTER_INFO asTransmitterInfo[1]; + ASIC_ENCODER_INFO asEncoderInfo[1]; +}ATOM_DISP_OUT_INFO_V2; -/* DispDevicePriorityInfo */ -typedef struct _ATOM_DISPLAY_DEVICE_PRIORITY_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; +// DispDevicePriorityInfo +typedef struct _ATOM_DISPLAY_DEVICE_PRIORITY_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; USHORT asDevicePriority[16]; -} ATOM_DISPLAY_DEVICE_PRIORITY_INFO; - -/* ProcessAuxChannelTransactionTable */ -typedef struct _PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS { - USHORT lpAuxRequest; - USHORT lpDataOut; - UCHAR ucChannelID; - union { - UCHAR ucReplyStatus; - UCHAR ucDelay; +}ATOM_DISPLAY_DEVICE_PRIORITY_INFO; + +//ProcessAuxChannelTransactionTable +typedef struct _PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS +{ + USHORT lpAuxRequest; + USHORT lpDataOut; + UCHAR ucChannelID; + union + { + UCHAR ucReplyStatus; + UCHAR ucDelay; + }; + UCHAR ucDataOutLen; + UCHAR ucReserved; +}PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS; + +//ProcessAuxChannelTransactionTable +typedef struct _PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS_V2 +{ + USHORT lpAuxRequest; + USHORT lpDataOut; + UCHAR ucChannelID; + union + { + UCHAR ucReplyStatus; + UCHAR ucDelay; }; - UCHAR ucDataOutLen; - UCHAR ucReserved; -} PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS; + UCHAR ucDataOutLen; + UCHAR ucHPD_ID; //=0: HPD1, =1: HPD2, =2: HPD3, =3: HPD4, =4: HPD5, =5: HPD6 +}PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS_V2; #define PROCESS_AUX_CHANNEL_TRANSACTION_PS_ALLOCATION PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS -/* GetSinkType */ +//GetSinkType -typedef struct _DP_ENCODER_SERVICE_PARAMETERS { +typedef struct _DP_ENCODER_SERVICE_PARAMETERS +{ USHORT ucLinkClock; - union { - UCHAR ucConfig; /* for DP training command */ - UCHAR ucI2cId; /* use for GET_SINK_TYPE command */ + union + { + UCHAR ucConfig; // for DP training command + UCHAR ucI2cId; // use for GET_SINK_TYPE command }; UCHAR ucAction; UCHAR ucStatus; UCHAR ucLaneNum; UCHAR ucReserved[2]; -} DP_ENCODER_SERVICE_PARAMETERS; +}DP_ENCODER_SERVICE_PARAMETERS; -/* ucAction */ +// ucAction #define ATOM_DP_ACTION_GET_SINK_TYPE 0x01 +/* obselete */ #define ATOM_DP_ACTION_TRAINING_START 0x02 #define ATOM_DP_ACTION_TRAINING_COMPLETE 0x03 #define ATOM_DP_ACTION_TRAINING_PATTERN_SEL 0x04 @@ -4318,7 +5273,7 @@ typedef struct _DP_ENCODER_SERVICE_PARAMETERS { #define ATOM_DP_ACTION_GET_VSWING_PREEMP 0x06 #define ATOM_DP_ACTION_BLANKING 0x07 -/* ucConfig */ +// ucConfig #define ATOM_DP_CONFIG_ENCODER_SEL_MASK 0x03 #define ATOM_DP_CONFIG_DIG1_ENCODER 0x00 #define ATOM_DP_CONFIG_DIG2_ENCODER 0x01 @@ -4326,14 +5281,14 @@ typedef struct _DP_ENCODER_SERVICE_PARAMETERS { #define ATOM_DP_CONFIG_LINK_SEL_MASK 0x04 #define ATOM_DP_CONFIG_LINK_A 0x00 #define ATOM_DP_CONFIG_LINK_B 0x04 - +/* /obselete */ #define DP_ENCODER_SERVICE_PS_ALLOCATION WRITE_ONE_BYTE_HW_I2C_DATA_PARAMETERS -/* DP_TRAINING_TABLE */ -#define DPCD_SET_LINKRATE_LANENUM_PATTERN1_TBL_ADDR ATOM_DP_TRAINING_TBL_ADDR +// DP_TRAINING_TABLE +#define DPCD_SET_LINKRATE_LANENUM_PATTERN1_TBL_ADDR ATOM_DP_TRAINING_TBL_ADDR #define DPCD_SET_SS_CNTL_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 8 ) -#define DPCD_SET_LANE_VSWING_PREEMP_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 16) -#define DPCD_SET_TRAINING_PATTERN0_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 24) +#define DPCD_SET_LANE_VSWING_PREEMP_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 16 ) +#define DPCD_SET_TRAINING_PATTERN0_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 24 ) #define DPCD_SET_TRAINING_PATTERN2_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 32) #define DPCD_GET_LINKRATE_LANENUM_SS_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 40) #define DPCD_GET_LANE_STATUS_ADJUST_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 48) @@ -4341,183 +5296,241 @@ typedef struct _DP_ENCODER_SERVICE_PARAMETERS { #define DP_I2C_AUX_DDC_WRITE_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 64) #define DP_I2C_AUX_DDC_READ_START_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 72) #define DP_I2C_AUX_DDC_READ_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 76) -#define DP_I2C_AUX_DDC_READ_END_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 80) +#define DP_I2C_AUX_DDC_WRITE_END_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 80) +#define DP_I2C_AUX_DDC_READ_END_TBL_ADDR (ATOM_DP_TRAINING_TBL_ADDR + 84) -typedef struct _PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS { - UCHAR ucI2CSpeed; - union { - UCHAR ucRegIndex; - UCHAR ucStatus; +typedef struct _PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS +{ + UCHAR ucI2CSpeed; + union + { + UCHAR ucRegIndex; + UCHAR ucStatus; }; - USHORT lpI2CDataOut; - UCHAR ucFlag; - UCHAR ucTransBytes; - UCHAR ucSlaveAddr; - UCHAR ucLineNumber; -} PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS; + USHORT lpI2CDataOut; + UCHAR ucFlag; + UCHAR ucTransBytes; + UCHAR ucSlaveAddr; + UCHAR ucLineNumber; +}PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS; #define PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION PROCESS_I2C_CHANNEL_TRANSACTION_PARAMETERS -/* ucFlag */ +//ucFlag #define HW_I2C_WRITE 1 #define HW_I2C_READ 0 +#define I2C_2BYTE_ADDR 0x02 +typedef struct _SET_HWBLOCK_INSTANCE_PARAMETER_V2 +{ + UCHAR ucHWBlkInst; // HW block instance, 0, 1, 2, ... + UCHAR ucReserved[3]; +}SET_HWBLOCK_INSTANCE_PARAMETER_V2; + +#define HWBLKINST_INSTANCE_MASK 0x07 +#define HWBLKINST_HWBLK_MASK 0xF0 +#define HWBLKINST_HWBLK_SHIFT 0x04 + +//ucHWBlock +#define SELECT_DISP_ENGINE 0 +#define SELECT_DISP_PLL 1 +#define SELECT_DCIO_UNIPHY_LINK0 2 +#define SELECT_DCIO_UNIPHY_LINK1 3 +#define SELECT_DCIO_IMPCAL 4 +#define SELECT_DCIO_DIG 6 +#define SELECT_CRTC_PIXEL_RATE 7 + +/****************************************************************************/ +//Portion VI: Definitinos for vbios MC scratch registers that driver used /****************************************************************************/ -/* Portion VI: Definitinos being oboselete */ + +#define MC_MISC0__MEMORY_TYPE_MASK 0xF0000000 +#define MC_MISC0__MEMORY_TYPE__GDDR1 0x10000000 +#define MC_MISC0__MEMORY_TYPE__DDR2 0x20000000 +#define MC_MISC0__MEMORY_TYPE__GDDR3 0x30000000 +#define MC_MISC0__MEMORY_TYPE__GDDR4 0x40000000 +#define MC_MISC0__MEMORY_TYPE__GDDR5 0x50000000 +#define MC_MISC0__MEMORY_TYPE__DDR3 0xB0000000 + +/****************************************************************************/ +//Portion VI: Definitinos being oboselete /****************************************************************************/ -/* ========================================================================================== */ -/* Remove the definitions below when driver is ready! */ -typedef struct _ATOM_DAC_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usMaxFrequency; /* in 10kHz unit */ - USHORT usReserved; -} ATOM_DAC_INFO; - -typedef struct _COMPASSIONATE_DATA { - ATOM_COMMON_TABLE_HEADER sHeader; - - /* ============================== DAC1 portion */ - UCHAR ucDAC1_BG_Adjustment; - UCHAR ucDAC1_DAC_Adjustment; - USHORT usDAC1_FORCE_Data; - /* ============================== DAC2 portion */ - UCHAR ucDAC2_CRT2_BG_Adjustment; - UCHAR ucDAC2_CRT2_DAC_Adjustment; - USHORT usDAC2_CRT2_FORCE_Data; - USHORT usDAC2_CRT2_MUX_RegisterIndex; - UCHAR ucDAC2_CRT2_MUX_RegisterInfo; /* Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low */ - UCHAR ucDAC2_NTSC_BG_Adjustment; - UCHAR ucDAC2_NTSC_DAC_Adjustment; - USHORT usDAC2_TV1_FORCE_Data; - USHORT usDAC2_TV1_MUX_RegisterIndex; - UCHAR ucDAC2_TV1_MUX_RegisterInfo; /* Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low */ - UCHAR ucDAC2_CV_BG_Adjustment; - UCHAR ucDAC2_CV_DAC_Adjustment; - USHORT usDAC2_CV_FORCE_Data; - USHORT usDAC2_CV_MUX_RegisterIndex; - UCHAR ucDAC2_CV_MUX_RegisterInfo; /* Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low */ - UCHAR ucDAC2_PAL_BG_Adjustment; - UCHAR ucDAC2_PAL_DAC_Adjustment; - USHORT usDAC2_TV2_FORCE_Data; -} COMPASSIONATE_DATA; +//========================================================================================== +//Remove the definitions below when driver is ready! +typedef struct _ATOM_DAC_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usMaxFrequency; // in 10kHz unit + USHORT usReserved; +}ATOM_DAC_INFO; + + +typedef struct _COMPASSIONATE_DATA +{ + ATOM_COMMON_TABLE_HEADER sHeader; + + //============================== DAC1 portion + UCHAR ucDAC1_BG_Adjustment; + UCHAR ucDAC1_DAC_Adjustment; + USHORT usDAC1_FORCE_Data; + //============================== DAC2 portion + UCHAR ucDAC2_CRT2_BG_Adjustment; + UCHAR ucDAC2_CRT2_DAC_Adjustment; + USHORT usDAC2_CRT2_FORCE_Data; + USHORT usDAC2_CRT2_MUX_RegisterIndex; + UCHAR ucDAC2_CRT2_MUX_RegisterInfo; //Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low + UCHAR ucDAC2_NTSC_BG_Adjustment; + UCHAR ucDAC2_NTSC_DAC_Adjustment; + USHORT usDAC2_TV1_FORCE_Data; + USHORT usDAC2_TV1_MUX_RegisterIndex; + UCHAR ucDAC2_TV1_MUX_RegisterInfo; //Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low + UCHAR ucDAC2_CV_BG_Adjustment; + UCHAR ucDAC2_CV_DAC_Adjustment; + USHORT usDAC2_CV_FORCE_Data; + USHORT usDAC2_CV_MUX_RegisterIndex; + UCHAR ucDAC2_CV_MUX_RegisterInfo; //Bit[4:0]=Bit position,Bit[7]=1:Active High;=0 Active Low + UCHAR ucDAC2_PAL_BG_Adjustment; + UCHAR ucDAC2_PAL_DAC_Adjustment; + USHORT usDAC2_TV2_FORCE_Data; +}COMPASSIONATE_DATA; /****************************Supported Device Info Table Definitions**********************/ -/* ucConnectInfo: */ -/* [7:4] - connector type */ -/* = 1 - VGA connector */ -/* = 2 - DVI-I */ -/* = 3 - DVI-D */ -/* = 4 - DVI-A */ -/* = 5 - SVIDEO */ -/* = 6 - COMPOSITE */ -/* = 7 - LVDS */ -/* = 8 - DIGITAL LINK */ -/* = 9 - SCART */ -/* = 0xA - HDMI_type A */ -/* = 0xB - HDMI_type B */ -/* = 0xE - Special case1 (DVI+DIN) */ -/* Others=TBD */ -/* [3:0] - DAC Associated */ -/* = 0 - no DAC */ -/* = 1 - DACA */ -/* = 2 - DACB */ -/* = 3 - External DAC */ -/* Others=TBD */ -/* */ - -typedef struct _ATOM_CONNECTOR_INFO { +// ucConnectInfo: +// [7:4] - connector type +// = 1 - VGA connector +// = 2 - DVI-I +// = 3 - DVI-D +// = 4 - DVI-A +// = 5 - SVIDEO +// = 6 - COMPOSITE +// = 7 - LVDS +// = 8 - DIGITAL LINK +// = 9 - SCART +// = 0xA - HDMI_type A +// = 0xB - HDMI_type B +// = 0xE - Special case1 (DVI+DIN) +// Others=TBD +// [3:0] - DAC Associated +// = 0 - no DAC +// = 1 - DACA +// = 2 - DACB +// = 3 - External DAC +// Others=TBD +// + +typedef struct _ATOM_CONNECTOR_INFO +{ #if ATOM_BIG_ENDIAN - UCHAR bfConnectorType:4; - UCHAR bfAssociatedDAC:4; + UCHAR bfConnectorType:4; + UCHAR bfAssociatedDAC:4; #else - UCHAR bfAssociatedDAC:4; - UCHAR bfConnectorType:4; + UCHAR bfAssociatedDAC:4; + UCHAR bfConnectorType:4; #endif -} ATOM_CONNECTOR_INFO; +}ATOM_CONNECTOR_INFO; + +typedef union _ATOM_CONNECTOR_INFO_ACCESS +{ + ATOM_CONNECTOR_INFO sbfAccess; + UCHAR ucAccess; +}ATOM_CONNECTOR_INFO_ACCESS; -typedef union _ATOM_CONNECTOR_INFO_ACCESS { - ATOM_CONNECTOR_INFO sbfAccess; - UCHAR ucAccess; -} ATOM_CONNECTOR_INFO_ACCESS; +typedef struct _ATOM_CONNECTOR_INFO_I2C +{ + ATOM_CONNECTOR_INFO_ACCESS sucConnectorInfo; + ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; +}ATOM_CONNECTOR_INFO_I2C; -typedef struct _ATOM_CONNECTOR_INFO_I2C { - ATOM_CONNECTOR_INFO_ACCESS sucConnectorInfo; - ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; -} ATOM_CONNECTOR_INFO_I2C; -typedef struct _ATOM_SUPPORTED_DEVICES_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usDeviceSupport; - ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO]; -} ATOM_SUPPORTED_DEVICES_INFO; +typedef struct _ATOM_SUPPORTED_DEVICES_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usDeviceSupport; + ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO]; +}ATOM_SUPPORTED_DEVICES_INFO; #define NO_INT_SRC_MAPPED 0xFF -typedef struct _ATOM_CONNECTOR_INC_SRC_BITMAP { - UCHAR ucIntSrcBitmap; -} ATOM_CONNECTOR_INC_SRC_BITMAP; - -typedef struct _ATOM_SUPPORTED_DEVICES_INFO_2 { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usDeviceSupport; - ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO_2]; - ATOM_CONNECTOR_INC_SRC_BITMAP - asIntSrcInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO_2]; -} ATOM_SUPPORTED_DEVICES_INFO_2; - -typedef struct _ATOM_SUPPORTED_DEVICES_INFO_2d1 { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usDeviceSupport; - ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE]; - ATOM_CONNECTOR_INC_SRC_BITMAP asIntSrcInfo[ATOM_MAX_SUPPORTED_DEVICE]; -} ATOM_SUPPORTED_DEVICES_INFO_2d1; +typedef struct _ATOM_CONNECTOR_INC_SRC_BITMAP +{ + UCHAR ucIntSrcBitmap; +}ATOM_CONNECTOR_INC_SRC_BITMAP; + +typedef struct _ATOM_SUPPORTED_DEVICES_INFO_2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usDeviceSupport; + ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO_2]; + ATOM_CONNECTOR_INC_SRC_BITMAP asIntSrcInfo[ATOM_MAX_SUPPORTED_DEVICE_INFO_2]; +}ATOM_SUPPORTED_DEVICES_INFO_2; + +typedef struct _ATOM_SUPPORTED_DEVICES_INFO_2d1 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usDeviceSupport; + ATOM_CONNECTOR_INFO_I2C asConnInfo[ATOM_MAX_SUPPORTED_DEVICE]; + ATOM_CONNECTOR_INC_SRC_BITMAP asIntSrcInfo[ATOM_MAX_SUPPORTED_DEVICE]; +}ATOM_SUPPORTED_DEVICES_INFO_2d1; #define ATOM_SUPPORTED_DEVICES_INFO_LAST ATOM_SUPPORTED_DEVICES_INFO_2d1 -typedef struct _ATOM_MISC_CONTROL_INFO { - USHORT usFrequency; - UCHAR ucPLL_ChargePump; /* PLL charge-pump gain control */ - UCHAR ucPLL_DutyCycle; /* PLL duty cycle control */ - UCHAR ucPLL_VCO_Gain; /* PLL VCO gain control */ - UCHAR ucPLL_VoltageSwing; /* PLL driver voltage swing control */ -} ATOM_MISC_CONTROL_INFO; + + +typedef struct _ATOM_MISC_CONTROL_INFO +{ + USHORT usFrequency; + UCHAR ucPLL_ChargePump; // PLL charge-pump gain control + UCHAR ucPLL_DutyCycle; // PLL duty cycle control + UCHAR ucPLL_VCO_Gain; // PLL VCO gain control + UCHAR ucPLL_VoltageSwing; // PLL driver voltage swing control +}ATOM_MISC_CONTROL_INFO; + #define ATOM_MAX_MISC_INFO 4 -typedef struct _ATOM_TMDS_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usMaxFrequency; /* in 10Khz */ - ATOM_MISC_CONTROL_INFO asMiscInfo[ATOM_MAX_MISC_INFO]; -} ATOM_TMDS_INFO; +typedef struct _ATOM_TMDS_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usMaxFrequency; // in 10Khz + ATOM_MISC_CONTROL_INFO asMiscInfo[ATOM_MAX_MISC_INFO]; +}ATOM_TMDS_INFO; + + +typedef struct _ATOM_ENCODER_ANALOG_ATTRIBUTE +{ + UCHAR ucTVStandard; //Same as TV standards defined above, + UCHAR ucPadding[1]; +}ATOM_ENCODER_ANALOG_ATTRIBUTE; -typedef struct _ATOM_ENCODER_ANALOG_ATTRIBUTE { - UCHAR ucTVStandard; /* Same as TV standards defined above, */ - UCHAR ucPadding[1]; -} ATOM_ENCODER_ANALOG_ATTRIBUTE; +typedef struct _ATOM_ENCODER_DIGITAL_ATTRIBUTE +{ + UCHAR ucAttribute; //Same as other digital encoder attributes defined above + UCHAR ucPadding[1]; +}ATOM_ENCODER_DIGITAL_ATTRIBUTE; -typedef struct _ATOM_ENCODER_DIGITAL_ATTRIBUTE { - UCHAR ucAttribute; /* Same as other digital encoder attributes defined above */ - UCHAR ucPadding[1]; -} ATOM_ENCODER_DIGITAL_ATTRIBUTE; +typedef union _ATOM_ENCODER_ATTRIBUTE +{ + ATOM_ENCODER_ANALOG_ATTRIBUTE sAlgAttrib; + ATOM_ENCODER_DIGITAL_ATTRIBUTE sDigAttrib; +}ATOM_ENCODER_ATTRIBUTE; -typedef union _ATOM_ENCODER_ATTRIBUTE { - ATOM_ENCODER_ANALOG_ATTRIBUTE sAlgAttrib; - ATOM_ENCODER_DIGITAL_ATTRIBUTE sDigAttrib; -} ATOM_ENCODER_ATTRIBUTE; -typedef struct _DVO_ENCODER_CONTROL_PARAMETERS { - USHORT usPixelClock; - USHORT usEncoderID; - UCHAR ucDeviceType; /* Use ATOM_DEVICE_xxx1_Index to indicate device type only. */ - UCHAR ucAction; /* ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT */ - ATOM_ENCODER_ATTRIBUTE usDevAttr; -} DVO_ENCODER_CONTROL_PARAMETERS; +typedef struct _DVO_ENCODER_CONTROL_PARAMETERS +{ + USHORT usPixelClock; + USHORT usEncoderID; + UCHAR ucDeviceType; //Use ATOM_DEVICE_xxx1_Index to indicate device type only. + UCHAR ucAction; //ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT + ATOM_ENCODER_ATTRIBUTE usDevAttr; +}DVO_ENCODER_CONTROL_PARAMETERS; + +typedef struct _DVO_ENCODER_CONTROL_PS_ALLOCATION +{ + DVO_ENCODER_CONTROL_PARAMETERS sDVOEncoder; + WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; //Caller doesn't need to init this portion +}DVO_ENCODER_CONTROL_PS_ALLOCATION; -typedef struct _DVO_ENCODER_CONTROL_PS_ALLOCATION { - DVO_ENCODER_CONTROL_PARAMETERS sDVOEncoder; - WRITE_ONE_BYTE_HW_I2C_DATA_PS_ALLOCATION sReserved; /* Caller doesn't need to init this portion */ -} DVO_ENCODER_CONTROL_PS_ALLOCATION; #define ATOM_XTMDS_ASIC_SI164_ID 1 #define ATOM_XTMDS_ASIC_SI178_ID 2 @@ -4526,27 +5539,30 @@ typedef struct _DVO_ENCODER_CONTROL_PS_ALLOCATION { #define ATOM_XTMDS_SUPPORTED_DUALLINK 0x00000002 #define ATOM_XTMDS_MVPU_FPGA 0x00000004 -typedef struct _ATOM_XTMDS_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - USHORT usSingleLinkMaxFrequency; - ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; /* Point the ID on which I2C is used to control external chip */ - UCHAR ucXtransimitterID; - UCHAR ucSupportedLink; /* Bit field, bit0=1, single link supported;bit1=1,dual link supported */ - UCHAR ucSequnceAlterID; /* Even with the same external TMDS asic, it's possible that the program seqence alters */ - /* due to design. This ID is used to alert driver that the sequence is not "standard"! */ - UCHAR ucMasterAddress; /* Address to control Master xTMDS Chip */ - UCHAR ucSlaveAddress; /* Address to control Slave xTMDS Chip */ -} ATOM_XTMDS_INFO; - -typedef struct _DFP_DPMS_STATUS_CHANGE_PARAMETERS { - UCHAR ucEnable; /* ATOM_ENABLE=On or ATOM_DISABLE=Off */ - UCHAR ucDevice; /* ATOM_DEVICE_DFP1_INDEX.... */ - UCHAR ucPadding[2]; -} DFP_DPMS_STATUS_CHANGE_PARAMETERS; + +typedef struct _ATOM_XTMDS_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + USHORT usSingleLinkMaxFrequency; + ATOM_I2C_ID_CONFIG_ACCESS sucI2cId; //Point the ID on which I2C is used to control external chip + UCHAR ucXtransimitterID; + UCHAR ucSupportedLink; // Bit field, bit0=1, single link supported;bit1=1,dual link supported + UCHAR ucSequnceAlterID; // Even with the same external TMDS asic, it's possible that the program seqence alters + // due to design. This ID is used to alert driver that the sequence is not "standard"! + UCHAR ucMasterAddress; // Address to control Master xTMDS Chip + UCHAR ucSlaveAddress; // Address to control Slave xTMDS Chip +}ATOM_XTMDS_INFO; + +typedef struct _DFP_DPMS_STATUS_CHANGE_PARAMETERS +{ + UCHAR ucEnable; // ATOM_ENABLE=On or ATOM_DISABLE=Off + UCHAR ucDevice; // ATOM_DEVICE_DFP1_INDEX.... + UCHAR ucPadding[2]; +}DFP_DPMS_STATUS_CHANGE_PARAMETERS; /****************************Legacy Power Play Table Definitions **********************/ -/* Definitions for ulPowerPlayMiscInfo */ +//Definitions for ulPowerPlayMiscInfo #define ATOM_PM_MISCINFO_SPLIT_CLOCK 0x00000000L #define ATOM_PM_MISCINFO_USING_MCLK_SRC 0x00000001L #define ATOM_PM_MISCINFO_USING_SCLK_SRC 0x00000002L @@ -4558,8 +5574,8 @@ typedef struct _DFP_DPMS_STATUS_CHANGE_PARAMETERS { #define ATOM_PM_MISCINFO_ENGINE_CLOCK_CONTRL_EN 0x00000020L #define ATOM_PM_MISCINFO_MEMORY_CLOCK_CONTRL_EN 0x00000040L -#define ATOM_PM_MISCINFO_PROGRAM_VOLTAGE 0x00000080L /* When this bit set, ucVoltageDropIndex is not an index for GPIO pin, but a voltage ID that SW needs program */ - +#define ATOM_PM_MISCINFO_PROGRAM_VOLTAGE 0x00000080L //When this bit set, ucVoltageDropIndex is not an index for GPIO pin, but a voltage ID that SW needs program + #define ATOM_PM_MISCINFO_ASIC_REDUCED_SPEED_SCLK_EN 0x00000100L #define ATOM_PM_MISCINFO_ASIC_DYNAMIC_VOLTAGE_EN 0x00000200L #define ATOM_PM_MISCINFO_ASIC_SLEEP_MODE_EN 0x00000400L @@ -4569,22 +5585,22 @@ typedef struct _DFP_DPMS_STATUS_CHANGE_PARAMETERS { #define ATOM_PM_MISCINFO_LOW_LCD_REFRESH_RATE 0x00004000L #define ATOM_PM_MISCINFO_DRIVER_DEFAULT_MODE 0x00008000L -#define ATOM_PM_MISCINFO_OVER_CLOCK_MODE 0x00010000L +#define ATOM_PM_MISCINFO_OVER_CLOCK_MODE 0x00010000L #define ATOM_PM_MISCINFO_OVER_DRIVE_MODE 0x00020000L #define ATOM_PM_MISCINFO_POWER_SAVING_MODE 0x00040000L #define ATOM_PM_MISCINFO_THERMAL_DIODE_MODE 0x00080000L -#define ATOM_PM_MISCINFO_FRAME_MODULATION_MASK 0x00300000L /* 0-FM Disable, 1-2 level FM, 2-4 level FM, 3-Reserved */ -#define ATOM_PM_MISCINFO_FRAME_MODULATION_SHIFT 20 +#define ATOM_PM_MISCINFO_FRAME_MODULATION_MASK 0x00300000L //0-FM Disable, 1-2 level FM, 2-4 level FM, 3-Reserved +#define ATOM_PM_MISCINFO_FRAME_MODULATION_SHIFT 20 #define ATOM_PM_MISCINFO_DYN_CLK_3D_IDLE 0x00400000L #define ATOM_PM_MISCINFO_DYNAMIC_CLOCK_DIVIDER_BY_2 0x00800000L #define ATOM_PM_MISCINFO_DYNAMIC_CLOCK_DIVIDER_BY_4 0x01000000L -#define ATOM_PM_MISCINFO_DYNAMIC_HDP_BLOCK_EN 0x02000000L /* When set, Dynamic */ -#define ATOM_PM_MISCINFO_DYNAMIC_MC_HOST_BLOCK_EN 0x04000000L /* When set, Dynamic */ -#define ATOM_PM_MISCINFO_3D_ACCELERATION_EN 0x08000000L /* When set, This mode is for acceleated 3D mode */ +#define ATOM_PM_MISCINFO_DYNAMIC_HDP_BLOCK_EN 0x02000000L //When set, Dynamic +#define ATOM_PM_MISCINFO_DYNAMIC_MC_HOST_BLOCK_EN 0x04000000L //When set, Dynamic +#define ATOM_PM_MISCINFO_3D_ACCELERATION_EN 0x08000000L //When set, This mode is for acceleated 3D mode -#define ATOM_PM_MISCINFO_POWERPLAY_SETTINGS_GROUP_MASK 0x70000000L /* 1-Optimal Battery Life Group, 2-High Battery, 3-Balanced, 4-High Performance, 5- Optimal Performance (Default state with Default clocks) */ +#define ATOM_PM_MISCINFO_POWERPLAY_SETTINGS_GROUP_MASK 0x70000000L //1-Optimal Battery Life Group, 2-High Battery, 3-Balanced, 4-High Performance, 5- Optimal Performance (Default state with Default clocks) #define ATOM_PM_MISCINFO_POWERPLAY_SETTINGS_GROUP_SHIFT 28 #define ATOM_PM_MISCINFO_ENABLE_BACK_BIAS 0x80000000L @@ -4594,55 +5610,59 @@ typedef struct _DFP_DPMS_STATUS_CHANGE_PARAMETERS { #define ATOM_PM_MISCINFO2_FS3D_OVERDRIVE_INFO 0x00000008L #define ATOM_PM_MISCINFO2_FORCEDLOWPWR_MODE 0x00000010L #define ATOM_PM_MISCINFO2_VDDCI_DYNAMIC_VOLTAGE_EN 0x00000020L -#define ATOM_PM_MISCINFO2_VIDEO_PLAYBACK_CAPABLE 0x00000040L /* If this bit is set in multi-pp mode, then driver will pack up one with the minior power consumption. */ - /* If it's not set in any pp mode, driver will use its default logic to pick a pp mode in video playback */ +#define ATOM_PM_MISCINFO2_VIDEO_PLAYBACK_CAPABLE 0x00000040L //If this bit is set in multi-pp mode, then driver will pack up one with the minior power consumption. + //If it's not set in any pp mode, driver will use its default logic to pick a pp mode in video playback #define ATOM_PM_MISCINFO2_NOT_VALID_ON_DC 0x00000080L #define ATOM_PM_MISCINFO2_STUTTER_MODE_EN 0x00000100L -#define ATOM_PM_MISCINFO2_UVD_SUPPORT_MODE 0x00000200L - -/* ucTableFormatRevision=1 */ -/* ucTableContentRevision=1 */ -typedef struct _ATOM_POWERMODE_INFO { - ULONG ulMiscInfo; /* The power level should be arranged in ascending order */ - ULONG ulReserved1; /* must set to 0 */ - ULONG ulReserved2; /* must set to 0 */ - USHORT usEngineClock; - USHORT usMemoryClock; - UCHAR ucVoltageDropIndex; /* index to GPIO table */ - UCHAR ucSelectedPanel_RefreshRate; /* panel refresh rate */ - UCHAR ucMinTemperature; - UCHAR ucMaxTemperature; - UCHAR ucNumPciELanes; /* number of PCIE lanes */ -} ATOM_POWERMODE_INFO; - -/* ucTableFormatRevision=2 */ -/* ucTableContentRevision=1 */ -typedef struct _ATOM_POWERMODE_INFO_V2 { - ULONG ulMiscInfo; /* The power level should be arranged in ascending order */ - ULONG ulMiscInfo2; - ULONG ulEngineClock; - ULONG ulMemoryClock; - UCHAR ucVoltageDropIndex; /* index to GPIO table */ - UCHAR ucSelectedPanel_RefreshRate; /* panel refresh rate */ - UCHAR ucMinTemperature; - UCHAR ucMaxTemperature; - UCHAR ucNumPciELanes; /* number of PCIE lanes */ -} ATOM_POWERMODE_INFO_V2; - -/* ucTableFormatRevision=2 */ -/* ucTableContentRevision=2 */ -typedef struct _ATOM_POWERMODE_INFO_V3 { - ULONG ulMiscInfo; /* The power level should be arranged in ascending order */ - ULONG ulMiscInfo2; - ULONG ulEngineClock; - ULONG ulMemoryClock; - UCHAR ucVoltageDropIndex; /* index to Core (VDDC) votage table */ - UCHAR ucSelectedPanel_RefreshRate; /* panel refresh rate */ - UCHAR ucMinTemperature; - UCHAR ucMaxTemperature; - UCHAR ucNumPciELanes; /* number of PCIE lanes */ - UCHAR ucVDDCI_VoltageDropIndex; /* index to VDDCI votage table */ -} ATOM_POWERMODE_INFO_V3; +#define ATOM_PM_MISCINFO2_UVD_SUPPORT_MODE 0x00000200L + +//ucTableFormatRevision=1 +//ucTableContentRevision=1 +typedef struct _ATOM_POWERMODE_INFO +{ + ULONG ulMiscInfo; //The power level should be arranged in ascending order + ULONG ulReserved1; // must set to 0 + ULONG ulReserved2; // must set to 0 + USHORT usEngineClock; + USHORT usMemoryClock; + UCHAR ucVoltageDropIndex; // index to GPIO table + UCHAR ucSelectedPanel_RefreshRate;// panel refresh rate + UCHAR ucMinTemperature; + UCHAR ucMaxTemperature; + UCHAR ucNumPciELanes; // number of PCIE lanes +}ATOM_POWERMODE_INFO; + +//ucTableFormatRevision=2 +//ucTableContentRevision=1 +typedef struct _ATOM_POWERMODE_INFO_V2 +{ + ULONG ulMiscInfo; //The power level should be arranged in ascending order + ULONG ulMiscInfo2; + ULONG ulEngineClock; + ULONG ulMemoryClock; + UCHAR ucVoltageDropIndex; // index to GPIO table + UCHAR ucSelectedPanel_RefreshRate;// panel refresh rate + UCHAR ucMinTemperature; + UCHAR ucMaxTemperature; + UCHAR ucNumPciELanes; // number of PCIE lanes +}ATOM_POWERMODE_INFO_V2; + +//ucTableFormatRevision=2 +//ucTableContentRevision=2 +typedef struct _ATOM_POWERMODE_INFO_V3 +{ + ULONG ulMiscInfo; //The power level should be arranged in ascending order + ULONG ulMiscInfo2; + ULONG ulEngineClock; + ULONG ulMemoryClock; + UCHAR ucVoltageDropIndex; // index to Core (VDDC) votage table + UCHAR ucSelectedPanel_RefreshRate;// panel refresh rate + UCHAR ucMinTemperature; + UCHAR ucMaxTemperature; + UCHAR ucNumPciELanes; // number of PCIE lanes + UCHAR ucVDDCI_VoltageDropIndex; // index to VDDCI votage table +}ATOM_POWERMODE_INFO_V3; + #define ATOM_MAX_NUMBEROF_POWER_BLOCK 8 @@ -4655,40 +5675,44 @@ typedef struct _ATOM_POWERMODE_INFO_V3 { #define ATOM_PP_OVERDRIVE_THERMALCONTROLLER_MUA6649 0x04 #define ATOM_PP_OVERDRIVE_THERMALCONTROLLER_LM64 0x05 #define ATOM_PP_OVERDRIVE_THERMALCONTROLLER_F75375 0x06 -#define ATOM_PP_OVERDRIVE_THERMALCONTROLLER_ASC7512 0x07 /* Andigilog */ - -typedef struct _ATOM_POWERPLAY_INFO { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucOverdriveThermalController; - UCHAR ucOverdriveI2cLine; - UCHAR ucOverdriveIntBitmap; - UCHAR ucOverdriveControllerAddress; - UCHAR ucSizeOfPowerModeEntry; - UCHAR ucNumOfPowerModeEntries; - ATOM_POWERMODE_INFO asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; -} ATOM_POWERPLAY_INFO; - -typedef struct _ATOM_POWERPLAY_INFO_V2 { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucOverdriveThermalController; - UCHAR ucOverdriveI2cLine; - UCHAR ucOverdriveIntBitmap; - UCHAR ucOverdriveControllerAddress; - UCHAR ucSizeOfPowerModeEntry; - UCHAR ucNumOfPowerModeEntries; - ATOM_POWERMODE_INFO_V2 asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; -} ATOM_POWERPLAY_INFO_V2; - -typedef struct _ATOM_POWERPLAY_INFO_V3 { - ATOM_COMMON_TABLE_HEADER sHeader; - UCHAR ucOverdriveThermalController; - UCHAR ucOverdriveI2cLine; - UCHAR ucOverdriveIntBitmap; - UCHAR ucOverdriveControllerAddress; - UCHAR ucSizeOfPowerModeEntry; - UCHAR ucNumOfPowerModeEntries; - ATOM_POWERMODE_INFO_V3 asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; -} ATOM_POWERPLAY_INFO_V3; +#define ATOM_PP_OVERDRIVE_THERMALCONTROLLER_ASC7512 0x07 // Andigilog + + +typedef struct _ATOM_POWERPLAY_INFO +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucOverdriveThermalController; + UCHAR ucOverdriveI2cLine; + UCHAR ucOverdriveIntBitmap; + UCHAR ucOverdriveControllerAddress; + UCHAR ucSizeOfPowerModeEntry; + UCHAR ucNumOfPowerModeEntries; + ATOM_POWERMODE_INFO asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; +}ATOM_POWERPLAY_INFO; + +typedef struct _ATOM_POWERPLAY_INFO_V2 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucOverdriveThermalController; + UCHAR ucOverdriveI2cLine; + UCHAR ucOverdriveIntBitmap; + UCHAR ucOverdriveControllerAddress; + UCHAR ucSizeOfPowerModeEntry; + UCHAR ucNumOfPowerModeEntries; + ATOM_POWERMODE_INFO_V2 asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; +}ATOM_POWERPLAY_INFO_V2; + +typedef struct _ATOM_POWERPLAY_INFO_V3 +{ + ATOM_COMMON_TABLE_HEADER sHeader; + UCHAR ucOverdriveThermalController; + UCHAR ucOverdriveI2cLine; + UCHAR ucOverdriveIntBitmap; + UCHAR ucOverdriveControllerAddress; + UCHAR ucSizeOfPowerModeEntry; + UCHAR ucNumOfPowerModeEntries; + ATOM_POWERMODE_INFO_V3 asPowerPlayInfo[ATOM_MAX_NUMBEROF_POWER_BLOCK]; +}ATOM_POWERPLAY_INFO_V3; /* New PPlib */ /**************************************************************************/ @@ -4873,40 +5897,42 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO UCHAR ucMaxHTLinkWidth; // From SBIOS - {2, 4, 8, 16} UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requriement. USHORT usHTLinkFreq; // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200). - ULONG ulFlags; + ULONG ulFlags; } ATOM_PPLIB_RS780_CLOCK_INFO; -#define ATOM_PPLIB_RS780_VOLTAGE_NONE 0 -#define ATOM_PPLIB_RS780_VOLTAGE_LOW 1 -#define ATOM_PPLIB_RS780_VOLTAGE_HIGH 2 -#define ATOM_PPLIB_RS780_VOLTAGE_VARIABLE 3 +#define ATOM_PPLIB_RS780_VOLTAGE_NONE 0 +#define ATOM_PPLIB_RS780_VOLTAGE_LOW 1 +#define ATOM_PPLIB_RS780_VOLTAGE_HIGH 2 +#define ATOM_PPLIB_RS780_VOLTAGE_VARIABLE 3 #define ATOM_PPLIB_RS780_SPMCLK_NONE 0 // We cannot change the side port memory clock, leave it as it is. #define ATOM_PPLIB_RS780_SPMCLK_LOW 1 #define ATOM_PPLIB_RS780_SPMCLK_HIGH 2 -#define ATOM_PPLIB_RS780_HTLINKFREQ_NONE 0 -#define ATOM_PPLIB_RS780_HTLINKFREQ_LOW 1 -#define ATOM_PPLIB_RS780_HTLINKFREQ_HIGH 2 +#define ATOM_PPLIB_RS780_HTLINKFREQ_NONE 0 +#define ATOM_PPLIB_RS780_HTLINKFREQ_LOW 1 +#define ATOM_PPLIB_RS780_HTLINKFREQ_HIGH 2 /**************************************************************************/ -/* Following definitions are for compatiblity issue in different SW components. */ + +// Following definitions are for compatiblity issue in different SW components. #define ATOM_MASTER_DATA_TABLE_REVISION 0x01 -#define Object_Info Object_Header +#define Object_Info Object_Header #define AdjustARB_SEQ MC_InitParameter #define VRAM_GPIO_DetectionInfo VoltageObjectInfo -#define ASIC_VDDCI_Info ASIC_ProfilingInfo +#define ASIC_VDDCI_Info ASIC_ProfilingInfo #define ASIC_MVDDQ_Info MemoryTrainingInfo -#define SS_Info PPLL_SS_Info +#define SS_Info PPLL_SS_Info #define ASIC_MVDDC_Info ASIC_InternalSS_Info #define DispDevicePriorityInfo SaveRestoreInfo #define DispOutInfo TV_VideoMode + #define ATOM_ENCODER_OBJECT_TABLE ATOM_OBJECT_TABLE #define ATOM_CONNECTOR_OBJECT_TABLE ATOM_OBJECT_TABLE -/* New device naming, remove them when both DAL/VBIOS is ready */ +//New device naming, remove them when both DAL/VBIOS is ready #define DFP2I_OUTPUT_CONTROL_PARAMETERS CRT1_OUTPUT_CONTROL_PARAMETERS #define DFP2I_OUTPUT_CONTROL_PS_ALLOCATION DFP2I_OUTPUT_CONTROL_PARAMETERS @@ -4921,7 +5947,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO #define ATOM_DEVICE_DFP1I_INDEX ATOM_DEVICE_DFP1_INDEX #define ATOM_DEVICE_DFP1X_INDEX ATOM_DEVICE_DFP2_INDEX - + #define ATOM_DEVICE_DFP2I_INDEX 0x00000009 #define ATOM_DEVICE_DFP2I_SUPPORT (0x1L << ATOM_DEVICE_DFP2I_INDEX) @@ -4939,7 +5965,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO #define ATOM_S3_DFP2I_ACTIVEb1 0x02 -#define ATOM_S3_DFP1I_ACTIVE ATOM_S3_DFP1_ACTIVE +#define ATOM_S3_DFP1I_ACTIVE ATOM_S3_DFP1_ACTIVE #define ATOM_S3_DFP1X_ACTIVE ATOM_S3_DFP2_ACTIVE #define ATOM_S3_DFP2I_ACTIVE 0x00000200L @@ -4958,14 +5984,14 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO #define ATOM_S6_ACC_REQ_DFP2Ib3 0x02 #define ATOM_S6_ACC_REQ_DFP2I 0x02000000L -#define TMDS1XEncoderControl DVOEncoderControl +#define TMDS1XEncoderControl DVOEncoderControl #define DFP1XOutputControl DVOOutputControl #define ExternalDFPOutputControl DFP1XOutputControl #define EnableExternalTMDS_Encoder TMDS1XEncoderControl #define DFP1IOutputControl TMDSAOutputControl -#define DFP2IOutputControl LVTMAOutputControl +#define DFP2IOutputControl LVTMAOutputControl #define DAC1_ENCODER_CONTROL_PARAMETERS DAC_ENCODER_CONTROL_PARAMETERS #define DAC1_ENCODER_CONTROL_PS_ALLOCATION DAC_ENCODER_CONTROL_PS_ALLOCATION @@ -4974,7 +6000,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO #define DAC2_ENCODER_CONTROL_PS_ALLOCATION DAC_ENCODER_CONTROL_PS_ALLOCATION #define ucDac1Standard ucDacStandard -#define ucDac2Standard ucDacStandard +#define ucDac2Standard ucDacStandard #define TMDS1EncoderControl TMDSAEncoderControl #define TMDS2EncoderControl LVTMAEncoderControl @@ -4984,12 +6010,56 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO #define CRT1OutputControl DAC1OutputControl #define CRT2OutputControl DAC2OutputControl -/* These two lines will be removed for sure in a few days, will follow up with Michael V. */ +//These two lines will be removed for sure in a few days, will follow up with Michael V. #define EnableLVDS_SS EnableSpreadSpectrumOnPPLL -#define ENABLE_LVDS_SS_PARAMETERS_V3 ENABLE_SPREAD_SPECTRUM_ON_PPLL +#define ENABLE_LVDS_SS_PARAMETERS_V3 ENABLE_SPREAD_SPECTRUM_ON_PPLL + +//#define ATOM_S2_CRT1_DPMS_STATE 0x00010000L +//#define ATOM_S2_LCD1_DPMS_STATE ATOM_S2_CRT1_DPMS_STATE +//#define ATOM_S2_TV1_DPMS_STATE ATOM_S2_CRT1_DPMS_STATE +//#define ATOM_S2_DFP1_DPMS_STATE ATOM_S2_CRT1_DPMS_STATE +//#define ATOM_S2_CRT2_DPMS_STATE ATOM_S2_CRT1_DPMS_STATE + +#define ATOM_S6_ACC_REQ_TV2 0x00400000L +#define ATOM_DEVICE_TV2_INDEX 0x00000006 +#define ATOM_DEVICE_TV2_SUPPORT (0x1L << ATOM_DEVICE_TV2_INDEX) +#define ATOM_S0_TV2 0x00100000L +#define ATOM_S3_TV2_ACTIVE ATOM_S3_DFP6_ACTIVE +#define ATOM_S3_TV2_CRTC_ACTIVE ATOM_S3_DFP6_CRTC_ACTIVE + +// +#define ATOM_S2_CRT1_DPMS_STATE 0x00010000L +#define ATOM_S2_LCD1_DPMS_STATE 0x00020000L +#define ATOM_S2_TV1_DPMS_STATE 0x00040000L +#define ATOM_S2_DFP1_DPMS_STATE 0x00080000L +#define ATOM_S2_CRT2_DPMS_STATE 0x00100000L +#define ATOM_S2_LCD2_DPMS_STATE 0x00200000L +#define ATOM_S2_TV2_DPMS_STATE 0x00400000L +#define ATOM_S2_DFP2_DPMS_STATE 0x00800000L +#define ATOM_S2_CV_DPMS_STATE 0x01000000L +#define ATOM_S2_DFP3_DPMS_STATE 0x02000000L +#define ATOM_S2_DFP4_DPMS_STATE 0x04000000L +#define ATOM_S2_DFP5_DPMS_STATE 0x08000000L + +#define ATOM_S2_CRT1_DPMS_STATEb2 0x01 +#define ATOM_S2_LCD1_DPMS_STATEb2 0x02 +#define ATOM_S2_TV1_DPMS_STATEb2 0x04 +#define ATOM_S2_DFP1_DPMS_STATEb2 0x08 +#define ATOM_S2_CRT2_DPMS_STATEb2 0x10 +#define ATOM_S2_LCD2_DPMS_STATEb2 0x20 +#define ATOM_S2_TV2_DPMS_STATEb2 0x40 +#define ATOM_S2_DFP2_DPMS_STATEb2 0x80 +#define ATOM_S2_CV_DPMS_STATEb3 0x01 +#define ATOM_S2_DFP3_DPMS_STATEb3 0x02 +#define ATOM_S2_DFP4_DPMS_STATEb3 0x04 +#define ATOM_S2_DFP5_DPMS_STATEb3 0x08 + +#define ATOM_S3_ASIC_GUI_ENGINE_HUNGb3 0x20 +#define ATOM_S3_ALLOW_FAST_PWR_SWITCHb3 0x40 +#define ATOM_S3_RQST_GPU_USE_MIN_PWRb3 0x80 /*********************************************************************************/ -#pragma pack() /* BIOS data must use byte aligment */ +#pragma pack() // BIOS data must use byte aligment #endif /* _ATOMBIOS_H */ diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index 260fcf59f00..dd9fdf56061 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -245,21 +245,25 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) switch (mode) { case DRM_MODE_DPMS_ON: - atombios_enable_crtc(crtc, 1); + atombios_enable_crtc(crtc, ATOM_ENABLE); if (ASIC_IS_DCE3(rdev)) - atombios_enable_crtc_memreq(crtc, 1); - atombios_blank_crtc(crtc, 0); - drm_vblank_post_modeset(dev, radeon_crtc->crtc_id); + atombios_enable_crtc_memreq(crtc, ATOM_ENABLE); + atombios_blank_crtc(crtc, ATOM_DISABLE); + /* XXX re-enable when interrupt support is added */ + if (!ASIC_IS_DCE4(rdev)) + drm_vblank_post_modeset(dev, radeon_crtc->crtc_id); radeon_crtc_load_lut(crtc); break; case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: case DRM_MODE_DPMS_OFF: - drm_vblank_pre_modeset(dev, radeon_crtc->crtc_id); - atombios_blank_crtc(crtc, 1); + /* XXX re-enable when interrupt support is added */ + if (!ASIC_IS_DCE4(rdev)) + drm_vblank_pre_modeset(dev, radeon_crtc->crtc_id); + atombios_blank_crtc(crtc, ATOM_ENABLE); if (ASIC_IS_DCE3(rdev)) - atombios_enable_crtc_memreq(crtc, 0); - atombios_enable_crtc(crtc, 0); + atombios_enable_crtc_memreq(crtc, ATOM_DISABLE); + atombios_enable_crtc(crtc, ATOM_DISABLE); break; } } @@ -307,7 +311,6 @@ atombios_set_crtc_dtd_timing(struct drm_crtc *crtc, args.susModeMiscInfo.usAccess = cpu_to_le16(misc); args.ucCRTC = radeon_crtc->crtc_id; - printk("executing set crtc dtd timing\n"); atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } @@ -347,10 +350,14 @@ static void atombios_crtc_set_timing(struct drm_crtc *crtc, args.susModeMiscInfo.usAccess = cpu_to_le16(misc); args.ucCRTC = radeon_crtc->crtc_id; - printk("executing set crtc timing\n"); atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } +union atom_enable_ss { + ENABLE_LVDS_SS_PARAMETERS legacy; + ENABLE_SPREAD_SPECTRUM_ON_PPLL_PS_ALLOCATION v1; +}; + static void atombios_set_ss(struct drm_crtc *crtc, int enable) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); @@ -360,11 +367,14 @@ static void atombios_set_ss(struct drm_crtc *crtc, int enable) struct radeon_encoder *radeon_encoder = NULL; struct radeon_encoder_atom_dig *dig = NULL; int index = GetIndexIntoMasterTable(COMMAND, EnableSpreadSpectrumOnPPLL); - ENABLE_SPREAD_SPECTRUM_ON_PPLL_PS_ALLOCATION args; - ENABLE_LVDS_SS_PARAMETERS legacy_args; + union atom_enable_ss args; uint16_t percentage = 0; uint8_t type = 0, step = 0, delay = 0, range = 0; + /* XXX add ss support for DCE4 */ + if (ASIC_IS_DCE4(rdev)) + return; + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) { radeon_encoder = to_radeon_encoder(encoder); @@ -388,80 +398,97 @@ static void atombios_set_ss(struct drm_crtc *crtc, int enable) if (!radeon_encoder) return; + memset(&args, 0, sizeof(args)); if (ASIC_IS_AVIVO(rdev)) { - memset(&args, 0, sizeof(args)); - args.usSpreadSpectrumPercentage = cpu_to_le16(percentage); - args.ucSpreadSpectrumType = type; - args.ucSpreadSpectrumStep = step; - args.ucSpreadSpectrumDelay = delay; - args.ucSpreadSpectrumRange = range; - args.ucPpll = radeon_crtc->crtc_id ? ATOM_PPLL2 : ATOM_PPLL1; - args.ucEnable = enable; - atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + args.v1.usSpreadSpectrumPercentage = cpu_to_le16(percentage); + args.v1.ucSpreadSpectrumType = type; + args.v1.ucSpreadSpectrumStep = step; + args.v1.ucSpreadSpectrumDelay = delay; + args.v1.ucSpreadSpectrumRange = range; + args.v1.ucPpll = radeon_crtc->crtc_id ? ATOM_PPLL2 : ATOM_PPLL1; + args.v1.ucEnable = enable; } else { - memset(&legacy_args, 0, sizeof(legacy_args)); - legacy_args.usSpreadSpectrumPercentage = cpu_to_le16(percentage); - legacy_args.ucSpreadSpectrumType = type; - legacy_args.ucSpreadSpectrumStepSize_Delay = (step & 3) << 2; - legacy_args.ucSpreadSpectrumStepSize_Delay |= (delay & 7) << 4; - legacy_args.ucEnable = enable; - atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&legacy_args); + args.legacy.usSpreadSpectrumPercentage = cpu_to_le16(percentage); + args.legacy.ucSpreadSpectrumType = type; + args.legacy.ucSpreadSpectrumStepSize_Delay = (step & 3) << 2; + args.legacy.ucSpreadSpectrumStepSize_Delay |= (delay & 7) << 4; + args.legacy.ucEnable = enable; } + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } -void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) +union adjust_pixel_clock { + ADJUST_DISPLAY_PLL_PS_ALLOCATION v1; + ADJUST_DISPLAY_PLL_PS_ALLOCATION_V3 v3; +}; + +static u32 atombios_adjust_pll(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct radeon_pll *pll) { - struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; struct radeon_device *rdev = dev->dev_private; struct drm_encoder *encoder = NULL; struct radeon_encoder *radeon_encoder = NULL; - uint8_t frev, crev; - int index; - SET_PIXEL_CLOCK_PS_ALLOCATION args; - PIXEL_CLOCK_PARAMETERS *spc1_ptr; - PIXEL_CLOCK_PARAMETERS_V2 *spc2_ptr; - PIXEL_CLOCK_PARAMETERS_V3 *spc3_ptr; - uint32_t pll_clock = mode->clock; - uint32_t adjusted_clock; - uint32_t ref_div = 0, fb_div = 0, frac_fb_div = 0, post_div = 0; - struct radeon_pll *pll; - int pll_flags = 0; + u32 adjusted_clock = mode->clock; + int encoder_mode = 0; - memset(&args, 0, sizeof(args)); + /* reset the pll flags */ + pll->flags = 0; + + /* select the PLL algo */ + if (ASIC_IS_AVIVO(rdev)) { + if (radeon_new_pll == 0) + pll->algo = PLL_ALGO_LEGACY; + else + pll->algo = PLL_ALGO_NEW; + } else { + if (radeon_new_pll == 1) + pll->algo = PLL_ALGO_NEW; + else + pll->algo = PLL_ALGO_LEGACY; + } if (ASIC_IS_AVIVO(rdev)) { if ((rdev->family == CHIP_RS600) || (rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740)) - pll_flags |= (RADEON_PLL_USE_FRAC_FB_DIV | - RADEON_PLL_PREFER_CLOSEST_LOWER); + pll->flags |= (RADEON_PLL_USE_FRAC_FB_DIV | + RADEON_PLL_PREFER_CLOSEST_LOWER); if (ASIC_IS_DCE32(rdev) && mode->clock > 200000) /* range limits??? */ - pll_flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; + pll->flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else - pll_flags |= RADEON_PLL_PREFER_LOW_REF_DIV; + pll->flags |= RADEON_PLL_PREFER_LOW_REF_DIV; } else { - pll_flags |= RADEON_PLL_LEGACY; + pll->flags |= RADEON_PLL_LEGACY; if (mode->clock > 200000) /* range limits??? */ - pll_flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; + pll->flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else - pll_flags |= RADEON_PLL_PREFER_LOW_REF_DIV; + pll->flags |= RADEON_PLL_PREFER_LOW_REF_DIV; } list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) { - if (!ASIC_IS_AVIVO(rdev)) { - if (encoder->encoder_type != - DRM_MODE_ENCODER_DAC) - pll_flags |= RADEON_PLL_NO_ODD_POST_DIV; - if (encoder->encoder_type == - DRM_MODE_ENCODER_LVDS) - pll_flags |= RADEON_PLL_USE_REF_DIV; - } radeon_encoder = to_radeon_encoder(encoder); + encoder_mode = atombios_get_encoder_mode(encoder); + if (ASIC_IS_AVIVO(rdev)) { + /* DVO wants 2x pixel clock if the DVO chip is in 12 bit mode */ + if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1) + adjusted_clock = mode->clock * 2; + /* LVDS PLL quirks */ + if (encoder->encoder_type == DRM_MODE_ENCODER_LVDS) { + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + pll->algo = dig->pll_algo; + } + } else { + if (encoder->encoder_type != DRM_MODE_ENCODER_DAC) + pll->flags |= RADEON_PLL_NO_ODD_POST_DIV; + if (encoder->encoder_type == DRM_MODE_ENCODER_LVDS) + pll->flags |= RADEON_PLL_USE_REF_DIV; + } break; } } @@ -471,46 +498,173 @@ void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) * special hw requirements. */ if (ASIC_IS_DCE3(rdev)) { - ADJUST_DISPLAY_PLL_PS_ALLOCATION adjust_pll_args; + union adjust_pixel_clock args; + u8 frev, crev; + int index; + + index = GetIndexIntoMasterTable(COMMAND, AdjustDisplayPll); + atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, + &crev); + + memset(&args, 0, sizeof(args)); + + switch (frev) { + case 1: + switch (crev) { + case 1: + case 2: + args.v1.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v1.ucTransmitterID = radeon_encoder->encoder_id; + args.v1.ucEncodeMode = encoder_mode; + + atom_execute_table(rdev->mode_info.atom_context, + index, (uint32_t *)&args); + adjusted_clock = le16_to_cpu(args.v1.usPixelClock) * 10; + break; + case 3: + args.v3.sInput.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v3.sInput.ucTransmitterID = radeon_encoder->encoder_id; + args.v3.sInput.ucEncodeMode = encoder_mode; + args.v3.sInput.ucDispPllConfig = 0; + if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + + if (encoder_mode == ATOM_ENCODER_MODE_DP) + args.v3.sInput.ucDispPllConfig |= + DISPPLL_CONFIG_COHERENT_MODE; + else { + if (dig->coherent_mode) + args.v3.sInput.ucDispPllConfig |= + DISPPLL_CONFIG_COHERENT_MODE; + if (mode->clock > 165000) + args.v3.sInput.ucDispPllConfig |= + DISPPLL_CONFIG_DUAL_LINK; + } + } else if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) { + /* may want to enable SS on DP/eDP eventually */ + args.v3.sInput.ucDispPllConfig |= + DISPPLL_CONFIG_SS_ENABLE; + if (mode->clock > 165000) + args.v3.sInput.ucDispPllConfig |= + DISPPLL_CONFIG_DUAL_LINK; + } + atom_execute_table(rdev->mode_info.atom_context, + index, (uint32_t *)&args); + adjusted_clock = le32_to_cpu(args.v3.sOutput.ulDispPllFreq) * 10; + if (args.v3.sOutput.ucRefDiv) { + pll->flags |= RADEON_PLL_USE_REF_DIV; + pll->reference_div = args.v3.sOutput.ucRefDiv; + } + if (args.v3.sOutput.ucPostDiv) { + pll->flags |= RADEON_PLL_USE_POST_DIV; + pll->post_div = args.v3.sOutput.ucPostDiv; + } + break; + default: + DRM_ERROR("Unknown table version %d %d\n", frev, crev); + return adjusted_clock; + } + break; + default: + DRM_ERROR("Unknown table version %d %d\n", frev, crev); + return adjusted_clock; + } + } + return adjusted_clock; +} + +union set_pixel_clock { + SET_PIXEL_CLOCK_PS_ALLOCATION base; + PIXEL_CLOCK_PARAMETERS v1; + PIXEL_CLOCK_PARAMETERS_V2 v2; + PIXEL_CLOCK_PARAMETERS_V3 v3; + PIXEL_CLOCK_PARAMETERS_V5 v5; +}; + +static void atombios_crtc_set_dcpll(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + u8 frev, crev; + int index; + union set_pixel_clock args; + + memset(&args, 0, sizeof(args)); + + index = GetIndexIntoMasterTable(COMMAND, SetPixelClock); + atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, + &crev); - if (!encoder) + switch (frev) { + case 1: + switch (crev) { + case 5: + /* if the default dcpll clock is specified, + * SetPixelClock provides the dividers + */ + args.v5.ucCRTC = ATOM_CRTC_INVALID; + args.v5.usPixelClock = rdev->clock.default_dispclk; + args.v5.ucPpll = ATOM_DCPLL; + break; + default: + DRM_ERROR("Unknown table version %d %d\n", frev, crev); return; + } + break; + default: + DRM_ERROR("Unknown table version %d %d\n", frev, crev); + return; + } + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); +} - memset(&adjust_pll_args, 0, sizeof(adjust_pll_args)); - adjust_pll_args.usPixelClock = cpu_to_le16(mode->clock / 10); - adjust_pll_args.ucTransmitterID = radeon_encoder->encoder_id; - adjust_pll_args.ucEncodeMode = atombios_get_encoder_mode(encoder); +static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) +{ + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct drm_encoder *encoder = NULL; + struct radeon_encoder *radeon_encoder = NULL; + u8 frev, crev; + int index; + union set_pixel_clock args; + u32 pll_clock = mode->clock; + u32 ref_div = 0, fb_div = 0, frac_fb_div = 0, post_div = 0; + struct radeon_pll *pll; + u32 adjusted_clock; + int encoder_mode = 0; - index = GetIndexIntoMasterTable(COMMAND, AdjustDisplayPll); - atom_execute_table(rdev->mode_info.atom_context, - index, (uint32_t *)&adjust_pll_args); - adjusted_clock = le16_to_cpu(adjust_pll_args.usPixelClock) * 10; - } else { - /* DVO wants 2x pixel clock if the DVO chip is in 12 bit mode */ - if (ASIC_IS_AVIVO(rdev) && - (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1)) - adjusted_clock = mode->clock * 2; - else - adjusted_clock = mode->clock; + memset(&args, 0, sizeof(args)); + + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { + if (encoder->crtc == crtc) { + radeon_encoder = to_radeon_encoder(encoder); + encoder_mode = atombios_get_encoder_mode(encoder); + break; + } } - if (radeon_crtc->crtc_id == 0) + if (!radeon_encoder) + return; + + switch (radeon_crtc->pll_id) { + case ATOM_PPLL1: pll = &rdev->clock.p1pll; - else + break; + case ATOM_PPLL2: pll = &rdev->clock.p2pll; + break; + case ATOM_DCPLL: + case ATOM_PPLL_INVALID: + pll = &rdev->clock.dcpll; + break; + } - if (ASIC_IS_AVIVO(rdev)) { - if (radeon_new_pll) - radeon_compute_pll_avivo(pll, adjusted_clock, &pll_clock, - &fb_div, &frac_fb_div, - &ref_div, &post_div, pll_flags); - else - radeon_compute_pll(pll, adjusted_clock, &pll_clock, - &fb_div, &frac_fb_div, - &ref_div, &post_div, pll_flags); - } else - radeon_compute_pll(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, - &ref_div, &post_div, pll_flags); + /* adjust pixel clock as needed */ + adjusted_clock = atombios_adjust_pll(crtc, mode, pll); + + radeon_compute_pll(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, + &ref_div, &post_div); index = GetIndexIntoMasterTable(COMMAND, SetPixelClock); atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, @@ -520,46 +674,47 @@ void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) case 1: switch (crev) { case 1: - spc1_ptr = (PIXEL_CLOCK_PARAMETERS *) & args.sPCLKInput; - spc1_ptr->usPixelClock = cpu_to_le16(mode->clock / 10); - spc1_ptr->usRefDiv = cpu_to_le16(ref_div); - spc1_ptr->usFbDiv = cpu_to_le16(fb_div); - spc1_ptr->ucFracFbDiv = frac_fb_div; - spc1_ptr->ucPostDiv = post_div; - spc1_ptr->ucPpll = - radeon_crtc->crtc_id ? ATOM_PPLL2 : ATOM_PPLL1; - spc1_ptr->ucCRTC = radeon_crtc->crtc_id; - spc1_ptr->ucRefDivSrc = 1; + args.v1.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v1.usRefDiv = cpu_to_le16(ref_div); + args.v1.usFbDiv = cpu_to_le16(fb_div); + args.v1.ucFracFbDiv = frac_fb_div; + args.v1.ucPostDiv = post_div; + args.v1.ucPpll = radeon_crtc->pll_id; + args.v1.ucCRTC = radeon_crtc->crtc_id; + args.v1.ucRefDivSrc = 1; break; case 2: - spc2_ptr = - (PIXEL_CLOCK_PARAMETERS_V2 *) & args.sPCLKInput; - spc2_ptr->usPixelClock = cpu_to_le16(mode->clock / 10); - spc2_ptr->usRefDiv = cpu_to_le16(ref_div); - spc2_ptr->usFbDiv = cpu_to_le16(fb_div); - spc2_ptr->ucFracFbDiv = frac_fb_div; - spc2_ptr->ucPostDiv = post_div; - spc2_ptr->ucPpll = - radeon_crtc->crtc_id ? ATOM_PPLL2 : ATOM_PPLL1; - spc2_ptr->ucCRTC = radeon_crtc->crtc_id; - spc2_ptr->ucRefDivSrc = 1; + args.v2.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v2.usRefDiv = cpu_to_le16(ref_div); + args.v2.usFbDiv = cpu_to_le16(fb_div); + args.v2.ucFracFbDiv = frac_fb_div; + args.v2.ucPostDiv = post_div; + args.v2.ucPpll = radeon_crtc->pll_id; + args.v2.ucCRTC = radeon_crtc->crtc_id; + args.v2.ucRefDivSrc = 1; break; case 3: - if (!encoder) - return; - spc3_ptr = - (PIXEL_CLOCK_PARAMETERS_V3 *) & args.sPCLKInput; - spc3_ptr->usPixelClock = cpu_to_le16(mode->clock / 10); - spc3_ptr->usRefDiv = cpu_to_le16(ref_div); - spc3_ptr->usFbDiv = cpu_to_le16(fb_div); - spc3_ptr->ucFracFbDiv = frac_fb_div; - spc3_ptr->ucPostDiv = post_div; - spc3_ptr->ucPpll = - radeon_crtc->crtc_id ? ATOM_PPLL2 : ATOM_PPLL1; - spc3_ptr->ucMiscInfo = (radeon_crtc->crtc_id << 2); - spc3_ptr->ucTransmitterId = radeon_encoder->encoder_id; - spc3_ptr->ucEncoderMode = - atombios_get_encoder_mode(encoder); + args.v3.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v3.usRefDiv = cpu_to_le16(ref_div); + args.v3.usFbDiv = cpu_to_le16(fb_div); + args.v3.ucFracFbDiv = frac_fb_div; + args.v3.ucPostDiv = post_div; + args.v3.ucPpll = radeon_crtc->pll_id; + args.v3.ucMiscInfo = (radeon_crtc->pll_id << 2); + args.v3.ucTransmitterId = radeon_encoder->encoder_id; + args.v3.ucEncoderMode = encoder_mode; + break; + case 5: + args.v5.ucCRTC = radeon_crtc->crtc_id; + args.v5.usPixelClock = cpu_to_le16(mode->clock / 10); + args.v5.ucRefDiv = ref_div; + args.v5.usFbDiv = cpu_to_le16(fb_div); + args.v5.ulFbDivDecFrac = cpu_to_le32(frac_fb_div * 100000); + args.v5.ucPostDiv = post_div; + args.v5.ucMiscInfo = 0; /* HDMI depth, etc. */ + args.v5.ucTransmitterID = radeon_encoder->encoder_id; + args.v5.ucEncoderMode = encoder_mode; + args.v5.ucPpll = radeon_crtc->pll_id; break; default: DRM_ERROR("Unknown table version %d %d\n", frev, crev); @@ -571,12 +726,145 @@ void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) return; } - printk("executing set pll\n"); atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } -int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, - struct drm_framebuffer *old_fb) +static int evergreen_crtc_set_base(struct drm_crtc *crtc, int x, int y, + struct drm_framebuffer *old_fb) +{ + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_framebuffer *radeon_fb; + struct drm_gem_object *obj; + struct radeon_bo *rbo; + uint64_t fb_location; + uint32_t fb_format, fb_pitch_pixels, tiling_flags; + int r; + + /* no fb bound */ + if (!crtc->fb) { + DRM_DEBUG("No FB bound\n"); + return 0; + } + + radeon_fb = to_radeon_framebuffer(crtc->fb); + + /* Pin framebuffer & get tilling informations */ + obj = radeon_fb->obj; + rbo = obj->driver_private; + r = radeon_bo_reserve(rbo, false); + if (unlikely(r != 0)) + return r; + r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &fb_location); + if (unlikely(r != 0)) { + radeon_bo_unreserve(rbo); + return -EINVAL; + } + radeon_bo_get_tiling_flags(rbo, &tiling_flags, NULL); + radeon_bo_unreserve(rbo); + + switch (crtc->fb->bits_per_pixel) { + case 8: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_8BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_INDEXED)); + break; + case 15: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB1555)); + break; + case 16: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB565)); + break; + case 24: + case 32: + fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) | + EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB8888)); + break; + default: + DRM_ERROR("Unsupported screen depth %d\n", + crtc->fb->bits_per_pixel); + return -EINVAL; + } + + switch (radeon_crtc->crtc_id) { + case 0: + WREG32(AVIVO_D1VGA_CONTROL, 0); + break; + case 1: + WREG32(AVIVO_D2VGA_CONTROL, 0); + break; + case 2: + WREG32(EVERGREEN_D3VGA_CONTROL, 0); + break; + case 3: + WREG32(EVERGREEN_D4VGA_CONTROL, 0); + break; + case 4: + WREG32(EVERGREEN_D5VGA_CONTROL, 0); + break; + case 5: + WREG32(EVERGREEN_D6VGA_CONTROL, 0); + break; + default: + break; + } + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset, + upper_32_bits(fb_location)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset, + upper_32_bits(fb_location)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + radeon_crtc->crtc_offset, + (u32)fb_location & EVERGREEN_GRPH_SURFACE_ADDRESS_MASK); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + radeon_crtc->crtc_offset, + (u32) fb_location & EVERGREEN_GRPH_SURFACE_ADDRESS_MASK); + WREG32(EVERGREEN_GRPH_CONTROL + radeon_crtc->crtc_offset, fb_format); + + WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_X + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_Y + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_GRPH_X_START + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_GRPH_Y_START + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_GRPH_X_END + radeon_crtc->crtc_offset, crtc->fb->width); + WREG32(EVERGREEN_GRPH_Y_END + radeon_crtc->crtc_offset, crtc->fb->height); + + fb_pitch_pixels = crtc->fb->pitch / (crtc->fb->bits_per_pixel / 8); + WREG32(EVERGREEN_GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels); + WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 1); + + WREG32(EVERGREEN_DESKTOP_HEIGHT + radeon_crtc->crtc_offset, + crtc->mode.vdisplay); + x &= ~3; + y &= ~1; + WREG32(EVERGREEN_VIEWPORT_START + radeon_crtc->crtc_offset, + (x << 16) | y); + WREG32(EVERGREEN_VIEWPORT_SIZE + radeon_crtc->crtc_offset, + (crtc->mode.hdisplay << 16) | crtc->mode.vdisplay); + + if (crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) + WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, + EVERGREEN_INTERLEAVE_EN); + else + WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0); + + if (old_fb && old_fb != crtc->fb) { + radeon_fb = to_radeon_framebuffer(old_fb); + rbo = radeon_fb->obj->driver_private; + r = radeon_bo_reserve(rbo, false); + if (unlikely(r != 0)) + return r; + radeon_bo_unpin(rbo); + radeon_bo_unreserve(rbo); + } + + /* Bytes per pixel may have changed */ + radeon_bandwidth_update(rdev); + + return 0; +} + +static int avivo_crtc_set_base(struct drm_crtc *crtc, int x, int y, + struct drm_framebuffer *old_fb) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; @@ -706,6 +994,84 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, return 0; } +int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + + if (ASIC_IS_DCE4(rdev)) + return evergreen_crtc_set_base(crtc, x, y, old_fb); + else if (ASIC_IS_AVIVO(rdev)) + return avivo_crtc_set_base(crtc, x, y, old_fb); + else + return radeon_crtc_set_base(crtc, x, y, old_fb); +} + +/* properly set additional regs when using atombios */ +static void radeon_legacy_atom_fixup(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + u32 disp_merge_cntl; + + switch (radeon_crtc->crtc_id) { + case 0: + disp_merge_cntl = RREG32(RADEON_DISP_MERGE_CNTL); + disp_merge_cntl &= ~RADEON_DISP_RGB_OFFSET_EN; + WREG32(RADEON_DISP_MERGE_CNTL, disp_merge_cntl); + break; + case 1: + disp_merge_cntl = RREG32(RADEON_DISP2_MERGE_CNTL); + disp_merge_cntl &= ~RADEON_DISP2_RGB_OFFSET_EN; + WREG32(RADEON_DISP2_MERGE_CNTL, disp_merge_cntl); + WREG32(RADEON_FP_H2_SYNC_STRT_WID, RREG32(RADEON_CRTC2_H_SYNC_STRT_WID)); + WREG32(RADEON_FP_V2_SYNC_STRT_WID, RREG32(RADEON_CRTC2_V_SYNC_STRT_WID)); + break; + } +} + +static int radeon_atom_pick_pll(struct drm_crtc *crtc) +{ + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + struct drm_encoder *test_encoder; + struct drm_crtc *test_crtc; + uint32_t pll_in_use = 0; + + if (ASIC_IS_DCE4(rdev)) { + /* if crtc is driving DP and we have an ext clock, use that */ + list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) { + if (test_encoder->crtc && (test_encoder->crtc == crtc)) { + if (atombios_get_encoder_mode(test_encoder) == ATOM_ENCODER_MODE_DP) { + if (rdev->clock.dp_extclk) + return ATOM_PPLL_INVALID; + } + } + } + + /* otherwise, pick one of the plls */ + list_for_each_entry(test_crtc, &dev->mode_config.crtc_list, head) { + struct radeon_crtc *radeon_test_crtc; + + if (crtc == test_crtc) + continue; + + radeon_test_crtc = to_radeon_crtc(test_crtc); + if ((radeon_test_crtc->pll_id >= ATOM_PPLL1) && + (radeon_test_crtc->pll_id <= ATOM_PPLL2)) + pll_in_use |= (1 << radeon_test_crtc->pll_id); + } + if (!(pll_in_use & 1)) + return ATOM_PPLL1; + return ATOM_PPLL2; + } else + return radeon_crtc->crtc_id; + +} + int atombios_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode, @@ -717,19 +1083,27 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc, /* TODO color tiling */ + /* pick pll */ + radeon_crtc->pll_id = radeon_atom_pick_pll(crtc); + atombios_set_ss(crtc, 0); + /* always set DCPLL */ + if (ASIC_IS_DCE4(rdev)) + atombios_crtc_set_dcpll(crtc); atombios_crtc_set_pll(crtc, adjusted_mode); atombios_set_ss(crtc, 1); - atombios_crtc_set_timing(crtc, adjusted_mode); - if (ASIC_IS_AVIVO(rdev)) - atombios_crtc_set_base(crtc, x, y, old_fb); + if (ASIC_IS_DCE4(rdev)) + atombios_set_crtc_dtd_timing(crtc, adjusted_mode); + else if (ASIC_IS_AVIVO(rdev)) + atombios_crtc_set_timing(crtc, adjusted_mode); else { + atombios_crtc_set_timing(crtc, adjusted_mode); if (radeon_crtc->crtc_id == 0) atombios_set_crtc_dtd_timing(crtc, adjusted_mode); - radeon_crtc_set_base(crtc, x, y, old_fb); - radeon_legacy_atom_set_surface(crtc); + radeon_legacy_atom_fixup(crtc); } + atombios_crtc_set_base(crtc, x, y, old_fb); atombios_overscan_setup(crtc, mode, adjusted_mode); atombios_scaler_setup(crtc); return 0; @@ -746,14 +1120,14 @@ static bool atombios_crtc_mode_fixup(struct drm_crtc *crtc, static void atombios_crtc_prepare(struct drm_crtc *crtc) { + atombios_lock_crtc(crtc, ATOM_ENABLE); atombios_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); - atombios_lock_crtc(crtc, 1); } static void atombios_crtc_commit(struct drm_crtc *crtc) { atombios_crtc_dpms(crtc, DRM_MODE_DPMS_ON); - atombios_lock_crtc(crtc, 0); + atombios_lock_crtc(crtc, ATOM_DISABLE); } static const struct drm_crtc_helper_funcs atombios_helper_funcs = { @@ -769,8 +1143,37 @@ static const struct drm_crtc_helper_funcs atombios_helper_funcs = { void radeon_atombios_init_crtc(struct drm_device *dev, struct radeon_crtc *radeon_crtc) { - if (radeon_crtc->crtc_id == 1) - radeon_crtc->crtc_offset = - AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL; + struct radeon_device *rdev = dev->dev_private; + + if (ASIC_IS_DCE4(rdev)) { + switch (radeon_crtc->crtc_id) { + case 0: + default: + radeon_crtc->crtc_offset = EVERGREEN_CRTC0_REGISTER_OFFSET; + break; + case 1: + radeon_crtc->crtc_offset = EVERGREEN_CRTC1_REGISTER_OFFSET; + break; + case 2: + radeon_crtc->crtc_offset = EVERGREEN_CRTC2_REGISTER_OFFSET; + break; + case 3: + radeon_crtc->crtc_offset = EVERGREEN_CRTC3_REGISTER_OFFSET; + break; + case 4: + radeon_crtc->crtc_offset = EVERGREEN_CRTC4_REGISTER_OFFSET; + break; + case 5: + radeon_crtc->crtc_offset = EVERGREEN_CRTC5_REGISTER_OFFSET; + break; + } + } else { + if (radeon_crtc->crtc_id == 1) + radeon_crtc->crtc_offset = + AVIVO_D2CRTC_H_TOTAL - AVIVO_D1CRTC_H_TOTAL; + else + radeon_crtc->crtc_offset = 0; + } + radeon_crtc->pll_id = -1; drm_crtc_helper_add(&radeon_crtc->base, &atombios_helper_funcs); } diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c index 3eb0ca5b3d7..8a133bda00a 100644 --- a/drivers/gpu/drm/radeon/atombios_dp.c +++ b/drivers/gpu/drm/radeon/atombios_dp.c @@ -321,6 +321,10 @@ static void dp_get_adjust_train(u8 link_status[DP_LINK_STATUS_SIZE], train_set[lane] = v | p; } +union aux_channel_transaction { + PROCESS_AUX_CHANNEL_TRANSACTION_PS_ALLOCATION v1; + PROCESS_AUX_CHANNEL_TRANSACTION_PARAMETERS_V2 v2; +}; /* radeon aux chan functions */ bool radeon_process_aux_ch(struct radeon_i2c_chan *chan, u8 *req_bytes, @@ -329,39 +333,45 @@ bool radeon_process_aux_ch(struct radeon_i2c_chan *chan, u8 *req_bytes, { struct drm_device *dev = chan->dev; struct radeon_device *rdev = dev->dev_private; - PROCESS_AUX_CHANNEL_TRANSACTION_PS_ALLOCATION args; + union aux_channel_transaction args; int index = GetIndexIntoMasterTable(COMMAND, ProcessAuxChannelTransaction); unsigned char *base; + int retry_count = 0; memset(&args, 0, sizeof(args)); base = (unsigned char *)rdev->mode_info.atom_context->scratch; +retry: memcpy(base, req_bytes, num_bytes); - args.lpAuxRequest = 0; - args.lpDataOut = 16; - args.ucDataOutLen = 0; - args.ucChannelID = chan->rec.i2c_id; - args.ucDelay = delay / 10; + args.v1.lpAuxRequest = 0; + args.v1.lpDataOut = 16; + args.v1.ucDataOutLen = 0; + args.v1.ucChannelID = chan->rec.i2c_id; + args.v1.ucDelay = delay / 10; + if (ASIC_IS_DCE4(rdev)) + args.v2.ucHPD_ID = chan->rec.hpd_id; atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); - if (args.ucReplyStatus) { - DRM_DEBUG("failed to get auxch %02x%02x %02x %02x 0x%02x %02x\n", + if (args.v1.ucReplyStatus && !args.v1.ucDataOutLen) { + if (args.v1.ucReplyStatus == 0x20 && retry_count++ < 10) + goto retry; + DRM_DEBUG("failed to get auxch %02x%02x %02x %02x 0x%02x %02x after %d retries\n", req_bytes[1], req_bytes[0], req_bytes[2], req_bytes[3], - chan->rec.i2c_id, args.ucReplyStatus); + chan->rec.i2c_id, args.v1.ucReplyStatus, retry_count); return false; } - if (args.ucDataOutLen && read_byte && read_buf_len) { - if (read_buf_len < args.ucDataOutLen) { + if (args.v1.ucDataOutLen && read_byte && read_buf_len) { + if (read_buf_len < args.v1.ucDataOutLen) { DRM_ERROR("Buffer to small for return answer %d %d\n", - read_buf_len, args.ucDataOutLen); + read_buf_len, args.v1.ucDataOutLen); return false; } { - int len = min(read_buf_len, args.ucDataOutLen); + int len = min(read_buf_len, args.v1.ucDataOutLen); memcpy(read_byte, base + 16, len); } } @@ -468,7 +478,7 @@ void radeon_dp_set_link_config(struct drm_connector *connector, struct radeon_connector *radeon_connector; struct radeon_connector_atom_dig *dig_connector; - if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) || + if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) && (connector->connector_type != DRM_MODE_CONNECTOR_eDP)) return; @@ -583,7 +593,7 @@ void dp_link_train(struct drm_encoder *encoder, u8 train_set[4]; int i; - if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) || + if ((connector->connector_type != DRM_MODE_CONNECTOR_DisplayPort) && (connector->connector_type != DRM_MODE_CONNECTOR_eDP)) return; @@ -596,21 +606,14 @@ void dp_link_train(struct drm_encoder *encoder, return; dig_connector = radeon_connector->con_priv; - if (ASIC_IS_DCE32(rdev)) { - if (dig->dig_block) - enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER; - else - enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER; - if (dig_connector->linkb) - enc_id |= ATOM_DP_CONFIG_LINK_B; - else - enc_id |= ATOM_DP_CONFIG_LINK_A; - } else { - if (dig_connector->linkb) - enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER | ATOM_DP_CONFIG_LINK_B; - else - enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER | ATOM_DP_CONFIG_LINK_A; - } + if (dig->dig_encoder) + enc_id |= ATOM_DP_CONFIG_DIG2_ENCODER; + else + enc_id |= ATOM_DP_CONFIG_DIG1_ENCODER; + if (dig_connector->linkb) + enc_id |= ATOM_DP_CONFIG_LINK_B; + else + enc_id |= ATOM_DP_CONFIG_LINK_A; memset(link_configuration, 0, DP_LINK_CONFIGURATION_SIZE); if (dig_connector->dp_clock == 270000) @@ -629,12 +632,19 @@ void dp_link_train(struct drm_encoder *encoder, dp_set_link_bw_lanes(radeon_connector, link_configuration); /* disable downspread on the sink */ dp_set_downspread(radeon_connector, 0); - /* start training on the source */ - radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_START, - dig_connector->dp_clock, enc_id, 0); - /* set training pattern 1 on the source */ - radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_PATTERN_SEL, - dig_connector->dp_clock, enc_id, 0); + if (ASIC_IS_DCE4(rdev)) { + /* start training on the source */ + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_LINK_TRAINING_START); + /* set training pattern 1 on the source */ + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_LINK_TRAINING_PATTERN1); + } else { + /* start training on the source */ + radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_START, + dig_connector->dp_clock, enc_id, 0); + /* set training pattern 1 on the source */ + radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_PATTERN_SEL, + dig_connector->dp_clock, enc_id, 0); + } /* set initial vs/emph */ memset(train_set, 0, 4); @@ -694,8 +704,11 @@ void dp_link_train(struct drm_encoder *encoder, /* set training pattern 2 on the sink */ dp_set_training(radeon_connector, DP_TRAINING_PATTERN_2); /* set training pattern 2 on the source */ - radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_PATTERN_SEL, - dig_connector->dp_clock, enc_id, 1); + if (ASIC_IS_DCE4(rdev)) + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_LINK_TRAINING_PATTERN2); + else + radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_PATTERN_SEL, + dig_connector->dp_clock, enc_id, 1); /* channel equalization loop */ tries = 0; @@ -732,7 +745,11 @@ void dp_link_train(struct drm_encoder *encoder, >> DP_TRAIN_PRE_EMPHASIS_SHIFT); /* disable the training pattern on the sink */ - dp_set_training(radeon_connector, DP_TRAINING_PATTERN_DISABLE); + if (ASIC_IS_DCE4(rdev)) + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_LINK_TRAINING_COMPLETE); + else + radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_COMPLETE, + dig_connector->dp_clock, enc_id, 0); radeon_dp_encoder_service(rdev, ATOM_DP_ACTION_TRAINING_COMPLETE, dig_connector->dp_clock, enc_id, 0); diff --git a/drivers/gpu/drm/radeon/avivod.h b/drivers/gpu/drm/radeon/avivod.h index d4e6e6e4a93..3c391e7e9fd 100644 --- a/drivers/gpu/drm/radeon/avivod.h +++ b/drivers/gpu/drm/radeon/avivod.h @@ -30,11 +30,13 @@ #define D1CRTC_CONTROL 0x6080 #define CRTC_EN (1 << 0) +#define D1CRTC_STATUS 0x609c #define D1CRTC_UPDATE_LOCK 0x60E8 #define D1GRPH_PRIMARY_SURFACE_ADDRESS 0x6110 #define D1GRPH_SECONDARY_SURFACE_ADDRESS 0x6118 #define D2CRTC_CONTROL 0x6880 +#define D2CRTC_STATUS 0x689c #define D2CRTC_UPDATE_LOCK 0x68E8 #define D2GRPH_PRIMARY_SURFACE_ADDRESS 0x6910 #define D2GRPH_SECONDARY_SURFACE_ADDRESS 0x6918 diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c new file mode 100644 index 00000000000..bd2e7aa85c1 --- /dev/null +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -0,0 +1,767 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include <linux/firmware.h> +#include <linux/platform_device.h> +#include "drmP.h" +#include "radeon.h" +#include "radeon_drm.h" +#include "rv770d.h" +#include "atom.h" +#include "avivod.h" +#include "evergreen_reg.h" + +static void evergreen_gpu_init(struct radeon_device *rdev); +void evergreen_fini(struct radeon_device *rdev); + +bool evergreen_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd) +{ + bool connected = false; + /* XXX */ + return connected; +} + +void evergreen_hpd_set_polarity(struct radeon_device *rdev, + enum radeon_hpd_id hpd) +{ + /* XXX */ +} + +void evergreen_hpd_init(struct radeon_device *rdev) +{ + /* XXX */ +} + + +void evergreen_bandwidth_update(struct radeon_device *rdev) +{ + /* XXX */ +} + +void evergreen_hpd_fini(struct radeon_device *rdev) +{ + /* XXX */ +} + +static int evergreen_mc_wait_for_idle(struct radeon_device *rdev) +{ + unsigned i; + u32 tmp; + + for (i = 0; i < rdev->usec_timeout; i++) { + /* read MC_STATUS */ + tmp = RREG32(SRBM_STATUS) & 0x1F00; + if (!tmp) + return 0; + udelay(1); + } + return -1; +} + +/* + * GART + */ +int evergreen_pcie_gart_enable(struct radeon_device *rdev) +{ + u32 tmp; + int r, i; + + if (rdev->gart.table.vram.robj == NULL) { + dev_err(rdev->dev, "No VRAM object for PCIE GART.\n"); + return -EINVAL; + } + r = radeon_gart_table_vram_pin(rdev); + if (r) + return r; + radeon_gart_restore(rdev); + /* Setup L2 cache */ + WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING | + ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | + EFFECTIVE_L2_QUEUE_SIZE(7)); + WREG32(VM_L2_CNTL2, 0); + WREG32(VM_L2_CNTL3, BANK_SELECT(0) | CACHE_UPDATE_MODE(2)); + /* Setup TLB control */ + tmp = ENABLE_L1_TLB | ENABLE_L1_FRAGMENT_PROCESSING | + SYSTEM_ACCESS_MODE_NOT_IN_SYS | + SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU | + EFFECTIVE_L1_TLB_SIZE(5) | EFFECTIVE_L1_QUEUE_SIZE(5); + WREG32(MC_VM_MD_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB3_CNTL, tmp); + WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); + WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12); + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR, rdev->gart.table_addr >> 12); + WREG32(VM_CONTEXT0_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) | + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT); + WREG32(VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR, + (u32)(rdev->dummy_page.addr >> 12)); + for (i = 1; i < 7; i++) + WREG32(VM_CONTEXT0_CNTL + (i * 4), 0); + + r600_pcie_gart_tlb_flush(rdev); + rdev->gart.ready = true; + return 0; +} + +void evergreen_pcie_gart_disable(struct radeon_device *rdev) +{ + u32 tmp; + int i, r; + + /* Disable all tables */ + for (i = 0; i < 7; i++) + WREG32(VM_CONTEXT0_CNTL + (i * 4), 0); + + /* Setup L2 cache */ + WREG32(VM_L2_CNTL, ENABLE_L2_FRAGMENT_PROCESSING | + EFFECTIVE_L2_QUEUE_SIZE(7)); + WREG32(VM_L2_CNTL2, 0); + WREG32(VM_L2_CNTL3, BANK_SELECT(0) | CACHE_UPDATE_MODE(2)); + /* Setup TLB control */ + tmp = EFFECTIVE_L1_TLB_SIZE(5) | EFFECTIVE_L1_QUEUE_SIZE(5); + WREG32(MC_VM_MD_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB3_CNTL, tmp); + if (rdev->gart.table.vram.robj) { + r = radeon_bo_reserve(rdev->gart.table.vram.robj, false); + if (likely(r == 0)) { + radeon_bo_kunmap(rdev->gart.table.vram.robj); + radeon_bo_unpin(rdev->gart.table.vram.robj); + radeon_bo_unreserve(rdev->gart.table.vram.robj); + } + } +} + +void evergreen_pcie_gart_fini(struct radeon_device *rdev) +{ + evergreen_pcie_gart_disable(rdev); + radeon_gart_table_vram_free(rdev); + radeon_gart_fini(rdev); +} + + +void evergreen_agp_enable(struct radeon_device *rdev) +{ + u32 tmp; + int i; + + /* Setup L2 cache */ + WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING | + ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | + EFFECTIVE_L2_QUEUE_SIZE(7)); + WREG32(VM_L2_CNTL2, 0); + WREG32(VM_L2_CNTL3, BANK_SELECT(0) | CACHE_UPDATE_MODE(2)); + /* Setup TLB control */ + tmp = ENABLE_L1_TLB | ENABLE_L1_FRAGMENT_PROCESSING | + SYSTEM_ACCESS_MODE_NOT_IN_SYS | + SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU | + EFFECTIVE_L1_TLB_SIZE(5) | EFFECTIVE_L1_QUEUE_SIZE(5); + WREG32(MC_VM_MD_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MD_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB0_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB1_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB2_CNTL, tmp); + WREG32(MC_VM_MB_L1_TLB3_CNTL, tmp); + for (i = 0; i < 7; i++) + WREG32(VM_CONTEXT0_CNTL + (i * 4), 0); +} + +static void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save) +{ + save->vga_control[0] = RREG32(D1VGA_CONTROL); + save->vga_control[1] = RREG32(D2VGA_CONTROL); + save->vga_control[2] = RREG32(EVERGREEN_D3VGA_CONTROL); + save->vga_control[3] = RREG32(EVERGREEN_D4VGA_CONTROL); + save->vga_control[4] = RREG32(EVERGREEN_D5VGA_CONTROL); + save->vga_control[5] = RREG32(EVERGREEN_D6VGA_CONTROL); + save->vga_render_control = RREG32(VGA_RENDER_CONTROL); + save->vga_hdp_control = RREG32(VGA_HDP_CONTROL); + save->crtc_control[0] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET); + save->crtc_control[1] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET); + save->crtc_control[2] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET); + save->crtc_control[3] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET); + save->crtc_control[4] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET); + save->crtc_control[5] = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET); + + /* Stop all video */ + WREG32(VGA_RENDER_CONTROL, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + + WREG32(D1VGA_CONTROL, 0); + WREG32(D2VGA_CONTROL, 0); + WREG32(EVERGREEN_D3VGA_CONTROL, 0); + WREG32(EVERGREEN_D4VGA_CONTROL, 0); + WREG32(EVERGREEN_D5VGA_CONTROL, 0); + WREG32(EVERGREEN_D6VGA_CONTROL, 0); +} + +static void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save) +{ + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC0_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC0_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC0_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC0_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC1_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC1_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC1_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC1_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC2_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC2_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC2_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC2_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC3_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC3_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC3_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC3_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC4_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC4_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC4_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC4_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC5_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + EVERGREEN_CRTC5_REGISTER_OFFSET, + upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + EVERGREEN_CRTC5_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + EVERGREEN_CRTC5_REGISTER_OFFSET, + (u32)rdev->mc.vram_start); + + WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH, upper_32_bits(rdev->mc.vram_start)); + WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS, (u32)rdev->mc.vram_start); + /* Unlock host access */ + WREG32(VGA_HDP_CONTROL, save->vga_hdp_control); + mdelay(1); + /* Restore video state */ + WREG32(D1VGA_CONTROL, save->vga_control[0]); + WREG32(D2VGA_CONTROL, save->vga_control[1]); + WREG32(EVERGREEN_D3VGA_CONTROL, save->vga_control[2]); + WREG32(EVERGREEN_D4VGA_CONTROL, save->vga_control[3]); + WREG32(EVERGREEN_D5VGA_CONTROL, save->vga_control[4]); + WREG32(EVERGREEN_D6VGA_CONTROL, save->vga_control[5]); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 1); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, save->crtc_control[0]); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, save->crtc_control[1]); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, save->crtc_control[2]); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, save->crtc_control[3]); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, save->crtc_control[4]); + WREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, save->crtc_control[5]); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + WREG32(VGA_RENDER_CONTROL, save->vga_render_control); +} + +static void evergreen_mc_program(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 tmp; + int i, j; + + /* Initialize HDP */ + for (i = 0, j = 0; i < 32; i++, j += 0x18) { + WREG32((0x2c14 + j), 0x00000000); + WREG32((0x2c18 + j), 0x00000000); + WREG32((0x2c1c + j), 0x00000000); + WREG32((0x2c20 + j), 0x00000000); + WREG32((0x2c24 + j), 0x00000000); + } + WREG32(HDP_REG_COHERENCY_FLUSH_CNTL, 0); + + evergreen_mc_stop(rdev, &save); + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + /* Lockout access through VGA aperture*/ + WREG32(VGA_HDP_CONTROL, VGA_MEMORY_DISABLE); + /* Update configuration */ + if (rdev->flags & RADEON_IS_AGP) { + if (rdev->mc.vram_start < rdev->mc.gtt_start) { + /* VRAM before AGP */ + WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR, + rdev->mc.vram_start >> 12); + WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR, + rdev->mc.gtt_end >> 12); + } else { + /* VRAM after AGP */ + WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR, + rdev->mc.gtt_start >> 12); + WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR, + rdev->mc.vram_end >> 12); + } + } else { + WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR, + rdev->mc.vram_start >> 12); + WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR, + rdev->mc.vram_end >> 12); + } + WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, 0); + tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16; + tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF); + WREG32(MC_VM_FB_LOCATION, tmp); + WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8)); + WREG32(HDP_NONSURFACE_INFO, (2 << 7)); + WREG32(HDP_NONSURFACE_SIZE, (rdev->mc.mc_vram_size - 1) | 0x3FF); + if (rdev->flags & RADEON_IS_AGP) { + WREG32(MC_VM_AGP_TOP, rdev->mc.gtt_end >> 16); + WREG32(MC_VM_AGP_BOT, rdev->mc.gtt_start >> 16); + WREG32(MC_VM_AGP_BASE, rdev->mc.agp_base >> 22); + } else { + WREG32(MC_VM_AGP_BASE, 0); + WREG32(MC_VM_AGP_TOP, 0x0FFFFFFF); + WREG32(MC_VM_AGP_BOT, 0x0FFFFFFF); + } + if (evergreen_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + evergreen_mc_resume(rdev, &save); + /* we need to own VRAM, so turn off the VGA renderer here + * to stop it overwriting our objects */ + rv515_vga_render_disable(rdev); +} + +#if 0 +/* + * CP. + */ +static void evergreen_cp_stop(struct radeon_device *rdev) +{ + /* XXX */ +} + + +static int evergreen_cp_load_microcode(struct radeon_device *rdev) +{ + /* XXX */ + + return 0; +} + + +/* + * Core functions + */ +static u32 evergreen_get_tile_pipe_to_backend_map(u32 num_tile_pipes, + u32 num_backends, + u32 backend_disable_mask) +{ + u32 backend_map = 0; + + return backend_map; +} +#endif + +static void evergreen_gpu_init(struct radeon_device *rdev) +{ + /* XXX */ +} + +int evergreen_mc_init(struct radeon_device *rdev) +{ + fixed20_12 a; + u32 tmp; + int chansize, numchan; + + /* Get VRAM informations */ + rdev->mc.vram_is_ddr = true; + tmp = RREG32(MC_ARB_RAMCFG); + if (tmp & CHANSIZE_OVERRIDE) { + chansize = 16; + } else if (tmp & CHANSIZE_MASK) { + chansize = 64; + } else { + chansize = 32; + } + tmp = RREG32(MC_SHARED_CHMAP); + switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) { + case 0: + default: + numchan = 1; + break; + case 1: + numchan = 2; + break; + case 2: + numchan = 4; + break; + case 3: + numchan = 8; + break; + } + rdev->mc.vram_width = numchan * chansize; + /* Could aper size report 0 ? */ + rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); + rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); + /* Setup GPU memory space */ + /* size in MB on evergreen */ + rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; + rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; + rdev->mc.visible_vram_size = rdev->mc.aper_size; + /* FIXME remove this once we support unmappable VRAM */ + if (rdev->mc.mc_vram_size > rdev->mc.aper_size) { + rdev->mc.mc_vram_size = rdev->mc.aper_size; + rdev->mc.real_vram_size = rdev->mc.aper_size; + } + r600_vram_gtt_location(rdev, &rdev->mc); + /* FIXME: we should enforce default clock in case GPU is not in + * default setup + */ + a.full = rfixed_const(100); + rdev->pm.sclk.full = rfixed_const(rdev->clock.default_sclk); + rdev->pm.sclk.full = rfixed_div(rdev->pm.sclk, a); + return 0; +} + +int evergreen_gpu_reset(struct radeon_device *rdev) +{ + /* FIXME: implement for evergreen */ + return 0; +} + +static int evergreen_startup(struct radeon_device *rdev) +{ +#if 0 + int r; + + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = r600_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } +#endif + evergreen_mc_program(rdev); +#if 0 + if (rdev->flags & RADEON_IS_AGP) { + evergreem_agp_enable(rdev); + } else { + r = evergreen_pcie_gart_enable(rdev); + if (r) + return r; + } +#endif + evergreen_gpu_init(rdev); +#if 0 + if (!rdev->r600_blit.shader_obj) { + r = r600_blit_init(rdev); + if (r) { + DRM_ERROR("radeon: failed blitter (%d).\n", r); + return r; + } + } + + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (unlikely(r != 0)) + return r; + r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->r600_blit.shader_gpu_addr); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); + if (r) { + DRM_ERROR("failed to pin blit object %d\n", r); + return r; + } + + /* Enable IRQ */ + r = r600_irq_init(rdev); + if (r) { + DRM_ERROR("radeon: IH init failed (%d).\n", r); + radeon_irq_kms_fini(rdev); + return r; + } + r600_irq_set(rdev); + + r = radeon_ring_init(rdev, rdev->cp.ring_size); + if (r) + return r; + r = evergreen_cp_load_microcode(rdev); + if (r) + return r; + r = r600_cp_resume(rdev); + if (r) + return r; + /* write back buffer are not vital so don't worry about failure */ + r600_wb_enable(rdev); +#endif + return 0; +} + +int evergreen_resume(struct radeon_device *rdev) +{ + int r; + + /* Do not reset GPU before posting, on rv770 hw unlike on r500 hw, + * posting will perform necessary task to bring back GPU into good + * shape. + */ + /* post card */ + atom_asic_init(rdev->mode_info.atom_context); + /* Initialize clocks */ + r = radeon_clocks_init(rdev); + if (r) { + return r; + } + + r = evergreen_startup(rdev); + if (r) { + DRM_ERROR("r600 startup failed on resume\n"); + return r; + } +#if 0 + r = r600_ib_test(rdev); + if (r) { + DRM_ERROR("radeon: failled testing IB (%d).\n", r); + return r; + } +#endif + return r; + +} + +int evergreen_suspend(struct radeon_device *rdev) +{ +#if 0 + int r; + + /* FIXME: we should wait for ring to be empty */ + r700_cp_stop(rdev); + rdev->cp.ready = false; + r600_wb_disable(rdev); + evergreen_pcie_gart_disable(rdev); + /* unpin shaders bo */ + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (likely(r == 0)) { + radeon_bo_unpin(rdev->r600_blit.shader_obj); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); + } +#endif + return 0; +} + +static bool evergreen_card_posted(struct radeon_device *rdev) +{ + u32 reg; + + /* first check CRTCs */ + reg = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET); + if (reg & EVERGREEN_CRTC_MASTER_EN) + return true; + + /* then check MEM_SIZE, in case the crtcs are off */ + if (RREG32(CONFIG_MEMSIZE)) + return true; + + return false; +} + +/* Plan is to move initialization in that function and use + * helper function so that radeon_device_init pretty much + * do nothing more than calling asic specific function. This + * should also allow to remove a bunch of callback function + * like vram_info. + */ +int evergreen_init(struct radeon_device *rdev) +{ + int r; + + r = radeon_dummy_page_init(rdev); + if (r) + return r; + /* This don't do much */ + r = radeon_gem_init(rdev); + if (r) + return r; + /* Read BIOS */ + if (!radeon_get_bios(rdev)) { + if (ASIC_IS_AVIVO(rdev)) + return -EINVAL; + } + /* Must be an ATOMBIOS */ + if (!rdev->is_atom_bios) { + dev_err(rdev->dev, "Expecting atombios for R600 GPU\n"); + return -EINVAL; + } + r = radeon_atombios_init(rdev); + if (r) + return r; + /* Post card if necessary */ + if (!evergreen_card_posted(rdev)) { + if (!rdev->bios) { + dev_err(rdev->dev, "Card not posted and no BIOS - ignoring\n"); + return -EINVAL; + } + DRM_INFO("GPU not posted. posting now...\n"); + atom_asic_init(rdev->mode_info.atom_context); + } + /* Initialize scratch registers */ + r600_scratch_init(rdev); + /* Initialize surface registers */ + radeon_surface_init(rdev); + /* Initialize clocks */ + radeon_get_clock_info(rdev->ddev); + r = radeon_clocks_init(rdev); + if (r) + return r; + /* Initialize power management */ + radeon_pm_init(rdev); + /* Fence driver */ + r = radeon_fence_driver_init(rdev); + if (r) + return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) + radeon_agp_disable(rdev); + } + /* initialize memory controller */ + r = evergreen_mc_init(rdev); + if (r) + return r; + /* Memory manager */ + r = radeon_bo_init(rdev); + if (r) + return r; +#if 0 + r = radeon_irq_kms_init(rdev); + if (r) + return r; + + rdev->cp.ring_obj = NULL; + r600_ring_init(rdev, 1024 * 1024); + + rdev->ih.ring_obj = NULL; + r600_ih_ring_init(rdev, 64 * 1024); + + r = r600_pcie_gart_init(rdev); + if (r) + return r; +#endif + rdev->accel_working = false; + r = evergreen_startup(rdev); + if (r) { + evergreen_suspend(rdev); + /*r600_wb_fini(rdev);*/ + /*radeon_ring_fini(rdev);*/ + /*evergreen_pcie_gart_fini(rdev);*/ + rdev->accel_working = false; + } + if (rdev->accel_working) { + r = radeon_ib_pool_init(rdev); + if (r) { + DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r); + rdev->accel_working = false; + } + r = r600_ib_test(rdev); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; + } + } + return 0; +} + +void evergreen_fini(struct radeon_device *rdev) +{ + evergreen_suspend(rdev); +#if 0 + r600_blit_fini(rdev); + r600_irq_fini(rdev); + radeon_irq_kms_fini(rdev); + radeon_ring_fini(rdev); + r600_wb_fini(rdev); + evergreen_pcie_gart_fini(rdev); +#endif + radeon_gem_fini(rdev); + radeon_fence_driver_fini(rdev); + radeon_clocks_fini(rdev); + radeon_agp_fini(rdev); + radeon_bo_fini(rdev); + radeon_atombios_fini(rdev); + kfree(rdev->bios); + rdev->bios = NULL; + radeon_dummy_page_fini(rdev); +} diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h new file mode 100644 index 00000000000..f7c7c964343 --- /dev/null +++ b/drivers/gpu/drm/radeon/evergreen_reg.h @@ -0,0 +1,176 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#ifndef __EVERGREEN_REG_H__ +#define __EVERGREEN_REG_H__ + +/* evergreen */ +#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS 0x310 +#define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH 0x324 +#define EVERGREEN_D3VGA_CONTROL 0x3e0 +#define EVERGREEN_D4VGA_CONTROL 0x3e4 +#define EVERGREEN_D5VGA_CONTROL 0x3e8 +#define EVERGREEN_D6VGA_CONTROL 0x3ec + +#define EVERGREEN_P1PLL_SS_CNTL 0x414 +#define EVERGREEN_P2PLL_SS_CNTL 0x454 +# define EVERGREEN_PxPLL_SS_EN (1 << 12) +/* GRPH blocks at 0x6800, 0x7400, 0x10000, 0x10c00, 0x11800, 0x12400 */ +#define EVERGREEN_GRPH_ENABLE 0x6800 +#define EVERGREEN_GRPH_CONTROL 0x6804 +# define EVERGREEN_GRPH_DEPTH(x) (((x) & 0x3) << 0) +# define EVERGREEN_GRPH_DEPTH_8BPP 0 +# define EVERGREEN_GRPH_DEPTH_16BPP 1 +# define EVERGREEN_GRPH_DEPTH_32BPP 2 +# define EVERGREEN_GRPH_FORMAT(x) (((x) & 0x7) << 8) +/* 8 BPP */ +# define EVERGREEN_GRPH_FORMAT_INDEXED 0 +/* 16 BPP */ +# define EVERGREEN_GRPH_FORMAT_ARGB1555 0 +# define EVERGREEN_GRPH_FORMAT_ARGB565 1 +# define EVERGREEN_GRPH_FORMAT_ARGB4444 2 +# define EVERGREEN_GRPH_FORMAT_AI88 3 +# define EVERGREEN_GRPH_FORMAT_MONO16 4 +# define EVERGREEN_GRPH_FORMAT_BGRA5551 5 +/* 32 BPP */ +# define EVERGREEN_GRPH_FORMAT_ARGB8888 0 +# define EVERGREEN_GRPH_FORMAT_ARGB2101010 1 +# define EVERGREEN_GRPH_FORMAT_32BPP_DIG 2 +# define EVERGREEN_GRPH_FORMAT_8B_ARGB2101010 3 +# define EVERGREEN_GRPH_FORMAT_BGRA1010102 4 +# define EVERGREEN_GRPH_FORMAT_8B_BGRA1010102 5 +# define EVERGREEN_GRPH_FORMAT_RGB111110 6 +# define EVERGREEN_GRPH_FORMAT_BGR101111 7 +#define EVERGREEN_GRPH_SWAP_CONTROL 0x680c +# define EVERGREEN_GRPH_ENDIAN_SWAP(x) (((x) & 0x3) << 0) +# define EVERGREEN_GRPH_ENDIAN_NONE 0 +# define EVERGREEN_GRPH_ENDIAN_8IN16 1 +# define EVERGREEN_GRPH_ENDIAN_8IN32 2 +# define EVERGREEN_GRPH_ENDIAN_8IN64 3 +# define EVERGREEN_GRPH_RED_CROSSBAR(x) (((x) & 0x3) << 4) +# define EVERGREEN_GRPH_RED_SEL_R 0 +# define EVERGREEN_GRPH_RED_SEL_G 1 +# define EVERGREEN_GRPH_RED_SEL_B 2 +# define EVERGREEN_GRPH_RED_SEL_A 3 +# define EVERGREEN_GRPH_GREEN_CROSSBAR(x) (((x) & 0x3) << 6) +# define EVERGREEN_GRPH_GREEN_SEL_G 0 +# define EVERGREEN_GRPH_GREEN_SEL_B 1 +# define EVERGREEN_GRPH_GREEN_SEL_A 2 +# define EVERGREEN_GRPH_GREEN_SEL_R 3 +# define EVERGREEN_GRPH_BLUE_CROSSBAR(x) (((x) & 0x3) << 8) +# define EVERGREEN_GRPH_BLUE_SEL_B 0 +# define EVERGREEN_GRPH_BLUE_SEL_A 1 +# define EVERGREEN_GRPH_BLUE_SEL_R 2 +# define EVERGREEN_GRPH_BLUE_SEL_G 3 +# define EVERGREEN_GRPH_ALPHA_CROSSBAR(x) (((x) & 0x3) << 10) +# define EVERGREEN_GRPH_ALPHA_SEL_A 0 +# define EVERGREEN_GRPH_ALPHA_SEL_R 1 +# define EVERGREEN_GRPH_ALPHA_SEL_G 2 +# define EVERGREEN_GRPH_ALPHA_SEL_B 3 +#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS 0x6810 +#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS 0x6814 +# define EVERGREEN_GRPH_DFQ_ENABLE (1 << 0) +# define EVERGREEN_GRPH_SURFACE_ADDRESS_MASK 0xffffff00 +#define EVERGREEN_GRPH_PITCH 0x6818 +#define EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH 0x681c +#define EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH 0x6820 +#define EVERGREEN_GRPH_SURFACE_OFFSET_X 0x6824 +#define EVERGREEN_GRPH_SURFACE_OFFSET_Y 0x6828 +#define EVERGREEN_GRPH_X_START 0x682c +#define EVERGREEN_GRPH_Y_START 0x6830 +#define EVERGREEN_GRPH_X_END 0x6834 +#define EVERGREEN_GRPH_Y_END 0x6838 + +/* CUR blocks at 0x6998, 0x7598, 0x10198, 0x10d98, 0x11998, 0x12598 */ +#define EVERGREEN_CUR_CONTROL 0x6998 +# define EVERGREEN_CURSOR_EN (1 << 0) +# define EVERGREEN_CURSOR_MODE(x) (((x) & 0x3) << 8) +# define EVERGREEN_CURSOR_MONO 0 +# define EVERGREEN_CURSOR_24_1 1 +# define EVERGREEN_CURSOR_24_8_PRE_MULT 2 +# define EVERGREEN_CURSOR_24_8_UNPRE_MULT 3 +# define EVERGREEN_CURSOR_2X_MAGNIFY (1 << 16) +# define EVERGREEN_CURSOR_FORCE_MC_ON (1 << 20) +# define EVERGREEN_CURSOR_URGENT_CONTROL(x) (((x) & 0x7) << 24) +# define EVERGREEN_CURSOR_URGENT_ALWAYS 0 +# define EVERGREEN_CURSOR_URGENT_1_8 1 +# define EVERGREEN_CURSOR_URGENT_1_4 2 +# define EVERGREEN_CURSOR_URGENT_3_8 3 +# define EVERGREEN_CURSOR_URGENT_1_2 4 +#define EVERGREEN_CUR_SURFACE_ADDRESS 0x699c +# define EVERGREEN_CUR_SURFACE_ADDRESS_MASK 0xfffff000 +#define EVERGREEN_CUR_SIZE 0x69a0 +#define EVERGREEN_CUR_SURFACE_ADDRESS_HIGH 0x69a4 +#define EVERGREEN_CUR_POSITION 0x69a8 +#define EVERGREEN_CUR_HOT_SPOT 0x69ac +#define EVERGREEN_CUR_COLOR1 0x69b0 +#define EVERGREEN_CUR_COLOR2 0x69b4 +#define EVERGREEN_CUR_UPDATE 0x69b8 +# define EVERGREEN_CURSOR_UPDATE_PENDING (1 << 0) +# define EVERGREEN_CURSOR_UPDATE_TAKEN (1 << 1) +# define EVERGREEN_CURSOR_UPDATE_LOCK (1 << 16) +# define EVERGREEN_CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24) + +/* LUT blocks at 0x69e0, 0x75e0, 0x101e0, 0x10de0, 0x119e0, 0x125e0 */ +#define EVERGREEN_DC_LUT_RW_MODE 0x69e0 +#define EVERGREEN_DC_LUT_RW_INDEX 0x69e4 +#define EVERGREEN_DC_LUT_SEQ_COLOR 0x69e8 +#define EVERGREEN_DC_LUT_PWL_DATA 0x69ec +#define EVERGREEN_DC_LUT_30_COLOR 0x69f0 +#define EVERGREEN_DC_LUT_VGA_ACCESS_ENABLE 0x69f4 +#define EVERGREEN_DC_LUT_WRITE_EN_MASK 0x69f8 +#define EVERGREEN_DC_LUT_AUTOFILL 0x69fc +#define EVERGREEN_DC_LUT_CONTROL 0x6a00 +#define EVERGREEN_DC_LUT_BLACK_OFFSET_BLUE 0x6a04 +#define EVERGREEN_DC_LUT_BLACK_OFFSET_GREEN 0x6a08 +#define EVERGREEN_DC_LUT_BLACK_OFFSET_RED 0x6a0c +#define EVERGREEN_DC_LUT_WHITE_OFFSET_BLUE 0x6a10 +#define EVERGREEN_DC_LUT_WHITE_OFFSET_GREEN 0x6a14 +#define EVERGREEN_DC_LUT_WHITE_OFFSET_RED 0x6a18 + +#define EVERGREEN_DATA_FORMAT 0x6b00 +# define EVERGREEN_INTERLEAVE_EN (1 << 0) +#define EVERGREEN_DESKTOP_HEIGHT 0x6b04 + +#define EVERGREEN_VIEWPORT_START 0x6d70 +#define EVERGREEN_VIEWPORT_SIZE 0x6d74 + +/* display controller offsets used for crtc/cur/lut/grph/viewport/etc. */ +#define EVERGREEN_CRTC0_REGISTER_OFFSET (0x6df0 - 0x6df0) +#define EVERGREEN_CRTC1_REGISTER_OFFSET (0x79f0 - 0x6df0) +#define EVERGREEN_CRTC2_REGISTER_OFFSET (0x105f0 - 0x6df0) +#define EVERGREEN_CRTC3_REGISTER_OFFSET (0x111f0 - 0x6df0) +#define EVERGREEN_CRTC4_REGISTER_OFFSET (0x11df0 - 0x6df0) +#define EVERGREEN_CRTC5_REGISTER_OFFSET (0x129f0 - 0x6df0) + +/* CRTC blocks at 0x6df0, 0x79f0, 0x105f0, 0x111f0, 0x11df0, 0x129f0 */ +#define EVERGREEN_CRTC_CONTROL 0x6e70 +# define EVERGREEN_CRTC_MASTER_EN (1 << 0) +#define EVERGREEN_CRTC_UPDATE_LOCK 0x6ed4 + +#define EVERGREEN_DC_GPIO_HPD_MASK 0x64b0 +#define EVERGREEN_DC_GPIO_HPD_A 0x64b4 +#define EVERGREEN_DC_GPIO_HPD_EN 0x64b8 +#define EVERGREEN_DC_GPIO_HPD_Y 0x64bc + +#endif diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 8760d66e058..91eb762eb3f 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -197,13 +197,13 @@ int r100_pci_gart_enable(struct radeon_device *rdev) { uint32_t tmp; + radeon_gart_restore(rdev); /* discard memory request outside of configured range */ tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS; WREG32(RADEON_AIC_CNTL, tmp); /* set address range for PCI address translate */ - WREG32(RADEON_AIC_LO_ADDR, rdev->mc.gtt_location); - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; - WREG32(RADEON_AIC_HI_ADDR, tmp); + WREG32(RADEON_AIC_LO_ADDR, rdev->mc.gtt_start); + WREG32(RADEON_AIC_HI_ADDR, rdev->mc.gtt_end); /* set PCI GART page-table base address */ WREG32(RADEON_AIC_PT_BASE, rdev->gart.table_addr); tmp = RREG32(RADEON_AIC_CNTL) | RADEON_PCIGART_TRANSLATE_EN; @@ -312,9 +312,11 @@ int r100_irq_process(struct radeon_device *rdev) /* Vertical blank interrupts */ if (status & RADEON_CRTC_VBLANK_STAT) { drm_handle_vblank(rdev->ddev, 0); + wake_up(&rdev->irq.vblank_queue); } if (status & RADEON_CRTC2_VBLANK_STAT) { drm_handle_vblank(rdev->ddev, 1); + wake_up(&rdev->irq.vblank_queue); } if (status & RADEON_FP_DETECT_STAT) { queue_hotplug = true; @@ -354,14 +356,20 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc) return RREG32(RADEON_CRTC2_CRNT_FRAME); } +/* Who ever call radeon_fence_emit should call ring_lock and ask + * for enough space (today caller are ib schedule and buffer move) */ void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { - /* Who ever call radeon_fence_emit should call ring_lock and ask - * for enough space (today caller are ib schedule and buffer move) */ + /* We have to make sure that caches are flushed before + * CPU might read something from VRAM. */ + radeon_ring_write(rdev, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(rdev, RADEON_RB3D_DC_FLUSH_ALL); + radeon_ring_write(rdev, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(rdev, RADEON_RB3D_ZC_FLUSH_ALL); /* Wait until IDLE & CLEAN */ - radeon_ring_write(rdev, PACKET0(0x1720, 0)); - radeon_ring_write(rdev, (1 << 16) | (1 << 17)); + radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(rdev, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); radeon_ring_write(rdev, rdev->config.r100.hdp_cntl | RADEON_HDP_READ_BUFFER_INVALIDATE); @@ -1504,6 +1512,7 @@ static int r100_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); return -EINVAL; } + track->vtx_size = r100_get_vtx_size(radeon_get_ib_value(p, idx + 0)); track->vap_vf_cntl = radeon_get_ib_value(p, idx + 1); track->immd_dwords = pkt->count - 1; r = r100_cs_track_check(p->rdev, track); @@ -1694,7 +1703,7 @@ int r100_gui_wait_for_idle(struct radeon_device *rdev) } for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(RADEON_RBBM_STATUS); - if (!(tmp & (1 << 31))) { + if (!(tmp & RADEON_RBBM_ACTIVE)) { return 0; } DRM_UDELAY(1); @@ -1709,8 +1718,8 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev) for (i = 0; i < rdev->usec_timeout; i++) { /* read MC_STATUS */ - tmp = RREG32(0x0150); - if (tmp & (1 << 2)) { + tmp = RREG32(RADEON_MC_STATUS); + if (tmp & RADEON_MC_IDLE) { return 0; } DRM_UDELAY(1); @@ -1783,7 +1792,7 @@ int r100_gpu_reset(struct radeon_device *rdev) } /* Check if GPU is idle */ status = RREG32(RADEON_RBBM_STATUS); - if (status & (1 << 31)) { + if (status & RADEON_RBBM_ACTIVE) { DRM_ERROR("Failed to reset GPU (RBBM_STATUS=0x%08X)\n", status); return -1; } @@ -1793,6 +1802,9 @@ int r100_gpu_reset(struct radeon_device *rdev) void r100_set_common_regs(struct radeon_device *rdev) { + struct drm_device *dev = rdev->ddev; + bool force_dac2 = false; + /* set these so they don't interfere with anything */ WREG32(RADEON_OV0_SCALE_CNTL, 0); WREG32(RADEON_SUBPIC_CNTL, 0); @@ -1801,6 +1813,68 @@ void r100_set_common_regs(struct radeon_device *rdev) WREG32(RADEON_DVI_I2C_CNTL_1, 0); WREG32(RADEON_CAP0_TRIG_CNTL, 0); WREG32(RADEON_CAP1_TRIG_CNTL, 0); + + /* always set up dac2 on rn50 and some rv100 as lots + * of servers seem to wire it up to a VGA port but + * don't report it in the bios connector + * table. + */ + switch (dev->pdev->device) { + /* RN50 */ + case 0x515e: + case 0x5969: + force_dac2 = true; + break; + /* RV100*/ + case 0x5159: + case 0x515a: + /* DELL triple head servers */ + if ((dev->pdev->subsystem_vendor == 0x1028 /* DELL */) && + ((dev->pdev->subsystem_device == 0x016c) || + (dev->pdev->subsystem_device == 0x016d) || + (dev->pdev->subsystem_device == 0x016e) || + (dev->pdev->subsystem_device == 0x016f) || + (dev->pdev->subsystem_device == 0x0170) || + (dev->pdev->subsystem_device == 0x017d) || + (dev->pdev->subsystem_device == 0x017e) || + (dev->pdev->subsystem_device == 0x0183) || + (dev->pdev->subsystem_device == 0x018a) || + (dev->pdev->subsystem_device == 0x019a))) + force_dac2 = true; + break; + } + + if (force_dac2) { + u32 disp_hw_debug = RREG32(RADEON_DISP_HW_DEBUG); + u32 tv_dac_cntl = RREG32(RADEON_TV_DAC_CNTL); + u32 dac2_cntl = RREG32(RADEON_DAC_CNTL2); + + /* For CRT on DAC2, don't turn it on if BIOS didn't + enable it, even it's detected. + */ + + /* force it to crtc0 */ + dac2_cntl &= ~RADEON_DAC2_DAC_CLK_SEL; + dac2_cntl |= RADEON_DAC2_DAC2_CLK_SEL; + disp_hw_debug |= RADEON_CRT2_DISP1_SEL; + + /* set up the TV DAC */ + tv_dac_cntl &= ~(RADEON_TV_DAC_PEDESTAL | + RADEON_TV_DAC_STD_MASK | + RADEON_TV_DAC_RDACPD | + RADEON_TV_DAC_GDACPD | + RADEON_TV_DAC_BDACPD | + RADEON_TV_DAC_BGADJ_MASK | + RADEON_TV_DAC_DACADJ_MASK); + tv_dac_cntl |= (RADEON_TV_DAC_NBLANK | + RADEON_TV_DAC_NHOLD | + RADEON_TV_DAC_STD_PS2 | + (0x58 << 16)); + + WREG32(RADEON_TV_DAC_CNTL, tv_dac_cntl); + WREG32(RADEON_DISP_HW_DEBUG, disp_hw_debug); + WREG32(RADEON_DAC_CNTL2, dac2_cntl); + } } /* @@ -1882,17 +1956,20 @@ static u32 r100_get_accessible_vram(struct radeon_device *rdev) void r100_vram_init_sizes(struct radeon_device *rdev) { u64 config_aper_size; - u32 accessible; + /* work out accessible VRAM */ + rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); + rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); + rdev->mc.visible_vram_size = r100_get_accessible_vram(rdev); + /* FIXME we don't use the second aperture yet when we could use it */ + if (rdev->mc.visible_vram_size > rdev->mc.aper_size) + rdev->mc.visible_vram_size = rdev->mc.aper_size; config_aper_size = RREG32(RADEON_CONFIG_APER_SIZE); - if (rdev->flags & RADEON_IS_IGP) { uint32_t tom; /* read NB_TOM to get the amount of ram stolen for the GPU */ tom = RREG32(RADEON_NB_TOM); rdev->mc.real_vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16); - /* for IGPs we need to keep VRAM where it was put by the BIOS */ - rdev->mc.vram_location = (tom & 0xffff) << 16; WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } else { @@ -1904,30 +1981,19 @@ void r100_vram_init_sizes(struct radeon_device *rdev) rdev->mc.real_vram_size = 8192 * 1024; WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.real_vram_size); } - /* let driver place VRAM */ - rdev->mc.vram_location = 0xFFFFFFFFUL; - /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - - * Novell bug 204882 + along with lots of ubuntu ones */ + /* Fix for RN50, M6, M7 with 8/16/32(??) MBs of VRAM - + * Novell bug 204882 + along with lots of ubuntu ones + */ if (config_aper_size > rdev->mc.real_vram_size) rdev->mc.mc_vram_size = config_aper_size; else rdev->mc.mc_vram_size = rdev->mc.real_vram_size; } - - /* work out accessible VRAM */ - accessible = r100_get_accessible_vram(rdev); - - rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); - rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); - - if (accessible > rdev->mc.aper_size) - accessible = rdev->mc.aper_size; - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) + /* FIXME remove this once we support unmappable VRAM */ + if (rdev->mc.mc_vram_size > rdev->mc.aper_size) { rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) rdev->mc.real_vram_size = rdev->mc.aper_size; + } } void r100_vga_set_state(struct radeon_device *rdev, bool state) @@ -1944,11 +2010,18 @@ void r100_vga_set_state(struct radeon_device *rdev, bool state) WREG32(RADEON_CONFIG_CNTL, temp); } -void r100_vram_info(struct radeon_device *rdev) +void r100_mc_init(struct radeon_device *rdev) { - r100_vram_get_type(rdev); + u64 base; + r100_vram_get_type(rdev); r100_vram_init_sizes(rdev); + base = rdev->mc.aper_base; + if (rdev->flags & RADEON_IS_IGP) + base = (RREG32(RADEON_NB_TOM) & 0xffff) << 16; + radeon_vram_location(rdev, &rdev->mc, base); + if (!(rdev->flags & RADEON_IS_AGP)) + radeon_gtt_location(rdev, &rdev->mc); } @@ -3219,10 +3292,9 @@ void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save) { /* Update base address for crtc */ - WREG32(R_00023C_DISPLAY_BASE_ADDR, rdev->mc.vram_location); + WREG32(R_00023C_DISPLAY_BASE_ADDR, rdev->mc.vram_start); if (!(rdev->flags & RADEON_SINGLE_CRTC)) { - WREG32(R_00033C_CRTC2_DISPLAY_BASE_ADDR, - rdev->mc.vram_location); + WREG32(R_00033C_CRTC2_DISPLAY_BASE_ADDR, rdev->mc.vram_start); } /* Restore CRTC registers */ WREG8(R_0003C2_GENMO_WT, save->GENMO_WT); @@ -3368,7 +3440,6 @@ int r100_suspend(struct radeon_device *rdev) void r100_fini(struct radeon_device *rdev) { - r100_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -3384,34 +3455,6 @@ void r100_fini(struct radeon_device *rdev) rdev->bios = NULL; } -int r100_mc_init(struct radeon_device *rdev) -{ - int r; - u32 tmp; - - /* Setup GPU memory space */ - rdev->mc.vram_location = 0xFFFFFFFFUL; - rdev->mc.gtt_location = 0xFFFFFFFFUL; - if (rdev->flags & RADEON_IS_IGP) { - tmp = G_00015C_MC_FB_START(RREG32(R_00015C_NB_TOM)); - rdev->mc.vram_location = tmp << 16; - } - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) { - printk(KERN_WARNING "[drm] Disabling AGP\n"); - rdev->flags &= ~RADEON_IS_AGP; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; - } else { - rdev->mc.gtt_location = rdev->mc.agp_base; - } - } - r = radeon_mc_setup(rdev); - if (r) - return r; - return 0; -} - int r100_init(struct radeon_device *rdev) { int r; @@ -3454,12 +3497,15 @@ int r100_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - r100_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r100_mc_init(rdev); - if (r) - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } + } + /* initialize VRAM */ + r100_mc_init(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); if (r) @@ -3482,13 +3528,12 @@ int r100_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - r100_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_fini(rdev); - radeon_irq_kms_fini(rdev); rdev->accel_working = false; } return 0; diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index 20942127c46..1146c9909c2 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -31,6 +31,7 @@ #include "radeon_reg.h" #include "radeon.h" +#include "r100d.h" #include "r200_reg_safe.h" #include "r100_track.h" @@ -79,6 +80,51 @@ static int r200_get_vtx_size_0(uint32_t vtx_fmt_0) return vtx_size; } +int r200_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_pages, + struct radeon_fence *fence) +{ + uint32_t size; + uint32_t cur_size; + int i, num_loops; + int r = 0; + + /* radeon pitch is /64 */ + size = num_pages << PAGE_SHIFT; + num_loops = DIV_ROUND_UP(size, 0x1FFFFF); + r = radeon_ring_lock(rdev, num_loops * 4 + 64); + if (r) { + DRM_ERROR("radeon: moving bo (%d).\n", r); + return r; + } + /* Must wait for 2D idle & clean before DMA or hangs might happen */ + radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(rdev, (1 << 16)); + for (i = 0; i < num_loops; i++) { + cur_size = size; + if (cur_size > 0x1FFFFF) { + cur_size = 0x1FFFFF; + } + size -= cur_size; + radeon_ring_write(rdev, PACKET0(0x720, 2)); + radeon_ring_write(rdev, src_offset); + radeon_ring_write(rdev, dst_offset); + radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30)); + src_offset += cur_size; + dst_offset += cur_size; + } + radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE); + if (fence) { + r = radeon_fence_emit(rdev, fence); + } + radeon_ring_unlock_commit(rdev); + return r; +} + + static int r200_get_vtx_size_1(uint32_t vtx_fmt_1) { int vtx_size, i, tex_size; @@ -371,13 +417,16 @@ int r200_packet0_check(struct radeon_cs_parser *p, case 5: case 6: case 7: + /* 1D/2D */ track->textures[i].tex_coord_type = 0; break; case 1: - track->textures[i].tex_coord_type = 1; + /* CUBE */ + track->textures[i].tex_coord_type = 2; break; case 2: - track->textures[i].tex_coord_type = 2; + /* 3D */ + track->textures[i].tex_coord_type = 1; break; } break; diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 0051d11b907..4cef90cd74e 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -117,18 +117,19 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev) r = radeon_gart_table_vram_pin(rdev); if (r) return r; + radeon_gart_restore(rdev); /* discard memory request outside of configured range */ tmp = RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD; WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp); - WREG32_PCIE(RADEON_PCIE_TX_GART_START_LO, rdev->mc.gtt_location); - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - RADEON_GPU_PAGE_SIZE; + WREG32_PCIE(RADEON_PCIE_TX_GART_START_LO, rdev->mc.gtt_start); + tmp = rdev->mc.gtt_end & ~RADEON_GPU_PAGE_MASK; WREG32_PCIE(RADEON_PCIE_TX_GART_END_LO, tmp); WREG32_PCIE(RADEON_PCIE_TX_GART_START_HI, 0); WREG32_PCIE(RADEON_PCIE_TX_GART_END_HI, 0); table_addr = rdev->gart.table_addr; WREG32_PCIE(RADEON_PCIE_TX_GART_BASE, table_addr); /* FIXME: setup default page */ - WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_LO, rdev->mc.vram_location); + WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_LO, rdev->mc.vram_start); WREG32_PCIE(RADEON_PCIE_TX_DISCARD_RD_ADDR_HI, 0); /* Clear error */ WREG32_PCIE(0x18, 0); @@ -174,18 +175,20 @@ void r300_fence_ring_emit(struct radeon_device *rdev, /* Who ever call radeon_fence_emit should call ring_lock and ask * for enough space (today caller are ib schedule and buffer move) */ /* Write SC register so SC & US assert idle */ - radeon_ring_write(rdev, PACKET0(0x43E0, 0)); + radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_TL, 0)); radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(0x43E4, 0)); + radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_BR, 0)); radeon_ring_write(rdev, 0); /* Flush 3D cache */ - radeon_ring_write(rdev, PACKET0(0x4E4C, 0)); - radeon_ring_write(rdev, (2 << 0)); - radeon_ring_write(rdev, PACKET0(0x4F18, 0)); - radeon_ring_write(rdev, (1 << 0)); + radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(rdev, R300_RB3D_DC_FLUSH); + radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(rdev, R300_ZC_FLUSH); /* Wait until IDLE & CLEAN */ - radeon_ring_write(rdev, PACKET0(0x1720, 0)); - radeon_ring_write(rdev, (1 << 17) | (1 << 16) | (1 << 9)); + radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(rdev, (RADEON_WAIT_3D_IDLECLEAN | + RADEON_WAIT_2D_IDLECLEAN | + RADEON_WAIT_DMA_GUI_IDLE)); radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); radeon_ring_write(rdev, rdev->config.r300.hdp_cntl | RADEON_HDP_READ_BUFFER_INVALIDATE); @@ -198,50 +201,6 @@ void r300_fence_ring_emit(struct radeon_device *rdev, radeon_ring_write(rdev, RADEON_SW_INT_FIRE); } -int r300_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_pages, - struct radeon_fence *fence) -{ - uint32_t size; - uint32_t cur_size; - int i, num_loops; - int r = 0; - - /* radeon pitch is /64 */ - size = num_pages << PAGE_SHIFT; - num_loops = DIV_ROUND_UP(size, 0x1FFFFF); - r = radeon_ring_lock(rdev, num_loops * 4 + 64); - if (r) { - DRM_ERROR("radeon: moving bo (%d).\n", r); - return r; - } - /* Must wait for 2D idle & clean before DMA or hangs might happen */ - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0 )); - radeon_ring_write(rdev, (1 << 16)); - for (i = 0; i < num_loops; i++) { - cur_size = size; - if (cur_size > 0x1FFFFF) { - cur_size = 0x1FFFFF; - } - size -= cur_size; - radeon_ring_write(rdev, PACKET0(0x720, 2)); - radeon_ring_write(rdev, src_offset); - radeon_ring_write(rdev, dst_offset); - radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30)); - src_offset += cur_size; - dst_offset += cur_size; - } - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE); - if (fence) { - r = radeon_fence_emit(rdev, fence); - } - radeon_ring_unlock_commit(rdev); - return r; -} - void r300_ring_start(struct radeon_device *rdev) { unsigned gb_tile_config; @@ -281,8 +240,8 @@ void r300_ring_start(struct radeon_device *rdev) radeon_ring_write(rdev, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(0x170C, 0)); - radeon_ring_write(rdev, 1 << 31); + radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0)); + radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG); radeon_ring_write(rdev, PACKET0(R300_GB_SELECT, 0)); radeon_ring_write(rdev, 0); radeon_ring_write(rdev, PACKET0(R300_GB_ENABLE, 0)); @@ -349,8 +308,8 @@ int r300_mc_wait_for_idle(struct radeon_device *rdev) for (i = 0; i < rdev->usec_timeout; i++) { /* read MC_STATUS */ - tmp = RREG32(0x0150); - if (tmp & (1 << 4)) { + tmp = RREG32(RADEON_MC_STATUS); + if (tmp & R300_MC_IDLE) { return 0; } DRM_UDELAY(1); @@ -395,8 +354,8 @@ void r300_gpu_init(struct radeon_device *rdev) "programming pipes. Bad things might happen.\n"); } - tmp = RREG32(0x170C); - WREG32(0x170C, tmp | (1 << 31)); + tmp = RREG32(R300_DST_PIPE_CONFIG); + WREG32(R300_DST_PIPE_CONFIG, tmp | R300_PIPE_AUTO_CONFIG); WREG32(R300_RB2D_DSTCACHE_MODE, R300_DC_AUTOFLUSH_ENABLE | @@ -437,8 +396,8 @@ int r300_ga_reset(struct radeon_device *rdev) /* GA still busy soft reset it */ WREG32(0x429C, 0x200); WREG32(R300_VAP_PVS_STATE_FLUSH_REG, 0); - WREG32(0x43E0, 0); - WREG32(0x43E4, 0); + WREG32(R300_RE_SCISSORS_TL, 0); + WREG32(R300_RE_SCISSORS_BR, 0); WREG32(0x24AC, 0); } /* Wait to prevent race in RBBM_STATUS */ @@ -488,7 +447,7 @@ int r300_gpu_reset(struct radeon_device *rdev) } /* Check if GPU is idle */ status = RREG32(RADEON_RBBM_STATUS); - if (status & (1 << 31)) { + if (status & RADEON_RBBM_ACTIVE) { DRM_ERROR("Failed to reset GPU (RBBM_STATUS=0x%08X)\n", status); return -1; } @@ -500,20 +459,28 @@ int r300_gpu_reset(struct radeon_device *rdev) /* * r300,r350,rv350,rv380 VRAM info */ -void r300_vram_info(struct radeon_device *rdev) +void r300_mc_init(struct radeon_device *rdev) { - uint32_t tmp; + u64 base; + u32 tmp; /* DDR for all card after R300 & IGP */ rdev->mc.vram_is_ddr = true; tmp = RREG32(RADEON_MEM_CNTL); - if (tmp & R300_MEM_NUM_CHANNELS_MASK) { - rdev->mc.vram_width = 128; - } else { - rdev->mc.vram_width = 64; + tmp &= R300_MEM_NUM_CHANNELS_MASK; + switch (tmp) { + case 0: rdev->mc.vram_width = 64; break; + case 1: rdev->mc.vram_width = 128; break; + case 2: rdev->mc.vram_width = 256; break; + default: rdev->mc.vram_width = 128; break; } - r100_vram_init_sizes(rdev); + base = rdev->mc.aper_base; + if (rdev->flags & RADEON_IS_IGP) + base = (RREG32(RADEON_NB_TOM) & 0xffff) << 16; + radeon_vram_location(rdev, &rdev->mc, base); + if (!(rdev->flags & RADEON_IS_AGP)) + radeon_gtt_location(rdev, &rdev->mc); } void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes) @@ -575,6 +542,40 @@ void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes) } +int rv370_get_pcie_lanes(struct radeon_device *rdev) +{ + u32 link_width_cntl; + + if (rdev->flags & RADEON_IS_IGP) + return 0; + + if (!(rdev->flags & RADEON_IS_PCIE)) + return 0; + + /* FIXME wait for idle */ + + if (rdev->family < CHIP_R600) + link_width_cntl = RREG32_PCIE(RADEON_PCIE_LC_LINK_WIDTH_CNTL); + else + link_width_cntl = RREG32_PCIE_P(RADEON_PCIE_LC_LINK_WIDTH_CNTL); + + switch ((link_width_cntl & RADEON_PCIE_LC_LINK_WIDTH_RD_MASK) >> RADEON_PCIE_LC_LINK_WIDTH_RD_SHIFT) { + case RADEON_PCIE_LC_LINK_WIDTH_X0: + return 0; + case RADEON_PCIE_LC_LINK_WIDTH_X1: + return 1; + case RADEON_PCIE_LC_LINK_WIDTH_X2: + return 2; + case RADEON_PCIE_LC_LINK_WIDTH_X4: + return 4; + case RADEON_PCIE_LC_LINK_WIDTH_X8: + return 8; + case RADEON_PCIE_LC_LINK_WIDTH_X16: + default: + return 16; + } +} + #if defined(CONFIG_DEBUG_FS) static int rv370_debugfs_pcie_gart_info(struct seq_file *m, void *data) { @@ -704,6 +705,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, tile_flags |= R300_TXO_MACRO_TILE; if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_TXO_MICRO_TILE; + else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + tile_flags |= R300_TXO_MICRO_TILE_SQUARE; tmp = idx_value + ((u32)reloc->lobj.gpu_offset); tmp |= tile_flags; @@ -754,6 +757,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, tile_flags |= R300_COLOR_TILE_ENABLE; if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) tile_flags |= R300_COLOR_MICROTILE_ENABLE; + else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE; tmp = idx_value & ~(0x7 << 16); tmp |= tile_flags; @@ -825,7 +830,9 @@ static int r300_packet0_check(struct radeon_cs_parser *p, if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) tile_flags |= R300_DEPTHMACROTILE_ENABLE; if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) - tile_flags |= R300_DEPTHMICROTILE_TILED;; + tile_flags |= R300_DEPTHMICROTILE_TILED; + else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE) + tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE; tmp = idx_value & ~(0x7 << 16); tmp |= tile_flags; @@ -1327,7 +1334,6 @@ int r300_suspend(struct radeon_device *rdev) void r300_fini(struct radeon_device *rdev) { - r300_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -1385,12 +1391,15 @@ int r300_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - r300_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r420_mc_init(rdev); - if (r) - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } + } + /* initialize memory controller */ + r300_mc_init(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); if (r) @@ -1418,15 +1427,15 @@ int r300_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - r300_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_fini(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_fini(rdev); - radeon_irq_kms_fini(rdev); + radeon_agp_fini(rdev); rdev->accel_working = false; } return 0; diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c index 34bffa0e4b7..ea46d558e8f 100644 --- a/drivers/gpu/drm/radeon/r300_cmdbuf.c +++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c @@ -33,6 +33,7 @@ #include "drmP.h" #include "drm.h" +#include "drm_buffer.h" #include "radeon_drm.h" #include "radeon_drv.h" #include "r300_reg.h" @@ -299,46 +300,42 @@ static __inline__ int r300_emit_carefully_checked_packet0(drm_radeon_private_t * int reg; int sz; int i; - int values[64]; + u32 *value; RING_LOCALS; sz = header.packet0.count; reg = (header.packet0.reghi << 8) | header.packet0.reglo; if ((sz > 64) || (sz < 0)) { - DRM_ERROR - ("Cannot emit more than 64 values at a time (reg=%04x sz=%d)\n", - reg, sz); + DRM_ERROR("Cannot emit more than 64 values at a time (reg=%04x sz=%d)\n", + reg, sz); return -EINVAL; } + for (i = 0; i < sz; i++) { - values[i] = ((int *)cmdbuf->buf)[i]; switch (r300_reg_flags[(reg >> 2) + i]) { case MARK_SAFE: break; case MARK_CHECK_OFFSET: - if (!radeon_check_offset(dev_priv, (u32) values[i])) { - DRM_ERROR - ("Offset failed range check (reg=%04x sz=%d)\n", - reg, sz); + value = drm_buffer_pointer_to_dword(cmdbuf->buffer, i); + if (!radeon_check_offset(dev_priv, *value)) { + DRM_ERROR("Offset failed range check (reg=%04x sz=%d)\n", + reg, sz); return -EINVAL; } break; default: DRM_ERROR("Register %04x failed check as flag=%02x\n", - reg + i * 4, r300_reg_flags[(reg >> 2) + i]); + reg + i * 4, r300_reg_flags[(reg >> 2) + i]); return -EINVAL; } } BEGIN_RING(1 + sz); OUT_RING(CP_PACKET0(reg, sz - 1)); - OUT_RING_TABLE(values, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * 4; - cmdbuf->bufsz -= sz * 4; - return 0; } @@ -362,7 +359,7 @@ static __inline__ int r300_emit_packet0(drm_radeon_private_t *dev_priv, if (!sz) return 0; - if (sz * 4 > cmdbuf->bufsz) + if (sz * 4 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; if (reg + sz * 4 >= 0x10000) { @@ -380,12 +377,9 @@ static __inline__ int r300_emit_packet0(drm_radeon_private_t *dev_priv, BEGIN_RING(1 + sz); OUT_RING(CP_PACKET0(reg, sz - 1)); - OUT_RING_TABLE((int *)cmdbuf->buf, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * 4; - cmdbuf->bufsz -= sz * 4; - return 0; } @@ -407,7 +401,7 @@ static __inline__ int r300_emit_vpu(drm_radeon_private_t *dev_priv, if (!sz) return 0; - if (sz * 16 > cmdbuf->bufsz) + if (sz * 16 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; /* VAP is very sensitive so we purge cache before we program it @@ -426,7 +420,7 @@ static __inline__ int r300_emit_vpu(drm_radeon_private_t *dev_priv, BEGIN_RING(3 + sz * 4); OUT_RING_REG(R300_VAP_PVS_UPLOAD_ADDRESS, addr); OUT_RING(CP_PACKET0_TABLE(R300_VAP_PVS_UPLOAD_DATA, sz * 4 - 1)); - OUT_RING_TABLE((int *)cmdbuf->buf, sz * 4); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz * 4); ADVANCE_RING(); BEGIN_RING(2); @@ -434,9 +428,6 @@ static __inline__ int r300_emit_vpu(drm_radeon_private_t *dev_priv, OUT_RING(0); ADVANCE_RING(); - cmdbuf->buf += sz * 16; - cmdbuf->bufsz -= sz * 16; - return 0; } @@ -449,14 +440,14 @@ static __inline__ int r300_emit_clear(drm_radeon_private_t *dev_priv, { RING_LOCALS; - if (8 * 4 > cmdbuf->bufsz) + if (8 * 4 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; BEGIN_RING(10); OUT_RING(CP_PACKET3(R200_3D_DRAW_IMMD_2, 8)); OUT_RING(R300_PRIM_TYPE_POINT | R300_PRIM_WALK_RING | (1 << R300_PRIM_NUM_VERTICES_SHIFT)); - OUT_RING_TABLE((int *)cmdbuf->buf, 8); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, 8); ADVANCE_RING(); BEGIN_RING(4); @@ -468,9 +459,6 @@ static __inline__ int r300_emit_clear(drm_radeon_private_t *dev_priv, /* set flush flag */ dev_priv->track_flush |= RADEON_FLUSH_EMITED; - cmdbuf->buf += 8 * 4; - cmdbuf->bufsz -= 8 * 4; - return 0; } @@ -480,28 +468,29 @@ static __inline__ int r300_emit_3d_load_vbpntr(drm_radeon_private_t *dev_priv, { int count, i, k; #define MAX_ARRAY_PACKET 64 - u32 payload[MAX_ARRAY_PACKET]; + u32 *data; u32 narrays; RING_LOCALS; - count = (header >> 16) & 0x3fff; + count = (header & RADEON_CP_PACKET_COUNT_MASK) >> 16; if ((count + 1) > MAX_ARRAY_PACKET) { DRM_ERROR("Too large payload in 3D_LOAD_VBPNTR (count=%d)\n", count); return -EINVAL; } - memset(payload, 0, MAX_ARRAY_PACKET * 4); - memcpy(payload, cmdbuf->buf + 4, (count + 1) * 4); - /* carefully check packet contents */ - narrays = payload[0]; + /* We have already read the header so advance the buffer. */ + drm_buffer_advance(cmdbuf->buffer, 4); + + narrays = *(u32 *)drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); k = 0; i = 1; while ((k < narrays) && (i < (count + 1))) { i++; /* skip attribute field */ - if (!radeon_check_offset(dev_priv, payload[i])) { + data = drm_buffer_pointer_to_dword(cmdbuf->buffer, i); + if (!radeon_check_offset(dev_priv, *data)) { DRM_ERROR ("Offset failed range check (k=%d i=%d) while processing 3D_LOAD_VBPNTR packet.\n", k, i); @@ -512,7 +501,8 @@ static __inline__ int r300_emit_3d_load_vbpntr(drm_radeon_private_t *dev_priv, if (k == narrays) break; /* have one more to process, they come in pairs */ - if (!radeon_check_offset(dev_priv, payload[i])) { + data = drm_buffer_pointer_to_dword(cmdbuf->buffer, i); + if (!radeon_check_offset(dev_priv, *data)) { DRM_ERROR ("Offset failed range check (k=%d i=%d) while processing 3D_LOAD_VBPNTR packet.\n", k, i); @@ -533,30 +523,30 @@ static __inline__ int r300_emit_3d_load_vbpntr(drm_radeon_private_t *dev_priv, BEGIN_RING(count + 2); OUT_RING(header); - OUT_RING_TABLE(payload, count + 1); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, count + 1); ADVANCE_RING(); - cmdbuf->buf += (count + 2) * 4; - cmdbuf->bufsz -= (count + 2) * 4; - return 0; } static __inline__ int r300_emit_bitblt_multi(drm_radeon_private_t *dev_priv, drm_radeon_kcmd_buffer_t *cmdbuf) { - u32 *cmd = (u32 *) cmdbuf->buf; + u32 *cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); int count, ret; RING_LOCALS; - count=(cmd[0]>>16) & 0x3fff; - if (cmd[0] & 0x8000) { - u32 offset; + count = (*cmd & RADEON_CP_PACKET_COUNT_MASK) >> 16; - if (cmd[1] & (RADEON_GMC_SRC_PITCH_OFFSET_CNTL + if (*cmd & 0x8000) { + u32 offset; + u32 *cmd1 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + if (*cmd1 & (RADEON_GMC_SRC_PITCH_OFFSET_CNTL | RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { - offset = cmd[2] << 10; + + u32 *cmd2 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 2); + offset = *cmd2 << 10; ret = !radeon_check_offset(dev_priv, offset); if (ret) { DRM_ERROR("Invalid bitblt first offset is %08X\n", offset); @@ -564,9 +554,10 @@ static __inline__ int r300_emit_bitblt_multi(drm_radeon_private_t *dev_priv, } } - if ((cmd[1] & RADEON_GMC_SRC_PITCH_OFFSET_CNTL) && - (cmd[1] & RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { - offset = cmd[3] << 10; + if ((*cmd1 & RADEON_GMC_SRC_PITCH_OFFSET_CNTL) && + (*cmd1 & RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { + u32 *cmd3 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 3); + offset = *cmd3 << 10; ret = !radeon_check_offset(dev_priv, offset); if (ret) { DRM_ERROR("Invalid bitblt second offset is %08X\n", offset); @@ -577,28 +568,25 @@ static __inline__ int r300_emit_bitblt_multi(drm_radeon_private_t *dev_priv, } BEGIN_RING(count+2); - OUT_RING(cmd[0]); - OUT_RING_TABLE((int *)(cmdbuf->buf + 4), count + 1); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, count + 2); ADVANCE_RING(); - cmdbuf->buf += (count+2)*4; - cmdbuf->bufsz -= (count+2)*4; - return 0; } static __inline__ int r300_emit_draw_indx_2(drm_radeon_private_t *dev_priv, drm_radeon_kcmd_buffer_t *cmdbuf) { - u32 *cmd; + u32 *cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); + u32 *cmd1 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); int count; int expected_count; RING_LOCALS; - cmd = (u32 *) cmdbuf->buf; - count = (cmd[0]>>16) & 0x3fff; - expected_count = cmd[1] >> 16; - if (!(cmd[1] & R300_VAP_VF_CNTL__INDEX_SIZE_32bit)) + count = (*cmd & RADEON_CP_PACKET_COUNT_MASK) >> 16; + + expected_count = *cmd1 >> 16; + if (!(*cmd1 & R300_VAP_VF_CNTL__INDEX_SIZE_32bit)) expected_count = (expected_count+1)/2; if (count && count != expected_count) { @@ -608,55 +596,53 @@ static __inline__ int r300_emit_draw_indx_2(drm_radeon_private_t *dev_priv, } BEGIN_RING(count+2); - OUT_RING(cmd[0]); - OUT_RING_TABLE((int *)(cmdbuf->buf + 4), count + 1); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, count + 2); ADVANCE_RING(); - cmdbuf->buf += (count+2)*4; - cmdbuf->bufsz -= (count+2)*4; - if (!count) { - drm_r300_cmd_header_t header; + drm_r300_cmd_header_t stack_header, *header; + u32 *cmd1, *cmd2, *cmd3; - if (cmdbuf->bufsz < 4*4 + sizeof(header)) { + if (drm_buffer_unprocessed(cmdbuf->buffer) + < 4*4 + sizeof(stack_header)) { DRM_ERROR("3D_DRAW_INDX_2: expect subsequent INDX_BUFFER, but stream is too short.\n"); return -EINVAL; } - header.u = *(unsigned int *)cmdbuf->buf; + header = drm_buffer_read_object(cmdbuf->buffer, + sizeof(stack_header), &stack_header); - cmdbuf->buf += sizeof(header); - cmdbuf->bufsz -= sizeof(header); - cmd = (u32 *) cmdbuf->buf; + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); + cmd1 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + cmd2 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 2); + cmd3 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 3); - if (header.header.cmd_type != R300_CMD_PACKET3 || - header.packet3.packet != R300_CMD_PACKET3_RAW || - cmd[0] != CP_PACKET3(RADEON_CP_INDX_BUFFER, 2)) { + if (header->header.cmd_type != R300_CMD_PACKET3 || + header->packet3.packet != R300_CMD_PACKET3_RAW || + *cmd != CP_PACKET3(RADEON_CP_INDX_BUFFER, 2)) { DRM_ERROR("3D_DRAW_INDX_2: expect subsequent INDX_BUFFER.\n"); return -EINVAL; } - if ((cmd[1] & 0x8000ffff) != 0x80000810) { - DRM_ERROR("Invalid indx_buffer reg address %08X\n", cmd[1]); + if ((*cmd1 & 0x8000ffff) != 0x80000810) { + DRM_ERROR("Invalid indx_buffer reg address %08X\n", + *cmd1); return -EINVAL; } - if (!radeon_check_offset(dev_priv, cmd[2])) { - DRM_ERROR("Invalid indx_buffer offset is %08X\n", cmd[2]); + if (!radeon_check_offset(dev_priv, *cmd2)) { + DRM_ERROR("Invalid indx_buffer offset is %08X\n", + *cmd2); return -EINVAL; } - if (cmd[3] != expected_count) { + if (*cmd3 != expected_count) { DRM_ERROR("INDX_BUFFER: buffer size %i, expected %i\n", - cmd[3], expected_count); + *cmd3, expected_count); return -EINVAL; } BEGIN_RING(4); - OUT_RING(cmd[0]); - OUT_RING_TABLE((int *)(cmdbuf->buf + 4), 3); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, 4); ADVANCE_RING(); - - cmdbuf->buf += 4*4; - cmdbuf->bufsz -= 4*4; } return 0; @@ -665,39 +651,39 @@ static __inline__ int r300_emit_draw_indx_2(drm_radeon_private_t *dev_priv, static __inline__ int r300_emit_raw_packet3(drm_radeon_private_t *dev_priv, drm_radeon_kcmd_buffer_t *cmdbuf) { - u32 header; + u32 *header; int count; RING_LOCALS; - if (4 > cmdbuf->bufsz) + if (4 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; /* Fixme !! This simply emits a packet without much checking. We need to be smarter. */ /* obtain first word - actual packet3 header */ - header = *(u32 *) cmdbuf->buf; + header = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); /* Is it packet 3 ? */ - if ((header >> 30) != 0x3) { - DRM_ERROR("Not a packet3 header (0x%08x)\n", header); + if ((*header >> 30) != 0x3) { + DRM_ERROR("Not a packet3 header (0x%08x)\n", *header); return -EINVAL; } - count = (header >> 16) & 0x3fff; + count = (*header >> 16) & 0x3fff; /* Check again now that we know how much data to expect */ - if ((count + 2) * 4 > cmdbuf->bufsz) { + if ((count + 2) * 4 > drm_buffer_unprocessed(cmdbuf->buffer)) { DRM_ERROR ("Expected packet3 of length %d but have only %d bytes left\n", - (count + 2) * 4, cmdbuf->bufsz); + (count + 2) * 4, drm_buffer_unprocessed(cmdbuf->buffer)); return -EINVAL; } /* Is it a packet type we know about ? */ - switch (header & 0xff00) { + switch (*header & 0xff00) { case RADEON_3D_LOAD_VBPNTR: /* load vertex array pointers */ - return r300_emit_3d_load_vbpntr(dev_priv, cmdbuf, header); + return r300_emit_3d_load_vbpntr(dev_priv, cmdbuf, *header); case RADEON_CNTL_BITBLT_MULTI: return r300_emit_bitblt_multi(dev_priv, cmdbuf); @@ -723,18 +709,14 @@ static __inline__ int r300_emit_raw_packet3(drm_radeon_private_t *dev_priv, /* these packets are safe */ break; default: - DRM_ERROR("Unknown packet3 header (0x%08x)\n", header); + DRM_ERROR("Unknown packet3 header (0x%08x)\n", *header); return -EINVAL; } BEGIN_RING(count + 2); - OUT_RING(header); - OUT_RING_TABLE((int *)(cmdbuf->buf + 4), count + 1); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, count + 2); ADVANCE_RING(); - cmdbuf->buf += (count + 2) * 4; - cmdbuf->bufsz -= (count + 2) * 4; - return 0; } @@ -748,8 +730,7 @@ static __inline__ int r300_emit_packet3(drm_radeon_private_t *dev_priv, { int n; int ret; - char *orig_buf = cmdbuf->buf; - int orig_bufsz = cmdbuf->bufsz; + int orig_iter = cmdbuf->buffer->iterator; /* This is a do-while-loop so that we run the interior at least once, * even if cmdbuf->nbox is 0. Compare r300_emit_cliprects for rationale. @@ -761,8 +742,7 @@ static __inline__ int r300_emit_packet3(drm_radeon_private_t *dev_priv, if (ret) return ret; - cmdbuf->buf = orig_buf; - cmdbuf->bufsz = orig_bufsz; + cmdbuf->buffer->iterator = orig_iter; } switch (header.packet3.packet) { @@ -785,9 +765,9 @@ static __inline__ int r300_emit_packet3(drm_radeon_private_t *dev_priv, break; default: - DRM_ERROR("bad packet3 type %i at %p\n", + DRM_ERROR("bad packet3 type %i at byte %d\n", header.packet3.packet, - cmdbuf->buf - sizeof(header)); + cmdbuf->buffer->iterator - (int)sizeof(header)); return -EINVAL; } @@ -923,12 +903,13 @@ static int r300_scratch(drm_radeon_private_t *dev_priv, drm_r300_cmd_header_t header) { u32 *ref_age_base; - u32 i, buf_idx, h_pending; - u64 ptr_addr; + u32 i, *buf_idx, h_pending; + u64 *ptr_addr; + u64 stack_ptr_addr; RING_LOCALS; - if (cmdbuf->bufsz < - (sizeof(u64) + header.scratch.n_bufs * sizeof(buf_idx))) { + if (drm_buffer_unprocessed(cmdbuf->buffer) < + (sizeof(u64) + header.scratch.n_bufs * sizeof(*buf_idx))) { return -EINVAL; } @@ -938,36 +919,35 @@ static int r300_scratch(drm_radeon_private_t *dev_priv, dev_priv->scratch_ages[header.scratch.reg]++; - ptr_addr = get_unaligned((u64 *)cmdbuf->buf); - ref_age_base = (u32 *)(unsigned long)ptr_addr; - - cmdbuf->buf += sizeof(u64); - cmdbuf->bufsz -= sizeof(u64); + ptr_addr = drm_buffer_read_object(cmdbuf->buffer, + sizeof(stack_ptr_addr), &stack_ptr_addr); + ref_age_base = (u32 *)(unsigned long)*ptr_addr; for (i=0; i < header.scratch.n_bufs; i++) { - buf_idx = *(u32 *)cmdbuf->buf; - buf_idx *= 2; /* 8 bytes per buf */ + buf_idx = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); + *buf_idx *= 2; /* 8 bytes per buf */ - if (DRM_COPY_TO_USER(ref_age_base + buf_idx, &dev_priv->scratch_ages[header.scratch.reg], sizeof(u32))) { + if (DRM_COPY_TO_USER(ref_age_base + *buf_idx, + &dev_priv->scratch_ages[header.scratch.reg], + sizeof(u32))) return -EINVAL; - } - if (DRM_COPY_FROM_USER(&h_pending, ref_age_base + buf_idx + 1, sizeof(u32))) { + if (DRM_COPY_FROM_USER(&h_pending, + ref_age_base + *buf_idx + 1, + sizeof(u32))) return -EINVAL; - } - if (h_pending == 0) { + if (h_pending == 0) return -EINVAL; - } h_pending--; - if (DRM_COPY_TO_USER(ref_age_base + buf_idx + 1, &h_pending, sizeof(u32))) { + if (DRM_COPY_TO_USER(ref_age_base + *buf_idx + 1, + &h_pending, + sizeof(u32))) return -EINVAL; - } - cmdbuf->buf += sizeof(buf_idx); - cmdbuf->bufsz -= sizeof(buf_idx); + drm_buffer_advance(cmdbuf->buffer, sizeof(*buf_idx)); } BEGIN_RING(2); @@ -1009,19 +989,16 @@ static inline int r300_emit_r500fp(drm_radeon_private_t *dev_priv, DRM_DEBUG("r500fp %d %d type: %d\n", sz, addr, type); if (!sz) return 0; - if (sz * stride * 4 > cmdbuf->bufsz) + if (sz * stride * 4 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; BEGIN_RING(3 + sz * stride); OUT_RING_REG(R500_GA_US_VECTOR_INDEX, addr); OUT_RING(CP_PACKET0_TABLE(R500_GA_US_VECTOR_DATA, sz * stride - 1)); - OUT_RING_TABLE((int *)cmdbuf->buf, sz * stride); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz * stride); ADVANCE_RING(); - cmdbuf->buf += sz * stride * 4; - cmdbuf->bufsz -= sz * stride * 4; - return 0; } @@ -1053,19 +1030,18 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, goto cleanup; } - while (cmdbuf->bufsz >= sizeof(drm_r300_cmd_header_t)) { + while (drm_buffer_unprocessed(cmdbuf->buffer) + >= sizeof(drm_r300_cmd_header_t)) { int idx; - drm_r300_cmd_header_t header; - - header.u = *(unsigned int *)cmdbuf->buf; + drm_r300_cmd_header_t *header, stack_header; - cmdbuf->buf += sizeof(header); - cmdbuf->bufsz -= sizeof(header); + header = drm_buffer_read_object(cmdbuf->buffer, + sizeof(stack_header), &stack_header); - switch (header.header.cmd_type) { + switch (header->header.cmd_type) { case R300_CMD_PACKET0: DRM_DEBUG("R300_CMD_PACKET0\n"); - ret = r300_emit_packet0(dev_priv, cmdbuf, header); + ret = r300_emit_packet0(dev_priv, cmdbuf, *header); if (ret) { DRM_ERROR("r300_emit_packet0 failed\n"); goto cleanup; @@ -1074,7 +1050,7 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, case R300_CMD_VPU: DRM_DEBUG("R300_CMD_VPU\n"); - ret = r300_emit_vpu(dev_priv, cmdbuf, header); + ret = r300_emit_vpu(dev_priv, cmdbuf, *header); if (ret) { DRM_ERROR("r300_emit_vpu failed\n"); goto cleanup; @@ -1083,7 +1059,7 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, case R300_CMD_PACKET3: DRM_DEBUG("R300_CMD_PACKET3\n"); - ret = r300_emit_packet3(dev_priv, cmdbuf, header); + ret = r300_emit_packet3(dev_priv, cmdbuf, *header); if (ret) { DRM_ERROR("r300_emit_packet3 failed\n"); goto cleanup; @@ -1117,8 +1093,8 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, int i; RING_LOCALS; - BEGIN_RING(header.delay.count); - for (i = 0; i < header.delay.count; i++) + BEGIN_RING(header->delay.count); + for (i = 0; i < header->delay.count; i++) OUT_RING(RADEON_CP_PACKET2); ADVANCE_RING(); } @@ -1126,7 +1102,7 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, case R300_CMD_DMA_DISCARD: DRM_DEBUG("RADEON_CMD_DMA_DISCARD\n"); - idx = header.dma.buf_idx; + idx = header->dma.buf_idx; if (idx < 0 || idx >= dma->buf_count) { DRM_ERROR("buffer index %d (of %d max)\n", idx, dma->buf_count - 1); @@ -1149,12 +1125,12 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, case R300_CMD_WAIT: DRM_DEBUG("R300_CMD_WAIT\n"); - r300_cmd_wait(dev_priv, header); + r300_cmd_wait(dev_priv, *header); break; case R300_CMD_SCRATCH: DRM_DEBUG("R300_CMD_SCRATCH\n"); - ret = r300_scratch(dev_priv, cmdbuf, header); + ret = r300_scratch(dev_priv, cmdbuf, *header); if (ret) { DRM_ERROR("r300_scratch failed\n"); goto cleanup; @@ -1168,16 +1144,16 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, goto cleanup; } DRM_DEBUG("R300_CMD_R500FP\n"); - ret = r300_emit_r500fp(dev_priv, cmdbuf, header); + ret = r300_emit_r500fp(dev_priv, cmdbuf, *header); if (ret) { DRM_ERROR("r300_emit_r500fp failed\n"); goto cleanup; } break; default: - DRM_ERROR("bad cmd_type %i at %p\n", - header.header.cmd_type, - cmdbuf->buf - sizeof(header)); + DRM_ERROR("bad cmd_type %i at byte %d\n", + header->header.cmd_type, + cmdbuf->buffer->iterator - (int)sizeof(*header)); ret = -EINVAL; goto cleanup; } diff --git a/drivers/gpu/drm/radeon/r300_reg.h b/drivers/gpu/drm/radeon/r300_reg.h index 1735a2b6958..1a0d5362cd7 100644 --- a/drivers/gpu/drm/radeon/r300_reg.h +++ b/drivers/gpu/drm/radeon/r300_reg.h @@ -952,6 +952,7 @@ # define R300_TXO_ENDIAN_HALFDW_SWAP (3 << 0) # define R300_TXO_MACRO_TILE (1 << 2) # define R300_TXO_MICRO_TILE (1 << 3) +# define R300_TXO_MICRO_TILE_SQUARE (2 << 3) # define R300_TXO_OFFSET_MASK 0xffffffe0 # define R300_TXO_OFFSET_SHIFT 5 /* END: Guess from R200 */ @@ -1360,6 +1361,7 @@ # define R300_COLORPITCH_MASK 0x00001FF8 /* GUESS */ # define R300_COLOR_TILE_ENABLE (1 << 16) /* GUESS */ # define R300_COLOR_MICROTILE_ENABLE (1 << 17) /* GUESS */ +# define R300_COLOR_MICROTILE_SQUARE_ENABLE (2 << 17) # define R300_COLOR_ENDIAN_NO_SWAP (0 << 18) /* GUESS */ # define R300_COLOR_ENDIAN_WORD_SWAP (1 << 18) /* GUESS */ # define R300_COLOR_ENDIAN_DWORD_SWAP (2 << 18) /* GUESS */ diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c index 053404e71a9..c7593b8f58e 100644 --- a/drivers/gpu/drm/radeon/r420.c +++ b/drivers/gpu/drm/radeon/r420.c @@ -40,30 +40,6 @@ static void r420_set_reg_safe(struct radeon_device *rdev) rdev->config.r300.reg_safe_bm_size = ARRAY_SIZE(r420_reg_safe_bm); } -int r420_mc_init(struct radeon_device *rdev) -{ - int r; - - /* Setup GPU memory space */ - rdev->mc.vram_location = 0xFFFFFFFFUL; - rdev->mc.gtt_location = 0xFFFFFFFFUL; - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) { - printk(KERN_WARNING "[drm] Disabling AGP\n"); - rdev->flags &= ~RADEON_IS_AGP; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; - } else { - rdev->mc.gtt_location = rdev->mc.agp_base; - } - } - r = radeon_mc_setup(rdev); - if (r) { - return r; - } - return 0; -} - void r420_pipes_init(struct radeon_device *rdev) { unsigned tmp; @@ -71,7 +47,8 @@ void r420_pipes_init(struct radeon_device *rdev) unsigned num_pipes; /* GA_ENHANCE workaround TCL deadlock issue */ - WREG32(0x4274, (1 << 0) | (1 << 1) | (1 << 2) | (1 << 3)); + WREG32(R300_GA_ENHANCE, R300_GA_DEADLOCK_CNTL | R300_GA_FASTSYNC_CNTL | + (1 << 2) | (1 << 3)); /* add idle wait as per freedesktop.org bug 24041 */ if (r100_gui_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait GUI idle while " @@ -99,17 +76,17 @@ void r420_pipes_init(struct radeon_device *rdev) tmp = (7 << 1); break; } - WREG32(0x42C8, (1 << num_pipes) - 1); + WREG32(R500_SU_REG_DEST, (1 << num_pipes) - 1); /* Sub pixel 1/12 so we can have 4K rendering according to doc */ - tmp |= (1 << 4) | (1 << 0); - WREG32(0x4018, tmp); + tmp |= R300_TILE_SIZE_16 | R300_ENABLE_TILING; + WREG32(R300_GB_TILE_CONFIG, tmp); if (r100_gui_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait GUI idle while " "programming pipes. Bad things might happen.\n"); } - tmp = RREG32(0x170C); - WREG32(0x170C, tmp | (1 << 31)); + tmp = RREG32(R300_DST_PIPE_CONFIG); + WREG32(R300_DST_PIPE_CONFIG, tmp | R300_PIPE_AUTO_CONFIG); WREG32(R300_RB2D_DSTCACHE_MODE, RREG32(R300_RB2D_DSTCACHE_MODE) | @@ -350,13 +327,15 @@ int r420_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - r300_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r420_mc_init(rdev); - if (r) { - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } } + /* initialize memory controller */ + r300_mc_init(rdev); r420_debugfs(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); @@ -391,16 +370,15 @@ int r420_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - r420_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_fini(rdev); if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_fini(rdev); radeon_agp_fini(rdev); - radeon_irq_kms_fini(rdev); rdev->accel_working = false; } return 0; diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h index 74ad89bdf2b..0cf2ad2a558 100644 --- a/drivers/gpu/drm/radeon/r500_reg.h +++ b/drivers/gpu/drm/radeon/r500_reg.h @@ -717,54 +717,62 @@ #define AVIVO_DVOA_BIT_DEPTH_CONTROL 0x7988 #define AVIVO_DC_GPIO_HPD_A 0x7e94 - -#define AVIVO_GPIO_0 0x7e30 -#define AVIVO_GPIO_1 0x7e40 -#define AVIVO_GPIO_2 0x7e50 -#define AVIVO_GPIO_3 0x7e60 - #define AVIVO_DC_GPIO_HPD_Y 0x7e9c -#define AVIVO_I2C_STATUS 0x7d30 -# define AVIVO_I2C_STATUS_DONE (1 << 0) -# define AVIVO_I2C_STATUS_NACK (1 << 1) -# define AVIVO_I2C_STATUS_HALT (1 << 2) -# define AVIVO_I2C_STATUS_GO (1 << 3) -# define AVIVO_I2C_STATUS_MASK 0x7 -/* If radeon_mm_i2c is to be believed, this is HALT, NACK, and maybe - * DONE? */ -# define AVIVO_I2C_STATUS_CMD_RESET 0x7 -# define AVIVO_I2C_STATUS_CMD_WAIT (1 << 3) -#define AVIVO_I2C_STOP 0x7d34 -#define AVIVO_I2C_START_CNTL 0x7d38 -# define AVIVO_I2C_START (1 << 8) -# define AVIVO_I2C_CONNECTOR0 (0 << 16) -# define AVIVO_I2C_CONNECTOR1 (1 << 16) -#define R520_I2C_START (1<<0) -#define R520_I2C_STOP (1<<1) -#define R520_I2C_RX (1<<2) -#define R520_I2C_EN (1<<8) -#define R520_I2C_DDC1 (0<<16) -#define R520_I2C_DDC2 (1<<16) -#define R520_I2C_DDC3 (2<<16) -#define R520_I2C_DDC_MASK (3<<16) -#define AVIVO_I2C_CONTROL2 0x7d3c -# define AVIVO_I2C_7D3C_SIZE_SHIFT 8 -# define AVIVO_I2C_7D3C_SIZE_MASK (0xf << 8) -#define AVIVO_I2C_CONTROL3 0x7d40 -/* Reading is done 4 bytes at a time: read the bottom 8 bits from - * 7d44, four times in a row. - * Writing is a little more complex. First write DATA with - * 0xnnnnnnzz, then 0xnnnnnnyy, where nnnnnn is some non-deterministic - * magic number, zz is, I think, the slave address, and yy is the byte - * you want to write. */ -#define AVIVO_I2C_DATA 0x7d44 -#define R520_I2C_ADDR_COUNT_MASK (0x7) -#define R520_I2C_DATA_COUNT_SHIFT (8) -#define R520_I2C_DATA_COUNT_MASK (0xF00) -#define AVIVO_I2C_CNTL 0x7d50 -# define AVIVO_I2C_EN (1 << 0) -# define AVIVO_I2C_RESET (1 << 8) +#define AVIVO_DC_I2C_STATUS1 0x7d30 +# define AVIVO_DC_I2C_DONE (1 << 0) +# define AVIVO_DC_I2C_NACK (1 << 1) +# define AVIVO_DC_I2C_HALT (1 << 2) +# define AVIVO_DC_I2C_GO (1 << 3) +#define AVIVO_DC_I2C_RESET 0x7d34 +# define AVIVO_DC_I2C_SOFT_RESET (1 << 0) +# define AVIVO_DC_I2C_ABORT (1 << 8) +#define AVIVO_DC_I2C_CONTROL1 0x7d38 +# define AVIVO_DC_I2C_START (1 << 0) +# define AVIVO_DC_I2C_STOP (1 << 1) +# define AVIVO_DC_I2C_RECEIVE (1 << 2) +# define AVIVO_DC_I2C_EN (1 << 8) +# define AVIVO_DC_I2C_PIN_SELECT(x) ((x) << 16) +# define AVIVO_SEL_DDC1 0 +# define AVIVO_SEL_DDC2 1 +# define AVIVO_SEL_DDC3 2 +#define AVIVO_DC_I2C_CONTROL2 0x7d3c +# define AVIVO_DC_I2C_ADDR_COUNT(x) ((x) << 0) +# define AVIVO_DC_I2C_DATA_COUNT(x) ((x) << 8) +#define AVIVO_DC_I2C_CONTROL3 0x7d40 +# define AVIVO_DC_I2C_DATA_DRIVE_EN (1 << 0) +# define AVIVO_DC_I2C_DATA_DRIVE_SEL (1 << 1) +# define AVIVO_DC_I2C_CLK_DRIVE_EN (1 << 7) +# define AVIVO_DC_I2C_RD_INTRA_BYTE_DELAY(x) ((x) << 8) +# define AVIVO_DC_I2C_WR_INTRA_BYTE_DELAY(x) ((x) << 16) +# define AVIVO_DC_I2C_TIME_LIMIT(x) ((x) << 24) +#define AVIVO_DC_I2C_DATA 0x7d44 +#define AVIVO_DC_I2C_INTERRUPT_CONTROL 0x7d48 +# define AVIVO_DC_I2C_INTERRUPT_STATUS (1 << 0) +# define AVIVO_DC_I2C_INTERRUPT_AK (1 << 8) +# define AVIVO_DC_I2C_INTERRUPT_ENABLE (1 << 16) +#define AVIVO_DC_I2C_ARBITRATION 0x7d50 +# define AVIVO_DC_I2C_SW_WANTS_TO_USE_I2C (1 << 0) +# define AVIVO_DC_I2C_SW_CAN_USE_I2C (1 << 1) +# define AVIVO_DC_I2C_SW_DONE_USING_I2C (1 << 8) +# define AVIVO_DC_I2C_HW_NEEDS_I2C (1 << 9) +# define AVIVO_DC_I2C_ABORT_HDCP_I2C (1 << 16) +# define AVIVO_DC_I2C_HW_USING_I2C (1 << 17) + +#define AVIVO_DC_GPIO_DDC1_MASK 0x7e40 +#define AVIVO_DC_GPIO_DDC1_A 0x7e44 +#define AVIVO_DC_GPIO_DDC1_EN 0x7e48 +#define AVIVO_DC_GPIO_DDC1_Y 0x7e4c + +#define AVIVO_DC_GPIO_DDC2_MASK 0x7e50 +#define AVIVO_DC_GPIO_DDC2_A 0x7e54 +#define AVIVO_DC_GPIO_DDC2_EN 0x7e58 +#define AVIVO_DC_GPIO_DDC2_Y 0x7e5c + +#define AVIVO_DC_GPIO_DDC3_MASK 0x7e60 +#define AVIVO_DC_GPIO_DDC3_A 0x7e64 +#define AVIVO_DC_GPIO_DDC3_EN 0x7e68 +#define AVIVO_DC_GPIO_DDC3_Y 0x7e6c #define AVIVO_DISP_INTERRUPT_STATUS 0x7edc # define AVIVO_D1_VBLANK_INTERRUPT (1 << 4) diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c index 9a189072f2b..2b8a5dd1351 100644 --- a/drivers/gpu/drm/radeon/r520.c +++ b/drivers/gpu/drm/radeon/r520.c @@ -119,13 +119,15 @@ static void r520_vram_get_type(struct radeon_device *rdev) rdev->mc.vram_width *= 2; } -void r520_vram_info(struct radeon_device *rdev) +void r520_mc_init(struct radeon_device *rdev) { fixed20_12 a; r520_vram_get_type(rdev); - r100_vram_init_sizes(rdev); + radeon_vram_location(rdev, &rdev->mc, 0); + if (!(rdev->flags & RADEON_IS_AGP)) + radeon_gtt_location(rdev, &rdev->mc); /* FIXME: we should enforce default clock in case GPU is not in * default setup */ @@ -267,12 +269,15 @@ int r520_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - r520_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r420_mc_init(rdev); - if (r) - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } + } + /* initialize memory controller */ + r520_mc_init(rdev); rv515_debugfs(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); @@ -294,13 +299,12 @@ int r520_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - rv515_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); rv370_pcie_gart_fini(rdev); radeon_agp_fini(rdev); - radeon_irq_kms_fini(rdev); rdev->accel_working = false; } return 0; diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index f5ff3490929..c5229019729 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -353,23 +353,14 @@ void r600_hpd_fini(struct radeon_device *rdev) /* * R600 PCIE GART */ -int r600_gart_clear_page(struct radeon_device *rdev, int i) -{ - void __iomem *ptr = (void *)rdev->gart.table.vram.ptr; - u64 pte; - - if (i < 0 || i > rdev->gart.num_gpu_pages) - return -EINVAL; - pte = 0; - writeq(pte, ((void __iomem *)ptr) + (i * 8)); - return 0; -} - void r600_pcie_gart_tlb_flush(struct radeon_device *rdev) { unsigned i; u32 tmp; + /* flush hdp cache so updates hit vram */ + WREG32(R_005480_HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + WREG32(VM_CONTEXT0_INVALIDATION_LOW_ADDR, rdev->mc.gtt_start >> 12); WREG32(VM_CONTEXT0_INVALIDATION_HIGH_ADDR, (rdev->mc.gtt_end - 1) >> 12); WREG32(VM_CONTEXT0_REQUEST_RESPONSE, REQUEST_TYPE(1)); @@ -416,6 +407,7 @@ int r600_pcie_gart_enable(struct radeon_device *rdev) r = radeon_gart_table_vram_pin(rdev); if (r) return r; + radeon_gart_restore(rdev); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING | @@ -619,12 +611,73 @@ static void r600_mc_program(struct radeon_device *rdev) rv515_vga_render_disable(rdev); } +/** + * r600_vram_gtt_location - try to find VRAM & GTT location + * @rdev: radeon device structure holding all necessary informations + * @mc: memory controller structure holding memory informations + * + * Function will place try to place VRAM at same place as in CPU (PCI) + * address space as some GPU seems to have issue when we reprogram at + * different address space. + * + * If there is not enough space to fit the unvisible VRAM after the + * aperture then we limit the VRAM size to the aperture. + * + * If we are using AGP then place VRAM adjacent to AGP aperture are we need + * them to be in one from GPU point of view so that we can program GPU to + * catch access outside them (weird GPU policy see ??). + * + * This function will never fails, worst case are limiting VRAM or GTT. + * + * Note: GTT start, end, size should be initialized before calling this + * function on AGP platform. + */ +void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) +{ + u64 size_bf, size_af; + + if (mc->mc_vram_size > 0xE0000000) { + /* leave room for at least 512M GTT */ + dev_warn(rdev->dev, "limiting VRAM\n"); + mc->real_vram_size = 0xE0000000; + mc->mc_vram_size = 0xE0000000; + } + if (rdev->flags & RADEON_IS_AGP) { + size_bf = mc->gtt_start; + size_af = 0xFFFFFFFF - mc->gtt_end + 1; + if (size_bf > size_af) { + if (mc->mc_vram_size > size_bf) { + dev_warn(rdev->dev, "limiting VRAM\n"); + mc->real_vram_size = size_bf; + mc->mc_vram_size = size_bf; + } + mc->vram_start = mc->gtt_start - mc->mc_vram_size; + } else { + if (mc->mc_vram_size > size_af) { + dev_warn(rdev->dev, "limiting VRAM\n"); + mc->real_vram_size = size_af; + mc->mc_vram_size = size_af; + } + mc->vram_start = mc->gtt_end; + } + mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; + dev_info(rdev->dev, "VRAM: %lluM 0x%08llX - 0x%08llX (%lluM used)\n", + mc->mc_vram_size >> 20, mc->vram_start, + mc->vram_end, mc->real_vram_size >> 20); + } else { + u64 base = 0; + if (rdev->flags & RADEON_IS_IGP) + base = (RREG32(MC_VM_FB_LOCATION) & 0xFFFF) << 24; + radeon_vram_location(rdev, &rdev->mc, base); + radeon_gtt_location(rdev, mc); + } +} + int r600_mc_init(struct radeon_device *rdev) { fixed20_12 a; u32 tmp; int chansize, numchan; - int r; /* Get VRAM informations */ rdev->mc.vram_is_ddr = true; @@ -659,78 +712,21 @@ int r600_mc_init(struct radeon_device *rdev) /* Setup GPU memory space */ rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE); rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE); - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) + rdev->mc.visible_vram_size = rdev->mc.aper_size; + /* FIXME remove this once we support unmappable VRAM */ + if (rdev->mc.mc_vram_size > rdev->mc.aper_size) { rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) rdev->mc.real_vram_size = rdev->mc.aper_size; - - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) - return r; - /* gtt_size is setup by radeon_agp_init */ - rdev->mc.gtt_location = rdev->mc.agp_base; - tmp = 0xFFFFFFFFUL - rdev->mc.agp_base - rdev->mc.gtt_size; - /* Try to put vram before or after AGP because we - * we want SYSTEM_APERTURE to cover both VRAM and - * AGP so that GPU can catch out of VRAM/AGP access - */ - if (rdev->mc.gtt_location > rdev->mc.mc_vram_size) { - /* Enough place before */ - rdev->mc.vram_location = rdev->mc.gtt_location - - rdev->mc.mc_vram_size; - } else if (tmp > rdev->mc.mc_vram_size) { - /* Enough place after */ - rdev->mc.vram_location = rdev->mc.gtt_location + - rdev->mc.gtt_size; - } else { - /* Try to setup VRAM then AGP might not - * not work on some card - */ - rdev->mc.vram_location = 0x00000000UL; - rdev->mc.gtt_location = rdev->mc.mc_vram_size; - } - } else { - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; - rdev->mc.vram_location = (RREG32(MC_VM_FB_LOCATION) & - 0xFFFF) << 24; - tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size; - if ((0xFFFFFFFFUL - tmp) >= rdev->mc.gtt_size) { - /* Enough place after vram */ - rdev->mc.gtt_location = tmp; - } else if (rdev->mc.vram_location >= rdev->mc.gtt_size) { - /* Enough place before vram */ - rdev->mc.gtt_location = 0; - } else { - /* Not enough place after or before shrink - * gart size - */ - if (rdev->mc.vram_location > (0xFFFFFFFFUL - tmp)) { - rdev->mc.gtt_location = 0; - rdev->mc.gtt_size = rdev->mc.vram_location; - } else { - rdev->mc.gtt_location = tmp; - rdev->mc.gtt_size = 0xFFFFFFFFUL - tmp; - } - } - rdev->mc.gtt_location = rdev->mc.mc_vram_size; } - rdev->mc.vram_start = rdev->mc.vram_location; - rdev->mc.vram_end = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1; - rdev->mc.gtt_start = rdev->mc.gtt_location; - rdev->mc.gtt_end = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; + r600_vram_gtt_location(rdev, &rdev->mc); /* FIXME: we should enforce default clock in case GPU is not in * default setup */ a.full = rfixed_const(100); rdev->pm.sclk.full = rfixed_const(rdev->clock.default_sclk); rdev->pm.sclk.full = rfixed_div(rdev->pm.sclk, a); - if (rdev->flags & RADEON_IS_IGP) rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); - return 0; } @@ -985,6 +981,9 @@ void r600_gpu_init(struct radeon_device *rdev) { u32 tiling_config; u32 ramcfg; + u32 backend_map; + u32 cc_rb_backend_disable; + u32 cc_gc_shader_pipe_config; u32 tmp; int i, j; u32 sq_config; @@ -1094,8 +1093,11 @@ void r600_gpu_init(struct radeon_device *rdev) default: break; } + rdev->config.r600.tiling_npipes = rdev->config.r600.max_tile_pipes; + rdev->config.r600.tiling_nbanks = 4 << ((ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT); tiling_config |= BANK_TILING((ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT); tiling_config |= GROUP_SIZE(0); + rdev->config.r600.tiling_group_size = 256; tmp = (ramcfg & NOOFROWS_MASK) >> NOOFROWS_SHIFT; if (tmp > 3) { tiling_config |= ROW_TILING(3); @@ -1105,24 +1107,33 @@ void r600_gpu_init(struct radeon_device *rdev) tiling_config |= SAMPLE_SPLIT(tmp); } tiling_config |= BANK_SWAPS(1); - tmp = r600_get_tile_pipe_to_backend_map(rdev->config.r600.max_tile_pipes, - rdev->config.r600.max_backends, - (0xff << rdev->config.r600.max_backends) & 0xff); - tiling_config |= BACKEND_MAP(tmp); + + cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000; + cc_rb_backend_disable |= + BACKEND_DISABLE((R6XX_MAX_BACKENDS_MASK << rdev->config.r600.max_backends) & R6XX_MAX_BACKENDS_MASK); + + cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; + cc_gc_shader_pipe_config |= + INACTIVE_QD_PIPES((R6XX_MAX_PIPES_MASK << rdev->config.r600.max_pipes) & R6XX_MAX_PIPES_MASK); + cc_gc_shader_pipe_config |= + INACTIVE_SIMDS((R6XX_MAX_SIMDS_MASK << rdev->config.r600.max_simds) & R6XX_MAX_SIMDS_MASK); + + backend_map = r600_get_tile_pipe_to_backend_map(rdev->config.r600.max_tile_pipes, + (R6XX_MAX_BACKENDS - + r600_count_pipe_bits((cc_rb_backend_disable & + R6XX_MAX_BACKENDS_MASK) >> 16)), + (cc_rb_backend_disable >> 16)); + + tiling_config |= BACKEND_MAP(backend_map); WREG32(GB_TILING_CONFIG, tiling_config); WREG32(DCP_TILING_CONFIG, tiling_config & 0xffff); WREG32(HDP_TILING_CONFIG, tiling_config & 0xffff); - tmp = BACKEND_DISABLE((R6XX_MAX_BACKENDS_MASK << rdev->config.r600.max_backends) & R6XX_MAX_BACKENDS_MASK); - WREG32(CC_RB_BACKEND_DISABLE, tmp); - /* Setup pipes */ - tmp = INACTIVE_QD_PIPES((R6XX_MAX_PIPES_MASK << rdev->config.r600.max_pipes) & R6XX_MAX_PIPES_MASK); - tmp |= INACTIVE_SIMDS((R6XX_MAX_SIMDS_MASK << rdev->config.r600.max_simds) & R6XX_MAX_SIMDS_MASK); - WREG32(CC_GC_SHADER_PIPE_CONFIG, tmp); - WREG32(GC_USER_SHADER_PIPE_CONFIG, tmp); + WREG32(CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); + WREG32(CC_GC_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); - tmp = R6XX_MAX_BACKENDS - r600_count_pipe_bits(tmp & INACTIVE_QD_PIPES_MASK); + tmp = R6XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config & INACTIVE_QD_PIPES_MASK) >> 8); WREG32(VGT_OUT_DEALLOC_CNTL, (tmp * 4) & DEALLOC_DIST_MASK); WREG32(VGT_VERTEX_REUSE_BLOCK_CNTL, ((tmp * 4) - 2) & VTX_REUSE_DEPTH_MASK); @@ -1658,6 +1669,12 @@ void r600_ring_init(struct radeon_device *rdev, unsigned ring_size) rdev->cp.align_mask = 16 - 1; } +void r600_cp_fini(struct radeon_device *rdev) +{ + r600_cp_stop(rdev); + radeon_ring_fini(rdev); +} + /* * GPU scratch registers helpers function. @@ -1781,34 +1798,40 @@ void r600_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { /* Also consider EVENT_WRITE_EOP. it handles the interrupts + timestamps + events */ + + radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE, 0)); + radeon_ring_write(rdev, CACHE_FLUSH_AND_INV_EVENT); + /* wait for 3D idle clean */ + radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(rdev, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(rdev, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit); /* Emit fence sequence & fire IRQ */ radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); radeon_ring_write(rdev, ((rdev->fence_drv.scratch_reg - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); radeon_ring_write(rdev, fence->seq); - radeon_ring_write(rdev, PACKET0(R_005480_HDP_MEM_COHERENCY_FLUSH_CNTL, 0)); - radeon_ring_write(rdev, 1); /* CP_INTERRUPT packet 3 no longer exists, use packet 0 */ radeon_ring_write(rdev, PACKET0(CP_INT_STATUS, 0)); radeon_ring_write(rdev, RB_INT_STAT); } -int r600_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_pages, - struct radeon_fence *fence) -{ - /* FIXME: implement */ - return 0; -} - int r600_copy_blit(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, unsigned num_pages, struct radeon_fence *fence) { - r600_blit_prepare_copy(rdev, num_pages * RADEON_GPU_PAGE_SIZE); + int r; + + mutex_lock(&rdev->r600_blit.mutex); + rdev->r600_blit.vb_ib = NULL; + r = r600_blit_prepare_copy(rdev, num_pages * RADEON_GPU_PAGE_SIZE); + if (r) { + if (rdev->r600_blit.vb_ib) + radeon_ib_free(rdev, &rdev->r600_blit.vb_ib); + mutex_unlock(&rdev->r600_blit.mutex); + return r; + } r600_kms_blit_copy(rdev, src_offset, dst_offset, num_pages * RADEON_GPU_PAGE_SIZE); r600_blit_done_copy(rdev, fence); + mutex_unlock(&rdev->r600_blit.mutex); return 0; } @@ -1864,26 +1887,25 @@ int r600_startup(struct radeon_device *rdev) return r; } r600_gpu_init(rdev); - - if (!rdev->r600_blit.shader_obj) { - r = r600_blit_init(rdev); + r = r600_blit_init(rdev); + if (r) { + r600_blit_fini(rdev); + rdev->asic->copy = NULL; + dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); + } + /* pin copy shader into vram */ + if (rdev->r600_blit.shader_obj) { + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (unlikely(r != 0)) + return r; + r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->r600_blit.shader_gpu_addr); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); if (r) { - DRM_ERROR("radeon: failed blitter (%d).\n", r); + dev_err(rdev->dev, "(%d) pin blit object failed\n", r); return r; } } - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->r600_blit.shader_gpu_addr); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - if (r) { - dev_err(rdev->dev, "(%d) pin blit object failed\n", r); - return r; - } - /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -1948,6 +1970,13 @@ int r600_resume(struct radeon_device *rdev) DRM_ERROR("radeon: failled testing IB (%d).\n", r); return r; } + + r = r600_audio_init(rdev); + if (r) { + DRM_ERROR("radeon: audio resume failed\n"); + return r; + } + return r; } @@ -1955,17 +1984,21 @@ int r600_suspend(struct radeon_device *rdev) { int r; + r600_audio_fini(rdev); /* FIXME: we should wait for ring to be empty */ r600_cp_stop(rdev); rdev->cp.ready = false; + r600_irq_suspend(rdev); r600_wb_disable(rdev); r600_pcie_gart_disable(rdev); /* unpin shaders bo */ - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - radeon_bo_unpin(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); + if (rdev->r600_blit.shader_obj) { + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (!r) { + radeon_bo_unpin(rdev->r600_blit.shader_obj); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); + } + } return 0; } @@ -2026,6 +2059,11 @@ int r600_init(struct radeon_device *rdev) r = radeon_fence_driver_init(rdev); if (r) return r; + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) + radeon_agp_disable(rdev); + } r = r600_mc_init(rdev); if (r) return r; @@ -2051,22 +2089,25 @@ int r600_init(struct radeon_device *rdev) rdev->accel_working = true; r = r600_startup(rdev); if (r) { - r600_suspend(rdev); + dev_err(rdev->dev, "disabling GPU acceleration\n"); + r600_cp_fini(rdev); r600_wb_fini(rdev); - radeon_ring_fini(rdev); + r600_irq_fini(rdev); + radeon_irq_kms_fini(rdev); r600_pcie_gart_fini(rdev); rdev->accel_working = false; } if (rdev->accel_working) { r = radeon_ib_pool_init(rdev); if (r) { - DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r); - rdev->accel_working = false; - } - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); rdev->accel_working = false; + } else { + r = r600_ib_test(rdev); + if (r) { + dev_err(rdev->dev, "IB test failed (%d).\n", r); + rdev->accel_working = false; + } } } @@ -2078,20 +2119,17 @@ int r600_init(struct radeon_device *rdev) void r600_fini(struct radeon_device *rdev) { - /* Suspend operations */ - r600_suspend(rdev); - r600_audio_fini(rdev); r600_blit_fini(rdev); + r600_cp_fini(rdev); + r600_wb_fini(rdev); r600_irq_fini(rdev); radeon_irq_kms_fini(rdev); - radeon_ring_fini(rdev); - r600_wb_fini(rdev); r600_pcie_gart_fini(rdev); + radeon_agp_fini(rdev); radeon_gem_fini(rdev); radeon_fence_driver_fini(rdev); radeon_clocks_fini(rdev); - radeon_agp_fini(rdev); radeon_bo_fini(rdev); radeon_atombios_fini(rdev); kfree(rdev->bios); @@ -2197,14 +2235,14 @@ void r600_ih_ring_init(struct radeon_device *rdev, unsigned ring_size) rb_bufsz = drm_order(ring_size / 4); ring_size = (1 << rb_bufsz) * 4; rdev->ih.ring_size = ring_size; - rdev->ih.align_mask = 4 - 1; + rdev->ih.ptr_mask = rdev->ih.ring_size - 1; + rdev->ih.rptr = 0; } -static int r600_ih_ring_alloc(struct radeon_device *rdev, unsigned ring_size) +static int r600_ih_ring_alloc(struct radeon_device *rdev) { int r; - rdev->ih.ring_size = ring_size; /* Allocate ring buffer */ if (rdev->ih.ring_obj == NULL) { r = radeon_bo_create(rdev, NULL, rdev->ih.ring_size, @@ -2234,9 +2272,6 @@ static int r600_ih_ring_alloc(struct radeon_device *rdev, unsigned ring_size) return r; } } - rdev->ih.ptr_mask = (rdev->cp.ring_size / 4) - 1; - rdev->ih.rptr = 0; - return 0; } @@ -2386,7 +2421,7 @@ int r600_irq_init(struct radeon_device *rdev) u32 interrupt_cntl, ih_cntl, ih_rb_cntl; /* allocate ring */ - ret = r600_ih_ring_alloc(rdev, rdev->ih.ring_size); + ret = r600_ih_ring_alloc(rdev); if (ret) return ret; @@ -2449,10 +2484,15 @@ int r600_irq_init(struct radeon_device *rdev) return ret; } -void r600_irq_fini(struct radeon_device *rdev) +void r600_irq_suspend(struct radeon_device *rdev) { r600_disable_interrupts(rdev); r600_rlc_stop(rdev); +} + +void r600_irq_fini(struct radeon_device *rdev) +{ + r600_irq_suspend(rdev); r600_ih_ring_fini(rdev); } @@ -2467,8 +2507,12 @@ int r600_irq_set(struct radeon_device *rdev) return -EINVAL; } /* don't enable anything if the ih is disabled */ - if (!rdev->ih.enabled) + if (!rdev->ih.enabled) { + r600_disable_interrupts(rdev); + /* force the active interrupt state to all disabled */ + r600_disable_interrupt_state(rdev); return 0; + } if (ASIC_IS_DCE3(rdev)) { hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN; @@ -2638,16 +2682,18 @@ static inline u32 r600_get_ih_wptr(struct radeon_device *rdev) wptr = RREG32(IH_RB_WPTR); if (wptr & RB_OVERFLOW) { - WARN_ON(1); - /* XXX deal with overflow */ - DRM_ERROR("IH RB overflow\n"); + /* When a ring buffer overflow happen start parsing interrupt + * from the last not overwritten vector (wptr + 16). Hopefully + * this should allow us to catchup. + */ + dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n", + wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask); + rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask; tmp = RREG32(IH_RB_CNTL); tmp |= IH_WPTR_OVERFLOW_CLEAR; WREG32(IH_RB_CNTL, tmp); } - wptr = wptr & WPTR_OFFSET_MASK; - - return wptr; + return (wptr & rdev->ih.ptr_mask); } /* r600 IV Ring @@ -2683,12 +2729,13 @@ int r600_irq_process(struct radeon_device *rdev) u32 wptr = r600_get_ih_wptr(rdev); u32 rptr = rdev->ih.rptr; u32 src_id, src_data; - u32 last_entry = rdev->ih.ring_size - 16; u32 ring_index, disp_int, disp_int_cont, disp_int_cont2; unsigned long flags; bool queue_hotplug = false; DRM_DEBUG("r600_irq_process start: rptr %d, wptr %d\n", rptr, wptr); + if (!rdev->ih.enabled) + return IRQ_NONE; spin_lock_irqsave(&rdev->ih.lock, flags); @@ -2718,6 +2765,7 @@ restart_ih: case 0: /* D1 vblank */ if (disp_int & LB_D1_VBLANK_INTERRUPT) { drm_handle_vblank(rdev->ddev, 0); + wake_up(&rdev->irq.vblank_queue); disp_int &= ~LB_D1_VBLANK_INTERRUPT; DRM_DEBUG("IH: D1 vblank\n"); } @@ -2738,6 +2786,7 @@ restart_ih: case 0: /* D2 vblank */ if (disp_int & LB_D2_VBLANK_INTERRUPT) { drm_handle_vblank(rdev->ddev, 1); + wake_up(&rdev->irq.vblank_queue); disp_int &= ~LB_D2_VBLANK_INTERRUPT; DRM_DEBUG("IH: D2 vblank\n"); } @@ -2817,10 +2866,8 @@ restart_ih: } /* wptr/rptr are in bytes! */ - if (rptr == last_entry) - rptr = 0; - else - rptr += 16; + rptr += 16; + rptr &= rdev->ih.ptr_mask; } /* make sure wptr hasn't changed while processing */ wptr = r600_get_ih_wptr(rdev); @@ -2888,3 +2935,18 @@ int r600_debugfs_mc_info_init(struct radeon_device *rdev) return 0; #endif } + +/** + * r600_ioctl_wait_idle - flush host path cache on wait idle ioctl + * rdev: radeon device structure + * bo: buffer object struct which userspace is waiting for idle + * + * Some R6XX/R7XX doesn't seems to take into account HDP flush performed + * through ring buffer, this leads to corruption in rendering, see + * http://bugzilla.kernel.org/show_bug.cgi?id=15186 to avoid this we + * directly perform HDP flush by writing register through MMIO. + */ +void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo) +{ + WREG32(R_005480_HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); +} diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c index 99e2c3891a7..db928016d03 100644 --- a/drivers/gpu/drm/radeon/r600_audio.c +++ b/drivers/gpu/drm/radeon/r600_audio.c @@ -147,15 +147,23 @@ static void r600_audio_update_hdmi(unsigned long param) } /* + * turn on/off audio engine + */ +static void r600_audio_engine_enable(struct radeon_device *rdev, bool enable) +{ + DRM_INFO("%s audio support", enable ? "Enabling" : "Disabling"); + WREG32_P(R600_AUDIO_ENABLE, enable ? 0x81000000 : 0x0, ~0x81000000); +} + +/* * initialize the audio vars and register the update timer */ int r600_audio_init(struct radeon_device *rdev) { - if (!r600_audio_chipset_supported(rdev)) + if (!radeon_audio || !r600_audio_chipset_supported(rdev)) return 0; - DRM_INFO("%s audio support", radeon_audio ? "Enabling" : "Disabling"); - WREG32_P(R600_AUDIO_ENABLE, radeon_audio ? 0x81000000 : 0x0, ~0x81000000); + r600_audio_engine_enable(rdev, true); rdev->audio_channels = -1; rdev->audio_rate = -1; @@ -258,10 +266,10 @@ void r600_audio_set_clock(struct drm_encoder *encoder, int clock) */ void r600_audio_fini(struct radeon_device *rdev) { - if (!r600_audio_chipset_supported(rdev)) + if (!radeon_audio || !r600_audio_chipset_supported(rdev)) return; - WREG32_P(R600_AUDIO_ENABLE, 0x0, ~0x81000000); - del_timer(&rdev->audio_timer); + + r600_audio_engine_enable(rdev, false); } diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c index 5ea43234758..f4fb88ece2b 100644 --- a/drivers/gpu/drm/radeon/r600_blit.c +++ b/drivers/gpu/drm/radeon/r600_blit.c @@ -49,7 +49,7 @@ set_render_target(drm_radeon_private_t *dev_priv, int format, int w, int h, u64 RING_LOCALS; DRM_DEBUG("\n"); - h = (h + 7) & ~7; + h = ALIGN(h, 8); if (h < 8) h = 8; diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c index 8787ea89dc6..f6c6c77db7e 100644 --- a/drivers/gpu/drm/radeon/r600_blit_kms.c +++ b/drivers/gpu/drm/radeon/r600_blit_kms.c @@ -25,7 +25,7 @@ set_render_target(struct radeon_device *rdev, int format, u32 cb_color_info; int pitch, slice; - h = (h + 7) & ~7; + h = ALIGN(h, 8); if (h < 8) h = 8; @@ -396,15 +396,13 @@ set_default_state(struct radeon_device *rdev) NUM_ES_STACK_ENTRIES(num_es_stack_entries)); /* emit an IB pointing at default state */ - dwords = (rdev->r600_blit.state_len + 0xf) & ~0xf; + dwords = ALIGN(rdev->r600_blit.state_len, 0x10); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); radeon_ring_write(rdev, gpu_addr & 0xFFFFFFFC); radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF); radeon_ring_write(rdev, dwords); - radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE, 0)); - radeon_ring_write(rdev, CACHE_FLUSH_AND_INV_EVENT); /* SQ config */ radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 6)); radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); @@ -449,6 +447,7 @@ int r600_blit_init(struct radeon_device *rdev) u32 packet2s[16]; int num_packet2s = 0; + mutex_init(&rdev->r600_blit.mutex); rdev->r600_blit.state_offset = 0; if (rdev->family >= CHIP_RV770) @@ -512,14 +511,16 @@ void r600_blit_fini(struct radeon_device *rdev) { int r; + if (rdev->r600_blit.shader_obj == NULL) + return; + /* If we can't reserve the bo, unref should be enough to destroy + * it when it becomes idle. + */ r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) { - dev_err(rdev->dev, "(%d) can't finish r600 blit\n", r); - goto out_unref; + if (!r) { + radeon_bo_unpin(rdev->r600_blit.shader_obj); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); } - radeon_bo_unpin(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); -out_unref: radeon_bo_unref(&rdev->r600_blit.shader_obj); } @@ -540,9 +541,6 @@ int r600_vb_ib_get(struct radeon_device *rdev) void r600_vb_ib_put(struct radeon_device *rdev) { radeon_fence_emit(rdev, rdev->r600_blit.vb_ib->fence); - mutex_lock(&rdev->ib_pool.mutex); - list_add_tail(&rdev->r600_blit.vb_ib->list, &rdev->ib_pool.scheduled_ibs); - mutex_unlock(&rdev->ib_pool.mutex); radeon_ib_free(rdev, &rdev->r600_blit.vb_ib); } @@ -555,7 +553,8 @@ int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes) int dwords_per_loop = 76, num_loops; r = r600_vb_ib_get(rdev); - WARN_ON(r); + if (r) + return r; /* set_render_target emits 2 extra dwords on rv6xx */ if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) @@ -577,11 +576,12 @@ int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes) ring_size = num_loops * dwords_per_loop; /* set default + shaders */ ring_size += 40; /* shaders + def state */ - ring_size += 7; /* fence emit for VB IB */ + ring_size += 10; /* fence emit for VB IB */ ring_size += 5; /* done copy */ - ring_size += 7; /* fence emit for done copy */ + ring_size += 10; /* fence emit for done copy */ r = radeon_ring_lock(rdev, ring_size); - WARN_ON(r); + if (r) + return r; set_default_state(rdev); /* 14 */ set_shaders(rdev); /* 26 */ @@ -592,13 +592,6 @@ void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence) { int r; - radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE, 0)); - radeon_ring_write(rdev, CACHE_FLUSH_AND_INV_EVENT); - /* wait for 3D idle clean */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(rdev, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit); - if (rdev->r600_blit.vb_ib) r600_vb_ib_put(rdev); diff --git a/drivers/gpu/drm/radeon/r600_blit_shaders.c b/drivers/gpu/drm/radeon/r600_blit_shaders.c index d745e815c2e..a112c59f9d8 100644 --- a/drivers/gpu/drm/radeon/r600_blit_shaders.c +++ b/drivers/gpu/drm/radeon/r600_blit_shaders.c @@ -9,11 +9,6 @@ const u32 r6xx_default_state[] = 0xc0012800, 0x80000000, 0x80000000, - 0xc0004600, - 0x00000016, - 0xc0016800, - 0x00000010, - 0x00028000, 0xc0016800, 0x00000010, 0x00008000, @@ -531,11 +526,6 @@ const u32 r7xx_default_state[] = 0xc0012800, 0x80000000, 0x80000000, - 0xc0004600, - 0x00000016, - 0xc0016800, - 0x00000010, - 0x00028000, 0xc0016800, 0x00000010, 0x00008000, diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c index 6d5a711c2e9..40416c068d9 100644 --- a/drivers/gpu/drm/radeon/r600_cp.c +++ b/drivers/gpu/drm/radeon/r600_cp.c @@ -734,8 +734,8 @@ static void r600_gfx_init(struct drm_device *dev, u32 hdp_host_path_cntl; u32 backend_map; u32 gb_tiling_config = 0; - u32 cc_rb_backend_disable = 0; - u32 cc_gc_shader_pipe_config = 0; + u32 cc_rb_backend_disable; + u32 cc_gc_shader_pipe_config; u32 ramcfg; /* setup chip specs */ @@ -857,29 +857,44 @@ static void r600_gfx_init(struct drm_device *dev, gb_tiling_config |= R600_BANK_SWAPS(1); - backend_map = r600_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes, - dev_priv->r600_max_backends, - (0xff << dev_priv->r600_max_backends) & 0xff); - gb_tiling_config |= R600_BACKEND_MAP(backend_map); + cc_rb_backend_disable = RADEON_READ(R600_CC_RB_BACKEND_DISABLE) & 0x00ff0000; + cc_rb_backend_disable |= + R600_BACKEND_DISABLE((R6XX_MAX_BACKENDS_MASK << dev_priv->r600_max_backends) & R6XX_MAX_BACKENDS_MASK); - cc_gc_shader_pipe_config = + cc_gc_shader_pipe_config = RADEON_READ(R600_CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; + cc_gc_shader_pipe_config |= R600_INACTIVE_QD_PIPES((R6XX_MAX_PIPES_MASK << dev_priv->r600_max_pipes) & R6XX_MAX_PIPES_MASK); cc_gc_shader_pipe_config |= R600_INACTIVE_SIMDS((R6XX_MAX_SIMDS_MASK << dev_priv->r600_max_simds) & R6XX_MAX_SIMDS_MASK); - cc_rb_backend_disable = - R600_BACKEND_DISABLE((R6XX_MAX_BACKENDS_MASK << dev_priv->r600_max_backends) & R6XX_MAX_BACKENDS_MASK); + backend_map = r600_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes, + (R6XX_MAX_BACKENDS - + r600_count_pipe_bits((cc_rb_backend_disable & + R6XX_MAX_BACKENDS_MASK) >> 16)), + (cc_rb_backend_disable >> 16)); + gb_tiling_config |= R600_BACKEND_MAP(backend_map); RADEON_WRITE(R600_GB_TILING_CONFIG, gb_tiling_config); RADEON_WRITE(R600_DCP_TILING_CONFIG, (gb_tiling_config & 0xffff)); RADEON_WRITE(R600_HDP_TILING_CONFIG, (gb_tiling_config & 0xffff)); + if (gb_tiling_config & 0xc0) { + dev_priv->r600_group_size = 512; + } else { + dev_priv->r600_group_size = 256; + } + dev_priv->r600_npipes = 1 << ((gb_tiling_config >> 1) & 0x7); + if (gb_tiling_config & 0x30) { + dev_priv->r600_nbanks = 8; + } else { + dev_priv->r600_nbanks = 4; + } RADEON_WRITE(R600_CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); RADEON_WRITE(R600_CC_GC_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); RADEON_WRITE(R600_GC_USER_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); num_qd_pipes = - R6XX_MAX_BACKENDS - r600_count_pipe_bits(cc_gc_shader_pipe_config & R600_INACTIVE_QD_PIPES_MASK); + R6XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config & R600_INACTIVE_QD_PIPES_MASK) >> 8); RADEON_WRITE(R600_VGT_OUT_DEALLOC_CNTL, (num_qd_pipes * 4) & R600_DEALLOC_DIST_MASK); RADEON_WRITE(R600_VGT_VERTEX_REUSE_BLOCK_CNTL, ((num_qd_pipes * 4) - 2) & R600_VTX_REUSE_DEPTH_MASK); @@ -1151,7 +1166,8 @@ static void r600_gfx_init(struct drm_device *dev, } -static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, +static u32 r700_get_tile_pipe_to_backend_map(drm_radeon_private_t *dev_priv, + u32 num_tile_pipes, u32 num_backends, u32 backend_disable_mask) { @@ -1162,6 +1178,7 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, u32 swizzle_pipe[R7XX_MAX_PIPES]; u32 cur_backend; u32 i; + bool force_no_swizzle; if (num_tile_pipes > R7XX_MAX_PIPES) num_tile_pipes = R7XX_MAX_PIPES; @@ -1191,6 +1208,18 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, if (enabled_backends_count != num_backends) num_backends = enabled_backends_count; + switch (dev_priv->flags & RADEON_FAMILY_MASK) { + case CHIP_RV770: + case CHIP_RV730: + force_no_swizzle = false; + break; + case CHIP_RV710: + case CHIP_RV740: + default: + force_no_swizzle = true; + break; + } + memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * R7XX_MAX_PIPES); switch (num_tile_pipes) { case 1: @@ -1201,49 +1230,100 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, swizzle_pipe[1] = 1; break; case 3: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 1; + } break; case 4: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 3; - swizzle_pipe[3] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 3; + swizzle_pipe[3] = 1; + } break; case 5: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 1; - swizzle_pipe[4] = 3; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 1; + swizzle_pipe[4] = 3; + } break; case 6: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 5; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 5; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + } break; case 7: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 5; + } break; case 8: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 7; - swizzle_pipe[7] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + swizzle_pipe[7] = 7; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 7; + swizzle_pipe[7] = 5; + } break; } @@ -1264,8 +1344,10 @@ static void r700_gfx_init(struct drm_device *dev, drm_radeon_private_t *dev_priv) { int i, j, num_qd_pipes; + u32 ta_aux_cntl; u32 sx_debug_1; u32 smx_dc_ctl0; + u32 db_debug3; u32 num_gs_verts_per_thread; u32 vgt_gs_per_es; u32 gs_prim_buffer_depth = 0; @@ -1276,8 +1358,8 @@ static void r700_gfx_init(struct drm_device *dev, u32 sq_dyn_gpr_size_simd_ab_0; u32 backend_map; u32 gb_tiling_config = 0; - u32 cc_rb_backend_disable = 0; - u32 cc_gc_shader_pipe_config = 0; + u32 cc_rb_backend_disable; + u32 cc_gc_shader_pipe_config; u32 mc_arb_ramcfg; u32 db_debug4; @@ -1428,35 +1510,51 @@ static void r700_gfx_init(struct drm_device *dev, gb_tiling_config |= R600_BANK_SWAPS(1); - backend_map = r700_get_tile_pipe_to_backend_map(dev_priv->r600_max_tile_pipes, - dev_priv->r600_max_backends, - (0xff << dev_priv->r600_max_backends) & 0xff); - gb_tiling_config |= R600_BACKEND_MAP(backend_map); + cc_rb_backend_disable = RADEON_READ(R600_CC_RB_BACKEND_DISABLE) & 0x00ff0000; + cc_rb_backend_disable |= + R600_BACKEND_DISABLE((R7XX_MAX_BACKENDS_MASK << dev_priv->r600_max_backends) & R7XX_MAX_BACKENDS_MASK); - cc_gc_shader_pipe_config = + cc_gc_shader_pipe_config = RADEON_READ(R600_CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; + cc_gc_shader_pipe_config |= R600_INACTIVE_QD_PIPES((R7XX_MAX_PIPES_MASK << dev_priv->r600_max_pipes) & R7XX_MAX_PIPES_MASK); cc_gc_shader_pipe_config |= R600_INACTIVE_SIMDS((R7XX_MAX_SIMDS_MASK << dev_priv->r600_max_simds) & R7XX_MAX_SIMDS_MASK); - cc_rb_backend_disable = - R600_BACKEND_DISABLE((R7XX_MAX_BACKENDS_MASK << dev_priv->r600_max_backends) & R7XX_MAX_BACKENDS_MASK); + if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV740) + backend_map = 0x28; + else + backend_map = r700_get_tile_pipe_to_backend_map(dev_priv, + dev_priv->r600_max_tile_pipes, + (R7XX_MAX_BACKENDS - + r600_count_pipe_bits((cc_rb_backend_disable & + R7XX_MAX_BACKENDS_MASK) >> 16)), + (cc_rb_backend_disable >> 16)); + gb_tiling_config |= R600_BACKEND_MAP(backend_map); RADEON_WRITE(R600_GB_TILING_CONFIG, gb_tiling_config); RADEON_WRITE(R600_DCP_TILING_CONFIG, (gb_tiling_config & 0xffff)); RADEON_WRITE(R600_HDP_TILING_CONFIG, (gb_tiling_config & 0xffff)); + if (gb_tiling_config & 0xc0) { + dev_priv->r600_group_size = 512; + } else { + dev_priv->r600_group_size = 256; + } + dev_priv->r600_npipes = 1 << ((gb_tiling_config >> 1) & 0x7); + if (gb_tiling_config & 0x30) { + dev_priv->r600_nbanks = 8; + } else { + dev_priv->r600_nbanks = 4; + } RADEON_WRITE(R600_CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); RADEON_WRITE(R600_CC_GC_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); - RADEON_WRITE(R600_GC_USER_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); RADEON_WRITE(R700_CC_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable); RADEON_WRITE(R700_CGTS_SYS_TCC_DISABLE, 0); RADEON_WRITE(R700_CGTS_TCC_DISABLE, 0); - RADEON_WRITE(R700_CGTS_USER_SYS_TCC_DISABLE, 0); - RADEON_WRITE(R700_CGTS_USER_TCC_DISABLE, 0); num_qd_pipes = - R7XX_MAX_BACKENDS - r600_count_pipe_bits(cc_gc_shader_pipe_config & R600_INACTIVE_QD_PIPES_MASK); + R7XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config & R600_INACTIVE_QD_PIPES_MASK) >> 8); RADEON_WRITE(R600_VGT_OUT_DEALLOC_CNTL, (num_qd_pipes * 4) & R600_DEALLOC_DIST_MASK); RADEON_WRITE(R600_VGT_VERTEX_REUSE_BLOCK_CNTL, ((num_qd_pipes * 4) - 2) & R600_VTX_REUSE_DEPTH_MASK); @@ -1466,10 +1564,8 @@ static void r700_gfx_init(struct drm_device *dev, RADEON_WRITE(R600_CP_MEQ_THRESHOLDS, R700_STQ_SPLIT(0x30)); - RADEON_WRITE(R600_TA_CNTL_AUX, (R600_DISABLE_CUBE_ANISO | - R600_SYNC_GRADIENT | - R600_SYNC_WALKER | - R600_SYNC_ALIGNER)); + ta_aux_cntl = RADEON_READ(R600_TA_CNTL_AUX); + RADEON_WRITE(R600_TA_CNTL_AUX, ta_aux_cntl | R600_DISABLE_CUBE_ANISO); sx_debug_1 = RADEON_READ(R700_SX_DEBUG_1); sx_debug_1 |= R700_ENABLE_NEW_SMX_ADDRESS; @@ -1480,14 +1576,28 @@ static void r700_gfx_init(struct drm_device *dev, smx_dc_ctl0 |= R700_CACHE_DEPTH((dev_priv->r700_sx_num_of_sets * 64) - 1); RADEON_WRITE(R600_SMX_DC_CTL0, smx_dc_ctl0); - RADEON_WRITE(R700_SMX_EVENT_CTL, (R700_ES_FLUSH_CTL(4) | - R700_GS_FLUSH_CTL(4) | - R700_ACK_FLUSH_CTL(3) | - R700_SYNC_FLUSH_CTL)); + if ((dev_priv->flags & RADEON_FAMILY_MASK) != CHIP_RV740) + RADEON_WRITE(R700_SMX_EVENT_CTL, (R700_ES_FLUSH_CTL(4) | + R700_GS_FLUSH_CTL(4) | + R700_ACK_FLUSH_CTL(3) | + R700_SYNC_FLUSH_CTL)); - if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV770) - RADEON_WRITE(R700_DB_DEBUG3, R700_DB_CLK_OFF_DELAY(0x1f)); - else { + db_debug3 = RADEON_READ(R700_DB_DEBUG3); + db_debug3 &= ~R700_DB_CLK_OFF_DELAY(0x1f); + switch (dev_priv->flags & RADEON_FAMILY_MASK) { + case CHIP_RV770: + case CHIP_RV740: + db_debug3 |= R700_DB_CLK_OFF_DELAY(0x1f); + break; + case CHIP_RV710: + case CHIP_RV730: + default: + db_debug3 |= R700_DB_CLK_OFF_DELAY(2); + break; + } + RADEON_WRITE(R700_DB_DEBUG3, db_debug3); + + if ((dev_priv->flags & RADEON_FAMILY_MASK) != CHIP_RV770) { db_debug4 = RADEON_READ(RV700_DB_DEBUG4); db_debug4 |= RV700_DISABLE_TILE_COVERED_FOR_PS_ITER; RADEON_WRITE(RV700_DB_DEBUG4, db_debug4); @@ -1516,10 +1626,10 @@ static void r700_gfx_init(struct drm_device *dev, R600_ALU_UPDATE_FIFO_HIWATER(0x8)); switch (dev_priv->flags & RADEON_FAMILY_MASK) { case CHIP_RV770: - sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x1); - break; case CHIP_RV730: case CHIP_RV710: + sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x1); + break; case CHIP_RV740: default: sq_ms_fifo_sizes |= R600_FETCH_FIFO_HIWATER(0x4); @@ -2526,3 +2636,12 @@ out: mutex_unlock(&dev_priv->cs_mutex); return r; } + +void r600_cs_legacy_get_tiling_conf(struct drm_device *dev, u32 *npipes, u32 *nbanks, u32 *group_size) +{ + struct drm_radeon_private *dev_priv = dev->dev_private; + + *npipes = dev_priv->r600_npipes; + *nbanks = dev_priv->r600_nbanks; + *group_size = dev_priv->r600_group_size; +} diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c index 44060b92d9e..cd2c63bce50 100644 --- a/drivers/gpu/drm/radeon/r600_cs.c +++ b/drivers/gpu/drm/radeon/r600_cs.c @@ -28,6 +28,7 @@ #include "drmP.h" #include "radeon.h" #include "r600d.h" +#include "r600_reg_safe.h" static int r600_cs_packet_next_reloc_mm(struct radeon_cs_parser *p, struct radeon_cs_reloc **cs_reloc); @@ -35,6 +36,312 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p, struct radeon_cs_reloc **cs_reloc); typedef int (*next_reloc_t)(struct radeon_cs_parser*, struct radeon_cs_reloc**); static next_reloc_t r600_cs_packet_next_reloc = &r600_cs_packet_next_reloc_mm; +extern void r600_cs_legacy_get_tiling_conf(struct drm_device *dev, u32 *npipes, u32 *nbanks, u32 *group_size); + + +struct r600_cs_track { + /* configuration we miror so that we use same code btw kms/ums */ + u32 group_size; + u32 nbanks; + u32 npipes; + /* value we track */ + u32 nsamples; + u32 cb_color_base_last[8]; + struct radeon_bo *cb_color_bo[8]; + u32 cb_color_bo_offset[8]; + struct radeon_bo *cb_color_frag_bo[8]; + struct radeon_bo *cb_color_tile_bo[8]; + u32 cb_color_info[8]; + u32 cb_color_size_idx[8]; + u32 cb_target_mask; + u32 cb_shader_mask; + u32 cb_color_size[8]; + u32 vgt_strmout_en; + u32 vgt_strmout_buffer_en; + u32 db_depth_control; + u32 db_depth_info; + u32 db_depth_size_idx; + u32 db_depth_view; + u32 db_depth_size; + u32 db_offset; + struct radeon_bo *db_bo; +}; + +static inline int r600_bpe_from_format(u32 *bpe, u32 format) +{ + switch (format) { + case V_038004_COLOR_8: + case V_038004_COLOR_4_4: + case V_038004_COLOR_3_3_2: + case V_038004_FMT_1: + *bpe = 1; + break; + case V_038004_COLOR_16: + case V_038004_COLOR_16_FLOAT: + case V_038004_COLOR_8_8: + case V_038004_COLOR_5_6_5: + case V_038004_COLOR_6_5_5: + case V_038004_COLOR_1_5_5_5: + case V_038004_COLOR_4_4_4_4: + case V_038004_COLOR_5_5_5_1: + *bpe = 2; + break; + case V_038004_FMT_8_8_8: + *bpe = 3; + break; + case V_038004_COLOR_32: + case V_038004_COLOR_32_FLOAT: + case V_038004_COLOR_16_16: + case V_038004_COLOR_16_16_FLOAT: + case V_038004_COLOR_8_24: + case V_038004_COLOR_8_24_FLOAT: + case V_038004_COLOR_24_8: + case V_038004_COLOR_24_8_FLOAT: + case V_038004_COLOR_10_11_11: + case V_038004_COLOR_10_11_11_FLOAT: + case V_038004_COLOR_11_11_10: + case V_038004_COLOR_11_11_10_FLOAT: + case V_038004_COLOR_2_10_10_10: + case V_038004_COLOR_8_8_8_8: + case V_038004_COLOR_10_10_10_2: + case V_038004_FMT_5_9_9_9_SHAREDEXP: + case V_038004_FMT_32_AS_8: + case V_038004_FMT_32_AS_8_8: + *bpe = 4; + break; + case V_038004_COLOR_X24_8_32_FLOAT: + case V_038004_COLOR_32_32: + case V_038004_COLOR_32_32_FLOAT: + case V_038004_COLOR_16_16_16_16: + case V_038004_COLOR_16_16_16_16_FLOAT: + *bpe = 8; + break; + case V_038004_FMT_16_16_16: + case V_038004_FMT_16_16_16_FLOAT: + *bpe = 6; + break; + case V_038004_FMT_32_32_32: + case V_038004_FMT_32_32_32_FLOAT: + *bpe = 12; + break; + case V_038004_COLOR_32_32_32_32: + case V_038004_COLOR_32_32_32_32_FLOAT: + *bpe = 16; + break; + case V_038004_FMT_GB_GR: + case V_038004_FMT_BG_RG: + case V_038004_COLOR_INVALID: + *bpe = 16; + return -EINVAL; + } + return 0; +} + +static void r600_cs_track_init(struct r600_cs_track *track) +{ + int i; + + for (i = 0; i < 8; i++) { + track->cb_color_base_last[i] = 0; + track->cb_color_size[i] = 0; + track->cb_color_size_idx[i] = 0; + track->cb_color_info[i] = 0; + track->cb_color_bo[i] = NULL; + track->cb_color_bo_offset[i] = 0xFFFFFFFF; + } + track->cb_target_mask = 0xFFFFFFFF; + track->cb_shader_mask = 0xFFFFFFFF; + track->db_bo = NULL; + /* assume the biggest format and that htile is enabled */ + track->db_depth_info = 7 | (1 << 25); + track->db_depth_view = 0xFFFFC000; + track->db_depth_size = 0xFFFFFFFF; + track->db_depth_size_idx = 0; + track->db_depth_control = 0xFFFFFFFF; +} + +static inline int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) +{ + struct r600_cs_track *track = p->track; + u32 bpe = 0, pitch, slice_tile_max, size, tmp, height; + volatile u32 *ib = p->ib->ptr; + + if (G_0280A0_TILE_MODE(track->cb_color_info[i])) { + dev_warn(p->dev, "FMASK or CMASK buffer are not supported by this kernel\n"); + return -EINVAL; + } + size = radeon_bo_size(track->cb_color_bo[i]); + if (r600_bpe_from_format(&bpe, G_0280A0_FORMAT(track->cb_color_info[i]))) { + dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08X)\n", + __func__, __LINE__, G_0280A0_FORMAT(track->cb_color_info[i]), + i, track->cb_color_info[i]); + return -EINVAL; + } + pitch = (G_028060_PITCH_TILE_MAX(track->cb_color_size[i]) + 1) << 3; + slice_tile_max = G_028060_SLICE_TILE_MAX(track->cb_color_size[i]) + 1; + if (!pitch) { + dev_warn(p->dev, "%s:%d cb pitch (%d) for %d invalid (0x%08X)\n", + __func__, __LINE__, pitch, i, track->cb_color_size[i]); + return -EINVAL; + } + height = size / (pitch * bpe); + if (height > 8192) + height = 8192; + switch (G_0280A0_ARRAY_MODE(track->cb_color_info[i])) { + case V_0280A0_ARRAY_LINEAR_GENERAL: + case V_0280A0_ARRAY_LINEAR_ALIGNED: + if (pitch & 0x3f) { + dev_warn(p->dev, "%s:%d cb pitch (%d x %d = %d) invalid\n", + __func__, __LINE__, pitch, bpe, pitch * bpe); + return -EINVAL; + } + if ((pitch * bpe) & (track->group_size - 1)) { + dev_warn(p->dev, "%s:%d cb pitch (%d) invalid\n", + __func__, __LINE__, pitch); + return -EINVAL; + } + break; + case V_0280A0_ARRAY_1D_TILED_THIN1: + if ((pitch * 8 * bpe * track->nsamples) & (track->group_size - 1)) { + dev_warn(p->dev, "%s:%d cb pitch (%d) invalid\n", + __func__, __LINE__, pitch); + return -EINVAL; + } + height &= ~0x7; + if (!height) + height = 8; + break; + case V_0280A0_ARRAY_2D_TILED_THIN1: + if (pitch & ((8 * track->nbanks) - 1)) { + dev_warn(p->dev, "%s:%d cb pitch (%d) invalid\n", + __func__, __LINE__, pitch); + return -EINVAL; + } + tmp = pitch * 8 * bpe * track->nsamples; + tmp = tmp / track->nbanks; + if (tmp & (track->group_size - 1)) { + dev_warn(p->dev, "%s:%d cb pitch (%d) invalid\n", + __func__, __LINE__, pitch); + return -EINVAL; + } + height &= ~((16 * track->npipes) - 1); + if (!height) + height = 16 * track->npipes; + break; + default: + dev_warn(p->dev, "%s invalid tiling %d for %d (0x%08X)\n", __func__, + G_0280A0_ARRAY_MODE(track->cb_color_info[i]), i, + track->cb_color_info[i]); + return -EINVAL; + } + /* check offset */ + tmp = height * pitch; + if ((tmp + track->cb_color_bo_offset[i]) > radeon_bo_size(track->cb_color_bo[i])) { + dev_warn(p->dev, "%s offset[%d] %d to big\n", __func__, i, track->cb_color_bo_offset[i]); + return -EINVAL; + } + /* limit max tile */ + tmp = (height * pitch) >> 6; + if (tmp < slice_tile_max) + slice_tile_max = tmp; + tmp = S_028060_PITCH_TILE_MAX((pitch >> 3) - 1) | + S_028060_SLICE_TILE_MAX(slice_tile_max - 1); + ib[track->cb_color_size_idx[i]] = tmp; + return 0; +} + +static int r600_cs_track_check(struct radeon_cs_parser *p) +{ + struct r600_cs_track *track = p->track; + u32 tmp; + int r, i; + volatile u32 *ib = p->ib->ptr; + + /* on legacy kernel we don't perform advanced check */ + if (p->rdev == NULL) + return 0; + /* we don't support out buffer yet */ + if (track->vgt_strmout_en || track->vgt_strmout_buffer_en) { + dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n"); + return -EINVAL; + } + /* check that we have a cb for each enabled target, we don't check + * shader_mask because it seems mesa isn't always setting it :( + */ + tmp = track->cb_target_mask; + for (i = 0; i < 8; i++) { + if ((tmp >> (i * 4)) & 0xF) { + /* at least one component is enabled */ + if (track->cb_color_bo[i] == NULL) { + dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", + __func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i); + return -EINVAL; + } + /* perform rewrite of CB_COLOR[0-7]_SIZE */ + r = r600_cs_track_validate_cb(p, i); + if (r) + return r; + } + } + /* Check depth buffer */ + if (G_028800_STENCIL_ENABLE(track->db_depth_control) || + G_028800_Z_ENABLE(track->db_depth_control)) { + u32 nviews, bpe, ntiles; + if (track->db_bo == NULL) { + dev_warn(p->dev, "z/stencil with no depth buffer\n"); + return -EINVAL; + } + if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) { + dev_warn(p->dev, "this kernel doesn't support z/stencil htile\n"); + return -EINVAL; + } + switch (G_028010_FORMAT(track->db_depth_info)) { + case V_028010_DEPTH_16: + bpe = 2; + break; + case V_028010_DEPTH_X8_24: + case V_028010_DEPTH_8_24: + case V_028010_DEPTH_X8_24_FLOAT: + case V_028010_DEPTH_8_24_FLOAT: + case V_028010_DEPTH_32_FLOAT: + bpe = 4; + break; + case V_028010_DEPTH_X24_8_32_FLOAT: + bpe = 8; + break; + default: + dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info)); + return -EINVAL; + } + if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) { + if (!track->db_depth_size_idx) { + dev_warn(p->dev, "z/stencil buffer size not set\n"); + return -EINVAL; + } + printk_once(KERN_WARNING "You have old & broken userspace please consider updating mesa\n"); + tmp = radeon_bo_size(track->db_bo) - track->db_offset; + tmp = (tmp / bpe) >> 6; + if (!tmp) { + dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n", + track->db_depth_size, bpe, track->db_offset, + radeon_bo_size(track->db_bo)); + return -EINVAL; + } + ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF); + } else { + ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1; + nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1; + tmp = ntiles * bpe * 64 * nviews; + if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) { + dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %d -> %d have %ld)\n", + track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset, + radeon_bo_size(track->db_bo)); + return -EINVAL; + } + } + } + return 0; +} /** * r600_cs_packet_parse() - parse cp packet and point ib index to next packet @@ -177,6 +484,28 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p, } /** + * r600_cs_packet_next_is_pkt3_nop() - test if next packet is packet3 nop for reloc + * @parser: parser structure holding parsing context. + * + * Check next packet is relocation packet3, do bo validation and compute + * GPU offset using the provided start. + **/ +static inline int r600_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p) +{ + struct radeon_cs_packet p3reloc; + int r; + + r = r600_cs_packet_parse(p, &p3reloc, p->idx); + if (r) { + return 0; + } + if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) { + return 0; + } + return 1; +} + +/** * r600_cs_packet_next_vline() - parse userspace VLINE packet * @parser: parser structure holding parsing context. * @@ -333,10 +662,339 @@ static int r600_cs_parse_packet0(struct radeon_cs_parser *p, return 0; } +/** + * r600_cs_check_reg() - check if register is authorized or not + * @parser: parser structure holding parsing context + * @reg: register we are testing + * @idx: index into the cs buffer + * + * This function will test against r600_reg_safe_bm and return 0 + * if register is safe. If register is not flag as safe this function + * will test it against a list of register needind special handling. + */ +static inline int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) +{ + struct r600_cs_track *track = (struct r600_cs_track *)p->track; + struct radeon_cs_reloc *reloc; + u32 last_reg = ARRAY_SIZE(r600_reg_safe_bm); + u32 m, i, tmp, *ib; + int r; + + i = (reg >> 7); + if (i > last_reg) { + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return -EINVAL; + } + m = 1 << ((reg >> 2) & 31); + if (!(r600_reg_safe_bm[i] & m)) + return 0; + ib = p->ib->ptr; + switch (reg) { + /* force following reg to 0 in an attemp to disable out buffer + * which will need us to better understand how it works to perform + * security check on it (Jerome) + */ + case R_0288A8_SQ_ESGS_RING_ITEMSIZE: + case R_008C44_SQ_ESGS_RING_SIZE: + case R_0288B0_SQ_ESTMP_RING_ITEMSIZE: + case R_008C54_SQ_ESTMP_RING_SIZE: + case R_0288C0_SQ_FBUF_RING_ITEMSIZE: + case R_008C74_SQ_FBUF_RING_SIZE: + case R_0288B4_SQ_GSTMP_RING_ITEMSIZE: + case R_008C5C_SQ_GSTMP_RING_SIZE: + case R_0288AC_SQ_GSVS_RING_ITEMSIZE: + case R_008C4C_SQ_GSVS_RING_SIZE: + case R_0288BC_SQ_PSTMP_RING_ITEMSIZE: + case R_008C6C_SQ_PSTMP_RING_SIZE: + case R_0288C4_SQ_REDUC_RING_ITEMSIZE: + case R_008C7C_SQ_REDUC_RING_SIZE: + case R_0288B8_SQ_VSTMP_RING_ITEMSIZE: + case R_008C64_SQ_VSTMP_RING_SIZE: + case R_0288C8_SQ_GS_VERT_ITEMSIZE: + /* get value to populate the IB don't remove */ + tmp =radeon_get_ib_value(p, idx); + ib[idx] = 0; + break; + case R_028800_DB_DEPTH_CONTROL: + track->db_depth_control = radeon_get_ib_value(p, idx); + break; + case R_028010_DB_DEPTH_INFO: + track->db_depth_info = radeon_get_ib_value(p, idx); + break; + case R_028004_DB_DEPTH_VIEW: + track->db_depth_view = radeon_get_ib_value(p, idx); + break; + case R_028000_DB_DEPTH_SIZE: + track->db_depth_size = radeon_get_ib_value(p, idx); + track->db_depth_size_idx = idx; + break; + case R_028AB0_VGT_STRMOUT_EN: + track->vgt_strmout_en = radeon_get_ib_value(p, idx); + break; + case R_028B20_VGT_STRMOUT_BUFFER_EN: + track->vgt_strmout_buffer_en = radeon_get_ib_value(p, idx); + break; + case R_028238_CB_TARGET_MASK: + track->cb_target_mask = radeon_get_ib_value(p, idx); + break; + case R_02823C_CB_SHADER_MASK: + track->cb_shader_mask = radeon_get_ib_value(p, idx); + break; + case R_028C04_PA_SC_AA_CONFIG: + tmp = G_028C04_MSAA_NUM_SAMPLES(radeon_get_ib_value(p, idx)); + track->nsamples = 1 << tmp; + break; + case R_0280A0_CB_COLOR0_INFO: + case R_0280A4_CB_COLOR1_INFO: + case R_0280A8_CB_COLOR2_INFO: + case R_0280AC_CB_COLOR3_INFO: + case R_0280B0_CB_COLOR4_INFO: + case R_0280B4_CB_COLOR5_INFO: + case R_0280B8_CB_COLOR6_INFO: + case R_0280BC_CB_COLOR7_INFO: + tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4; + track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); + break; + case R_028060_CB_COLOR0_SIZE: + case R_028064_CB_COLOR1_SIZE: + case R_028068_CB_COLOR2_SIZE: + case R_02806C_CB_COLOR3_SIZE: + case R_028070_CB_COLOR4_SIZE: + case R_028074_CB_COLOR5_SIZE: + case R_028078_CB_COLOR6_SIZE: + case R_02807C_CB_COLOR7_SIZE: + tmp = (reg - R_028060_CB_COLOR0_SIZE) / 4; + track->cb_color_size[tmp] = radeon_get_ib_value(p, idx); + track->cb_color_size_idx[tmp] = idx; + break; + /* This register were added late, there is userspace + * which does provide relocation for those but set + * 0 offset. In order to avoid breaking old userspace + * we detect this and set address to point to last + * CB_COLOR0_BASE, note that if userspace doesn't set + * CB_COLOR0_BASE before this register we will report + * error. Old userspace always set CB_COLOR0_BASE + * before any of this. + */ + case R_0280E0_CB_COLOR0_FRAG: + case R_0280E4_CB_COLOR1_FRAG: + case R_0280E8_CB_COLOR2_FRAG: + case R_0280EC_CB_COLOR3_FRAG: + case R_0280F0_CB_COLOR4_FRAG: + case R_0280F4_CB_COLOR5_FRAG: + case R_0280F8_CB_COLOR6_FRAG: + case R_0280FC_CB_COLOR7_FRAG: + tmp = (reg - R_0280E0_CB_COLOR0_FRAG) / 4; + if (!r600_cs_packet_next_is_pkt3_nop(p)) { + if (!track->cb_color_base_last[tmp]) { + dev_err(p->dev, "Broken old userspace ? no cb_color0_base supplied before trying to write 0x%08X\n", reg); + return -EINVAL; + } + ib[idx] = track->cb_color_base_last[tmp]; + printk_once(KERN_WARNING "You have old & broken userspace " + "please consider updating mesa & xf86-video-ati\n"); + track->cb_color_frag_bo[tmp] = track->cb_color_bo[tmp]; + } else { + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_frag_bo[tmp] = reloc->robj; + } + break; + case R_0280C0_CB_COLOR0_TILE: + case R_0280C4_CB_COLOR1_TILE: + case R_0280C8_CB_COLOR2_TILE: + case R_0280CC_CB_COLOR3_TILE: + case R_0280D0_CB_COLOR4_TILE: + case R_0280D4_CB_COLOR5_TILE: + case R_0280D8_CB_COLOR6_TILE: + case R_0280DC_CB_COLOR7_TILE: + tmp = (reg - R_0280C0_CB_COLOR0_TILE) / 4; + if (!r600_cs_packet_next_is_pkt3_nop(p)) { + if (!track->cb_color_base_last[tmp]) { + dev_err(p->dev, "Broken old userspace ? no cb_color0_base supplied before trying to write 0x%08X\n", reg); + return -EINVAL; + } + ib[idx] = track->cb_color_base_last[tmp]; + printk_once(KERN_WARNING "You have old & broken userspace " + "please consider updating mesa & xf86-video-ati\n"); + track->cb_color_tile_bo[tmp] = track->cb_color_bo[tmp]; + } else { + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_tile_bo[tmp] = reloc->robj; + } + break; + case CB_COLOR0_BASE: + case CB_COLOR1_BASE: + case CB_COLOR2_BASE: + case CB_COLOR3_BASE: + case CB_COLOR4_BASE: + case CB_COLOR5_BASE: + case CB_COLOR6_BASE: + case CB_COLOR7_BASE: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = (reg - CB_COLOR0_BASE) / 4; + track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->cb_color_base_last[tmp] = ib[idx]; + track->cb_color_bo[tmp] = reloc->robj; + break; + case DB_DEPTH_BASE: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->db_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->db_bo = reloc->robj; + break; + case DB_HTILE_DATA_BASE: + case SQ_PGM_START_FS: + case SQ_PGM_START_ES: + case SQ_PGM_START_VS: + case SQ_PGM_START_GS: + case SQ_PGM_START_PS: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + break; + default: + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return -EINVAL; + } + return 0; +} + +static inline unsigned minify(unsigned size, unsigned levels) +{ + size = size >> levels; + if (size < 1) + size = 1; + return size; +} + +static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned nlevels, + unsigned w0, unsigned h0, unsigned d0, unsigned bpe, + unsigned *l0_size, unsigned *mipmap_size) +{ + unsigned offset, i, level, face; + unsigned width, height, depth, rowstride, size; + + w0 = minify(w0, 0); + h0 = minify(h0, 0); + d0 = minify(d0, 0); + for(i = 0, offset = 0, level = blevel; i < nlevels; i++, level++) { + width = minify(w0, i); + height = minify(h0, i); + depth = minify(d0, i); + for(face = 0; face < nfaces; face++) { + rowstride = ((width * bpe) + 255) & ~255; + size = height * rowstride * depth; + offset += size; + offset = (offset + 0x1f) & ~0x1f; + } + } + *l0_size = (((w0 * bpe) + 255) & ~255) * h0 * d0; + *mipmap_size = offset; + if (!blevel) + *mipmap_size -= *l0_size; + if (!nlevels) + *mipmap_size = *l0_size; +} + +/** + * r600_check_texture_resource() - check if register is authorized or not + * @p: parser structure holding parsing context + * @idx: index into the cs buffer + * @texture: texture's bo structure + * @mipmap: mipmap's bo structure + * + * This function will check that the resource has valid field and that + * the texture and mipmap bo object are big enough to cover this resource. + */ +static inline int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, + struct radeon_bo *texture, + struct radeon_bo *mipmap) +{ + u32 nfaces, nlevels, blevel, w0, h0, d0, bpe = 0; + u32 word0, word1, l0_size, mipmap_size; + + /* on legacy kernel we don't perform advanced check */ + if (p->rdev == NULL) + return 0; + word0 = radeon_get_ib_value(p, idx + 0); + word1 = radeon_get_ib_value(p, idx + 1); + w0 = G_038000_TEX_WIDTH(word0) + 1; + h0 = G_038004_TEX_HEIGHT(word1) + 1; + d0 = G_038004_TEX_DEPTH(word1); + nfaces = 1; + switch (G_038000_DIM(word0)) { + case V_038000_SQ_TEX_DIM_1D: + case V_038000_SQ_TEX_DIM_2D: + case V_038000_SQ_TEX_DIM_3D: + break; + case V_038000_SQ_TEX_DIM_CUBEMAP: + nfaces = 6; + break; + case V_038000_SQ_TEX_DIM_1D_ARRAY: + case V_038000_SQ_TEX_DIM_2D_ARRAY: + case V_038000_SQ_TEX_DIM_2D_MSAA: + case V_038000_SQ_TEX_DIM_2D_ARRAY_MSAA: + default: + dev_warn(p->dev, "this kernel doesn't support %d texture dim\n", G_038000_DIM(word0)); + return -EINVAL; + } + if (r600_bpe_from_format(&bpe, G_038004_DATA_FORMAT(word1))) { + dev_warn(p->dev, "%s:%d texture invalid format %d\n", + __func__, __LINE__, G_038004_DATA_FORMAT(word1)); + return -EINVAL; + } + word0 = radeon_get_ib_value(p, idx + 4); + word1 = radeon_get_ib_value(p, idx + 5); + blevel = G_038010_BASE_LEVEL(word0); + nlevels = G_038014_LAST_LEVEL(word1); + r600_texture_size(nfaces, blevel, nlevels, w0, h0, d0, bpe, &l0_size, &mipmap_size); + /* using get ib will give us the offset into the texture bo */ + word0 = radeon_get_ib_value(p, idx + 2); + if ((l0_size + word0) > radeon_bo_size(texture)) { + dev_warn(p->dev, "texture bo too small (%d %d %d %d -> %d have %ld)\n", + w0, h0, bpe, word0, l0_size, radeon_bo_size(texture)); + return -EINVAL; + } + /* using get ib will give us the offset into the mipmap bo */ + word0 = radeon_get_ib_value(p, idx + 3); + if ((mipmap_size + word0) > radeon_bo_size(mipmap)) { + dev_warn(p->dev, "mipmap bo too small (%d %d %d %d %d %d -> %d have %ld)\n", + w0, h0, bpe, blevel, nlevels, word0, mipmap_size, radeon_bo_size(texture)); + return -EINVAL; + } + return 0; +} + static int r600_packet3_check(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { struct radeon_cs_reloc *reloc; + struct r600_cs_track *track; volatile u32 *ib; unsigned idx; unsigned i; @@ -344,6 +1002,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, int r; u32 idx_value; + track = (struct r600_cs_track *)p->track; ib = p->ib->ptr; idx = pkt->idx + 1; idx_value = radeon_get_ib_value(p, idx); @@ -380,12 +1039,22 @@ static int r600_packet3_check(struct radeon_cs_parser *p, } ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + r = r600_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } break; case PACKET3_DRAW_INDEX_AUTO: if (pkt->count != 1) { DRM_ERROR("bad DRAW_INDEX_AUTO\n"); return -EINVAL; } + r = r600_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream %d\n", __func__, __LINE__, idx); + return r; + } break; case PACKET3_DRAW_INDEX_IMMD_BE: case PACKET3_DRAW_INDEX_IMMD: @@ -393,6 +1062,11 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DRAW_INDEX_IMMD\n"); return -EINVAL; } + r = r600_cs_track_check(p); + if (r) { + dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); + return r; + } break; case PACKET3_WAIT_REG_MEM: if (pkt->count != 5) { @@ -465,30 +1139,9 @@ static int r600_packet3_check(struct radeon_cs_parser *p, } for (i = 0; i < pkt->count; i++) { reg = start_reg + (4 * i); - switch (reg) { - case SQ_ESGS_RING_BASE: - case SQ_GSVS_RING_BASE: - case SQ_ESTMP_RING_BASE: - case SQ_GSTMP_RING_BASE: - case SQ_VSTMP_RING_BASE: - case SQ_PSTMP_RING_BASE: - case SQ_FBUF_RING_BASE: - case SQ_REDUC_RING_BASE: - case SX_MEMORY_EXPORT_BASE: - r = r600_cs_packet_next_reloc(p, &reloc); - if (r) { - DRM_ERROR("bad SET_CONFIG_REG " - "0x%04X\n", reg); - return -EINVAL; - } - ib[idx+1+i] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - break; - case CP_COHER_BASE: - /* use PACKET3_SURFACE_SYNC */ - return -EINVAL; - default: - break; - } + r = r600_cs_check_reg(p, reg, idx+1+i); + if (r) + return r; } break; case PACKET3_SET_CONTEXT_REG: @@ -502,55 +1155,9 @@ static int r600_packet3_check(struct radeon_cs_parser *p, } for (i = 0; i < pkt->count; i++) { reg = start_reg + (4 * i); - switch (reg) { - case DB_DEPTH_BASE: - case DB_HTILE_DATA_BASE: - case CB_COLOR0_BASE: - case CB_COLOR1_BASE: - case CB_COLOR2_BASE: - case CB_COLOR3_BASE: - case CB_COLOR4_BASE: - case CB_COLOR5_BASE: - case CB_COLOR6_BASE: - case CB_COLOR7_BASE: - case SQ_PGM_START_FS: - case SQ_PGM_START_ES: - case SQ_PGM_START_VS: - case SQ_PGM_START_GS: - case SQ_PGM_START_PS: - r = r600_cs_packet_next_reloc(p, &reloc); - if (r) { - DRM_ERROR("bad SET_CONTEXT_REG " - "0x%04X\n", reg); - return -EINVAL; - } - ib[idx+1+i] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - break; - case VGT_DMA_BASE: - case VGT_DMA_BASE_HI: - /* These should be handled by DRAW_INDEX packet 3 */ - case VGT_STRMOUT_BASE_OFFSET_0: - case VGT_STRMOUT_BASE_OFFSET_1: - case VGT_STRMOUT_BASE_OFFSET_2: - case VGT_STRMOUT_BASE_OFFSET_3: - case VGT_STRMOUT_BASE_OFFSET_HI_0: - case VGT_STRMOUT_BASE_OFFSET_HI_1: - case VGT_STRMOUT_BASE_OFFSET_HI_2: - case VGT_STRMOUT_BASE_OFFSET_HI_3: - case VGT_STRMOUT_BUFFER_BASE_0: - case VGT_STRMOUT_BUFFER_BASE_1: - case VGT_STRMOUT_BUFFER_BASE_2: - case VGT_STRMOUT_BUFFER_BASE_3: - case VGT_STRMOUT_BUFFER_OFFSET_0: - case VGT_STRMOUT_BUFFER_OFFSET_1: - case VGT_STRMOUT_BUFFER_OFFSET_2: - case VGT_STRMOUT_BUFFER_OFFSET_3: - /* These should be handled by STRMOUT_BUFFER packet 3 */ - DRM_ERROR("bad context reg: 0x%08x\n", reg); - return -EINVAL; - default: - break; - } + r = r600_cs_check_reg(p, reg, idx+1+i); + if (r) + return r; } break; case PACKET3_SET_RESOURCE: @@ -567,6 +1174,9 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } for (i = 0; i < (pkt->count / 7); i++) { + struct radeon_bo *texture, *mipmap; + u32 size, offset; + switch (G__SQ_VTX_CONSTANT_TYPE(radeon_get_ib_value(p, idx+(i*7)+6+1))) { case SQ_TEX_VTX_VALID_TEXTURE: /* tex base */ @@ -576,6 +1186,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } ib[idx+1+(i*7)+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + texture = reloc->robj; /* tex mip base */ r = r600_cs_packet_next_reloc(p, &reloc); if (r) { @@ -583,6 +1194,11 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } ib[idx+1+(i*7)+3] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + mipmap = reloc->robj; + r = r600_check_texture_resource(p, idx+(i*7)+1, + texture, mipmap); + if (r) + return r; break; case SQ_TEX_VTX_VALID_BUFFER: /* vtx base */ @@ -591,6 +1207,13 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE\n"); return -EINVAL; } + offset = radeon_get_ib_value(p, idx+1+(i*7)+0); + size = radeon_get_ib_value(p, idx+1+(i*7)+1); + if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) { + /* force size to size of the buffer */ + dev_warn(p->dev, "vbo resource seems too big for the bo\n"); + ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj); + } ib[idx+1+(i*7)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff); ib[idx+1+(i*7)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; break; @@ -678,11 +1301,31 @@ static int r600_packet3_check(struct radeon_cs_parser *p, int r600_cs_parse(struct radeon_cs_parser *p) { struct radeon_cs_packet pkt; + struct r600_cs_track *track; int r; + if (p->track == NULL) { + /* initialize tracker, we are in kms */ + track = kzalloc(sizeof(*track), GFP_KERNEL); + if (track == NULL) + return -ENOMEM; + r600_cs_track_init(track); + if (p->rdev->family < CHIP_RV770) { + track->npipes = p->rdev->config.r600.tiling_npipes; + track->nbanks = p->rdev->config.r600.tiling_nbanks; + track->group_size = p->rdev->config.r600.tiling_group_size; + } else if (p->rdev->family <= CHIP_RV740) { + track->npipes = p->rdev->config.rv770.tiling_npipes; + track->nbanks = p->rdev->config.rv770.tiling_nbanks; + track->group_size = p->rdev->config.rv770.tiling_group_size; + } + p->track = track; + } do { r = r600_cs_packet_parse(p, &pkt, p->idx); if (r) { + kfree(p->track); + p->track = NULL; return r; } p->idx += pkt.count + 2; @@ -697,9 +1340,13 @@ int r600_cs_parse(struct radeon_cs_parser *p) break; default: DRM_ERROR("Unknown packet type %d !\n", pkt.type); + kfree(p->track); + p->track = NULL; return -EINVAL; } if (r) { + kfree(p->track); + p->track = NULL; return r; } } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); @@ -709,6 +1356,8 @@ int r600_cs_parse(struct radeon_cs_parser *p) mdelay(1); } #endif + kfree(p->track); + p->track = NULL; return 0; } @@ -751,15 +1400,24 @@ int r600_cs_legacy(struct drm_device *dev, void *data, struct drm_file *filp, { struct radeon_cs_parser parser; struct radeon_cs_chunk *ib_chunk; - struct radeon_ib fake_ib; + struct radeon_ib fake_ib; + struct r600_cs_track *track; int r; + /* initialize tracker */ + track = kzalloc(sizeof(*track), GFP_KERNEL); + if (track == NULL) + return -ENOMEM; + r600_cs_track_init(track); + r600_cs_legacy_get_tiling_conf(dev, &track->npipes, &track->nbanks, &track->group_size); /* initialize parser */ memset(&parser, 0, sizeof(struct radeon_cs_parser)); parser.filp = filp; + parser.dev = &dev->pdev->dev; parser.rdev = NULL; parser.family = family; parser.ib = &fake_ib; + parser.track = track; fake_ib.ptr = ib; r = radeon_cs_parser_init(&parser, data); if (r) { diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h index 05894edadab..5b2e4d44282 100644 --- a/drivers/gpu/drm/radeon/r600d.h +++ b/drivers/gpu/drm/radeon/r600d.h @@ -882,4 +882,494 @@ #define S_000E60_SOFT_RESET_VMC(x) (((x) & 1) << 17) #define R_005480_HDP_MEM_COHERENCY_FLUSH_CNTL 0x5480 + +#define R_028C04_PA_SC_AA_CONFIG 0x028C04 +#define S_028C04_MSAA_NUM_SAMPLES(x) (((x) & 0x3) << 0) +#define G_028C04_MSAA_NUM_SAMPLES(x) (((x) >> 0) & 0x3) +#define C_028C04_MSAA_NUM_SAMPLES 0xFFFFFFFC +#define S_028C04_AA_MASK_CENTROID_DTMN(x) (((x) & 0x1) << 4) +#define G_028C04_AA_MASK_CENTROID_DTMN(x) (((x) >> 4) & 0x1) +#define C_028C04_AA_MASK_CENTROID_DTMN 0xFFFFFFEF +#define S_028C04_MAX_SAMPLE_DIST(x) (((x) & 0xF) << 13) +#define G_028C04_MAX_SAMPLE_DIST(x) (((x) >> 13) & 0xF) +#define C_028C04_MAX_SAMPLE_DIST 0xFFFE1FFF +#define R_0280E0_CB_COLOR0_FRAG 0x0280E0 +#define S_0280E0_BASE_256B(x) (((x) & 0xFFFFFFFF) << 0) +#define G_0280E0_BASE_256B(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_0280E0_BASE_256B 0x00000000 +#define R_0280E4_CB_COLOR1_FRAG 0x0280E4 +#define R_0280E8_CB_COLOR2_FRAG 0x0280E8 +#define R_0280EC_CB_COLOR3_FRAG 0x0280EC +#define R_0280F0_CB_COLOR4_FRAG 0x0280F0 +#define R_0280F4_CB_COLOR5_FRAG 0x0280F4 +#define R_0280F8_CB_COLOR6_FRAG 0x0280F8 +#define R_0280FC_CB_COLOR7_FRAG 0x0280FC +#define R_0280C0_CB_COLOR0_TILE 0x0280C0 +#define S_0280C0_BASE_256B(x) (((x) & 0xFFFFFFFF) << 0) +#define G_0280C0_BASE_256B(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_0280C0_BASE_256B 0x00000000 +#define R_0280C4_CB_COLOR1_TILE 0x0280C4 +#define R_0280C8_CB_COLOR2_TILE 0x0280C8 +#define R_0280CC_CB_COLOR3_TILE 0x0280CC +#define R_0280D0_CB_COLOR4_TILE 0x0280D0 +#define R_0280D4_CB_COLOR5_TILE 0x0280D4 +#define R_0280D8_CB_COLOR6_TILE 0x0280D8 +#define R_0280DC_CB_COLOR7_TILE 0x0280DC +#define R_0280A0_CB_COLOR0_INFO 0x0280A0 +#define S_0280A0_ENDIAN(x) (((x) & 0x3) << 0) +#define G_0280A0_ENDIAN(x) (((x) >> 0) & 0x3) +#define C_0280A0_ENDIAN 0xFFFFFFFC +#define S_0280A0_FORMAT(x) (((x) & 0x3F) << 2) +#define G_0280A0_FORMAT(x) (((x) >> 2) & 0x3F) +#define C_0280A0_FORMAT 0xFFFFFF03 +#define V_0280A0_COLOR_INVALID 0x00000000 +#define V_0280A0_COLOR_8 0x00000001 +#define V_0280A0_COLOR_4_4 0x00000002 +#define V_0280A0_COLOR_3_3_2 0x00000003 +#define V_0280A0_COLOR_16 0x00000005 +#define V_0280A0_COLOR_16_FLOAT 0x00000006 +#define V_0280A0_COLOR_8_8 0x00000007 +#define V_0280A0_COLOR_5_6_5 0x00000008 +#define V_0280A0_COLOR_6_5_5 0x00000009 +#define V_0280A0_COLOR_1_5_5_5 0x0000000A +#define V_0280A0_COLOR_4_4_4_4 0x0000000B +#define V_0280A0_COLOR_5_5_5_1 0x0000000C +#define V_0280A0_COLOR_32 0x0000000D +#define V_0280A0_COLOR_32_FLOAT 0x0000000E +#define V_0280A0_COLOR_16_16 0x0000000F +#define V_0280A0_COLOR_16_16_FLOAT 0x00000010 +#define V_0280A0_COLOR_8_24 0x00000011 +#define V_0280A0_COLOR_8_24_FLOAT 0x00000012 +#define V_0280A0_COLOR_24_8 0x00000013 +#define V_0280A0_COLOR_24_8_FLOAT 0x00000014 +#define V_0280A0_COLOR_10_11_11 0x00000015 +#define V_0280A0_COLOR_10_11_11_FLOAT 0x00000016 +#define V_0280A0_COLOR_11_11_10 0x00000017 +#define V_0280A0_COLOR_11_11_10_FLOAT 0x00000018 +#define V_0280A0_COLOR_2_10_10_10 0x00000019 +#define V_0280A0_COLOR_8_8_8_8 0x0000001A +#define V_0280A0_COLOR_10_10_10_2 0x0000001B +#define V_0280A0_COLOR_X24_8_32_FLOAT 0x0000001C +#define V_0280A0_COLOR_32_32 0x0000001D +#define V_0280A0_COLOR_32_32_FLOAT 0x0000001E +#define V_0280A0_COLOR_16_16_16_16 0x0000001F +#define V_0280A0_COLOR_16_16_16_16_FLOAT 0x00000020 +#define V_0280A0_COLOR_32_32_32_32 0x00000022 +#define V_0280A0_COLOR_32_32_32_32_FLOAT 0x00000023 +#define S_0280A0_ARRAY_MODE(x) (((x) & 0xF) << 8) +#define G_0280A0_ARRAY_MODE(x) (((x) >> 8) & 0xF) +#define C_0280A0_ARRAY_MODE 0xFFFFF0FF +#define V_0280A0_ARRAY_LINEAR_GENERAL 0x00000000 +#define V_0280A0_ARRAY_LINEAR_ALIGNED 0x00000001 +#define V_0280A0_ARRAY_1D_TILED_THIN1 0x00000002 +#define V_0280A0_ARRAY_2D_TILED_THIN1 0x00000004 +#define S_0280A0_NUMBER_TYPE(x) (((x) & 0x7) << 12) +#define G_0280A0_NUMBER_TYPE(x) (((x) >> 12) & 0x7) +#define C_0280A0_NUMBER_TYPE 0xFFFF8FFF +#define S_0280A0_READ_SIZE(x) (((x) & 0x1) << 15) +#define G_0280A0_READ_SIZE(x) (((x) >> 15) & 0x1) +#define C_0280A0_READ_SIZE 0xFFFF7FFF +#define S_0280A0_COMP_SWAP(x) (((x) & 0x3) << 16) +#define G_0280A0_COMP_SWAP(x) (((x) >> 16) & 0x3) +#define C_0280A0_COMP_SWAP 0xFFFCFFFF +#define S_0280A0_TILE_MODE(x) (((x) & 0x3) << 18) +#define G_0280A0_TILE_MODE(x) (((x) >> 18) & 0x3) +#define C_0280A0_TILE_MODE 0xFFF3FFFF +#define S_0280A0_BLEND_CLAMP(x) (((x) & 0x1) << 20) +#define G_0280A0_BLEND_CLAMP(x) (((x) >> 20) & 0x1) +#define C_0280A0_BLEND_CLAMP 0xFFEFFFFF +#define S_0280A0_CLEAR_COLOR(x) (((x) & 0x1) << 21) +#define G_0280A0_CLEAR_COLOR(x) (((x) >> 21) & 0x1) +#define C_0280A0_CLEAR_COLOR 0xFFDFFFFF +#define S_0280A0_BLEND_BYPASS(x) (((x) & 0x1) << 22) +#define G_0280A0_BLEND_BYPASS(x) (((x) >> 22) & 0x1) +#define C_0280A0_BLEND_BYPASS 0xFFBFFFFF +#define S_0280A0_BLEND_FLOAT32(x) (((x) & 0x1) << 23) +#define G_0280A0_BLEND_FLOAT32(x) (((x) >> 23) & 0x1) +#define C_0280A0_BLEND_FLOAT32 0xFF7FFFFF +#define S_0280A0_SIMPLE_FLOAT(x) (((x) & 0x1) << 24) +#define G_0280A0_SIMPLE_FLOAT(x) (((x) >> 24) & 0x1) +#define C_0280A0_SIMPLE_FLOAT 0xFEFFFFFF +#define S_0280A0_ROUND_MODE(x) (((x) & 0x1) << 25) +#define G_0280A0_ROUND_MODE(x) (((x) >> 25) & 0x1) +#define C_0280A0_ROUND_MODE 0xFDFFFFFF +#define S_0280A0_TILE_COMPACT(x) (((x) & 0x1) << 26) +#define G_0280A0_TILE_COMPACT(x) (((x) >> 26) & 0x1) +#define C_0280A0_TILE_COMPACT 0xFBFFFFFF +#define S_0280A0_SOURCE_FORMAT(x) (((x) & 0x1) << 27) +#define G_0280A0_SOURCE_FORMAT(x) (((x) >> 27) & 0x1) +#define C_0280A0_SOURCE_FORMAT 0xF7FFFFFF +#define R_0280A4_CB_COLOR1_INFO 0x0280A4 +#define R_0280A8_CB_COLOR2_INFO 0x0280A8 +#define R_0280AC_CB_COLOR3_INFO 0x0280AC +#define R_0280B0_CB_COLOR4_INFO 0x0280B0 +#define R_0280B4_CB_COLOR5_INFO 0x0280B4 +#define R_0280B8_CB_COLOR6_INFO 0x0280B8 +#define R_0280BC_CB_COLOR7_INFO 0x0280BC +#define R_028060_CB_COLOR0_SIZE 0x028060 +#define S_028060_PITCH_TILE_MAX(x) (((x) & 0x3FF) << 0) +#define G_028060_PITCH_TILE_MAX(x) (((x) >> 0) & 0x3FF) +#define C_028060_PITCH_TILE_MAX 0xFFFFFC00 +#define S_028060_SLICE_TILE_MAX(x) (((x) & 0xFFFFF) << 10) +#define G_028060_SLICE_TILE_MAX(x) (((x) >> 10) & 0xFFFFF) +#define C_028060_SLICE_TILE_MAX 0xC00003FF +#define R_028064_CB_COLOR1_SIZE 0x028064 +#define R_028068_CB_COLOR2_SIZE 0x028068 +#define R_02806C_CB_COLOR3_SIZE 0x02806C +#define R_028070_CB_COLOR4_SIZE 0x028070 +#define R_028074_CB_COLOR5_SIZE 0x028074 +#define R_028078_CB_COLOR6_SIZE 0x028078 +#define R_02807C_CB_COLOR7_SIZE 0x02807C +#define R_028238_CB_TARGET_MASK 0x028238 +#define S_028238_TARGET0_ENABLE(x) (((x) & 0xF) << 0) +#define G_028238_TARGET0_ENABLE(x) (((x) >> 0) & 0xF) +#define C_028238_TARGET0_ENABLE 0xFFFFFFF0 +#define S_028238_TARGET1_ENABLE(x) (((x) & 0xF) << 4) +#define G_028238_TARGET1_ENABLE(x) (((x) >> 4) & 0xF) +#define C_028238_TARGET1_ENABLE 0xFFFFFF0F +#define S_028238_TARGET2_ENABLE(x) (((x) & 0xF) << 8) +#define G_028238_TARGET2_ENABLE(x) (((x) >> 8) & 0xF) +#define C_028238_TARGET2_ENABLE 0xFFFFF0FF +#define S_028238_TARGET3_ENABLE(x) (((x) & 0xF) << 12) +#define G_028238_TARGET3_ENABLE(x) (((x) >> 12) & 0xF) +#define C_028238_TARGET3_ENABLE 0xFFFF0FFF +#define S_028238_TARGET4_ENABLE(x) (((x) & 0xF) << 16) +#define G_028238_TARGET4_ENABLE(x) (((x) >> 16) & 0xF) +#define C_028238_TARGET4_ENABLE 0xFFF0FFFF +#define S_028238_TARGET5_ENABLE(x) (((x) & 0xF) << 20) +#define G_028238_TARGET5_ENABLE(x) (((x) >> 20) & 0xF) +#define C_028238_TARGET5_ENABLE 0xFF0FFFFF +#define S_028238_TARGET6_ENABLE(x) (((x) & 0xF) << 24) +#define G_028238_TARGET6_ENABLE(x) (((x) >> 24) & 0xF) +#define C_028238_TARGET6_ENABLE 0xF0FFFFFF +#define S_028238_TARGET7_ENABLE(x) (((x) & 0xF) << 28) +#define G_028238_TARGET7_ENABLE(x) (((x) >> 28) & 0xF) +#define C_028238_TARGET7_ENABLE 0x0FFFFFFF +#define R_02823C_CB_SHADER_MASK 0x02823C +#define S_02823C_OUTPUT0_ENABLE(x) (((x) & 0xF) << 0) +#define G_02823C_OUTPUT0_ENABLE(x) (((x) >> 0) & 0xF) +#define C_02823C_OUTPUT0_ENABLE 0xFFFFFFF0 +#define S_02823C_OUTPUT1_ENABLE(x) (((x) & 0xF) << 4) +#define G_02823C_OUTPUT1_ENABLE(x) (((x) >> 4) & 0xF) +#define C_02823C_OUTPUT1_ENABLE 0xFFFFFF0F +#define S_02823C_OUTPUT2_ENABLE(x) (((x) & 0xF) << 8) +#define G_02823C_OUTPUT2_ENABLE(x) (((x) >> 8) & 0xF) +#define C_02823C_OUTPUT2_ENABLE 0xFFFFF0FF +#define S_02823C_OUTPUT3_ENABLE(x) (((x) & 0xF) << 12) +#define G_02823C_OUTPUT3_ENABLE(x) (((x) >> 12) & 0xF) +#define C_02823C_OUTPUT3_ENABLE 0xFFFF0FFF +#define S_02823C_OUTPUT4_ENABLE(x) (((x) & 0xF) << 16) +#define G_02823C_OUTPUT4_ENABLE(x) (((x) >> 16) & 0xF) +#define C_02823C_OUTPUT4_ENABLE 0xFFF0FFFF +#define S_02823C_OUTPUT5_ENABLE(x) (((x) & 0xF) << 20) +#define G_02823C_OUTPUT5_ENABLE(x) (((x) >> 20) & 0xF) +#define C_02823C_OUTPUT5_ENABLE 0xFF0FFFFF +#define S_02823C_OUTPUT6_ENABLE(x) (((x) & 0xF) << 24) +#define G_02823C_OUTPUT6_ENABLE(x) (((x) >> 24) & 0xF) +#define C_02823C_OUTPUT6_ENABLE 0xF0FFFFFF +#define S_02823C_OUTPUT7_ENABLE(x) (((x) & 0xF) << 28) +#define G_02823C_OUTPUT7_ENABLE(x) (((x) >> 28) & 0xF) +#define C_02823C_OUTPUT7_ENABLE 0x0FFFFFFF +#define R_028AB0_VGT_STRMOUT_EN 0x028AB0 +#define S_028AB0_STREAMOUT(x) (((x) & 0x1) << 0) +#define G_028AB0_STREAMOUT(x) (((x) >> 0) & 0x1) +#define C_028AB0_STREAMOUT 0xFFFFFFFE +#define R_028B20_VGT_STRMOUT_BUFFER_EN 0x028B20 +#define S_028B20_BUFFER_0_EN(x) (((x) & 0x1) << 0) +#define G_028B20_BUFFER_0_EN(x) (((x) >> 0) & 0x1) +#define C_028B20_BUFFER_0_EN 0xFFFFFFFE +#define S_028B20_BUFFER_1_EN(x) (((x) & 0x1) << 1) +#define G_028B20_BUFFER_1_EN(x) (((x) >> 1) & 0x1) +#define C_028B20_BUFFER_1_EN 0xFFFFFFFD +#define S_028B20_BUFFER_2_EN(x) (((x) & 0x1) << 2) +#define G_028B20_BUFFER_2_EN(x) (((x) >> 2) & 0x1) +#define C_028B20_BUFFER_2_EN 0xFFFFFFFB +#define S_028B20_BUFFER_3_EN(x) (((x) & 0x1) << 3) +#define G_028B20_BUFFER_3_EN(x) (((x) >> 3) & 0x1) +#define C_028B20_BUFFER_3_EN 0xFFFFFFF7 +#define S_028B20_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_028B20_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_028B20_SIZE 0x00000000 +#define R_038000_SQ_TEX_RESOURCE_WORD0_0 0x038000 +#define S_038000_DIM(x) (((x) & 0x7) << 0) +#define G_038000_DIM(x) (((x) >> 0) & 0x7) +#define C_038000_DIM 0xFFFFFFF8 +#define V_038000_SQ_TEX_DIM_1D 0x00000000 +#define V_038000_SQ_TEX_DIM_2D 0x00000001 +#define V_038000_SQ_TEX_DIM_3D 0x00000002 +#define V_038000_SQ_TEX_DIM_CUBEMAP 0x00000003 +#define V_038000_SQ_TEX_DIM_1D_ARRAY 0x00000004 +#define V_038000_SQ_TEX_DIM_2D_ARRAY 0x00000005 +#define V_038000_SQ_TEX_DIM_2D_MSAA 0x00000006 +#define V_038000_SQ_TEX_DIM_2D_ARRAY_MSAA 0x00000007 +#define S_038000_TILE_MODE(x) (((x) & 0xF) << 3) +#define G_038000_TILE_MODE(x) (((x) >> 3) & 0xF) +#define C_038000_TILE_MODE 0xFFFFFF87 +#define S_038000_TILE_TYPE(x) (((x) & 0x1) << 7) +#define G_038000_TILE_TYPE(x) (((x) >> 7) & 0x1) +#define C_038000_TILE_TYPE 0xFFFFFF7F +#define S_038000_PITCH(x) (((x) & 0x7FF) << 8) +#define G_038000_PITCH(x) (((x) >> 8) & 0x7FF) +#define C_038000_PITCH 0xFFF800FF +#define S_038000_TEX_WIDTH(x) (((x) & 0x1FFF) << 19) +#define G_038000_TEX_WIDTH(x) (((x) >> 19) & 0x1FFF) +#define C_038000_TEX_WIDTH 0x0007FFFF +#define R_038004_SQ_TEX_RESOURCE_WORD1_0 0x038004 +#define S_038004_TEX_HEIGHT(x) (((x) & 0x1FFF) << 0) +#define G_038004_TEX_HEIGHT(x) (((x) >> 0) & 0x1FFF) +#define C_038004_TEX_HEIGHT 0xFFFFE000 +#define S_038004_TEX_DEPTH(x) (((x) & 0x1FFF) << 13) +#define G_038004_TEX_DEPTH(x) (((x) >> 13) & 0x1FFF) +#define C_038004_TEX_DEPTH 0xFC001FFF +#define S_038004_DATA_FORMAT(x) (((x) & 0x3F) << 26) +#define G_038004_DATA_FORMAT(x) (((x) >> 26) & 0x3F) +#define C_038004_DATA_FORMAT 0x03FFFFFF +#define V_038004_COLOR_INVALID 0x00000000 +#define V_038004_COLOR_8 0x00000001 +#define V_038004_COLOR_4_4 0x00000002 +#define V_038004_COLOR_3_3_2 0x00000003 +#define V_038004_COLOR_16 0x00000005 +#define V_038004_COLOR_16_FLOAT 0x00000006 +#define V_038004_COLOR_8_8 0x00000007 +#define V_038004_COLOR_5_6_5 0x00000008 +#define V_038004_COLOR_6_5_5 0x00000009 +#define V_038004_COLOR_1_5_5_5 0x0000000A +#define V_038004_COLOR_4_4_4_4 0x0000000B +#define V_038004_COLOR_5_5_5_1 0x0000000C +#define V_038004_COLOR_32 0x0000000D +#define V_038004_COLOR_32_FLOAT 0x0000000E +#define V_038004_COLOR_16_16 0x0000000F +#define V_038004_COLOR_16_16_FLOAT 0x00000010 +#define V_038004_COLOR_8_24 0x00000011 +#define V_038004_COLOR_8_24_FLOAT 0x00000012 +#define V_038004_COLOR_24_8 0x00000013 +#define V_038004_COLOR_24_8_FLOAT 0x00000014 +#define V_038004_COLOR_10_11_11 0x00000015 +#define V_038004_COLOR_10_11_11_FLOAT 0x00000016 +#define V_038004_COLOR_11_11_10 0x00000017 +#define V_038004_COLOR_11_11_10_FLOAT 0x00000018 +#define V_038004_COLOR_2_10_10_10 0x00000019 +#define V_038004_COLOR_8_8_8_8 0x0000001A +#define V_038004_COLOR_10_10_10_2 0x0000001B +#define V_038004_COLOR_X24_8_32_FLOAT 0x0000001C +#define V_038004_COLOR_32_32 0x0000001D +#define V_038004_COLOR_32_32_FLOAT 0x0000001E +#define V_038004_COLOR_16_16_16_16 0x0000001F +#define V_038004_COLOR_16_16_16_16_FLOAT 0x00000020 +#define V_038004_COLOR_32_32_32_32 0x00000022 +#define V_038004_COLOR_32_32_32_32_FLOAT 0x00000023 +#define V_038004_FMT_1 0x00000025 +#define V_038004_FMT_GB_GR 0x00000027 +#define V_038004_FMT_BG_RG 0x00000028 +#define V_038004_FMT_32_AS_8 0x00000029 +#define V_038004_FMT_32_AS_8_8 0x0000002A +#define V_038004_FMT_5_9_9_9_SHAREDEXP 0x0000002B +#define V_038004_FMT_8_8_8 0x0000002C +#define V_038004_FMT_16_16_16 0x0000002D +#define V_038004_FMT_16_16_16_FLOAT 0x0000002E +#define V_038004_FMT_32_32_32 0x0000002F +#define V_038004_FMT_32_32_32_FLOAT 0x00000030 +#define R_038010_SQ_TEX_RESOURCE_WORD4_0 0x038010 +#define S_038010_FORMAT_COMP_X(x) (((x) & 0x3) << 0) +#define G_038010_FORMAT_COMP_X(x) (((x) >> 0) & 0x3) +#define C_038010_FORMAT_COMP_X 0xFFFFFFFC +#define S_038010_FORMAT_COMP_Y(x) (((x) & 0x3) << 2) +#define G_038010_FORMAT_COMP_Y(x) (((x) >> 2) & 0x3) +#define C_038010_FORMAT_COMP_Y 0xFFFFFFF3 +#define S_038010_FORMAT_COMP_Z(x) (((x) & 0x3) << 4) +#define G_038010_FORMAT_COMP_Z(x) (((x) >> 4) & 0x3) +#define C_038010_FORMAT_COMP_Z 0xFFFFFFCF +#define S_038010_FORMAT_COMP_W(x) (((x) & 0x3) << 6) +#define G_038010_FORMAT_COMP_W(x) (((x) >> 6) & 0x3) +#define C_038010_FORMAT_COMP_W 0xFFFFFF3F +#define S_038010_NUM_FORMAT_ALL(x) (((x) & 0x3) << 8) +#define G_038010_NUM_FORMAT_ALL(x) (((x) >> 8) & 0x3) +#define C_038010_NUM_FORMAT_ALL 0xFFFFFCFF +#define S_038010_SRF_MODE_ALL(x) (((x) & 0x1) << 10) +#define G_038010_SRF_MODE_ALL(x) (((x) >> 10) & 0x1) +#define C_038010_SRF_MODE_ALL 0xFFFFFBFF +#define S_038010_FORCE_DEGAMMA(x) (((x) & 0x1) << 11) +#define G_038010_FORCE_DEGAMMA(x) (((x) >> 11) & 0x1) +#define C_038010_FORCE_DEGAMMA 0xFFFFF7FF +#define S_038010_ENDIAN_SWAP(x) (((x) & 0x3) << 12) +#define G_038010_ENDIAN_SWAP(x) (((x) >> 12) & 0x3) +#define C_038010_ENDIAN_SWAP 0xFFFFCFFF +#define S_038010_REQUEST_SIZE(x) (((x) & 0x3) << 14) +#define G_038010_REQUEST_SIZE(x) (((x) >> 14) & 0x3) +#define C_038010_REQUEST_SIZE 0xFFFF3FFF +#define S_038010_DST_SEL_X(x) (((x) & 0x7) << 16) +#define G_038010_DST_SEL_X(x) (((x) >> 16) & 0x7) +#define C_038010_DST_SEL_X 0xFFF8FFFF +#define S_038010_DST_SEL_Y(x) (((x) & 0x7) << 19) +#define G_038010_DST_SEL_Y(x) (((x) >> 19) & 0x7) +#define C_038010_DST_SEL_Y 0xFFC7FFFF +#define S_038010_DST_SEL_Z(x) (((x) & 0x7) << 22) +#define G_038010_DST_SEL_Z(x) (((x) >> 22) & 0x7) +#define C_038010_DST_SEL_Z 0xFE3FFFFF +#define S_038010_DST_SEL_W(x) (((x) & 0x7) << 25) +#define G_038010_DST_SEL_W(x) (((x) >> 25) & 0x7) +#define C_038010_DST_SEL_W 0xF1FFFFFF +#define S_038010_BASE_LEVEL(x) (((x) & 0xF) << 28) +#define G_038010_BASE_LEVEL(x) (((x) >> 28) & 0xF) +#define C_038010_BASE_LEVEL 0x0FFFFFFF +#define R_038014_SQ_TEX_RESOURCE_WORD5_0 0x038014 +#define S_038014_LAST_LEVEL(x) (((x) & 0xF) << 0) +#define G_038014_LAST_LEVEL(x) (((x) >> 0) & 0xF) +#define C_038014_LAST_LEVEL 0xFFFFFFF0 +#define S_038014_BASE_ARRAY(x) (((x) & 0x1FFF) << 4) +#define G_038014_BASE_ARRAY(x) (((x) >> 4) & 0x1FFF) +#define C_038014_BASE_ARRAY 0xFFFE000F +#define S_038014_LAST_ARRAY(x) (((x) & 0x1FFF) << 17) +#define G_038014_LAST_ARRAY(x) (((x) >> 17) & 0x1FFF) +#define C_038014_LAST_ARRAY 0xC001FFFF +#define R_0288A8_SQ_ESGS_RING_ITEMSIZE 0x0288A8 +#define S_0288A8_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288A8_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288A8_ITEMSIZE 0xFFFF8000 +#define R_008C44_SQ_ESGS_RING_SIZE 0x008C44 +#define S_008C44_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C44_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C44_MEM_SIZE 0x00000000 +#define R_0288B0_SQ_ESTMP_RING_ITEMSIZE 0x0288B0 +#define S_0288B0_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288B0_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288B0_ITEMSIZE 0xFFFF8000 +#define R_008C54_SQ_ESTMP_RING_SIZE 0x008C54 +#define S_008C54_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C54_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C54_MEM_SIZE 0x00000000 +#define R_0288C0_SQ_FBUF_RING_ITEMSIZE 0x0288C0 +#define S_0288C0_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288C0_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288C0_ITEMSIZE 0xFFFF8000 +#define R_008C74_SQ_FBUF_RING_SIZE 0x008C74 +#define S_008C74_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C74_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C74_MEM_SIZE 0x00000000 +#define R_0288B4_SQ_GSTMP_RING_ITEMSIZE 0x0288B4 +#define S_0288B4_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288B4_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288B4_ITEMSIZE 0xFFFF8000 +#define R_008C5C_SQ_GSTMP_RING_SIZE 0x008C5C +#define S_008C5C_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C5C_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C5C_MEM_SIZE 0x00000000 +#define R_0288AC_SQ_GSVS_RING_ITEMSIZE 0x0288AC +#define S_0288AC_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288AC_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288AC_ITEMSIZE 0xFFFF8000 +#define R_008C4C_SQ_GSVS_RING_SIZE 0x008C4C +#define S_008C4C_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C4C_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C4C_MEM_SIZE 0x00000000 +#define R_0288BC_SQ_PSTMP_RING_ITEMSIZE 0x0288BC +#define S_0288BC_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288BC_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288BC_ITEMSIZE 0xFFFF8000 +#define R_008C6C_SQ_PSTMP_RING_SIZE 0x008C6C +#define S_008C6C_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C6C_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C6C_MEM_SIZE 0x00000000 +#define R_0288C4_SQ_REDUC_RING_ITEMSIZE 0x0288C4 +#define S_0288C4_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288C4_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288C4_ITEMSIZE 0xFFFF8000 +#define R_008C7C_SQ_REDUC_RING_SIZE 0x008C7C +#define S_008C7C_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C7C_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C7C_MEM_SIZE 0x00000000 +#define R_0288B8_SQ_VSTMP_RING_ITEMSIZE 0x0288B8 +#define S_0288B8_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288B8_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288B8_ITEMSIZE 0xFFFF8000 +#define R_008C64_SQ_VSTMP_RING_SIZE 0x008C64 +#define S_008C64_MEM_SIZE(x) (((x) & 0xFFFFFFFF) << 0) +#define G_008C64_MEM_SIZE(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_008C64_MEM_SIZE 0x00000000 +#define R_0288C8_SQ_GS_VERT_ITEMSIZE 0x0288C8 +#define S_0288C8_ITEMSIZE(x) (((x) & 0x7FFF) << 0) +#define G_0288C8_ITEMSIZE(x) (((x) >> 0) & 0x7FFF) +#define C_0288C8_ITEMSIZE 0xFFFF8000 +#define R_028010_DB_DEPTH_INFO 0x028010 +#define S_028010_FORMAT(x) (((x) & 0x7) << 0) +#define G_028010_FORMAT(x) (((x) >> 0) & 0x7) +#define C_028010_FORMAT 0xFFFFFFF8 +#define V_028010_DEPTH_INVALID 0x00000000 +#define V_028010_DEPTH_16 0x00000001 +#define V_028010_DEPTH_X8_24 0x00000002 +#define V_028010_DEPTH_8_24 0x00000003 +#define V_028010_DEPTH_X8_24_FLOAT 0x00000004 +#define V_028010_DEPTH_8_24_FLOAT 0x00000005 +#define V_028010_DEPTH_32_FLOAT 0x00000006 +#define V_028010_DEPTH_X24_8_32_FLOAT 0x00000007 +#define S_028010_READ_SIZE(x) (((x) & 0x1) << 3) +#define G_028010_READ_SIZE(x) (((x) >> 3) & 0x1) +#define C_028010_READ_SIZE 0xFFFFFFF7 +#define S_028010_ARRAY_MODE(x) (((x) & 0xF) << 15) +#define G_028010_ARRAY_MODE(x) (((x) >> 15) & 0xF) +#define C_028010_ARRAY_MODE 0xFFF87FFF +#define S_028010_TILE_SURFACE_ENABLE(x) (((x) & 0x1) << 25) +#define G_028010_TILE_SURFACE_ENABLE(x) (((x) >> 25) & 0x1) +#define C_028010_TILE_SURFACE_ENABLE 0xFDFFFFFF +#define S_028010_TILE_COMPACT(x) (((x) & 0x1) << 26) +#define G_028010_TILE_COMPACT(x) (((x) >> 26) & 0x1) +#define C_028010_TILE_COMPACT 0xFBFFFFFF +#define S_028010_ZRANGE_PRECISION(x) (((x) & 0x1) << 31) +#define G_028010_ZRANGE_PRECISION(x) (((x) >> 31) & 0x1) +#define C_028010_ZRANGE_PRECISION 0x7FFFFFFF +#define R_028000_DB_DEPTH_SIZE 0x028000 +#define S_028000_PITCH_TILE_MAX(x) (((x) & 0x3FF) << 0) +#define G_028000_PITCH_TILE_MAX(x) (((x) >> 0) & 0x3FF) +#define C_028000_PITCH_TILE_MAX 0xFFFFFC00 +#define S_028000_SLICE_TILE_MAX(x) (((x) & 0xFFFFF) << 10) +#define G_028000_SLICE_TILE_MAX(x) (((x) >> 10) & 0xFFFFF) +#define C_028000_SLICE_TILE_MAX 0xC00003FF +#define R_028004_DB_DEPTH_VIEW 0x028004 +#define S_028004_SLICE_START(x) (((x) & 0x7FF) << 0) +#define G_028004_SLICE_START(x) (((x) >> 0) & 0x7FF) +#define C_028004_SLICE_START 0xFFFFF800 +#define S_028004_SLICE_MAX(x) (((x) & 0x7FF) << 13) +#define G_028004_SLICE_MAX(x) (((x) >> 13) & 0x7FF) +#define C_028004_SLICE_MAX 0xFF001FFF +#define R_028800_DB_DEPTH_CONTROL 0x028800 +#define S_028800_STENCIL_ENABLE(x) (((x) & 0x1) << 0) +#define G_028800_STENCIL_ENABLE(x) (((x) >> 0) & 0x1) +#define C_028800_STENCIL_ENABLE 0xFFFFFFFE +#define S_028800_Z_ENABLE(x) (((x) & 0x1) << 1) +#define G_028800_Z_ENABLE(x) (((x) >> 1) & 0x1) +#define C_028800_Z_ENABLE 0xFFFFFFFD +#define S_028800_Z_WRITE_ENABLE(x) (((x) & 0x1) << 2) +#define G_028800_Z_WRITE_ENABLE(x) (((x) >> 2) & 0x1) +#define C_028800_Z_WRITE_ENABLE 0xFFFFFFFB +#define S_028800_ZFUNC(x) (((x) & 0x7) << 4) +#define G_028800_ZFUNC(x) (((x) >> 4) & 0x7) +#define C_028800_ZFUNC 0xFFFFFF8F +#define S_028800_BACKFACE_ENABLE(x) (((x) & 0x1) << 7) +#define G_028800_BACKFACE_ENABLE(x) (((x) >> 7) & 0x1) +#define C_028800_BACKFACE_ENABLE 0xFFFFFF7F +#define S_028800_STENCILFUNC(x) (((x) & 0x7) << 8) +#define G_028800_STENCILFUNC(x) (((x) >> 8) & 0x7) +#define C_028800_STENCILFUNC 0xFFFFF8FF +#define S_028800_STENCILFAIL(x) (((x) & 0x7) << 11) +#define G_028800_STENCILFAIL(x) (((x) >> 11) & 0x7) +#define C_028800_STENCILFAIL 0xFFFFC7FF +#define S_028800_STENCILZPASS(x) (((x) & 0x7) << 14) +#define G_028800_STENCILZPASS(x) (((x) >> 14) & 0x7) +#define C_028800_STENCILZPASS 0xFFFE3FFF +#define S_028800_STENCILZFAIL(x) (((x) & 0x7) << 17) +#define G_028800_STENCILZFAIL(x) (((x) >> 17) & 0x7) +#define C_028800_STENCILZFAIL 0xFFF1FFFF +#define S_028800_STENCILFUNC_BF(x) (((x) & 0x7) << 20) +#define G_028800_STENCILFUNC_BF(x) (((x) >> 20) & 0x7) +#define C_028800_STENCILFUNC_BF 0xFF8FFFFF +#define S_028800_STENCILFAIL_BF(x) (((x) & 0x7) << 23) +#define G_028800_STENCILFAIL_BF(x) (((x) >> 23) & 0x7) +#define C_028800_STENCILFAIL_BF 0xFC7FFFFF +#define S_028800_STENCILZPASS_BF(x) (((x) & 0x7) << 26) +#define G_028800_STENCILZPASS_BF(x) (((x) >> 26) & 0x7) +#define C_028800_STENCILZPASS_BF 0xE3FFFFFF +#define S_028800_STENCILZFAIL_BF(x) (((x) & 0x7) << 29) +#define G_028800_STENCILZFAIL_BF(x) (((x) >> 29) & 0x7) +#define C_028800_STENCILZFAIL_BF 0x1FFFFFFF + #endif diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index eb5f99b9469..829e26e8a4b 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -89,6 +89,7 @@ extern int radeon_testing; extern int radeon_connector_table; extern int radeon_tv; extern int radeon_new_pll; +extern int radeon_dynpm; extern int radeon_audio; /* @@ -96,6 +97,7 @@ extern int radeon_audio; * symbol; */ #define RADEON_MAX_USEC_TIMEOUT 100000 /* 100 ms */ +/* RADEON_IB_POOL_SIZE must be a power of 2 */ #define RADEON_IB_POOL_SIZE 16 #define RADEON_DEBUGFS_MAX_NUM_FILES 32 #define RADEONFB_CONN_LIMIT 4 @@ -117,6 +119,21 @@ struct radeon_device; /* * BIOS. */ +#define ATRM_BIOS_PAGE 4096 + +#if defined(CONFIG_VGA_SWITCHEROO) +bool radeon_atrm_supported(struct pci_dev *pdev); +int radeon_atrm_get_bios_chunk(uint8_t *bios, int offset, int len); +#else +static inline bool radeon_atrm_supported(struct pci_dev *pdev) +{ + return false; +} + +static inline int radeon_atrm_get_bios_chunk(uint8_t *bios, int offset, int len){ + return -EINVAL; +} +#endif bool radeon_get_bios(struct radeon_device *rdev); @@ -137,17 +154,23 @@ void radeon_dummy_page_fini(struct radeon_device *rdev); struct radeon_clock { struct radeon_pll p1pll; struct radeon_pll p2pll; + struct radeon_pll dcpll; struct radeon_pll spll; struct radeon_pll mpll; /* 10 Khz units */ uint32_t default_mclk; uint32_t default_sclk; + uint32_t default_dispclk; + uint32_t dp_extclk; }; /* * Power management */ int radeon_pm_init(struct radeon_device *rdev); +void radeon_pm_compute_clocks(struct radeon_device *rdev); +void radeon_combios_get_power_modes(struct radeon_device *rdev); +void radeon_atombios_get_power_modes(struct radeon_device *rdev); /* * Fences. @@ -274,6 +297,7 @@ union radeon_gart_table { }; #define RADEON_GPU_PAGE_SIZE 4096 +#define RADEON_GPU_PAGE_MASK (RADEON_GPU_PAGE_SIZE - 1) struct radeon_gart { dma_addr_t table_addr; @@ -308,21 +332,19 @@ struct radeon_mc { /* for some chips with <= 32MB we need to lie * about vram size near mc fb location */ u64 mc_vram_size; - u64 gtt_location; + u64 visible_vram_size; u64 gtt_size; u64 gtt_start; u64 gtt_end; - u64 vram_location; u64 vram_start; u64 vram_end; unsigned vram_width; u64 real_vram_size; int vram_mtrr; bool vram_is_ddr; - bool igp_sideport_enabled; + bool igp_sideport_enabled; }; -int radeon_mc_setup(struct radeon_device *rdev); bool radeon_combios_sideport_present(struct radeon_device *rdev); bool radeon_atombios_sideport_present(struct radeon_device *rdev); @@ -347,6 +369,7 @@ struct radeon_irq { bool sw_int; /* FIXME: use a define max crtc rather than hardcode it */ bool crtc_vblank_int[2]; + wait_queue_head_t vblank_queue; /* FIXME: use defines for max hpd/dacs */ bool hpd[6]; spinlock_t sw_lock; @@ -363,11 +386,12 @@ void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev); */ struct radeon_ib { struct list_head list; - unsigned long idx; + unsigned idx; uint64_t gpu_addr; struct radeon_fence *fence; - uint32_t *ptr; + uint32_t *ptr; uint32_t length_dw; + bool free; }; /* @@ -377,10 +401,10 @@ struct radeon_ib { struct radeon_ib_pool { struct mutex mutex; struct radeon_bo *robj; - struct list_head scheduled_ibs; + struct list_head bogus_ib; struct radeon_ib ibs[RADEON_IB_POOL_SIZE]; bool ready; - DECLARE_BITMAP(alloc_bm, RADEON_IB_POOL_SIZE); + unsigned head_id; }; struct radeon_cp { @@ -410,13 +434,13 @@ struct r600_ih { unsigned wptr_old; unsigned ring_size; uint64_t gpu_addr; - uint32_t align_mask; uint32_t ptr_mask; spinlock_t lock; bool enabled; }; struct r600_blit { + struct mutex mutex; struct radeon_bo *shader_obj; u64 shader_gpu_addr; u32 vs_offset, ps_offset; @@ -432,6 +456,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib); int radeon_ib_pool_init(struct radeon_device *rdev); void radeon_ib_pool_fini(struct radeon_device *rdev); int radeon_ib_test(struct radeon_device *rdev); +extern void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib); /* Ring access between begin & end cannot sleep */ void radeon_ring_free_size(struct radeon_device *rdev); int radeon_ring_lock(struct radeon_device *rdev, unsigned ndw); @@ -465,6 +490,7 @@ struct radeon_cs_chunk { }; struct radeon_cs_parser { + struct device *dev; struct radeon_device *rdev; struct drm_file *filp; /* chunks */ @@ -568,7 +594,99 @@ struct radeon_wb { * Equation between gpu/memory clock and available bandwidth is hw dependent * (type of memory, bus size, efficiency, ...) */ +enum radeon_pm_state { + PM_STATE_DISABLED, + PM_STATE_MINIMUM, + PM_STATE_PAUSED, + PM_STATE_ACTIVE +}; +enum radeon_pm_action { + PM_ACTION_NONE, + PM_ACTION_MINIMUM, + PM_ACTION_DOWNCLOCK, + PM_ACTION_UPCLOCK +}; + +enum radeon_voltage_type { + VOLTAGE_NONE = 0, + VOLTAGE_GPIO, + VOLTAGE_VDDC, + VOLTAGE_SW +}; + +enum radeon_pm_state_type { + POWER_STATE_TYPE_DEFAULT, + POWER_STATE_TYPE_POWERSAVE, + POWER_STATE_TYPE_BATTERY, + POWER_STATE_TYPE_BALANCED, + POWER_STATE_TYPE_PERFORMANCE, +}; + +enum radeon_pm_clock_mode_type { + POWER_MODE_TYPE_DEFAULT, + POWER_MODE_TYPE_LOW, + POWER_MODE_TYPE_MID, + POWER_MODE_TYPE_HIGH, +}; + +struct radeon_voltage { + enum radeon_voltage_type type; + /* gpio voltage */ + struct radeon_gpio_rec gpio; + u32 delay; /* delay in usec from voltage drop to sclk change */ + bool active_high; /* voltage drop is active when bit is high */ + /* VDDC voltage */ + u8 vddc_id; /* index into vddc voltage table */ + u8 vddci_id; /* index into vddci voltage table */ + bool vddci_enabled; + /* r6xx+ sw */ + u32 voltage; +}; + +struct radeon_pm_non_clock_info { + /* pcie lanes */ + int pcie_lanes; + /* standardized non-clock flags */ + u32 flags; +}; + +struct radeon_pm_clock_info { + /* memory clock */ + u32 mclk; + /* engine clock */ + u32 sclk; + /* voltage info */ + struct radeon_voltage voltage; + /* standardized clock flags - not sure we'll need these */ + u32 flags; +}; + +struct radeon_power_state { + enum radeon_pm_state_type type; + /* XXX: use a define for num clock modes */ + struct radeon_pm_clock_info clock_info[8]; + /* number of valid clock modes in this power state */ + int num_clock_modes; + struct radeon_pm_clock_info *default_clock_mode; + /* non clock info about this state */ + struct radeon_pm_non_clock_info non_clock_info; + bool voltage_drop_active; +}; + +/* + * Some modes are overclocked by very low value, accept them + */ +#define RADEON_MODE_OVERCLOCK_MARGIN 500 /* 5 MHz */ + struct radeon_pm { + struct mutex mutex; + struct delayed_work idle_work; + enum radeon_pm_state state; + enum radeon_pm_action planned_action; + unsigned long action_timeout; + bool downclocked; + int active_crtcs; + int req_vblank; fixed20_12 max_bandwidth; fixed20_12 igp_sideport_mclk; fixed20_12 igp_system_mclk; @@ -580,6 +698,15 @@ struct radeon_pm { fixed20_12 core_bandwidth; fixed20_12 sclk; fixed20_12 needed_bandwidth; + /* XXX: use a define for num power modes */ + struct radeon_power_state power_state[8]; + /* number of valid power states */ + int num_power_states; + struct radeon_power_state *current_power_state; + struct radeon_pm_clock_info *current_clock_mode; + struct radeon_power_state *requested_power_state; + struct radeon_pm_clock_info *requested_clock_mode; + struct radeon_power_state *default_power_state; }; @@ -649,6 +776,7 @@ struct radeon_asic { void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock); uint32_t (*get_memory_clock)(struct radeon_device *rdev); void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock); + int (*get_pcie_lanes)(struct radeon_device *rdev); void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes); void (*set_clock_gating)(struct radeon_device *rdev, int enable); int (*set_surface_reg)(struct radeon_device *rdev, int reg, @@ -660,6 +788,13 @@ struct radeon_asic { void (*hpd_fini)(struct radeon_device *rdev); bool (*hpd_sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd); void (*hpd_set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd); + /* ioctl hw specific callback. Some hw might want to perform special + * operation on specific ioctl. For instance on wait idle some hw + * might want to perform and HDP flush through MMIO as it seems that + * some R6XX/R7XX hw doesn't take HDP flush into account if programmed + * through ring. + */ + void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo); }; /* @@ -692,6 +827,9 @@ struct r600_asic { unsigned sx_max_export_pos_size; unsigned sx_max_export_smx_size; unsigned sq_num_cf_insts; + unsigned tiling_nbanks; + unsigned tiling_npipes; + unsigned tiling_group_size; }; struct rv770_asic { @@ -712,6 +850,9 @@ struct rv770_asic { unsigned sc_prim_fifo_size; unsigned sc_hiz_tile_fifo_size; unsigned sc_earlyz_tile_fifo_fize; + unsigned tiling_nbanks; + unsigned tiling_npipes; + unsigned tiling_group_size; }; union radeon_asic_config { @@ -821,6 +962,8 @@ struct radeon_device { struct r600_ih ih; /* r6/700 interrupt ring */ struct workqueue_struct *wq; struct work_struct hotplug_work; + int num_crtc; /* number of crtcs */ + struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */ /* audio stuff */ struct timer_list audio_timer; @@ -829,6 +972,8 @@ struct radeon_device { int audio_bits_per_sample; uint8_t audio_status_bits; uint8_t audio_category_code; + + bool powered_down; }; int radeon_device_init(struct radeon_device *rdev, @@ -847,7 +992,7 @@ void r600_kms_blit_copy(struct radeon_device *rdev, static inline uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg) { - if (reg < 0x10000) + if (reg < rdev->rmmio_size) return readl(((void __iomem *)rdev->rmmio) + reg); else { writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); @@ -857,7 +1002,7 @@ static inline uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg) static inline void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) { - if (reg < 0x10000) + if (reg < rdev->rmmio_size) writel(v, ((void __iomem *)rdev->rmmio) + reg); else { writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); @@ -886,6 +1031,8 @@ static inline void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32 #define WREG32_MC(reg, v) rdev->mc_wreg(rdev, (reg), (v)) #define RREG32_PCIE(reg) rv370_pcie_rreg(rdev, (reg)) #define WREG32_PCIE(reg, v) rv370_pcie_wreg(rdev, (reg), (v)) +#define RREG32_PCIE_P(reg) rdev->pciep_rreg(rdev, (reg)) +#define WREG32_PCIE_P(reg, v) rdev->pciep_wreg(rdev, (reg), (v)) #define WREG32_P(reg, val, mask) \ do { \ uint32_t tmp_ = RREG32(reg); \ @@ -947,7 +1094,7 @@ void r100_pll_errata_after_index(struct radeon_device *rdev); #define ASIC_IS_AVIVO(rdev) ((rdev->family >= CHIP_RS600)) #define ASIC_IS_DCE3(rdev) ((rdev->family >= CHIP_RV620)) #define ASIC_IS_DCE32(rdev) ((rdev->family >= CHIP_RV730)) - +#define ASIC_IS_DCE4(rdev) ((rdev->family >= CHIP_CEDAR)) /* * BIOS helpers. @@ -1006,6 +1153,7 @@ static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v) #define radeon_set_engine_clock(rdev, e) (rdev)->asic->set_engine_clock((rdev), (e)) #define radeon_get_memory_clock(rdev) (rdev)->asic->get_memory_clock((rdev)) #define radeon_set_memory_clock(rdev, e) (rdev)->asic->set_memory_clock((rdev), (e)) +#define radeon_get_pcie_lanes(rdev) (rdev)->asic->get_pcie_lanes((rdev)) #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->set_pcie_lanes((rdev), (l)) #define radeon_set_clock_gating(rdev, e) (rdev)->asic->set_clock_gating((rdev), (e)) #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->set_surface_reg((rdev), (r), (f), (p), (o), (s))) @@ -1017,7 +1165,10 @@ static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v) #define radeon_hpd_set_polarity(rdev, hpd) (rdev)->asic->hpd_set_polarity((rdev), (hpd)) /* Common functions */ +/* AGP */ +extern void radeon_agp_disable(struct radeon_device *rdev); extern int radeon_gart_table_vram_pin(struct radeon_device *rdev); +extern void radeon_gart_restore(struct radeon_device *rdev); extern int radeon_modeset_init(struct radeon_device *rdev); extern void radeon_modeset_fini(struct radeon_device *rdev); extern bool radeon_card_posted(struct radeon_device *rdev); @@ -1031,6 +1182,10 @@ extern void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enabl extern void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable); extern void radeon_ttm_placement_from_domain(struct radeon_bo *rbo, u32 domain); extern bool radeon_ttm_bo_is_radeon_bo(struct ttm_buffer_object *bo); +extern void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base); +extern void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc); +extern int radeon_resume_kms(struct drm_device *dev); +extern int radeon_suspend_kms(struct drm_device *dev, pm_message_t state); /* r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 */ struct r100_mc_save { @@ -1085,7 +1240,7 @@ extern void r200_set_safe_registers(struct radeon_device *rdev); /* r300,r350,rv350,rv370,rv380 */ extern void r300_set_reg_safe(struct radeon_device *rdev); extern void r300_mc_program(struct radeon_device *rdev); -extern void r300_vram_info(struct radeon_device *rdev); +extern void r300_mc_init(struct radeon_device *rdev); extern void r300_clock_startup(struct radeon_device *rdev); extern int r300_mc_wait_for_idle(struct radeon_device *rdev); extern int rv370_pcie_gart_init(struct radeon_device *rdev); @@ -1094,7 +1249,6 @@ extern int rv370_pcie_gart_enable(struct radeon_device *rdev); extern void rv370_pcie_gart_disable(struct radeon_device *rdev); /* r420,r423,rv410 */ -extern int r420_mc_init(struct radeon_device *rdev); extern u32 r420_mc_rreg(struct radeon_device *rdev, u32 reg); extern void r420_mc_wreg(struct radeon_device *rdev, u32 reg, u32 v); extern int r420_debugfs_pipes_info_init(struct radeon_device *rdev); @@ -1136,12 +1290,13 @@ extern void rs690_line_buffer_adjust(struct radeon_device *rdev, struct drm_display_mode *mode2); /* r600, rv610, rv630, rv620, rv635, rv670, rs780, rs880 */ +extern void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc); extern bool r600_card_posted(struct radeon_device *rdev); extern void r600_cp_stop(struct radeon_device *rdev); extern void r600_ring_init(struct radeon_device *rdev, unsigned ring_size); extern int r600_cp_resume(struct radeon_device *rdev); +extern void r600_cp_fini(struct radeon_device *rdev); extern int r600_count_pipe_bits(uint32_t val); -extern int r600_gart_clear_page(struct radeon_device *rdev, int i); extern int r600_mc_wait_for_idle(struct radeon_device *rdev); extern int r600_pcie_gart_init(struct radeon_device *rdev); extern void r600_pcie_gart_tlb_flush(struct radeon_device *rdev); @@ -1160,7 +1315,8 @@ extern int r600_irq_init(struct radeon_device *rdev); extern void r600_irq_fini(struct radeon_device *rdev); extern void r600_ih_ring_init(struct radeon_device *rdev, unsigned ring_size); extern int r600_irq_set(struct radeon_device *rdev); - +extern void r600_irq_suspend(struct radeon_device *rdev); +/* r600 audio */ extern int r600_audio_init(struct radeon_device *rdev); extern int r600_audio_tmds_index(struct drm_encoder *encoder); extern void r600_audio_set_clock(struct drm_encoder *encoder, int clock); @@ -1176,6 +1332,14 @@ extern void r600_hdmi_update_audio_settings(struct drm_encoder *encoder, uint8_t status_bits, uint8_t category_code); +/* evergreen */ +struct evergreen_mc_save { + u32 vga_control[6]; + u32 vga_render_control; + u32 vga_hdp_control; + u32 crtc_control[6]; +}; + #include "radeon_object.h" #endif diff --git a/drivers/gpu/drm/radeon/radeon_agp.c b/drivers/gpu/drm/radeon/radeon_agp.c index 220f454ea9f..c4457791dff 100644 --- a/drivers/gpu/drm/radeon/radeon_agp.c +++ b/drivers/gpu/drm/radeon/radeon_agp.c @@ -144,9 +144,19 @@ int radeon_agp_init(struct radeon_device *rdev) ret = drm_agp_info(rdev->ddev, &info); if (ret) { + drm_agp_release(rdev->ddev); DRM_ERROR("Unable to get AGP info: %d\n", ret); return ret; } + + if (rdev->ddev->agp->agp_info.aper_size < 32) { + drm_agp_release(rdev->ddev); + dev_warn(rdev->dev, "AGP aperture too small (%zuM) " + "need at least 32M, disabling AGP\n", + rdev->ddev->agp->agp_info.aper_size); + return -EINVAL; + } + mode.mode = info.mode; agp_status = (RREG32(RADEON_AGP_STATUS) | RADEON_AGPv3_MODE) & mode.mode; is_v3 = !!(agp_status & RADEON_AGPv3_MODE); @@ -221,11 +231,16 @@ int radeon_agp_init(struct radeon_device *rdev) ret = drm_agp_enable(rdev->ddev, mode); if (ret) { DRM_ERROR("Unable to enable AGP (mode = 0x%lx)\n", mode.mode); + drm_agp_release(rdev->ddev); return ret; } rdev->mc.agp_base = rdev->ddev->agp->agp_info.aper_base; rdev->mc.gtt_size = rdev->ddev->agp->agp_info.aper_size << 20; + rdev->mc.gtt_start = rdev->mc.agp_base; + rdev->mc.gtt_end = rdev->mc.gtt_start + rdev->mc.gtt_size - 1; + dev_info(rdev->dev, "GTT: %lluM 0x%08llX - 0x%08llX\n", + rdev->mc.gtt_size >> 20, rdev->mc.gtt_start, rdev->mc.gtt_end); /* workaround some hw issues */ if (rdev->family < CHIP_R200) { diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index f2fbd2e4e9d..d3a157b2bcb 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -43,7 +43,7 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev, uint32_t mem_clock void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable); /* - * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 + * r100,rv100,rs100,rv200,rs200 */ extern int r100_init(struct radeon_device *rdev); extern void r100_fini(struct radeon_device *rdev); @@ -108,6 +108,7 @@ static struct radeon_asic r100_asic = { .set_engine_clock = &radeon_legacy_set_engine_clock, .get_memory_clock = &radeon_legacy_get_memory_clock, .set_memory_clock = NULL, + .get_pcie_lanes = NULL, .set_pcie_lanes = NULL, .set_clock_gating = &radeon_legacy_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -117,6 +118,52 @@ static struct radeon_asic r100_asic = { .hpd_fini = &r100_hpd_fini, .hpd_sense = &r100_hpd_sense, .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, +}; + +/* + * r200,rv250,rs300,rv280 + */ +extern int r200_copy_dma(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_pages, + struct radeon_fence *fence); +static struct radeon_asic r200_asic = { + .init = &r100_init, + .fini = &r100_fini, + .suspend = &r100_suspend, + .resume = &r100_resume, + .vga_set_state = &r100_vga_set_state, + .gpu_reset = &r100_gpu_reset, + .gart_tlb_flush = &r100_pci_gart_tlb_flush, + .gart_set_page = &r100_pci_gart_set_page, + .cp_commit = &r100_cp_commit, + .ring_start = &r100_ring_start, + .ring_test = &r100_ring_test, + .ring_ib_execute = &r100_ring_ib_execute, + .irq_set = &r100_irq_set, + .irq_process = &r100_irq_process, + .get_vblank_counter = &r100_get_vblank_counter, + .fence_ring_emit = &r100_fence_ring_emit, + .cs_parse = &r100_cs_parse, + .copy_blit = &r100_copy_blit, + .copy_dma = &r200_copy_dma, + .copy = &r100_copy_blit, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_legacy_set_clock_gating, + .set_surface_reg = r100_set_surface_reg, + .clear_surface_reg = r100_clear_surface_reg, + .bandwidth_update = &r100_bandwidth_update, + .hpd_init = &r100_hpd_init, + .hpd_fini = &r100_hpd_fini, + .hpd_sense = &r100_hpd_sense, + .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -137,11 +184,8 @@ extern int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t extern uint32_t rv370_pcie_rreg(struct radeon_device *rdev, uint32_t reg); extern void rv370_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); extern void rv370_set_pcie_lanes(struct radeon_device *rdev, int lanes); -extern int r300_copy_dma(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_pages, - struct radeon_fence *fence); +extern int rv370_get_pcie_lanes(struct radeon_device *rdev); + static struct radeon_asic r300_asic = { .init = &r300_init, .fini = &r300_fini, @@ -161,12 +205,13 @@ static struct radeon_asic r300_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_legacy_get_engine_clock, .set_engine_clock = &radeon_legacy_set_engine_clock, .get_memory_clock = &radeon_legacy_get_memory_clock, .set_memory_clock = NULL, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = &rv370_set_pcie_lanes, .set_clock_gating = &radeon_legacy_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -176,6 +221,45 @@ static struct radeon_asic r300_asic = { .hpd_fini = &r100_hpd_fini, .hpd_sense = &r100_hpd_sense, .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, +}; + + +static struct radeon_asic r300_asic_pcie = { + .init = &r300_init, + .fini = &r300_fini, + .suspend = &r300_suspend, + .resume = &r300_resume, + .vga_set_state = &r100_vga_set_state, + .gpu_reset = &r300_gpu_reset, + .gart_tlb_flush = &rv370_pcie_gart_tlb_flush, + .gart_set_page = &rv370_pcie_gart_set_page, + .cp_commit = &r100_cp_commit, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ring_ib_execute = &r100_ring_ib_execute, + .irq_set = &r100_irq_set, + .irq_process = &r100_irq_process, + .get_vblank_counter = &r100_get_vblank_counter, + .fence_ring_emit = &r300_fence_ring_emit, + .cs_parse = &r300_cs_parse, + .copy_blit = &r100_copy_blit, + .copy_dma = &r200_copy_dma, + .copy = &r100_copy_blit, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_legacy_set_clock_gating, + .set_surface_reg = r100_set_surface_reg, + .clear_surface_reg = r100_clear_surface_reg, + .bandwidth_update = &r100_bandwidth_update, + .hpd_init = &r100_hpd_init, + .hpd_fini = &r100_hpd_fini, + .hpd_sense = &r100_hpd_sense, + .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; /* @@ -204,12 +288,13 @@ static struct radeon_asic r420_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_atom_get_engine_clock, .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = &rv370_set_pcie_lanes, .set_clock_gating = &radeon_atom_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -219,6 +304,7 @@ static struct radeon_asic r420_asic = { .hpd_fini = &r100_hpd_fini, .hpd_sense = &r100_hpd_sense, .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -252,12 +338,13 @@ static struct radeon_asic rs400_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_legacy_get_engine_clock, .set_engine_clock = &radeon_legacy_set_engine_clock, .get_memory_clock = &radeon_legacy_get_memory_clock, .set_memory_clock = NULL, + .get_pcie_lanes = NULL, .set_pcie_lanes = NULL, .set_clock_gating = &radeon_legacy_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -267,6 +354,7 @@ static struct radeon_asic rs400_asic = { .hpd_fini = &r100_hpd_fini, .hpd_sense = &r100_hpd_sense, .hpd_set_polarity = &r100_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -310,19 +398,23 @@ static struct radeon_asic rs600_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_atom_get_engine_clock, .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, .set_pcie_lanes = NULL, .set_clock_gating = &radeon_atom_set_clock_gating, + .set_surface_reg = r100_set_surface_reg, + .clear_surface_reg = r100_clear_surface_reg, .bandwidth_update = &rs600_bandwidth_update, .hpd_init = &rs600_hpd_init, .hpd_fini = &rs600_hpd_fini, .hpd_sense = &rs600_hpd_sense, .hpd_set_polarity = &rs600_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -355,12 +447,13 @@ static struct radeon_asic rs690_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, - .copy = &r300_copy_dma, + .copy_dma = &r200_copy_dma, + .copy = &r200_copy_dma, .get_engine_clock = &radeon_atom_get_engine_clock, .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, .set_pcie_lanes = NULL, .set_clock_gating = &radeon_atom_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -370,6 +463,7 @@ static struct radeon_asic rs690_asic = { .hpd_fini = &rs600_hpd_fini, .hpd_sense = &rs600_hpd_sense, .hpd_set_polarity = &rs600_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -406,12 +500,13 @@ static struct radeon_asic rv515_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_atom_get_engine_clock, .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = &rv370_set_pcie_lanes, .set_clock_gating = &radeon_atom_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -421,6 +516,7 @@ static struct radeon_asic rv515_asic = { .hpd_fini = &rs600_hpd_fini, .hpd_sense = &rs600_hpd_sense, .hpd_set_polarity = &rs600_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; @@ -448,12 +544,13 @@ static struct radeon_asic r520_asic = { .fence_ring_emit = &r300_fence_ring_emit, .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, - .copy_dma = &r300_copy_dma, + .copy_dma = &r200_copy_dma, .copy = &r100_copy_blit, .get_engine_clock = &radeon_atom_get_engine_clock, .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = &rv370_set_pcie_lanes, .set_clock_gating = &radeon_atom_set_clock_gating, .set_surface_reg = r100_set_surface_reg, @@ -463,6 +560,7 @@ static struct radeon_asic r520_asic = { .hpd_fini = &rs600_hpd_fini, .hpd_sense = &rs600_hpd_sense, .hpd_set_polarity = &rs600_hpd_set_polarity, + .ioctl_wait_idle = NULL, }; /* @@ -504,6 +602,7 @@ void r600_hpd_fini(struct radeon_device *rdev); bool r600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd); void r600_hpd_set_polarity(struct radeon_device *rdev, enum radeon_hpd_id hpd); +extern void r600_ioctl_wait_idle(struct radeon_device *rdev, struct radeon_bo *bo); static struct radeon_asic r600_asic = { .init = &r600_init, @@ -529,8 +628,9 @@ static struct radeon_asic r600_asic = { .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_atom_set_clock_gating, + .set_clock_gating = NULL, .set_surface_reg = r600_set_surface_reg, .clear_surface_reg = r600_clear_surface_reg, .bandwidth_update = &rv515_bandwidth_update, @@ -538,6 +638,7 @@ static struct radeon_asic r600_asic = { .hpd_fini = &r600_hpd_fini, .hpd_sense = &r600_hpd_sense, .hpd_set_polarity = &r600_hpd_set_polarity, + .ioctl_wait_idle = r600_ioctl_wait_idle, }; /* @@ -573,6 +674,7 @@ static struct radeon_asic rv770_asic = { .set_engine_clock = &radeon_atom_set_engine_clock, .get_memory_clock = &radeon_atom_get_memory_clock, .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, .set_pcie_lanes = NULL, .set_clock_gating = &radeon_atom_set_clock_gating, .set_surface_reg = r600_set_surface_reg, @@ -582,6 +684,57 @@ static struct radeon_asic rv770_asic = { .hpd_fini = &r600_hpd_fini, .hpd_sense = &r600_hpd_sense, .hpd_set_polarity = &r600_hpd_set_polarity, + .ioctl_wait_idle = r600_ioctl_wait_idle, +}; + +/* + * evergreen + */ +int evergreen_init(struct radeon_device *rdev); +void evergreen_fini(struct radeon_device *rdev); +int evergreen_suspend(struct radeon_device *rdev); +int evergreen_resume(struct radeon_device *rdev); +int evergreen_gpu_reset(struct radeon_device *rdev); +void evergreen_bandwidth_update(struct radeon_device *rdev); +void evergreen_hpd_init(struct radeon_device *rdev); +void evergreen_hpd_fini(struct radeon_device *rdev); +bool evergreen_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd); +void evergreen_hpd_set_polarity(struct radeon_device *rdev, + enum radeon_hpd_id hpd); + +static struct radeon_asic evergreen_asic = { + .init = &evergreen_init, + .fini = &evergreen_fini, + .suspend = &evergreen_suspend, + .resume = &evergreen_resume, + .cp_commit = NULL, + .gpu_reset = &evergreen_gpu_reset, + .vga_set_state = &r600_vga_set_state, + .gart_tlb_flush = &r600_pcie_gart_tlb_flush, + .gart_set_page = &rs600_gart_set_page, + .ring_test = NULL, + .ring_ib_execute = NULL, + .irq_set = NULL, + .irq_process = NULL, + .get_vblank_counter = NULL, + .fence_ring_emit = NULL, + .cs_parse = NULL, + .copy_blit = NULL, + .copy_dma = NULL, + .copy = NULL, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + .set_surface_reg = r600_set_surface_reg, + .clear_surface_reg = r600_clear_surface_reg, + .bandwidth_update = &evergreen_bandwidth_update, + .hpd_init = &evergreen_hpd_init, + .hpd_fini = &evergreen_hpd_fini, + .hpd_sense = &evergreen_hpd_sense, + .hpd_set_polarity = &evergreen_hpd_set_polarity, }; #endif diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index fa82ca74324..93783b15c81 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -159,8 +159,15 @@ static struct radeon_hpd radeon_atom_get_hpd_info_from_gpio(struct radeon_device struct radeon_gpio_rec *gpio) { struct radeon_hpd hpd; + u32 reg; + + if (ASIC_IS_DCE4(rdev)) + reg = EVERGREEN_DC_GPIO_HPD_A; + else + reg = AVIVO_DC_GPIO_HPD_A; + hpd.gpio = *gpio; - if (gpio->reg == AVIVO_DC_GPIO_HPD_A) { + if (gpio->reg == reg) { switch(gpio->mask) { case (1 << 0): hpd.hpd = RADEON_HPD_1; @@ -206,6 +213,15 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev, *connector_type = DRM_MODE_CONNECTOR_DVID; } + /* Asrock RS600 board lists the DVI port as HDMI */ + if ((dev->pdev->device == 0x7941) && + (dev->pdev->subsystem_vendor == 0x1849) && + (dev->pdev->subsystem_device == 0x7941)) { + if ((*connector_type == DRM_MODE_CONNECTOR_HDMIA) && + (supported_device == ATOM_DEVICE_DFP3_SUPPORT)) + *connector_type = DRM_MODE_CONNECTOR_DVID; + } + /* a-bit f-i90hd - ciaranm on #radeonhd - this board has no DVI */ if ((dev->pdev->device == 0x7941) && (dev->pdev->subsystem_vendor == 0x147b) && @@ -287,6 +303,15 @@ static bool radeon_atom_apply_quirks(struct drm_device *dev, *connector_type = DRM_MODE_CONNECTOR_DVID; } + /* XFX Pine Group device rv730 reports no VGA DDC lines + * even though they are wired up to record 0x93 + */ + if ((dev->pdev->device == 0x9498) && + (dev->pdev->subsystem_vendor == 0x1682) && + (dev->pdev->subsystem_device == 0x2452)) { + struct radeon_device *rdev = dev->dev_private; + *i2c_bus = radeon_lookup_i2c_gpio(rdev, 0x93); + } return true; } @@ -556,6 +581,9 @@ bool radeon_get_atom_connector_info_from_object_table(struct drm_device *dev) ddc_bus.valid = false; } + /* needed for aux chan transactions */ + ddc_bus.hpd_id = hpd.hpd ? (hpd.hpd - 1) : 0; + conn_id = le16_to_cpu(path->usConnObjectId); if (!radeon_atom_apply_quirks @@ -820,6 +848,7 @@ union firmware_info { ATOM_FIRMWARE_INFO_V1_2 info_12; ATOM_FIRMWARE_INFO_V1_3 info_13; ATOM_FIRMWARE_INFO_V1_4 info_14; + ATOM_FIRMWARE_INFO_V2_1 info_21; }; bool radeon_atom_get_clock_info(struct drm_device *dev) @@ -831,6 +860,7 @@ bool radeon_atom_get_clock_info(struct drm_device *dev) uint8_t frev, crev; struct radeon_pll *p1pll = &rdev->clock.p1pll; struct radeon_pll *p2pll = &rdev->clock.p2pll; + struct radeon_pll *dcpll = &rdev->clock.dcpll; struct radeon_pll *spll = &rdev->clock.spll; struct radeon_pll *mpll = &rdev->clock.mpll; uint16_t data_offset; @@ -933,8 +963,19 @@ bool radeon_atom_get_clock_info(struct drm_device *dev) rdev->clock.default_mclk = le32_to_cpu(firmware_info->info.ulDefaultMemoryClock); + if (ASIC_IS_DCE4(rdev)) { + rdev->clock.default_dispclk = + le32_to_cpu(firmware_info->info_21.ulDefaultDispEngineClkFreq); + if (rdev->clock.default_dispclk == 0) + rdev->clock.default_dispclk = 60000; /* 600 Mhz */ + rdev->clock.dp_extclk = + le16_to_cpu(firmware_info->info_21.usUniphyDPModeExtClkFreq); + } + *dcpll = *p1pll; + return true; } + return false; } @@ -1073,6 +1114,30 @@ static struct radeon_atom_ss *radeon_atombios_get_ss_info(struct return ss; } +static void radeon_atom_apply_lvds_quirks(struct drm_device *dev, + struct radeon_encoder_atom_dig *lvds) +{ + + /* Toshiba A300-1BU laptop panel doesn't like new pll divider algo */ + if ((dev->pdev->device == 0x95c4) && + (dev->pdev->subsystem_vendor == 0x1179) && + (dev->pdev->subsystem_device == 0xff50)) { + if ((lvds->native_mode.hdisplay == 1280) && + (lvds->native_mode.vdisplay == 800)) + lvds->pll_algo = PLL_ALGO_LEGACY; + } + + /* Dell Studio 15 laptop panel doesn't like new pll divider algo */ + if ((dev->pdev->device == 0x95c4) && + (dev->pdev->subsystem_vendor == 0x1028) && + (dev->pdev->subsystem_device == 0x029f)) { + if ((lvds->native_mode.hdisplay == 1280) && + (lvds->native_mode.vdisplay == 800)) + lvds->pll_algo = PLL_ALGO_LEGACY; + } + +} + union lvds_info { struct _ATOM_LVDS_INFO info; struct _ATOM_LVDS_INFO_V12 info_12; @@ -1143,6 +1208,21 @@ struct radeon_encoder_atom_dig *radeon_atombios_get_lvds_info(struct lvds->ss = radeon_atombios_get_ss_info(encoder, lvds_info->info.ucSS_Id); + if (ASIC_IS_AVIVO(rdev)) { + if (radeon_new_pll == 0) + lvds->pll_algo = PLL_ALGO_LEGACY; + else + lvds->pll_algo = PLL_ALGO_NEW; + } else { + if (radeon_new_pll == 1) + lvds->pll_algo = PLL_ALGO_NEW; + else + lvds->pll_algo = PLL_ALGO_LEGACY; + } + + /* LVDS quirks */ + radeon_atom_apply_lvds_quirks(dev, lvds); + encoder->native_mode = lvds->native_mode; } return lvds; @@ -1367,20 +1447,375 @@ radeon_atombios_get_tv_dac_info(struct radeon_encoder *encoder) return tv_dac; } -void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable) +union power_info { + struct _ATOM_POWERPLAY_INFO info; + struct _ATOM_POWERPLAY_INFO_V2 info_2; + struct _ATOM_POWERPLAY_INFO_V3 info_3; + struct _ATOM_PPLIB_POWERPLAYTABLE info_4; +}; + +void radeon_atombios_get_power_modes(struct radeon_device *rdev) { - DYNAMIC_CLOCK_GATING_PS_ALLOCATION args; - int index = GetIndexIntoMasterTable(COMMAND, DynamicClockGating); + struct radeon_mode_info *mode_info = &rdev->mode_info; + int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); + u16 data_offset; + u8 frev, crev; + u32 misc, misc2 = 0, sclk, mclk; + union power_info *power_info; + struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info; + struct _ATOM_PPLIB_STATE *power_state; + int num_modes = 0, i, j; + int state_index = 0, mode_index = 0; - args.ucEnable = enable; + atom_parse_data_header(mode_info->atom_context, index, NULL, &frev, &crev, &data_offset); - atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); + + rdev->pm.default_power_state = NULL; + + if (power_info) { + if (frev < 4) { + num_modes = power_info->info.ucNumOfPowerModeEntries; + if (num_modes > ATOM_MAX_NUMBEROF_POWER_BLOCK) + num_modes = ATOM_MAX_NUMBEROF_POWER_BLOCK; + for (i = 0; i < num_modes; i++) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; + switch (frev) { + case 1: + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = + le16_to_cpu(power_info->info.asPowerPlayInfo[i].usMemoryClock); + rdev->pm.power_state[state_index].clock_info[0].sclk = + le16_to_cpu(power_info->info.asPowerPlayInfo[i].usEngineClock); + /* skip invalid modes */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk == 0) || + (rdev->pm.power_state[state_index].clock_info[0].sclk == 0)) + continue; + /* skip overclock modes for now */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk > + rdev->clock.default_mclk + RADEON_MODE_OVERCLOCK_MARGIN) || + (rdev->pm.power_state[state_index].clock_info[0].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN)) + continue; + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = + power_info->info.asPowerPlayInfo[i].ucNumPciELanes; + misc = le32_to_cpu(power_info->info.asPowerPlayInfo[i].ulMiscInfo); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_GPIO; + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = + radeon_lookup_gpio(rdev, + power_info->info.asPowerPlayInfo[i].ucVoltageDropIndex); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH) + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + true; + else + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + false; + } else if (misc & ATOM_PM_MISCINFO_PROGRAM_VOLTAGE) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_VDDC; + rdev->pm.power_state[state_index].clock_info[0].voltage.vddc_id = + power_info->info.asPowerPlayInfo[i].ucVoltageDropIndex; + } + /* order matters! */ + if (misc & ATOM_PM_MISCINFO_POWER_SAVING_MODE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_POWERSAVE; + if (misc & ATOM_PM_MISCINFO_DEFAULT_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_DEFAULT_LOW_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_LOAD_BALANCE_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + if (misc & ATOM_PM_MISCINFO_3D_ACCELERATION_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_PERFORMANCE; + if (misc & ATOM_PM_MISCINFO_DRIVER_DEFAULT_MODE) { + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + rdev->pm.power_state[state_index].default_clock_mode = + &rdev->pm.power_state[state_index].clock_info[0]; + } + state_index++; + break; + case 2: + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = + le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMemoryClock); + rdev->pm.power_state[state_index].clock_info[0].sclk = + le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulEngineClock); + /* skip invalid modes */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk == 0) || + (rdev->pm.power_state[state_index].clock_info[0].sclk == 0)) + continue; + /* skip overclock modes for now */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk > + rdev->clock.default_mclk + RADEON_MODE_OVERCLOCK_MARGIN) || + (rdev->pm.power_state[state_index].clock_info[0].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN)) + continue; + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = + power_info->info_2.asPowerPlayInfo[i].ucNumPciELanes; + misc = le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMiscInfo); + misc2 = le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMiscInfo2); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_GPIO; + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = + radeon_lookup_gpio(rdev, + power_info->info_2.asPowerPlayInfo[i].ucVoltageDropIndex); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH) + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + true; + else + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + false; + } else if (misc & ATOM_PM_MISCINFO_PROGRAM_VOLTAGE) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_VDDC; + rdev->pm.power_state[state_index].clock_info[0].voltage.vddc_id = + power_info->info_2.asPowerPlayInfo[i].ucVoltageDropIndex; + } + /* order matters! */ + if (misc & ATOM_PM_MISCINFO_POWER_SAVING_MODE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_POWERSAVE; + if (misc & ATOM_PM_MISCINFO_DEFAULT_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_DEFAULT_LOW_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_LOAD_BALANCE_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + if (misc & ATOM_PM_MISCINFO_3D_ACCELERATION_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_PERFORMANCE; + if (misc2 & ATOM_PM_MISCINFO2_SYSTEM_AC_LITE_MODE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + if (misc & ATOM_PM_MISCINFO_DRIVER_DEFAULT_MODE) { + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + rdev->pm.power_state[state_index].default_clock_mode = + &rdev->pm.power_state[state_index].clock_info[0]; + } + state_index++; + break; + case 3: + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = + le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMemoryClock); + rdev->pm.power_state[state_index].clock_info[0].sclk = + le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulEngineClock); + /* skip invalid modes */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk == 0) || + (rdev->pm.power_state[state_index].clock_info[0].sclk == 0)) + continue; + /* skip overclock modes for now */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk > + rdev->clock.default_mclk + RADEON_MODE_OVERCLOCK_MARGIN) || + (rdev->pm.power_state[state_index].clock_info[0].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN)) + continue; + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = + power_info->info_3.asPowerPlayInfo[i].ucNumPciELanes; + misc = le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMiscInfo); + misc2 = le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMiscInfo2); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_SUPPORT) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_GPIO; + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio = + radeon_lookup_gpio(rdev, + power_info->info_3.asPowerPlayInfo[i].ucVoltageDropIndex); + if (misc & ATOM_PM_MISCINFO_VOLTAGE_DROP_ACTIVE_HIGH) + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + true; + else + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + false; + } else if (misc & ATOM_PM_MISCINFO_PROGRAM_VOLTAGE) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = + VOLTAGE_VDDC; + rdev->pm.power_state[state_index].clock_info[0].voltage.vddc_id = + power_info->info_3.asPowerPlayInfo[i].ucVoltageDropIndex; + if (misc2 & ATOM_PM_MISCINFO2_VDDCI_DYNAMIC_VOLTAGE_EN) { + rdev->pm.power_state[state_index].clock_info[0].voltage.vddci_enabled = + true; + rdev->pm.power_state[state_index].clock_info[0].voltage.vddci_id = + power_info->info_3.asPowerPlayInfo[i].ucVDDCI_VoltageDropIndex; + } + } + /* order matters! */ + if (misc & ATOM_PM_MISCINFO_POWER_SAVING_MODE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_POWERSAVE; + if (misc & ATOM_PM_MISCINFO_DEFAULT_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_DEFAULT_LOW_DC_STATE_ENTRY_TRUE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + if (misc & ATOM_PM_MISCINFO_LOAD_BALANCE_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + if (misc & ATOM_PM_MISCINFO_3D_ACCELERATION_EN) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_PERFORMANCE; + if (misc2 & ATOM_PM_MISCINFO2_SYSTEM_AC_LITE_MODE) + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + if (misc & ATOM_PM_MISCINFO_DRIVER_DEFAULT_MODE) { + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + rdev->pm.power_state[state_index].default_clock_mode = + &rdev->pm.power_state[state_index].clock_info[0]; + } + state_index++; + break; + } + } + } else if (frev == 4) { + for (i = 0; i < power_info->info_4.ucNumStates; i++) { + mode_index = 0; + power_state = (struct _ATOM_PPLIB_STATE *) + (mode_info->atom_context->bios + + data_offset + + le16_to_cpu(power_info->info_4.usStateArrayOffset) + + i * power_info->info_4.ucStateEntrySize); + non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) + (mode_info->atom_context->bios + + data_offset + + le16_to_cpu(power_info->info_4.usNonClockInfoArrayOffset) + + (power_state->ucNonClockStateIndex * + power_info->info_4.ucNonClockSize)); + for (j = 0; j < (power_info->info_4.ucStateEntrySize - 1); j++) { + if (rdev->flags & RADEON_IS_IGP) { + struct _ATOM_PPLIB_RS780_CLOCK_INFO *clock_info = + (struct _ATOM_PPLIB_RS780_CLOCK_INFO *) + (mode_info->atom_context->bios + + data_offset + + le16_to_cpu(power_info->info_4.usClockInfoArrayOffset) + + (power_state->ucClockStateIndices[j] * + power_info->info_4.ucClockInfoSize)); + sclk = le16_to_cpu(clock_info->usLowEngineClockLow); + sclk |= clock_info->ucLowEngineClockHigh << 16; + rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk; + /* skip invalid modes */ + if (rdev->pm.power_state[state_index].clock_info[mode_index].sclk == 0) + continue; + /* skip overclock modes for now */ + if (rdev->pm.power_state[state_index].clock_info[mode_index].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN) + continue; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = + VOLTAGE_SW; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = + clock_info->usVDDC; + mode_index++; + } else { + struct _ATOM_PPLIB_R600_CLOCK_INFO *clock_info = + (struct _ATOM_PPLIB_R600_CLOCK_INFO *) + (mode_info->atom_context->bios + + data_offset + + le16_to_cpu(power_info->info_4.usClockInfoArrayOffset) + + (power_state->ucClockStateIndices[j] * + power_info->info_4.ucClockInfoSize)); + sclk = le16_to_cpu(clock_info->usEngineClockLow); + sclk |= clock_info->ucEngineClockHigh << 16; + mclk = le16_to_cpu(clock_info->usMemoryClockLow); + mclk |= clock_info->ucMemoryClockHigh << 16; + rdev->pm.power_state[state_index].clock_info[mode_index].mclk = mclk; + rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk; + /* skip invalid modes */ + if ((rdev->pm.power_state[state_index].clock_info[mode_index].mclk == 0) || + (rdev->pm.power_state[state_index].clock_info[mode_index].sclk == 0)) + continue; + /* skip overclock modes for now */ + if ((rdev->pm.power_state[state_index].clock_info[mode_index].mclk > + rdev->clock.default_mclk + RADEON_MODE_OVERCLOCK_MARGIN) || + (rdev->pm.power_state[state_index].clock_info[mode_index].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN)) + continue; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = + VOLTAGE_SW; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = + clock_info->usVDDC; + mode_index++; + } + } + rdev->pm.power_state[state_index].num_clock_modes = mode_index; + if (mode_index) { + misc = le32_to_cpu(non_clock_info->ulCapsAndSettings); + misc2 = le16_to_cpu(non_clock_info->usClassification); + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = + ((misc & ATOM_PPLIB_PCIE_LINK_WIDTH_MASK) >> + ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT) + 1; + switch (misc2 & ATOM_PPLIB_CLASSIFICATION_UI_MASK) { + case ATOM_PPLIB_CLASSIFICATION_UI_BATTERY: + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + break; + case ATOM_PPLIB_CLASSIFICATION_UI_BALANCED: + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BALANCED; + break; + case ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE: + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_PERFORMANCE; + break; + } + if (misc2 & ATOM_PPLIB_CLASSIFICATION_BOOT) { + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + rdev->pm.power_state[state_index].default_clock_mode = + &rdev->pm.power_state[state_index].clock_info[mode_index - 1]; + } + state_index++; + } + } + } + } else { + /* XXX figure out some good default low power mode for cards w/out power tables */ + } + + if (rdev->pm.default_power_state == NULL) { + /* add the default mode */ + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = rdev->clock.default_mclk; + rdev->pm.power_state[state_index].clock_info[0].sclk = rdev->clock.default_sclk; + rdev->pm.power_state[state_index].default_clock_mode = + &rdev->pm.power_state[state_index].clock_info[0]; + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; + if (rdev->asic->get_pcie_lanes) + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = radeon_get_pcie_lanes(rdev); + else + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = 16; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + state_index++; + } + rdev->pm.num_power_states = state_index; + + rdev->pm.current_power_state = rdev->pm.default_power_state; + rdev->pm.current_clock_mode = + rdev->pm.default_power_state->default_clock_mode; } -void radeon_atom_static_pwrmgt_setup(struct radeon_device *rdev, int enable) +void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable) { - ENABLE_ASIC_STATIC_PWR_MGT_PS_ALLOCATION args; - int index = GetIndexIntoMasterTable(COMMAND, EnableASIC_StaticPwrMgt); + DYNAMIC_CLOCK_GATING_PS_ALLOCATION args; + int index = GetIndexIntoMasterTable(COMMAND, DynamicClockGating); args.ucEnable = enable; diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c new file mode 100644 index 00000000000..3f557c4151e --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c @@ -0,0 +1,257 @@ +/* + * Copyright (c) 2010 Red Hat Inc. + * Author : Dave Airlie <airlied@redhat.com> + * + * Licensed under GPLv2 + * + * ATPX support for both Intel/ATI + */ +#include <linux/vga_switcheroo.h> +#include <acpi/acpi.h> +#include <acpi/acpi_bus.h> +#include <linux/pci.h> + +#define ATPX_VERSION 0 +#define ATPX_GPU_PWR 2 +#define ATPX_MUX_SELECT 3 + +#define ATPX_INTEGRATED 0 +#define ATPX_DISCRETE 1 + +#define ATPX_MUX_IGD 0 +#define ATPX_MUX_DISCRETE 1 + +static struct radeon_atpx_priv { + bool atpx_detected; + /* handle for device - and atpx */ + acpi_handle dhandle; + acpi_handle atpx_handle; + acpi_handle atrm_handle; +} radeon_atpx_priv; + +/* retrieve the ROM in 4k blocks */ +static int radeon_atrm_call(acpi_handle atrm_handle, uint8_t *bios, + int offset, int len) +{ + acpi_status status; + union acpi_object atrm_arg_elements[2], *obj; + struct acpi_object_list atrm_arg; + struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL}; + + atrm_arg.count = 2; + atrm_arg.pointer = &atrm_arg_elements[0]; + + atrm_arg_elements[0].type = ACPI_TYPE_INTEGER; + atrm_arg_elements[0].integer.value = offset; + + atrm_arg_elements[1].type = ACPI_TYPE_INTEGER; + atrm_arg_elements[1].integer.value = len; + + status = acpi_evaluate_object(atrm_handle, NULL, &atrm_arg, &buffer); + if (ACPI_FAILURE(status)) { + printk("failed to evaluate ATRM got %s\n", acpi_format_exception(status)); + return -ENODEV; + } + + obj = (union acpi_object *)buffer.pointer; + memcpy(bios+offset, obj->buffer.pointer, len); + kfree(buffer.pointer); + return len; +} + +bool radeon_atrm_supported(struct pci_dev *pdev) +{ + /* get the discrete ROM only via ATRM */ + if (!radeon_atpx_priv.atpx_detected) + return false; + + if (radeon_atpx_priv.dhandle == DEVICE_ACPI_HANDLE(&pdev->dev)) + return false; + return true; +} + + +int radeon_atrm_get_bios_chunk(uint8_t *bios, int offset, int len) +{ + return radeon_atrm_call(radeon_atpx_priv.atrm_handle, bios, offset, len); +} + +static int radeon_atpx_get_version(acpi_handle handle) +{ + acpi_status status; + union acpi_object atpx_arg_elements[2], *obj; + struct acpi_object_list atpx_arg; + struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; + + atpx_arg.count = 2; + atpx_arg.pointer = &atpx_arg_elements[0]; + + atpx_arg_elements[0].type = ACPI_TYPE_INTEGER; + atpx_arg_elements[0].integer.value = ATPX_VERSION; + + atpx_arg_elements[1].type = ACPI_TYPE_INTEGER; + atpx_arg_elements[1].integer.value = ATPX_VERSION; + + status = acpi_evaluate_object(handle, NULL, &atpx_arg, &buffer); + if (ACPI_FAILURE(status)) { + printk("%s: failed to call ATPX: %s\n", __func__, acpi_format_exception(status)); + return -ENOSYS; + } + obj = (union acpi_object *)buffer.pointer; + if (obj && (obj->type == ACPI_TYPE_BUFFER)) + printk(KERN_INFO "radeon atpx: version is %d\n", *((u8 *)(obj->buffer.pointer) + 2)); + kfree(buffer.pointer); + return 0; +} + +static int radeon_atpx_execute(acpi_handle handle, int cmd_id, u16 value) +{ + acpi_status status; + union acpi_object atpx_arg_elements[2]; + struct acpi_object_list atpx_arg; + struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL }; + uint8_t buf[4] = {0}; + + if (!handle) + return -EINVAL; + + atpx_arg.count = 2; + atpx_arg.pointer = &atpx_arg_elements[0]; + + atpx_arg_elements[0].type = ACPI_TYPE_INTEGER; + atpx_arg_elements[0].integer.value = cmd_id; + + buf[2] = value & 0xff; + buf[3] = (value >> 8) & 0xff; + + atpx_arg_elements[1].type = ACPI_TYPE_BUFFER; + atpx_arg_elements[1].buffer.length = 4; + atpx_arg_elements[1].buffer.pointer = buf; + + status = acpi_evaluate_object(handle, NULL, &atpx_arg, &buffer); + if (ACPI_FAILURE(status)) { + printk("%s: failed to call ATPX: %s\n", __func__, acpi_format_exception(status)); + return -ENOSYS; + } + kfree(buffer.pointer); + + return 0; +} + +static int radeon_atpx_set_discrete_state(acpi_handle handle, int state) +{ + return radeon_atpx_execute(handle, ATPX_GPU_PWR, state); +} + +static int radeon_atpx_switch_mux(acpi_handle handle, int mux_id) +{ + return radeon_atpx_execute(handle, ATPX_MUX_SELECT, mux_id); +} + + +static int radeon_atpx_switchto(enum vga_switcheroo_client_id id) +{ + if (id == VGA_SWITCHEROO_IGD) + radeon_atpx_switch_mux(radeon_atpx_priv.atpx_handle, 0); + else + radeon_atpx_switch_mux(radeon_atpx_priv.atpx_handle, 1); + return 0; +} + +static int radeon_atpx_power_state(enum vga_switcheroo_client_id id, + enum vga_switcheroo_state state) +{ + /* on w500 ACPI can't change intel gpu state */ + if (id == VGA_SWITCHEROO_IGD) + return 0; + + radeon_atpx_set_discrete_state(radeon_atpx_priv.atpx_handle, state); + return 0; +} + +static bool radeon_atpx_pci_probe_handle(struct pci_dev *pdev) +{ + acpi_handle dhandle, atpx_handle, atrm_handle; + acpi_status status; + + dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); + if (!dhandle) + return false; + + status = acpi_get_handle(dhandle, "ATPX", &atpx_handle); + if (ACPI_FAILURE(status)) + return false; + + status = acpi_get_handle(dhandle, "ATRM", &atrm_handle); + if (ACPI_FAILURE(status)) + return false; + + radeon_atpx_priv.dhandle = dhandle; + radeon_atpx_priv.atpx_handle = atpx_handle; + radeon_atpx_priv.atrm_handle = atrm_handle; + return true; +} + +static int radeon_atpx_init(void) +{ + /* set up the ATPX handle */ + + radeon_atpx_get_version(radeon_atpx_priv.atpx_handle); + return 0; +} + +static int radeon_atpx_get_client_id(struct pci_dev *pdev) +{ + if (radeon_atpx_priv.dhandle == DEVICE_ACPI_HANDLE(&pdev->dev)) + return VGA_SWITCHEROO_IGD; + else + return VGA_SWITCHEROO_DIS; +} + +static struct vga_switcheroo_handler radeon_atpx_handler = { + .switchto = radeon_atpx_switchto, + .power_state = radeon_atpx_power_state, + .init = radeon_atpx_init, + .get_client_id = radeon_atpx_get_client_id, +}; + +static bool radeon_atpx_detect(void) +{ + char acpi_method_name[255] = { 0 }; + struct acpi_buffer buffer = {sizeof(acpi_method_name), acpi_method_name}; + struct pci_dev *pdev = NULL; + bool has_atpx = false; + int vga_count = 0; + + while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) { + vga_count++; + + has_atpx |= (radeon_atpx_pci_probe_handle(pdev) == true); + } + + if (has_atpx && vga_count == 2) { + acpi_get_name(radeon_atpx_priv.atpx_handle, ACPI_FULL_PATHNAME, &buffer); + printk(KERN_INFO "VGA switcheroo: detected switching method %s handle\n", + acpi_method_name); + radeon_atpx_priv.atpx_detected = true; + return true; + } + return false; +} + +void radeon_register_atpx_handler(void) +{ + bool r; + + /* detect if we have any ATPX + 2 VGA in the system */ + r = radeon_atpx_detect(); + if (!r) + return; + + vga_switcheroo_register_handler(&radeon_atpx_handler); +} + +void radeon_unregister_atpx_handler(void) +{ + vga_switcheroo_unregister_handler(); +} diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c index 4ddfd4b5bc5..7932dc4d6b9 100644 --- a/drivers/gpu/drm/radeon/radeon_benchmark.c +++ b/drivers/gpu/drm/radeon/radeon_benchmark.c @@ -65,31 +65,42 @@ void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize, if (r) { goto out_cleanup; } - start_jiffies = jiffies; - for (i = 0; i < n; i++) { - r = radeon_fence_create(rdev, &fence); - if (r) { - goto out_cleanup; + + /* r100 doesn't have dma engine so skip the test */ + if (rdev->asic->copy_dma) { + + start_jiffies = jiffies; + for (i = 0; i < n; i++) { + r = radeon_fence_create(rdev, &fence); + if (r) { + goto out_cleanup; + } + + r = radeon_copy_dma(rdev, saddr, daddr, + size / RADEON_GPU_PAGE_SIZE, fence); + + if (r) { + goto out_cleanup; + } + r = radeon_fence_wait(fence, false); + if (r) { + goto out_cleanup; + } + radeon_fence_unref(&fence); } - r = radeon_copy_dma(rdev, saddr, daddr, size / RADEON_GPU_PAGE_SIZE, fence); - if (r) { - goto out_cleanup; + end_jiffies = jiffies; + time = end_jiffies - start_jiffies; + time = jiffies_to_msecs(time); + if (time > 0) { + i = ((n * size) >> 10) / time; + printk(KERN_INFO "radeon: dma %u bo moves of %ukb from" + " %d to %d in %lums (%ukb/ms %ukb/s %uM/s)\n", + n, size >> 10, + sdomain, ddomain, time, + i, i * 1000, (i * 1000) / 1024); } - r = radeon_fence_wait(fence, false); - if (r) { - goto out_cleanup; - } - radeon_fence_unref(&fence); - } - end_jiffies = jiffies; - time = end_jiffies - start_jiffies; - time = jiffies_to_msecs(time); - if (time > 0) { - i = ((n * size) >> 10) / time; - printk(KERN_INFO "radeon: dma %u bo moves of %ukb from %d to %d" - " in %lums (%ukb/ms %ukb/s %uM/s)\n", n, size >> 10, - sdomain, ddomain, time, i, i * 1000, (i * 1000) / 1024); } + start_jiffies = jiffies; for (i = 0; i < n; i++) { r = radeon_fence_create(rdev, &fence); diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c index 906921740c6..55724046052 100644 --- a/drivers/gpu/drm/radeon/radeon_bios.c +++ b/drivers/gpu/drm/radeon/radeon_bios.c @@ -30,6 +30,7 @@ #include "radeon.h" #include "atom.h" +#include <linux/vga_switcheroo.h> /* * BIOS. */ @@ -62,7 +63,7 @@ static bool igp_read_bios_from_vram(struct radeon_device *rdev) iounmap(bios); return false; } - memcpy(rdev->bios, bios, size); + memcpy_fromio(rdev->bios, bios, size); iounmap(bios); return true; } @@ -93,6 +94,38 @@ static bool radeon_read_bios(struct radeon_device *rdev) return true; } +/* ATRM is used to get the BIOS on the discrete cards in + * dual-gpu systems. + */ +static bool radeon_atrm_get_bios(struct radeon_device *rdev) +{ + int ret; + int size = 64 * 1024; + int i; + + if (!radeon_atrm_supported(rdev->pdev)) + return false; + + rdev->bios = kmalloc(size, GFP_KERNEL); + if (!rdev->bios) { + DRM_ERROR("Unable to allocate bios\n"); + return false; + } + + for (i = 0; i < size / ATRM_BIOS_PAGE; i++) { + ret = radeon_atrm_get_bios_chunk(rdev->bios, + (i * ATRM_BIOS_PAGE), + ATRM_BIOS_PAGE); + if (ret <= 0) + break; + } + + if (i == 0 || rdev->bios[0] != 0x55 || rdev->bios[1] != 0xaa) { + kfree(rdev->bios); + return false; + } + return true; +} static bool r700_read_disabled_bios(struct radeon_device *rdev) { uint32_t viph_control; @@ -388,16 +421,16 @@ static bool radeon_read_disabled_bios(struct radeon_device *rdev) return legacy_read_disabled_bios(rdev); } + bool radeon_get_bios(struct radeon_device *rdev) { bool r; uint16_t tmp; - if (rdev->flags & RADEON_IS_IGP) { + r = radeon_atrm_get_bios(rdev); + if (r == false) r = igp_read_bios_from_vram(rdev); - if (r == false) - r = radeon_read_bios(rdev); - } else + if (r == false) r = radeon_read_bios(rdev); if (r == false) { r = radeon_read_disabled_bios(rdev); @@ -408,6 +441,13 @@ bool radeon_get_bios(struct radeon_device *rdev) return false; } if (rdev->bios[0] != 0x55 || rdev->bios[1] != 0xaa) { + printk("BIOS signature incorrect %x %x\n", rdev->bios[0], rdev->bios[1]); + goto free_bios; + } + + tmp = RBIOS16(0x18); + if (RBIOS8(tmp + 0x14) != 0x0) { + DRM_INFO("Not an x86 BIOS ROM, not using.\n"); goto free_bios; } diff --git a/drivers/gpu/drm/radeon/radeon_clocks.c b/drivers/gpu/drm/radeon/radeon_clocks.c index 812f24dbc2a..f64936cc4dd 100644 --- a/drivers/gpu/drm/radeon/radeon_clocks.c +++ b/drivers/gpu/drm/radeon/radeon_clocks.c @@ -56,7 +56,7 @@ uint32_t radeon_legacy_get_engine_clock(struct radeon_device *rdev) else if (post_div == 3) sclk >>= 2; else if (post_div == 4) - sclk >>= 4; + sclk >>= 3; return sclk; } @@ -86,7 +86,7 @@ uint32_t radeon_legacy_get_memory_clock(struct radeon_device *rdev) else if (post_div == 3) mclk >>= 2; else if (post_div == 4) - mclk >>= 4; + mclk >>= 3; return mclk; } @@ -96,6 +96,7 @@ void radeon_get_clock_info(struct drm_device *dev) struct radeon_device *rdev = dev->dev_private; struct radeon_pll *p1pll = &rdev->clock.p1pll; struct radeon_pll *p2pll = &rdev->clock.p2pll; + struct radeon_pll *dcpll = &rdev->clock.dcpll; struct radeon_pll *spll = &rdev->clock.spll; struct radeon_pll *mpll = &rdev->clock.mpll; int ret; @@ -204,6 +205,17 @@ void radeon_get_clock_info(struct drm_device *dev) p2pll->max_frac_feedback_div = 0; } + /* dcpll is DCE4 only */ + dcpll->min_post_div = 2; + dcpll->max_post_div = 0x7f; + dcpll->min_frac_feedback_div = 0; + dcpll->max_frac_feedback_div = 9; + dcpll->min_ref_div = 2; + dcpll->max_ref_div = 0x3ff; + dcpll->min_feedback_div = 4; + dcpll->max_feedback_div = 0xfff; + dcpll->best_vco = 0; + p1pll->min_ref_div = 2; p1pll->max_ref_div = 0x3ff; p1pll->min_feedback_div = 4; @@ -846,8 +858,10 @@ int radeon_static_clocks_init(struct drm_device *dev) /* XXX make sure engine is idle */ if (radeon_dynclks != -1) { - if (radeon_dynclks) - radeon_set_clock_gating(rdev, 1); + if (radeon_dynclks) { + if (rdev->asic->set_clock_gating) + radeon_set_clock_gating(rdev, 1); + } } radeon_apply_clock_quirks(rdev); return 0; diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index 579c8920e08..e9ea38ece37 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -150,6 +150,9 @@ static uint16_t combios_get_table_offset(struct drm_device *dev, int rev; uint16_t offset = 0, check_offset; + if (!rdev->bios) + return 0; + switch (table) { /* absolute offset tables */ case COMBIOS_ASIC_INIT_1_TABLE: @@ -443,6 +446,39 @@ static uint16_t combios_get_table_offset(struct drm_device *dev, } +bool radeon_combios_check_hardcoded_edid(struct radeon_device *rdev) +{ + int edid_info; + struct edid *edid; + edid_info = combios_get_table_offset(rdev->ddev, COMBIOS_HARDCODED_EDID_TABLE); + if (!edid_info) + return false; + + edid = kmalloc(EDID_LENGTH * (DRM_MAX_EDID_EXT_NUM + 1), + GFP_KERNEL); + if (edid == NULL) + return false; + + memcpy((unsigned char *)edid, + (unsigned char *)(rdev->bios + edid_info), EDID_LENGTH); + + if (!drm_edid_is_valid(edid)) { + kfree(edid); + return false; + } + + rdev->mode_info.bios_hardcoded_edid = edid; + return true; +} + +struct edid * +radeon_combios_get_hardcoded_edid(struct radeon_device *rdev) +{ + if (rdev->mode_info.bios_hardcoded_edid) + return rdev->mode_info.bios_hardcoded_edid; + return NULL; +} + static struct radeon_i2c_bus_rec combios_setup_i2c_bus(struct radeon_device *rdev, int ddc_line) { @@ -486,9 +522,65 @@ static struct radeon_i2c_bus_rec combios_setup_i2c_bus(struct radeon_device *rde i2c.y_data_reg = ddc_line; } - if (rdev->family < CHIP_R200) - i2c.hw_capable = false; - else { + switch (rdev->family) { + case CHIP_R100: + case CHIP_RV100: + case CHIP_RS100: + case CHIP_RV200: + case CHIP_RS200: + case CHIP_RS300: + switch (ddc_line) { + case RADEON_GPIO_DVI_DDC: + /* in theory this should be hw capable, + * but it doesn't seem to work + */ + i2c.hw_capable = false; + break; + default: + i2c.hw_capable = false; + break; + } + break; + case CHIP_R200: + switch (ddc_line) { + case RADEON_GPIO_DVI_DDC: + case RADEON_GPIO_MONID: + i2c.hw_capable = true; + break; + default: + i2c.hw_capable = false; + break; + } + break; + case CHIP_RV250: + case CHIP_RV280: + switch (ddc_line) { + case RADEON_GPIO_VGA_DDC: + case RADEON_GPIO_DVI_DDC: + case RADEON_GPIO_CRT2_DDC: + i2c.hw_capable = true; + break; + default: + i2c.hw_capable = false; + break; + } + break; + case CHIP_R300: + case CHIP_R350: + switch (ddc_line) { + case RADEON_GPIO_VGA_DDC: + case RADEON_GPIO_DVI_DDC: + i2c.hw_capable = true; + break; + default: + i2c.hw_capable = false; + break; + } + break; + case CHIP_RV350: + case CHIP_RV380: + case CHIP_RS400: + case CHIP_RS480: switch (ddc_line) { case RADEON_GPIO_VGA_DDC: case RADEON_GPIO_DVI_DDC: @@ -504,9 +596,14 @@ static struct radeon_i2c_bus_rec combios_setup_i2c_bus(struct radeon_device *rde i2c.hw_capable = false; break; } + break; + default: + i2c.hw_capable = false; + break; } i2c.mm_i2c = false; i2c.i2c_id = 0; + i2c.hpd_id = 0; if (ddc_line) i2c.valid = true; @@ -527,9 +624,6 @@ bool radeon_combios_get_clock_info(struct drm_device *dev) int8_t rev; uint16_t sclk, mclk; - if (rdev->bios == NULL) - return false; - pll_info = combios_get_table_offset(dev, COMBIOS_PLL_INFO_TABLE); if (pll_info) { rev = RBIOS8(pll_info); @@ -654,9 +748,6 @@ struct radeon_encoder_primary_dac *radeon_combios_get_primary_dac_info(struct if (!p_dac) return NULL; - if (rdev->bios == NULL) - goto out; - /* check CRT table */ dac_info = combios_get_table_offset(dev, COMBIOS_CRT_INFO_TABLE); if (dac_info) { @@ -673,7 +764,6 @@ struct radeon_encoder_primary_dac *radeon_combios_get_primary_dac_info(struct found = 1; } -out: if (!found) /* fallback to defaults */ radeon_legacy_get_primary_dac_info_from_table(rdev, p_dac); @@ -687,9 +777,6 @@ radeon_combios_get_tv_info(struct radeon_device *rdev) uint16_t tv_info; enum radeon_tv_std tv_std = TV_STD_NTSC; - if (rdev->bios == NULL) - return tv_std; - tv_info = combios_get_table_offset(dev, COMBIOS_TV_INFO_TABLE); if (tv_info) { if (RBIOS8(tv_info + 6) == 'T') { @@ -793,9 +880,6 @@ struct radeon_encoder_tv_dac *radeon_combios_get_tv_dac_info(struct if (!tv_dac) return NULL; - if (rdev->bios == NULL) - goto out; - /* first check TV table */ dac_info = combios_get_table_offset(dev, COMBIOS_TV_INFO_TABLE); if (dac_info) { @@ -857,7 +941,6 @@ struct radeon_encoder_tv_dac *radeon_combios_get_tv_dac_info(struct } } -out: if (!found) /* fallback to defaults */ radeon_legacy_get_tv_dac_info_from_table(rdev, tv_dac); @@ -945,11 +1028,6 @@ struct radeon_encoder_lvds *radeon_combios_get_lvds_info(struct radeon_encoder int tmp, i; struct radeon_encoder_lvds *lvds = NULL; - if (rdev->bios == NULL) { - lvds = radeon_legacy_get_lvds_info_from_regs(rdev); - goto out; - } - lcd_info = combios_get_table_offset(dev, COMBIOS_LCD_INFO_TABLE); if (lcd_info) { @@ -971,8 +1049,7 @@ struct radeon_encoder_lvds *radeon_combios_get_lvds_info(struct radeon_encoder lvds->native_mode.vdisplay); lvds->panel_vcc_delay = RBIOS16(lcd_info + 0x2c); - if (lvds->panel_vcc_delay > 2000 || lvds->panel_vcc_delay < 0) - lvds->panel_vcc_delay = 2000; + lvds->panel_vcc_delay = min_t(u16, lvds->panel_vcc_delay, 2000); lvds->panel_pwr_delay = RBIOS8(lcd_info + 0x24); lvds->panel_digon_delay = RBIOS16(lcd_info + 0x38) & 0xf; @@ -1051,7 +1128,7 @@ struct radeon_encoder_lvds *radeon_combios_get_lvds_info(struct radeon_encoder DRM_INFO("No panel info found in BIOS\n"); lvds = radeon_legacy_get_lvds_info_from_regs(rdev); } -out: + if (lvds) encoder->native_mode = lvds->native_mode; return lvds; @@ -1103,9 +1180,6 @@ bool radeon_legacy_get_tmds_info_from_combios(struct radeon_encoder *encoder, int i, n; uint8_t ver; - if (rdev->bios == NULL) - return false; - tmds_info = combios_get_table_offset(dev, COMBIOS_DFP_INFO_TABLE); if (tmds_info) { @@ -1185,9 +1259,6 @@ bool radeon_legacy_get_ext_tmds_info_from_combios(struct radeon_encoder *encoder enum radeon_combios_ddc gpio; struct radeon_i2c_bus_rec i2c_bus; - if (rdev->bios == NULL) - return false; - tmds->i2c_bus = NULL; if (rdev->flags & RADEON_IS_IGP) { offset = combios_get_table_offset(dev, COMBIOS_I2C_INFO_TABLE); @@ -1254,7 +1325,10 @@ bool radeon_legacy_get_ext_tmds_info_from_combios(struct radeon_encoder *encoder tmds->i2c_bus = radeon_i2c_create(dev, &i2c_bus, "DVO"); break; case DDC_LCD: /* MM i2c */ - DRM_ERROR("MM i2c requires hw i2c engine\n"); + i2c_bus.valid = true; + i2c_bus.hw_capable = true; + i2c_bus.mm_i2c = true; + tmds->i2c_bus = radeon_i2c_create(dev, &i2c_bus, "DVO"); break; default: DRM_ERROR("Unsupported gpio %d\n", gpio); @@ -1280,47 +1354,47 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) rdev->mode_info.connector_table = radeon_connector_table; if (rdev->mode_info.connector_table == CT_NONE) { #ifdef CONFIG_PPC_PMAC - if (machine_is_compatible("PowerBook3,3")) { + if (of_machine_is_compatible("PowerBook3,3")) { /* powerbook with VGA */ rdev->mode_info.connector_table = CT_POWERBOOK_VGA; - } else if (machine_is_compatible("PowerBook3,4") || - machine_is_compatible("PowerBook3,5")) { + } else if (of_machine_is_compatible("PowerBook3,4") || + of_machine_is_compatible("PowerBook3,5")) { /* powerbook with internal tmds */ rdev->mode_info.connector_table = CT_POWERBOOK_INTERNAL; - } else if (machine_is_compatible("PowerBook5,1") || - machine_is_compatible("PowerBook5,2") || - machine_is_compatible("PowerBook5,3") || - machine_is_compatible("PowerBook5,4") || - machine_is_compatible("PowerBook5,5")) { + } else if (of_machine_is_compatible("PowerBook5,1") || + of_machine_is_compatible("PowerBook5,2") || + of_machine_is_compatible("PowerBook5,3") || + of_machine_is_compatible("PowerBook5,4") || + of_machine_is_compatible("PowerBook5,5")) { /* powerbook with external single link tmds (sil164) */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook5,6")) { + } else if (of_machine_is_compatible("PowerBook5,6")) { /* powerbook with external dual or single link tmds */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook5,7") || - machine_is_compatible("PowerBook5,8") || - machine_is_compatible("PowerBook5,9")) { + } else if (of_machine_is_compatible("PowerBook5,7") || + of_machine_is_compatible("PowerBook5,8") || + of_machine_is_compatible("PowerBook5,9")) { /* PowerBook6,2 ? */ /* powerbook with external dual link tmds (sil1178?) */ rdev->mode_info.connector_table = CT_POWERBOOK_EXTERNAL; - } else if (machine_is_compatible("PowerBook4,1") || - machine_is_compatible("PowerBook4,2") || - machine_is_compatible("PowerBook4,3") || - machine_is_compatible("PowerBook6,3") || - machine_is_compatible("PowerBook6,5") || - machine_is_compatible("PowerBook6,7")) { + } else if (of_machine_is_compatible("PowerBook4,1") || + of_machine_is_compatible("PowerBook4,2") || + of_machine_is_compatible("PowerBook4,3") || + of_machine_is_compatible("PowerBook6,3") || + of_machine_is_compatible("PowerBook6,5") || + of_machine_is_compatible("PowerBook6,7")) { /* ibook */ rdev->mode_info.connector_table = CT_IBOOK; - } else if (machine_is_compatible("PowerMac4,4")) { + } else if (of_machine_is_compatible("PowerMac4,4")) { /* emac */ rdev->mode_info.connector_table = CT_EMAC; - } else if (machine_is_compatible("PowerMac10,1")) { + } else if (of_machine_is_compatible("PowerMac10,1")) { /* mini with internal tmds */ rdev->mode_info.connector_table = CT_MINI_INTERNAL; - } else if (machine_is_compatible("PowerMac10,2")) { + } else if (of_machine_is_compatible("PowerMac10,2")) { /* mini with external tmds */ rdev->mode_info.connector_table = CT_MINI_EXTERNAL; - } else if (machine_is_compatible("PowerMac12,1")) { + } else if (of_machine_is_compatible("PowerMac12,1")) { /* PowerMac8,1 ? */ /* imac g5 isight */ rdev->mode_info.connector_table = CT_IMAC_G5_ISIGHT; @@ -1910,9 +1984,6 @@ bool radeon_get_legacy_connector_info_from_bios(struct drm_device *dev) struct radeon_i2c_bus_rec ddc_i2c; struct radeon_hpd hpd; - if (rdev->bios == NULL) - return false; - conn_info = combios_get_table_offset(dev, COMBIOS_CONNECTOR_INFO_TABLE); if (conn_info) { for (i = 0; i < 4; i++) { @@ -2279,6 +2350,115 @@ bool radeon_get_legacy_connector_info_from_bios(struct drm_device *dev) return true; } +void radeon_combios_get_power_modes(struct radeon_device *rdev) +{ + struct drm_device *dev = rdev->ddev; + u16 offset, misc, misc2 = 0; + u8 rev, blocks, tmp; + int state_index = 0; + + rdev->pm.default_power_state = NULL; + + if (rdev->flags & RADEON_IS_MOBILITY) { + offset = combios_get_table_offset(dev, COMBIOS_POWERPLAY_INFO_TABLE); + if (offset) { + rev = RBIOS8(offset); + blocks = RBIOS8(offset + 0x2); + /* power mode 0 tends to be the only valid one */ + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = RBIOS32(offset + 0x5 + 0x2); + rdev->pm.power_state[state_index].clock_info[0].sclk = RBIOS32(offset + 0x5 + 0x6); + if ((rdev->pm.power_state[state_index].clock_info[0].mclk == 0) || + (rdev->pm.power_state[state_index].clock_info[0].sclk == 0)) + goto default_mode; + /* skip overclock modes for now */ + if ((rdev->pm.power_state[state_index].clock_info[0].mclk > + rdev->clock.default_mclk + RADEON_MODE_OVERCLOCK_MARGIN) || + (rdev->pm.power_state[state_index].clock_info[0].sclk > + rdev->clock.default_sclk + RADEON_MODE_OVERCLOCK_MARGIN)) + goto default_mode; + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_BATTERY; + misc = RBIOS16(offset + 0x5 + 0x0); + if (rev > 4) + misc2 = RBIOS16(offset + 0x5 + 0xe); + if (misc & 0x4) { + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_GPIO; + if (misc & 0x8) + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + true; + else + rdev->pm.power_state[state_index].clock_info[0].voltage.active_high = + false; + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.valid = true; + if (rev < 6) { + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.reg = + RBIOS16(offset + 0x5 + 0xb) * 4; + tmp = RBIOS8(offset + 0x5 + 0xd); + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.mask = (1 << tmp); + } else { + u8 entries = RBIOS8(offset + 0x5 + 0xb); + u16 voltage_table_offset = RBIOS16(offset + 0x5 + 0xc); + if (entries && voltage_table_offset) { + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.reg = + RBIOS16(voltage_table_offset) * 4; + tmp = RBIOS8(voltage_table_offset + 0x2); + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.mask = (1 << tmp); + } else + rdev->pm.power_state[state_index].clock_info[0].voltage.gpio.valid = false; + } + switch ((misc2 & 0x700) >> 8) { + case 0: + default: + rdev->pm.power_state[state_index].clock_info[0].voltage.delay = 0; + break; + case 1: + rdev->pm.power_state[state_index].clock_info[0].voltage.delay = 33; + break; + case 2: + rdev->pm.power_state[state_index].clock_info[0].voltage.delay = 66; + break; + case 3: + rdev->pm.power_state[state_index].clock_info[0].voltage.delay = 99; + break; + case 4: + rdev->pm.power_state[state_index].clock_info[0].voltage.delay = 132; + break; + } + } else + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; + if (rev > 6) + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = + RBIOS8(offset + 0x5 + 0x10); + state_index++; + } else { + /* XXX figure out some good default low power mode for mobility cards w/out power tables */ + } + } else { + /* XXX figure out some good default low power mode for desktop cards */ + } + +default_mode: + /* add the default mode */ + rdev->pm.power_state[state_index].type = + POWER_STATE_TYPE_DEFAULT; + rdev->pm.power_state[state_index].num_clock_modes = 1; + rdev->pm.power_state[state_index].clock_info[0].mclk = rdev->clock.default_mclk; + rdev->pm.power_state[state_index].clock_info[0].sclk = rdev->clock.default_sclk; + rdev->pm.power_state[state_index].default_clock_mode = &rdev->pm.power_state[state_index].clock_info[0]; + rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE; + if (rdev->asic->get_pcie_lanes) + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = radeon_get_pcie_lanes(rdev); + else + rdev->pm.power_state[state_index].non_clock_info.pcie_lanes = 16; + rdev->pm.default_power_state = &rdev->pm.power_state[state_index]; + rdev->pm.num_power_states = state_index + 1; + + rdev->pm.current_power_state = rdev->pm.default_power_state; + rdev->pm.current_clock_mode = + rdev->pm.default_power_state->default_clock_mode; +} + void radeon_external_tmds_setup(struct drm_encoder *encoder) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); @@ -2290,23 +2470,21 @@ void radeon_external_tmds_setup(struct drm_encoder *encoder) switch (tmds->dvo_chip) { case DVO_SIL164: /* sil 164 */ - radeon_i2c_do_lock(tmds->i2c_bus, 1); - radeon_i2c_sw_put_byte(tmds->i2c_bus, - tmds->slave_addr, - 0x08, 0x30); - radeon_i2c_sw_put_byte(tmds->i2c_bus, + radeon_i2c_put_byte(tmds->i2c_bus, + tmds->slave_addr, + 0x08, 0x30); + radeon_i2c_put_byte(tmds->i2c_bus, tmds->slave_addr, 0x09, 0x00); - radeon_i2c_sw_put_byte(tmds->i2c_bus, - tmds->slave_addr, - 0x0a, 0x90); - radeon_i2c_sw_put_byte(tmds->i2c_bus, - tmds->slave_addr, - 0x0c, 0x89); - radeon_i2c_sw_put_byte(tmds->i2c_bus, + radeon_i2c_put_byte(tmds->i2c_bus, + tmds->slave_addr, + 0x0a, 0x90); + radeon_i2c_put_byte(tmds->i2c_bus, + tmds->slave_addr, + 0x0c, 0x89); + radeon_i2c_put_byte(tmds->i2c_bus, tmds->slave_addr, 0x08, 0x3b); - radeon_i2c_do_lock(tmds->i2c_bus, 0); break; case DVO_SIL1178: /* sil 1178 - untested */ @@ -2339,9 +2517,6 @@ bool radeon_combios_external_tmds_setup(struct drm_encoder *encoder) uint32_t reg, val, and_mask, or_mask; struct radeon_encoder_ext_tmds *tmds = radeon_encoder->enc_priv; - if (rdev->bios == NULL) - return false; - if (!tmds) return false; @@ -2391,11 +2566,9 @@ bool radeon_combios_external_tmds_setup(struct drm_encoder *encoder) index++; val = RBIOS8(index); index++; - radeon_i2c_do_lock(tmds->i2c_bus, 1); - radeon_i2c_sw_put_byte(tmds->i2c_bus, - slave_addr, - reg, val); - radeon_i2c_do_lock(tmds->i2c_bus, 0); + radeon_i2c_put_byte(tmds->i2c_bus, + slave_addr, + reg, val); break; default: DRM_ERROR("Unknown id %d\n", id >> 13); @@ -2448,11 +2621,9 @@ bool radeon_combios_external_tmds_setup(struct drm_encoder *encoder) reg = id & 0x1fff; val = RBIOS8(index); index += 1; - radeon_i2c_do_lock(tmds->i2c_bus, 1); - radeon_i2c_sw_put_byte(tmds->i2c_bus, - tmds->slave_addr, - reg, val); - radeon_i2c_do_lock(tmds->i2c_bus, 0); + radeon_i2c_put_byte(tmds->i2c_bus, + tmds->slave_addr, + reg, val); break; default: DRM_ERROR("Unknown id %d\n", id >> 13); diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 55266416fa4..ee0083f982d 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -479,10 +479,8 @@ static enum drm_connector_status radeon_lvds_detect(struct drm_connector *connec ret = connector_status_connected; else { if (radeon_connector->ddc_bus) { - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); if (radeon_connector->edid) ret = connector_status_connected; } @@ -580,24 +578,21 @@ static enum drm_connector_status radeon_vga_detect(struct drm_connector *connect struct radeon_connector *radeon_connector = to_radeon_connector(connector); struct drm_encoder *encoder; struct drm_encoder_helper_funcs *encoder_funcs; - bool dret; + bool dret = false; enum drm_connector_status ret = connector_status_disconnected; encoder = radeon_best_single_encoder(connector); if (!encoder) ret = connector_status_disconnected; - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); - dret = radeon_ddc_probe(radeon_connector); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); + if (radeon_connector->ddc_bus) + dret = radeon_ddc_probe(radeon_connector); if (dret) { if (radeon_connector->edid) { kfree(radeon_connector->edid); radeon_connector->edid = NULL; } - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); if (!radeon_connector->edid) { DRM_ERROR("%s: probed a monitor but no|invalid EDID\n", @@ -740,19 +735,16 @@ static enum drm_connector_status radeon_dvi_detect(struct drm_connector *connect struct drm_mode_object *obj; int i; enum drm_connector_status ret = connector_status_disconnected; - bool dret; + bool dret = false; - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); - dret = radeon_ddc_probe(radeon_connector); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); + if (radeon_connector->ddc_bus) + dret = radeon_ddc_probe(radeon_connector); if (dret) { if (radeon_connector->edid) { kfree(radeon_connector->edid); radeon_connector->edid = NULL; } - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); if (!radeon_connector->edid) { DRM_ERROR("%s: probed a monitor but no|invalid EDID\n", @@ -776,7 +768,7 @@ static enum drm_connector_status radeon_dvi_detect(struct drm_connector *connect * connected and the DVI port disconnected. If the edid doesn't * say HDMI, vice versa. */ - if (radeon_connector->shared_ddc && connector_status_connected) { + if (radeon_connector->shared_ddc && (ret == connector_status_connected)) { struct drm_device *dev = connector->dev; struct drm_connector *list_connector; struct radeon_connector *list_radeon_connector; @@ -948,7 +940,7 @@ static void radeon_dp_connector_destroy(struct drm_connector *connector) if (radeon_connector->edid) kfree(radeon_connector->edid); if (radeon_dig_connector->dp_i2c_bus) - radeon_i2c_destroy(radeon_dig_connector->dp_i2c_bus); + radeon_i2c_destroy_dp(radeon_dig_connector->dp_i2c_bus); kfree(radeon_connector->con_priv); drm_sysfs_connector_remove(connector); drm_connector_cleanup(connector); @@ -984,12 +976,10 @@ static enum drm_connector_status radeon_dp_detect(struct drm_connector *connecto ret = connector_status_connected; } } else { - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); if (radeon_ddc_probe(radeon_connector)) { radeon_dig_connector->dp_sink_type = sink_type; ret = connector_status_connected; } - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); } return ret; @@ -1056,8 +1046,7 @@ radeon_add_atom_connector(struct drm_device *dev, return; } if (radeon_connector->ddc_bus && i2c_bus->valid) { - if (memcmp(&radeon_connector->ddc_bus->rec, i2c_bus, - sizeof(struct radeon_i2c_bus_rec)) == 0) { + if (radeon_connector->ddc_bus->rec.i2c_id == i2c_bus->i2c_id) { radeon_connector->shared_ddc = true; shared_ddc = true; } @@ -1343,7 +1332,7 @@ radeon_add_legacy_connector(struct drm_device *dev, radeon_connector->dac_load_detect = false; drm_connector_attach_property(&radeon_connector->base, rdev->mode_info.load_detect_property, - 1); + radeon_connector->dac_load_detect); drm_connector_attach_property(&radeon_connector->base, rdev->mode_info.tv_std_property, radeon_combios_get_tv_info(rdev)); diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c index 06123ba31d3..dc6eba6b96d 100644 --- a/drivers/gpu/drm/radeon/radeon_cp.c +++ b/drivers/gpu/drm/radeon/radeon_cp.c @@ -1644,6 +1644,7 @@ static int radeon_do_resume_cp(struct drm_device *dev, struct drm_file *file_pri radeon_cp_load_microcode(dev_priv); radeon_cp_init_ring_buffer(dev, dev_priv, file_priv); + dev_priv->have_z_offset = 0; radeon_do_engine_reset(dev); radeon_irq_set_state(dev, RADEON_SW_INT_ENABLE, 1); diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 65590a0f1d9..70ba02ed772 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -86,7 +86,7 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p) &p->validated); } } - return radeon_bo_list_validate(&p->validated, p->ib->fence); + return radeon_bo_list_validate(&p->validated); } int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) @@ -189,18 +189,13 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error) { unsigned i; - if (error) { - radeon_bo_list_unvalidate(&parser->validated, - parser->ib->fence); - } else { - radeon_bo_list_unreserve(&parser->validated); + if (!error && parser->ib) { + radeon_bo_list_fence(&parser->validated, parser->ib->fence); } + radeon_bo_list_unreserve(&parser->validated); for (i = 0; i < parser->nrelocs; i++) { - if (parser->relocs[i].gobj) { - mutex_lock(&parser->rdev->ddev->struct_mutex); - drm_gem_object_unreference(parser->relocs[i].gobj); - mutex_unlock(&parser->rdev->ddev->struct_mutex); - } + if (parser->relocs[i].gobj) + drm_gem_object_unreference_unlocked(parser->relocs[i].gobj); } kfree(parser->track); kfree(parser->relocs); @@ -231,6 +226,7 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) memset(&parser, 0, sizeof(struct radeon_cs_parser)); parser.filp = filp; parser.rdev = rdev; + parser.dev = rdev->dev; r = radeon_cs_parser_init(&parser, data); if (r) { DRM_ERROR("Failed to initialize parser !\n"); diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c index 28772a37009..b7023fff89e 100644 --- a/drivers/gpu/drm/radeon/radeon_cursor.c +++ b/drivers/gpu/drm/radeon/radeon_cursor.c @@ -36,7 +36,14 @@ static void radeon_lock_cursor(struct drm_crtc *crtc, bool lock) struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); uint32_t cur_lock; - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + cur_lock = RREG32(EVERGREEN_CUR_UPDATE + radeon_crtc->crtc_offset); + if (lock) + cur_lock |= EVERGREEN_CURSOR_UPDATE_LOCK; + else + cur_lock &= ~EVERGREEN_CURSOR_UPDATE_LOCK; + WREG32(EVERGREEN_CUR_UPDATE + radeon_crtc->crtc_offset, cur_lock); + } else if (ASIC_IS_AVIVO(rdev)) { cur_lock = RREG32(AVIVO_D1CUR_UPDATE + radeon_crtc->crtc_offset); if (lock) cur_lock |= AVIVO_D1CURSOR_UPDATE_LOCK; @@ -58,7 +65,10 @@ static void radeon_hide_cursor(struct drm_crtc *crtc) struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct radeon_device *rdev = crtc->dev->dev_private; - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + WREG32(RADEON_MM_INDEX, EVERGREEN_CUR_CONTROL + radeon_crtc->crtc_offset); + WREG32(RADEON_MM_DATA, EVERGREEN_CURSOR_MODE(EVERGREEN_CURSOR_24_8_PRE_MULT)); + } else if (ASIC_IS_AVIVO(rdev)) { WREG32(RADEON_MM_INDEX, AVIVO_D1CUR_CONTROL + radeon_crtc->crtc_offset); WREG32(RADEON_MM_DATA, (AVIVO_D1CURSOR_MODE_24BPP << AVIVO_D1CURSOR_MODE_SHIFT)); } else { @@ -81,10 +91,14 @@ static void radeon_show_cursor(struct drm_crtc *crtc) struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct radeon_device *rdev = crtc->dev->dev_private; - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + WREG32(RADEON_MM_INDEX, EVERGREEN_CUR_CONTROL + radeon_crtc->crtc_offset); + WREG32(RADEON_MM_DATA, EVERGREEN_CURSOR_EN | + EVERGREEN_CURSOR_MODE(EVERGREEN_CURSOR_24_8_PRE_MULT)); + } else if (ASIC_IS_AVIVO(rdev)) { WREG32(RADEON_MM_INDEX, AVIVO_D1CUR_CONTROL + radeon_crtc->crtc_offset); WREG32(RADEON_MM_DATA, AVIVO_D1CURSOR_EN | - (AVIVO_D1CURSOR_MODE_24BPP << AVIVO_D1CURSOR_MODE_SHIFT)); + (AVIVO_D1CURSOR_MODE_24BPP << AVIVO_D1CURSOR_MODE_SHIFT)); } else { switch (radeon_crtc->crtc_id) { case 0: @@ -109,7 +123,10 @@ static void radeon_set_cursor(struct drm_crtc *crtc, struct drm_gem_object *obj, struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct radeon_device *rdev = crtc->dev->dev_private; - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + WREG32(EVERGREEN_CUR_SURFACE_ADDRESS_HIGH + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_CUR_SURFACE_ADDRESS + radeon_crtc->crtc_offset, gpu_addr); + } else if (ASIC_IS_AVIVO(rdev)) { if (rdev->family >= CHIP_RV770) { if (radeon_crtc->crtc_id) WREG32(R700_D2CUR_SURFACE_ADDRESS_HIGH, 0); @@ -169,17 +186,13 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc, unpin: if (radeon_crtc->cursor_bo) { radeon_gem_object_unpin(radeon_crtc->cursor_bo); - mutex_lock(&crtc->dev->struct_mutex); - drm_gem_object_unreference(radeon_crtc->cursor_bo); - mutex_unlock(&crtc->dev->struct_mutex); + drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo); } radeon_crtc->cursor_bo = obj; return 0; fail: - mutex_lock(&crtc->dev->struct_mutex); - drm_gem_object_unreference(obj); - mutex_unlock(&crtc->dev->struct_mutex); + drm_gem_object_unreference_unlocked(obj); return 0; } @@ -201,7 +214,20 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc, yorigin = CURSOR_HEIGHT - 1; radeon_lock_cursor(crtc, true); - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + /* cursors are offset into the total surface */ + x += crtc->x; + y += crtc->y; + DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y); + + /* XXX: check if evergreen has the same issues as avivo chips */ + WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset, + ((xorigin ? 0 : x) << 16) | + (yorigin ? 0 : y)); + WREG32(EVERGREEN_CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin); + WREG32(EVERGREEN_CUR_SIZE + radeon_crtc->crtc_offset, + ((radeon_crtc->cursor_width - 1) << 16) | (radeon_crtc->cursor_height - 1)); + } else if (ASIC_IS_AVIVO(rdev)) { int w = radeon_crtc->cursor_width; int i = 0; struct drm_crtc *crtc_p; diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 0c51f8e4661..e28e4ed5f72 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -30,6 +30,7 @@ #include <drm/drm_crtc_helper.h> #include <drm/radeon_drm.h> #include <linux/vgaarb.h> +#include <linux/vga_switcheroo.h> #include "radeon_reg.h" #include "radeon.h" #include "radeon_asic.h" @@ -100,80 +101,103 @@ void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg) } } -/* - * MC common functions +/** + * radeon_vram_location - try to find VRAM location + * @rdev: radeon device structure holding all necessary informations + * @mc: memory controller structure holding memory informations + * @base: base address at which to put VRAM + * + * Function will place try to place VRAM at base address provided + * as parameter (which is so far either PCI aperture address or + * for IGP TOM base address). + * + * If there is not enough space to fit the unvisible VRAM in the 32bits + * address space then we limit the VRAM size to the aperture. + * + * If we are using AGP and if the AGP aperture doesn't allow us to have + * room for all the VRAM than we restrict the VRAM to the PCI aperture + * size and print a warning. + * + * This function will never fails, worst case are limiting VRAM. + * + * Note: GTT start, end, size should be initialized before calling this + * function on AGP platform. + * + * Note: We don't explictly enforce VRAM start to be aligned on VRAM size, + * this shouldn't be a problem as we are using the PCI aperture as a reference. + * Otherwise this would be needed for rv280, all r3xx, and all r4xx, but + * not IGP. + * + * Note: we use mc_vram_size as on some board we need to program the mc to + * cover the whole aperture even if VRAM size is inferior to aperture size + * Novell bug 204882 + along with lots of ubuntu ones + * + * Note: when limiting vram it's safe to overwritte real_vram_size because + * we are not in case where real_vram_size is inferior to mc_vram_size (ie + * note afected by bogus hw of Novell bug 204882 + along with lots of ubuntu + * ones) + * + * Note: IGP TOM addr should be the same as the aperture addr, we don't + * explicitly check for that thought. + * + * FIXME: when reducing VRAM size align new size on power of 2. */ -int radeon_mc_setup(struct radeon_device *rdev) +void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64 base) +{ + mc->vram_start = base; + if (mc->mc_vram_size > (0xFFFFFFFF - base + 1)) { + dev_warn(rdev->dev, "limiting VRAM to PCI aperture size\n"); + mc->real_vram_size = mc->aper_size; + mc->mc_vram_size = mc->aper_size; + } + mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; + if (rdev->flags & RADEON_IS_AGP && mc->vram_end > mc->gtt_start && mc->vram_end <= mc->gtt_end) { + dev_warn(rdev->dev, "limiting VRAM to PCI aperture size\n"); + mc->real_vram_size = mc->aper_size; + mc->mc_vram_size = mc->aper_size; + } + mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; + dev_info(rdev->dev, "VRAM: %lluM 0x%08llX - 0x%08llX (%lluM used)\n", + mc->mc_vram_size >> 20, mc->vram_start, + mc->vram_end, mc->real_vram_size >> 20); +} + +/** + * radeon_gtt_location - try to find GTT location + * @rdev: radeon device structure holding all necessary informations + * @mc: memory controller structure holding memory informations + * + * Function will place try to place GTT before or after VRAM. + * + * If GTT size is bigger than space left then we ajust GTT size. + * Thus function will never fails. + * + * FIXME: when reducing GTT size align new size on power of 2. + */ +void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) { - uint32_t tmp; + u64 size_af, size_bf; - /* Some chips have an "issue" with the memory controller, the - * location must be aligned to the size. We just align it down, - * too bad if we walk over the top of system memory, we don't - * use DMA without a remapped anyway. - * Affected chips are rv280, all r3xx, and all r4xx, but not IGP - */ - /* FGLRX seems to setup like this, VRAM a 0, then GART. - */ - /* - * Note: from R6xx the address space is 40bits but here we only - * use 32bits (still have to see a card which would exhaust 4G - * address space). - */ - if (rdev->mc.vram_location != 0xFFFFFFFFUL) { - /* vram location was already setup try to put gtt after - * if it fits */ - tmp = rdev->mc.vram_location + rdev->mc.mc_vram_size; - tmp = (tmp + rdev->mc.gtt_size - 1) & ~(rdev->mc.gtt_size - 1); - if ((0xFFFFFFFFUL - tmp) >= rdev->mc.gtt_size) { - rdev->mc.gtt_location = tmp; - } else { - if (rdev->mc.gtt_size >= rdev->mc.vram_location) { - printk(KERN_ERR "[drm] GTT too big to fit " - "before or after vram location.\n"); - return -EINVAL; - } - rdev->mc.gtt_location = 0; - } - } else if (rdev->mc.gtt_location != 0xFFFFFFFFUL) { - /* gtt location was already setup try to put vram before - * if it fits */ - if (rdev->mc.mc_vram_size < rdev->mc.gtt_location) { - rdev->mc.vram_location = 0; - } else { - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size; - tmp += (rdev->mc.mc_vram_size - 1); - tmp &= ~(rdev->mc.mc_vram_size - 1); - if ((0xFFFFFFFFUL - tmp) >= rdev->mc.mc_vram_size) { - rdev->mc.vram_location = tmp; - } else { - printk(KERN_ERR "[drm] vram too big to fit " - "before or after GTT location.\n"); - return -EINVAL; - } + size_af = 0xFFFFFFFF - mc->vram_end; + size_bf = mc->vram_start; + if (size_bf > size_af) { + if (mc->gtt_size > size_bf) { + dev_warn(rdev->dev, "limiting GTT\n"); + mc->gtt_size = size_bf; } + mc->gtt_start = mc->vram_start - mc->gtt_size; } else { - rdev->mc.vram_location = 0; - tmp = rdev->mc.mc_vram_size; - tmp = (tmp + rdev->mc.gtt_size - 1) & ~(rdev->mc.gtt_size - 1); - rdev->mc.gtt_location = tmp; - } - rdev->mc.vram_start = rdev->mc.vram_location; - rdev->mc.vram_end = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1; - rdev->mc.gtt_start = rdev->mc.gtt_location; - rdev->mc.gtt_end = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; - DRM_INFO("radeon: VRAM %uM\n", (unsigned)(rdev->mc.mc_vram_size >> 20)); - DRM_INFO("radeon: VRAM from 0x%08X to 0x%08X\n", - (unsigned)rdev->mc.vram_location, - (unsigned)(rdev->mc.vram_location + rdev->mc.mc_vram_size - 1)); - DRM_INFO("radeon: GTT %uM\n", (unsigned)(rdev->mc.gtt_size >> 20)); - DRM_INFO("radeon: GTT from 0x%08X to 0x%08X\n", - (unsigned)rdev->mc.gtt_location, - (unsigned)(rdev->mc.gtt_location + rdev->mc.gtt_size - 1)); - return 0; + if (mc->gtt_size > size_af) { + dev_warn(rdev->dev, "limiting GTT\n"); + mc->gtt_size = size_af; + } + mc->gtt_start = mc->vram_end + 1; + } + mc->gtt_end = mc->gtt_start + mc->gtt_size - 1; + dev_info(rdev->dev, "GTT: %lluM 0x%08llX - 0x%08llX\n", + mc->gtt_size >> 20, mc->gtt_start, mc->gtt_end); } - /* * GPU helpers function. */ @@ -182,7 +206,16 @@ bool radeon_card_posted(struct radeon_device *rdev) uint32_t reg; /* first check CRTCs */ - if (ASIC_IS_AVIVO(rdev)) { + if (ASIC_IS_DCE4(rdev)) { + reg = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET) | + RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET); + if (reg & EVERGREEN_CRTC_MASTER_EN) + return true; + } else if (ASIC_IS_AVIVO(rdev)) { reg = RREG32(AVIVO_D1CRTC_CONTROL) | RREG32(AVIVO_D2CRTC_CONTROL); if (reg & AVIVO_CRTC_EN) { @@ -229,6 +262,8 @@ bool radeon_boot_test_post_card(struct radeon_device *rdev) int radeon_dummy_page_init(struct radeon_device *rdev) { + if (rdev->dummy_page.page) + return 0; rdev->dummy_page.page = alloc_page(GFP_DMA32 | GFP_KERNEL | __GFP_ZERO); if (rdev->dummy_page.page == NULL) return -ENOMEM; @@ -310,7 +345,7 @@ void radeon_register_accessor_init(struct radeon_device *rdev) rdev->mc_rreg = &rs600_mc_rreg; rdev->mc_wreg = &rs600_mc_wreg; } - if (rdev->family >= CHIP_R600) { + if ((rdev->family >= CHIP_R600) && (rdev->family <= CHIP_RV740)) { rdev->pciep_rreg = &r600_pciep_rreg; rdev->pciep_wreg = &r600_pciep_wreg; } @@ -329,21 +364,22 @@ int radeon_asic_init(struct radeon_device *rdev) case CHIP_RS100: case CHIP_RV200: case CHIP_RS200: + rdev->asic = &r100_asic; + break; case CHIP_R200: case CHIP_RV250: case CHIP_RS300: case CHIP_RV280: - rdev->asic = &r100_asic; + rdev->asic = &r200_asic; break; case CHIP_R300: case CHIP_R350: case CHIP_RV350: case CHIP_RV380: - rdev->asic = &r300_asic; - if (rdev->flags & RADEON_IS_PCIE) { - rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush; - rdev->asic->gart_set_page = &rv370_pcie_gart_set_page; - } + if (rdev->flags & RADEON_IS_PCIE) + rdev->asic = &r300_asic_pcie; + else + rdev->asic = &r300_asic; break; case CHIP_R420: case CHIP_R423: @@ -387,6 +423,13 @@ int radeon_asic_init(struct radeon_device *rdev) case CHIP_RV740: rdev->asic = &rv770_asic; break; + case CHIP_CEDAR: + case CHIP_REDWOOD: + case CHIP_JUNIPER: + case CHIP_CYPRESS: + case CHIP_HEMLOCK: + rdev->asic = &evergreen_asic; + break; default: /* FIXME: not supported yet */ return -EINVAL; @@ -544,6 +587,7 @@ void radeon_agp_disable(struct radeon_device *rdev) rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush; rdev->asic->gart_set_page = &r100_pci_gart_set_page; } + rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; } void radeon_check_arguments(struct radeon_device *rdev) @@ -612,6 +656,36 @@ void radeon_check_arguments(struct radeon_device *rdev) } } +static void radeon_switcheroo_set_state(struct pci_dev *pdev, enum vga_switcheroo_state state) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + struct radeon_device *rdev = dev->dev_private; + pm_message_t pmm = { .event = PM_EVENT_SUSPEND }; + if (state == VGA_SWITCHEROO_ON) { + printk(KERN_INFO "radeon: switched on\n"); + /* don't suspend or resume card normally */ + rdev->powered_down = false; + radeon_resume_kms(dev); + } else { + printk(KERN_INFO "radeon: switched off\n"); + radeon_suspend_kms(dev, pmm); + /* don't suspend or resume card normally */ + rdev->powered_down = true; + } +} + +static bool radeon_switcheroo_can_switch(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + bool can_switch; + + spin_lock(&dev->count_lock); + can_switch = (dev->open_count == 0); + spin_unlock(&dev->count_lock); + return can_switch; +} + + int radeon_device_init(struct radeon_device *rdev, struct drm_device *ddev, struct pci_dev *pdev, @@ -637,11 +711,14 @@ int radeon_device_init(struct radeon_device *rdev, mutex_init(&rdev->cs_mutex); mutex_init(&rdev->ib_pool.mutex); mutex_init(&rdev->cp.mutex); + mutex_init(&rdev->dc_hw_i2c_mutex); if (rdev->family >= CHIP_R600) spin_lock_init(&rdev->ih.lock); mutex_init(&rdev->gem.mutex); + mutex_init(&rdev->pm.mutex); rwlock_init(&rdev->fence_drv.lock); INIT_LIST_HEAD(&rdev->gem.objects); + init_waitqueue_head(&rdev->irq.vblank_queue); /* setup workqueue */ rdev->wq = create_workqueue("radeon"); @@ -691,6 +768,9 @@ int radeon_device_init(struct radeon_device *rdev, /* this will fail for cards that aren't VGA class devices, just * ignore it */ vga_client_register(rdev->pdev, rdev, NULL, radeon_vga_set_decode); + vga_switcheroo_register_client(rdev->pdev, + radeon_switcheroo_set_state, + radeon_switcheroo_can_switch); r = radeon_init(rdev); if (r) @@ -722,6 +802,7 @@ void radeon_device_fini(struct radeon_device *rdev) rdev->shutdown = true; radeon_fini(rdev); destroy_workqueue(rdev->wq); + vga_switcheroo_unregister_client(rdev->pdev); vga_client_register(rdev->pdev, NULL, NULL, NULL); iounmap(rdev->rmmio); rdev->rmmio = NULL; @@ -745,6 +826,8 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) } rdev = dev->dev_private; + if (rdev->powered_down) + return 0; /* unpin the front buffers */ list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct radeon_framebuffer *rfb = to_radeon_framebuffer(crtc->fb); @@ -790,6 +873,9 @@ int radeon_resume_kms(struct drm_device *dev) { struct radeon_device *rdev = dev->dev_private; + if (rdev->powered_down) + return 0; + acquire_console_sem(); pci_set_power_state(dev->pdev, PCI_D0); pci_restore_state(dev->pdev); diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 0ec491ead2f..ba8d806dcf3 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -68,6 +68,36 @@ static void avivo_crtc_load_lut(struct drm_crtc *crtc) WREG32(AVIVO_D1GRPH_LUT_SEL + radeon_crtc->crtc_offset, radeon_crtc->crtc_id); } +static void evergreen_crtc_load_lut(struct drm_crtc *crtc) +{ + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + int i; + + DRM_DEBUG("%d\n", radeon_crtc->crtc_id); + WREG32(EVERGREEN_DC_LUT_CONTROL + radeon_crtc->crtc_offset, 0); + + WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_BLUE + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_GREEN + radeon_crtc->crtc_offset, 0); + WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_RED + radeon_crtc->crtc_offset, 0); + + WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_BLUE + radeon_crtc->crtc_offset, 0xffff); + WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_GREEN + radeon_crtc->crtc_offset, 0xffff); + WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_RED + radeon_crtc->crtc_offset, 0xffff); + + WREG32(EVERGREEN_DC_LUT_RW_MODE, radeon_crtc->crtc_id); + WREG32(EVERGREEN_DC_LUT_WRITE_EN_MASK, 0x00000007); + + WREG32(EVERGREEN_DC_LUT_RW_INDEX, 0); + for (i = 0; i < 256; i++) { + WREG32(EVERGREEN_DC_LUT_30_COLOR, + (radeon_crtc->lut_r[i] << 20) | + (radeon_crtc->lut_g[i] << 10) | + (radeon_crtc->lut_b[i] << 0)); + } +} + static void legacy_crtc_load_lut(struct drm_crtc *crtc) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); @@ -100,7 +130,9 @@ void radeon_crtc_load_lut(struct drm_crtc *crtc) if (!crtc->enabled) return; - if (ASIC_IS_AVIVO(rdev)) + if (ASIC_IS_DCE4(rdev)) + evergreen_crtc_load_lut(crtc); + else if (ASIC_IS_AVIVO(rdev)) avivo_crtc_load_lut(crtc); else legacy_crtc_load_lut(crtc); @@ -278,7 +310,7 @@ static void radeon_print_display_setup(struct drm_device *dev) DRM_INFO(" %s\n", connector_names[connector->connector_type]); if (radeon_connector->hpd.hpd != RADEON_HPD_NONE) DRM_INFO(" %s\n", hpd_names[radeon_connector->hpd.hpd]); - if (radeon_connector->ddc_bus) + if (radeon_connector->ddc_bus) { DRM_INFO(" DDC: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x\n", radeon_connector->ddc_bus->rec.mask_clk_reg, radeon_connector->ddc_bus->rec.mask_data_reg, @@ -288,6 +320,15 @@ static void radeon_print_display_setup(struct drm_device *dev) radeon_connector->ddc_bus->rec.en_data_reg, radeon_connector->ddc_bus->rec.y_clk_reg, radeon_connector->ddc_bus->rec.y_data_reg); + } else { + if (connector->connector_type == DRM_MODE_CONNECTOR_VGA || + connector->connector_type == DRM_MODE_CONNECTOR_DVII || + connector->connector_type == DRM_MODE_CONNECTOR_DVID || + connector->connector_type == DRM_MODE_CONNECTOR_DVIA || + connector->connector_type == DRM_MODE_CONNECTOR_HDMIA || + connector->connector_type == DRM_MODE_CONNECTOR_HDMIB) + DRM_INFO(" DDC: no ddc bus - possible BIOS bug - please report to xorg-driver-ati@lists.x.org\n"); + } DRM_INFO(" Encoders:\n"); list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { radeon_encoder = to_radeon_encoder(encoder); @@ -352,22 +393,25 @@ static bool radeon_setup_enc_conn(struct drm_device *dev) int radeon_ddc_get_modes(struct radeon_connector *radeon_connector) { + struct drm_device *dev = radeon_connector->base.dev; + struct radeon_device *rdev = dev->dev_private; int ret = 0; if ((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) || (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)) { struct radeon_connector_atom_dig *dig = radeon_connector->con_priv; - if (dig->dp_i2c_bus) + if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT || + dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) && dig->dp_i2c_bus) radeon_connector->edid = drm_get_edid(&radeon_connector->base, &dig->dp_i2c_bus->adapter); } if (!radeon_connector->ddc_bus) return -1; if (!radeon_connector->edid) { - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); } - + /* some servers provide a hardcoded edid in rom for KVMs */ + if (!radeon_connector->edid) + radeon_connector->edid = radeon_combios_get_hardcoded_edid(rdev); if (radeon_connector->edid) { drm_mode_connector_update_edid_property(&radeon_connector->base, radeon_connector->edid); ret = drm_add_edid_modes(&radeon_connector->base, radeon_connector->edid); @@ -385,9 +429,7 @@ static int radeon_ddc_dump(struct drm_connector *connector) if (!radeon_connector->ddc_bus) return -1; - radeon_i2c_do_lock(radeon_connector->ddc_bus, 1); edid = drm_get_edid(connector, &radeon_connector->ddc_bus->adapter); - radeon_i2c_do_lock(radeon_connector->ddc_bus, 0); if (edid) { kfree(edid); } @@ -404,17 +446,18 @@ static inline uint32_t radeon_div(uint64_t n, uint32_t d) return n; } -void radeon_compute_pll(struct radeon_pll *pll, - uint64_t freq, - uint32_t *dot_clock_p, - uint32_t *fb_div_p, - uint32_t *frac_fb_div_p, - uint32_t *ref_div_p, - uint32_t *post_div_p, - int flags) +static void radeon_compute_pll_legacy(struct radeon_pll *pll, + uint64_t freq, + uint32_t *dot_clock_p, + uint32_t *fb_div_p, + uint32_t *frac_fb_div_p, + uint32_t *ref_div_p, + uint32_t *post_div_p) { uint32_t min_ref_div = pll->min_ref_div; uint32_t max_ref_div = pll->max_ref_div; + uint32_t min_post_div = pll->min_post_div; + uint32_t max_post_div = pll->max_post_div; uint32_t min_fractional_feed_div = 0; uint32_t max_fractional_feed_div = 0; uint32_t best_vco = pll->best_vco; @@ -430,7 +473,7 @@ void radeon_compute_pll(struct radeon_pll *pll, DRM_DEBUG("PLL freq %llu %u %u\n", freq, pll->min_ref_div, pll->max_ref_div); freq = freq * 1000; - if (flags & RADEON_PLL_USE_REF_DIV) + if (pll->flags & RADEON_PLL_USE_REF_DIV) min_ref_div = max_ref_div = pll->reference_div; else { while (min_ref_div < max_ref_div-1) { @@ -445,19 +488,22 @@ void radeon_compute_pll(struct radeon_pll *pll, } } - if (flags & RADEON_PLL_USE_FRAC_FB_DIV) { + if (pll->flags & RADEON_PLL_USE_POST_DIV) + min_post_div = max_post_div = pll->post_div; + + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV) { min_fractional_feed_div = pll->min_frac_feedback_div; max_fractional_feed_div = pll->max_frac_feedback_div; } - for (post_div = pll->min_post_div; post_div <= pll->max_post_div; ++post_div) { + for (post_div = min_post_div; post_div <= max_post_div; ++post_div) { uint32_t ref_div; - if ((flags & RADEON_PLL_NO_ODD_POST_DIV) && (post_div & 1)) + if ((pll->flags & RADEON_PLL_NO_ODD_POST_DIV) && (post_div & 1)) continue; /* legacy radeons only have a few post_divs */ - if (flags & RADEON_PLL_LEGACY) { + if (pll->flags & RADEON_PLL_LEGACY) { if ((post_div == 5) || (post_div == 7) || (post_div == 9) || @@ -504,7 +550,7 @@ void radeon_compute_pll(struct radeon_pll *pll, tmp += (uint64_t)pll->reference_freq * 1000 * frac_feedback_div; current_freq = radeon_div(tmp, ref_div * post_div); - if (flags & RADEON_PLL_PREFER_CLOSEST_LOWER) { + if (pll->flags & RADEON_PLL_PREFER_CLOSEST_LOWER) { error = freq - current_freq; error = error < 0 ? 0xffffffff : error; } else @@ -531,12 +577,12 @@ void radeon_compute_pll(struct radeon_pll *pll, best_freq = current_freq; best_error = error; best_vco_diff = vco_diff; - } else if (((flags & RADEON_PLL_PREFER_LOW_REF_DIV) && (ref_div < best_ref_div)) || - ((flags & RADEON_PLL_PREFER_HIGH_REF_DIV) && (ref_div > best_ref_div)) || - ((flags & RADEON_PLL_PREFER_LOW_FB_DIV) && (feedback_div < best_feedback_div)) || - ((flags & RADEON_PLL_PREFER_HIGH_FB_DIV) && (feedback_div > best_feedback_div)) || - ((flags & RADEON_PLL_PREFER_LOW_POST_DIV) && (post_div < best_post_div)) || - ((flags & RADEON_PLL_PREFER_HIGH_POST_DIV) && (post_div > best_post_div))) { + } else if (((pll->flags & RADEON_PLL_PREFER_LOW_REF_DIV) && (ref_div < best_ref_div)) || + ((pll->flags & RADEON_PLL_PREFER_HIGH_REF_DIV) && (ref_div > best_ref_div)) || + ((pll->flags & RADEON_PLL_PREFER_LOW_FB_DIV) && (feedback_div < best_feedback_div)) || + ((pll->flags & RADEON_PLL_PREFER_HIGH_FB_DIV) && (feedback_div > best_feedback_div)) || + ((pll->flags & RADEON_PLL_PREFER_LOW_POST_DIV) && (post_div < best_post_div)) || + ((pll->flags & RADEON_PLL_PREFER_HIGH_POST_DIV) && (post_div > best_post_div))) { best_post_div = post_div; best_ref_div = ref_div; best_feedback_div = feedback_div; @@ -566,96 +612,194 @@ void radeon_compute_pll(struct radeon_pll *pll, *post_div_p = best_post_div; } -void radeon_compute_pll_avivo(struct radeon_pll *pll, - uint64_t freq, - uint32_t *dot_clock_p, - uint32_t *fb_div_p, - uint32_t *frac_fb_div_p, - uint32_t *ref_div_p, - uint32_t *post_div_p, - int flags) +static bool +calc_fb_div(struct radeon_pll *pll, + uint32_t freq, + uint32_t post_div, + uint32_t ref_div, + uint32_t *fb_div, + uint32_t *fb_div_frac) { - fixed20_12 m, n, frac_n, p, f_vco, f_pclk, best_freq; - fixed20_12 pll_out_max, pll_out_min; - fixed20_12 pll_in_max, pll_in_min; - fixed20_12 reference_freq; - fixed20_12 error, ffreq, a, b; - - pll_out_max.full = rfixed_const(pll->pll_out_max); - pll_out_min.full = rfixed_const(pll->pll_out_min); - pll_in_max.full = rfixed_const(pll->pll_in_max); - pll_in_min.full = rfixed_const(pll->pll_in_min); - reference_freq.full = rfixed_const(pll->reference_freq); - do_div(freq, 10); + fixed20_12 feedback_divider, a, b; + u32 vco_freq; + + vco_freq = freq * post_div; + /* feedback_divider = vco_freq * ref_div / pll->reference_freq; */ + a.full = rfixed_const(pll->reference_freq); + feedback_divider.full = rfixed_const(vco_freq); + feedback_divider.full = rfixed_div(feedback_divider, a); + a.full = rfixed_const(ref_div); + feedback_divider.full = rfixed_mul(feedback_divider, a); + + if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV) { + /* feedback_divider = floor((feedback_divider * 10.0) + 0.5) * 0.1; */ + a.full = rfixed_const(10); + feedback_divider.full = rfixed_mul(feedback_divider, a); + feedback_divider.full += rfixed_const_half(0); + feedback_divider.full = rfixed_floor(feedback_divider); + feedback_divider.full = rfixed_div(feedback_divider, a); + + /* *fb_div = floor(feedback_divider); */ + a.full = rfixed_floor(feedback_divider); + *fb_div = rfixed_trunc(a); + /* *fb_div_frac = fmod(feedback_divider, 1.0) * 10.0; */ + a.full = rfixed_const(10); + b.full = rfixed_mul(feedback_divider, a); + + feedback_divider.full = rfixed_floor(feedback_divider); + feedback_divider.full = rfixed_mul(feedback_divider, a); + feedback_divider.full = b.full - feedback_divider.full; + *fb_div_frac = rfixed_trunc(feedback_divider); + } else { + /* *fb_div = floor(feedback_divider + 0.5); */ + feedback_divider.full += rfixed_const_half(0); + feedback_divider.full = rfixed_floor(feedback_divider); + + *fb_div = rfixed_trunc(feedback_divider); + *fb_div_frac = 0; + } + + if (((*fb_div) < pll->min_feedback_div) || ((*fb_div) > pll->max_feedback_div)) + return false; + else + return true; +} + +static bool +calc_fb_ref_div(struct radeon_pll *pll, + uint32_t freq, + uint32_t post_div, + uint32_t *fb_div, + uint32_t *fb_div_frac, + uint32_t *ref_div) +{ + fixed20_12 ffreq, max_error, error, pll_out, a; + u32 vco; + ffreq.full = rfixed_const(freq); - error.full = rfixed_const(100 * 100); + /* max_error = ffreq * 0.0025; */ + a.full = rfixed_const(400); + max_error.full = rfixed_div(ffreq, a); + + for ((*ref_div) = pll->min_ref_div; (*ref_div) < pll->max_ref_div; ++(*ref_div)) { + if (calc_fb_div(pll, freq, post_div, (*ref_div), fb_div, fb_div_frac)) { + vco = pll->reference_freq * (((*fb_div) * 10) + (*fb_div_frac)); + vco = vco / ((*ref_div) * 10); + + if ((vco < pll->pll_out_min) || (vco > pll->pll_out_max)) + continue; - /* max p */ - p.full = rfixed_div(pll_out_max, ffreq); - p.full = rfixed_floor(p); + /* pll_out = vco / post_div; */ + a.full = rfixed_const(post_div); + pll_out.full = rfixed_const(vco); + pll_out.full = rfixed_div(pll_out, a); - /* min m */ - m.full = rfixed_div(reference_freq, pll_in_max); - m.full = rfixed_ceil(m); + if (pll_out.full >= ffreq.full) { + error.full = pll_out.full - ffreq.full; + if (error.full <= max_error.full) + return true; + } + } + } + return false; +} - while (1) { - n.full = rfixed_div(ffreq, reference_freq); - n.full = rfixed_mul(n, m); - n.full = rfixed_mul(n, p); +static void radeon_compute_pll_new(struct radeon_pll *pll, + uint64_t freq, + uint32_t *dot_clock_p, + uint32_t *fb_div_p, + uint32_t *frac_fb_div_p, + uint32_t *ref_div_p, + uint32_t *post_div_p) +{ + u32 fb_div = 0, fb_div_frac = 0, post_div = 0, ref_div = 0; + u32 best_freq = 0, vco_frequency; - f_vco.full = rfixed_div(n, m); - f_vco.full = rfixed_mul(f_vco, reference_freq); + /* freq = freq / 10; */ + do_div(freq, 10); - f_pclk.full = rfixed_div(f_vco, p); + if (pll->flags & RADEON_PLL_USE_POST_DIV) { + post_div = pll->post_div; + if ((post_div < pll->min_post_div) || (post_div > pll->max_post_div)) + goto done; + + vco_frequency = freq * post_div; + if ((vco_frequency < pll->pll_out_min) || (vco_frequency > pll->pll_out_max)) + goto done; + + if (pll->flags & RADEON_PLL_USE_REF_DIV) { + ref_div = pll->reference_div; + if ((ref_div < pll->min_ref_div) || (ref_div > pll->max_ref_div)) + goto done; + if (!calc_fb_div(pll, freq, post_div, ref_div, &fb_div, &fb_div_frac)) + goto done; + } + } else { + for (post_div = pll->max_post_div; post_div >= pll->min_post_div; --post_div) { + if (pll->flags & RADEON_PLL_LEGACY) { + if ((post_div == 5) || + (post_div == 7) || + (post_div == 9) || + (post_div == 10) || + (post_div == 11)) + continue; + } - if (f_pclk.full > ffreq.full) - error.full = f_pclk.full - ffreq.full; - else - error.full = ffreq.full - f_pclk.full; - error.full = rfixed_div(error, f_pclk); - a.full = rfixed_const(100 * 100); - error.full = rfixed_mul(error, a); - - a.full = rfixed_mul(m, p); - a.full = rfixed_div(n, a); - best_freq.full = rfixed_mul(reference_freq, a); - - if (rfixed_trunc(error) < 25) - break; - - a.full = rfixed_const(1); - m.full = m.full + a.full; - a.full = rfixed_div(reference_freq, m); - if (a.full >= pll_in_min.full) - continue; + if ((pll->flags & RADEON_PLL_NO_ODD_POST_DIV) && (post_div & 1)) + continue; - m.full = rfixed_div(reference_freq, pll_in_max); - m.full = rfixed_ceil(m); - a.full= rfixed_const(1); - p.full = p.full - a.full; - a.full = rfixed_mul(p, ffreq); - if (a.full >= pll_out_min.full) - continue; - else { - DRM_ERROR("Unable to find pll dividers\n"); - break; + vco_frequency = freq * post_div; + if ((vco_frequency < pll->pll_out_min) || (vco_frequency > pll->pll_out_max)) + continue; + if (pll->flags & RADEON_PLL_USE_REF_DIV) { + ref_div = pll->reference_div; + if ((ref_div < pll->min_ref_div) || (ref_div > pll->max_ref_div)) + goto done; + if (calc_fb_div(pll, freq, post_div, ref_div, &fb_div, &fb_div_frac)) + break; + } else { + if (calc_fb_ref_div(pll, freq, post_div, &fb_div, &fb_div_frac, &ref_div)) + break; + } } } - a.full = rfixed_const(10); - b.full = rfixed_mul(n, a); + best_freq = pll->reference_freq * 10 * fb_div; + best_freq += pll->reference_freq * fb_div_frac; + best_freq = best_freq / (ref_div * post_div); - frac_n.full = rfixed_floor(n); - frac_n.full = rfixed_mul(frac_n, a); - frac_n.full = b.full - frac_n.full; +done: + if (best_freq == 0) + DRM_ERROR("Couldn't find valid PLL dividers\n"); - *dot_clock_p = rfixed_trunc(best_freq); - *fb_div_p = rfixed_trunc(n); - *frac_fb_div_p = rfixed_trunc(frac_n); - *ref_div_p = rfixed_trunc(m); - *post_div_p = rfixed_trunc(p); + *dot_clock_p = best_freq / 10; + *fb_div_p = fb_div; + *frac_fb_div_p = fb_div_frac; + *ref_div_p = ref_div; + *post_div_p = post_div; + + DRM_DEBUG("%u %d.%d, %d, %d\n", *dot_clock_p, *fb_div_p, *frac_fb_div_p, *ref_div_p, *post_div_p); +} - DRM_DEBUG("%u %d.%d, %d, %d\n", *dot_clock_p * 10, *fb_div_p, *frac_fb_div_p, *ref_div_p, *post_div_p); +void radeon_compute_pll(struct radeon_pll *pll, + uint64_t freq, + uint32_t *dot_clock_p, + uint32_t *fb_div_p, + uint32_t *frac_fb_div_p, + uint32_t *ref_div_p, + uint32_t *post_div_p) +{ + switch (pll->algo) { + case PLL_ALGO_NEW: + radeon_compute_pll_new(pll, freq, dot_clock_p, fb_div_p, + frac_fb_div_p, ref_div_p, post_div_p); + break; + case PLL_ALGO_LEGACY: + default: + radeon_compute_pll_legacy(pll, freq, dot_clock_p, fb_div_p, + frac_fb_div_p, ref_div_p, post_div_p); + break; + } } static void radeon_user_framebuffer_destroy(struct drm_framebuffer *fb) @@ -666,12 +810,8 @@ static void radeon_user_framebuffer_destroy(struct drm_framebuffer *fb) if (fb->fbdev) radeonfb_remove(dev, fb); - if (radeon_fb->obj) { - radeon_gem_object_unpin(radeon_fb->obj); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(radeon_fb->obj); - mutex_unlock(&dev->struct_mutex); - } + if (radeon_fb->obj) + drm_gem_object_unreference_unlocked(radeon_fb->obj); drm_framebuffer_cleanup(fb); kfree(radeon_fb); } @@ -715,7 +855,11 @@ radeon_user_framebuffer_create(struct drm_device *dev, struct drm_gem_object *obj; obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle); - + if (obj == NULL) { + dev_err(&dev->pdev->dev, "No GEM object associated to handle 0x%08X, " + "can't create framebuffer\n", mode_cmd->handle); + return NULL; + } return radeon_framebuffer_create(dev, mode_cmd, obj); } @@ -803,7 +947,7 @@ static int radeon_modeset_create_props(struct radeon_device *rdev) int radeon_modeset_init(struct radeon_device *rdev) { - int num_crtc = 2, i; + int i; int ret; drm_mode_config_init(rdev->ddev); @@ -826,11 +970,23 @@ int radeon_modeset_init(struct radeon_device *rdev) return ret; } + /* check combios for a valid hardcoded EDID - Sun servers */ + if (!rdev->is_atom_bios) { + /* check for hardcoded EDID in BIOS */ + radeon_combios_check_hardcoded_edid(rdev); + } + if (rdev->flags & RADEON_SINGLE_CRTC) - num_crtc = 1; + rdev->num_crtc = 1; + else { + if (ASIC_IS_DCE4(rdev)) + rdev->num_crtc = 6; + else + rdev->num_crtc = 2; + } /* allocate crtcs */ - for (i = 0; i < num_crtc; i++) { + for (i = 0; i < rdev->num_crtc; i++) { radeon_crtc_init(rdev->ddev, i); } @@ -847,6 +1003,8 @@ int radeon_modeset_init(struct radeon_device *rdev) void radeon_modeset_fini(struct radeon_device *rdev) { + kfree(rdev->mode_info.bios_hardcoded_edid); + if (rdev->mode_info.mode_config_initialized) { radeon_hpd_fini(rdev); drm_mode_config_cleanup(rdev->ddev); diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 8ba3de7994d..6eec0ece6a6 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -40,9 +40,11 @@ /* * KMS wrapper. + * - 2.0.0 - initial interface + * - 2.1.0 - add square tiling interface */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 0 +#define KMS_DRIVER_MINOR 1 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); @@ -86,7 +88,8 @@ int radeon_benchmarking = 0; int radeon_testing = 0; int radeon_connector_table = 0; int radeon_tv = 1; -int radeon_new_pll = 1; +int radeon_new_pll = -1; +int radeon_dynpm = -1; int radeon_audio = 1; MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers"); @@ -122,9 +125,12 @@ module_param_named(connector_table, radeon_connector_table, int, 0444); MODULE_PARM_DESC(tv, "TV enable (0 = disable)"); module_param_named(tv, radeon_tv, int, 0444); -MODULE_PARM_DESC(new_pll, "Select new PLL code for AVIVO chips"); +MODULE_PARM_DESC(new_pll, "Select new PLL code"); module_param_named(new_pll, radeon_new_pll, int, 0444); +MODULE_PARM_DESC(dynpm, "Disable/Enable dynamic power management (1 = enable)"); +module_param_named(dynpm, radeon_dynpm, int, 0444); + MODULE_PARM_DESC(audio, "Audio enable (0 = disable)"); module_param_named(audio, radeon_audio, int, 0444); @@ -339,6 +345,7 @@ static int __init radeon_init(void) driver = &kms_driver; driver->driver_features |= DRIVER_MODESET; driver->num_ioctls = radeon_max_kms_ioctl; + radeon_register_atpx_handler(); } /* if the vga console setting is enabled still * let modprobe override it */ @@ -348,6 +355,7 @@ static int __init radeon_init(void) static void __exit radeon_exit(void) { drm_exit(driver); + radeon_unregister_atpx_handler(); } module_init(radeon_init); diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h index e13785282a8..ec55f2b23c2 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.h +++ b/drivers/gpu/drm/radeon/radeon_drv.h @@ -106,9 +106,10 @@ * 1.29- R500 3D cmd buffer support * 1.30- Add support for occlusion queries * 1.31- Add support for num Z pipes from GET_PARAM + * 1.32- fixes for rv740 setup */ #define DRIVER_MAJOR 1 -#define DRIVER_MINOR 31 +#define DRIVER_MINOR 32 #define DRIVER_PATCHLEVEL 0 enum radeon_cp_microcode_version { @@ -267,6 +268,8 @@ typedef struct drm_radeon_private { u32 scratch_ages[5]; + int have_z_offset; + /* starting from here on, data is preserved accross an open */ uint32_t flags; /* see radeon_chip_flags */ resource_size_t fb_aper_offset; @@ -294,6 +297,9 @@ typedef struct drm_radeon_private { int r700_sc_prim_fifo_size; int r700_sc_hiz_tile_fifo_size; int r700_sc_earlyz_tile_fifo_fize; + int r600_group_size; + int r600_npipes; + int r600_nbanks; struct mutex cs_mutex; u32 cs_id_scnt; @@ -309,9 +315,11 @@ typedef struct drm_radeon_buf_priv { u32 age; } drm_radeon_buf_priv_t; +struct drm_buffer; + typedef struct drm_radeon_kcmd_buffer { int bufsz; - char *buf; + struct drm_buffer *buffer; int nbox; struct drm_clip_rect __user *boxes; } drm_radeon_kcmd_buffer_t; @@ -454,6 +462,15 @@ extern void r600_blit_swap(struct drm_device *dev, int sx, int sy, int dx, int dy, int w, int h, int src_pitch, int dst_pitch, int cpp); +/* atpx handler */ +#if defined(CONFIG_VGA_SWITCHEROO) +void radeon_register_atpx_handler(void); +void radeon_unregister_atpx_handler(void); +#else +static inline void radeon_register_atpx_handler(void) {} +static inline void radeon_unregister_atpx_handler(void) {} +#endif + /* Flags for stats.boxes */ #define RADEON_BOX_DMA_IDLE 0x1 @@ -2121,4 +2138,32 @@ extern void radeon_commit_ring(drm_radeon_private_t *dev_priv); write &= mask; \ } while (0) +/** + * Copy given number of dwords from drm buffer to the ring buffer. + */ +#define OUT_RING_DRM_BUFFER(buf, sz) do { \ + int _size = (sz) * 4; \ + struct drm_buffer *_buf = (buf); \ + int _part_size; \ + while (_size > 0) { \ + _part_size = _size; \ + \ + if (write + _part_size/4 > mask) \ + _part_size = ((mask + 1) - write)*4; \ + \ + if (drm_buffer_index(_buf) + _part_size > PAGE_SIZE) \ + _part_size = PAGE_SIZE - drm_buffer_index(_buf);\ + \ + \ + \ + memcpy(ring + write, &_buf->data[drm_buffer_page(_buf)] \ + [drm_buffer_index(_buf)], _part_size); \ + \ + _size -= _part_size; \ + write = (write + _part_size/4) & mask; \ + drm_buffer_advance(_buf, _part_size); \ + } \ +} while (0) + + #endif /* __RADEON_DRV_H__ */ diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c index 82eb551970b..bc926ea0a53 100644 --- a/drivers/gpu/drm/radeon/radeon_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_encoders.c @@ -53,7 +53,7 @@ static uint32_t radeon_encoder_clones(struct drm_encoder *encoder) /* DVO requires 2x ppll clocks depending on tmds chip */ if (radeon_encoder->devices & ATOM_DEVICE_DFP2_SUPPORT) return index_mask; - + count = -1; list_for_each_entry(clone_encoder, &dev->mode_config.encoder_list, head) { struct radeon_encoder *radeon_clone = to_radeon_encoder(clone_encoder); @@ -156,6 +156,26 @@ radeon_get_encoder_id(struct drm_device *dev, uint32_t supported_device, uint8_t return ret; } +static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder) +{ + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_LVDS: + case ENCODER_OBJECT_ID_INTERNAL_TMDS1: + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1: + case ENCODER_OBJECT_ID_INTERNAL_LVTM1: + case ENCODER_OBJECT_ID_INTERNAL_DVO1: + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1: + case ENCODER_OBJECT_ID_INTERNAL_DDI: + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + return true; + default: + return false; + } +} void radeon_link_encoder_connector(struct drm_device *dev) { @@ -202,12 +222,38 @@ radeon_get_connector_for_encoder(struct drm_encoder *encoder) list_for_each_entry(connector, &dev->mode_config.connector_list, head) { radeon_connector = to_radeon_connector(connector); - if (radeon_encoder->devices & radeon_connector->devices) + if (radeon_encoder->active_device & radeon_connector->devices) return connector; } return NULL; } +static struct radeon_connector_atom_dig * +radeon_get_atom_connector_priv_from_encoder(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector; + struct radeon_connector_atom_dig *dig_connector; + + if (!rdev->is_atom_bios) + return NULL; + + connector = radeon_get_connector_for_encoder(encoder); + if (!connector) + return NULL; + + radeon_connector = to_radeon_connector(connector); + + if (!radeon_connector->con_priv) + return NULL; + + dig_connector = radeon_connector->con_priv; + + return dig_connector; +} + static bool radeon_atom_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -216,6 +262,9 @@ static bool radeon_atom_mode_fixup(struct drm_encoder *encoder, struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; + /* adjust pm to upcoming mode change */ + radeon_pm_compute_clocks(rdev); + /* set the active encoder to connector routing */ radeon_encoder_set_active_device(encoder); drm_mode_set_crtcinfo(adjusted_mode, 0); @@ -438,34 +487,20 @@ atombios_digital_setup(struct drm_encoder *encoder, int action) struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct radeon_connector_atom_dig *dig_connector = + radeon_get_atom_connector_priv_from_encoder(encoder); union lvds_encoder_control args; int index = 0; int hdmi_detected = 0; uint8_t frev, crev; - struct radeon_encoder_atom_dig *dig; - struct drm_connector *connector; - struct radeon_connector *radeon_connector; - struct radeon_connector_atom_dig *dig_connector; - - connector = radeon_get_connector_for_encoder(encoder); - if (!connector) - return; - - radeon_connector = to_radeon_connector(connector); - if (!radeon_encoder->enc_priv) + if (!dig || !dig_connector) return; - dig = radeon_encoder->enc_priv; - - if (!radeon_connector->con_priv) - return; - - if (drm_detect_hdmi_monitor(radeon_connector->edid)) + if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) hdmi_detected = 1; - dig_connector = radeon_connector->con_priv; - memset(&args, 0, sizeof(args)); switch (radeon_encoder->encoder_id) { @@ -566,7 +601,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) { struct drm_connector *connector; struct radeon_connector *radeon_connector; - struct radeon_connector_atom_dig *radeon_dig_connector; + struct radeon_connector_atom_dig *dig_connector; connector = radeon_get_connector_for_encoder(encoder); if (!connector) @@ -597,9 +632,9 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) break; case DRM_MODE_CONNECTOR_DisplayPort: case DRM_MODE_CONNECTOR_eDP: - radeon_dig_connector = radeon_connector->con_priv; - if ((radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || - (radeon_dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) + dig_connector = radeon_connector->con_priv; + if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || + (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) return ATOM_ENCODER_MODE_DP; else if (drm_detect_hdmi_monitor(radeon_connector->edid)) return ATOM_ENCODER_MODE_HDMI; @@ -636,6 +671,18 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) * - 2 DIG encoder blocks. * DIG1/2 can drive UNIPHY0/1/2 link A or link B * + * DCE 4.0 + * - 3 DIG transmitter blocks UNPHY0/1/2 (links A and B). + * Supports up to 6 digital outputs + * - 6 DIG encoder blocks. + * - DIG to PHY mapping is hardcoded + * DIG1 drives UNIPHY0 link A, A+B + * DIG2 drives UNIPHY0 link B + * DIG3 drives UNIPHY1 link A, A+B + * DIG4 drives UNIPHY1 link B + * DIG5 drives UNIPHY2 link A, A+B + * DIG6 drives UNIPHY2 link B + * * Routing * crtc -> dig encoder -> UNIPHY/LVTMA (1 or 2 links) * Examples: @@ -644,108 +691,78 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) * crtc0 -> dig1 -> UNIPHY2 link A -> LVDS * crtc1 -> dig2 -> UNIPHY1 link B+A -> TMDS/HDMI */ -static void + +union dig_encoder_control { + DIG_ENCODER_CONTROL_PS_ALLOCATION v1; + DIG_ENCODER_CONTROL_PARAMETERS_V2 v2; + DIG_ENCODER_CONTROL_PARAMETERS_V3 v3; +}; + +void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action) { struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); - DIG_ENCODER_CONTROL_PS_ALLOCATION args; + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct radeon_connector_atom_dig *dig_connector = + radeon_get_atom_connector_priv_from_encoder(encoder); + union dig_encoder_control args; int index = 0, num = 0; uint8_t frev, crev; - struct radeon_encoder_atom_dig *dig; - struct drm_connector *connector; - struct radeon_connector *radeon_connector; - struct radeon_connector_atom_dig *dig_connector; - connector = radeon_get_connector_for_encoder(encoder); - if (!connector) + if (!dig || !dig_connector) return; - radeon_connector = to_radeon_connector(connector); - - if (!radeon_connector->con_priv) - return; - - dig_connector = radeon_connector->con_priv; - - if (!radeon_encoder->enc_priv) - return; - - dig = radeon_encoder->enc_priv; - memset(&args, 0, sizeof(args)); - if (ASIC_IS_DCE32(rdev)) { - if (dig->dig_block) + if (ASIC_IS_DCE4(rdev)) + index = GetIndexIntoMasterTable(COMMAND, DIGxEncoderControl); + else { + if (dig->dig_encoder) index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl); else index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl); - num = dig->dig_block + 1; - } else { - switch (radeon_encoder->encoder_id) { - case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: - /* XXX doesn't really matter which dig encoder we pick as long as it's - * not already in use - */ - if (dig_connector->linkb) - index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl); - else - index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl); - num = 1; - break; - case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: - /* Only dig2 encoder can drive LVTMA */ - index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl); - num = 2; - break; - } } + num = dig->dig_encoder + 1; atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev); - args.ucAction = action; - args.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); + args.v1.ucAction = action; + args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); + args.v1.ucEncoderMode = atombios_get_encoder_mode(encoder); - if (ASIC_IS_DCE32(rdev)) { + if (args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP) { + if (dig_connector->dp_clock == 270000) + args.v1.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ; + args.v1.ucLaneNum = dig_connector->dp_lane_count; + } else if (radeon_encoder->pixel_clock > 165000) + args.v1.ucLaneNum = 8; + else + args.v1.ucLaneNum = 4; + + if (ASIC_IS_DCE4(rdev)) { + args.v3.acConfig.ucDigSel = dig->dig_encoder; + args.v3.ucBitPerColor = PANEL_8BIT_PER_COLOR; + } else { switch (radeon_encoder->encoder_id) { case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: - args.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER1; + args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER1; break; case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: - args.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER2; + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: + args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER2; break; case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: - args.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER3; - break; - } - } else { - switch (radeon_encoder->encoder_id) { - case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: - args.ucConfig = ATOM_ENCODER_CONFIG_TRANSMITTER1; - break; - case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: - args.ucConfig = ATOM_ENCODER_CONFIG_TRANSMITTER2; + args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER3; break; } + if (dig_connector->linkb) + args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKB; + else + args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKA; } - args.ucEncoderMode = atombios_get_encoder_mode(encoder); - - if (args.ucEncoderMode == ATOM_ENCODER_MODE_DP) { - if (dig_connector->dp_clock == 270000) - args.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ; - args.ucLaneNum = dig_connector->dp_lane_count; - } else if (radeon_encoder->pixel_clock > 165000) - args.ucLaneNum = 8; - else - args.ucLaneNum = 4; - - if (dig_connector->linkb) - args.ucConfig |= ATOM_ENCODER_CONFIG_LINKB; - else - args.ucConfig |= ATOM_ENCODER_CONFIG_LINKA; - atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } @@ -753,6 +770,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action) union dig_transmitter_control { DIG_TRANSMITTER_CONTROL_PS_ALLOCATION v1; DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 v2; + DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 v3; }; void @@ -761,37 +779,29 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t struct drm_device *dev = encoder->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct radeon_connector_atom_dig *dig_connector = + radeon_get_atom_connector_priv_from_encoder(encoder); + struct drm_connector *connector; + struct radeon_connector *radeon_connector; union dig_transmitter_control args; int index = 0, num = 0; uint8_t frev, crev; - struct radeon_encoder_atom_dig *dig; - struct drm_connector *connector; - struct radeon_connector *radeon_connector; - struct radeon_connector_atom_dig *dig_connector; bool is_dp = false; + int pll_id = 0; - connector = radeon_get_connector_for_encoder(encoder); - if (!connector) + if (!dig || !dig_connector) return; + connector = radeon_get_connector_for_encoder(encoder); radeon_connector = to_radeon_connector(connector); - if (!radeon_encoder->enc_priv) - return; - - dig = radeon_encoder->enc_priv; - - if (!radeon_connector->con_priv) - return; - - dig_connector = radeon_connector->con_priv; - if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_DP) is_dp = true; memset(&args, 0, sizeof(args)); - if (ASIC_IS_DCE32(rdev)) + if (ASIC_IS_DCE32(rdev) || ASIC_IS_DCE4(rdev)) index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl); else { switch (radeon_encoder->encoder_id) { @@ -821,8 +831,55 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t else args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); } - if (ASIC_IS_DCE32(rdev)) { - if (dig->dig_block) + if (ASIC_IS_DCE4(rdev)) { + if (is_dp) + args.v3.ucLaneNum = dig_connector->dp_lane_count; + else if (radeon_encoder->pixel_clock > 165000) + args.v3.ucLaneNum = 8; + else + args.v3.ucLaneNum = 4; + + if (dig_connector->linkb) { + args.v3.acConfig.ucLinkSel = 1; + args.v3.acConfig.ucEncoderSel = 1; + } + + /* Select the PLL for the PHY + * DP PHY should be clocked from external src if there is + * one. + */ + if (encoder->crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + pll_id = radeon_crtc->pll_id; + } + if (is_dp && rdev->clock.dp_extclk) + args.v3.acConfig.ucRefClkSource = 2; /* external src */ + else + args.v3.acConfig.ucRefClkSource = pll_id; + + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: + args.v3.acConfig.ucTransmitterSel = 0; + num = 0; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: + args.v3.acConfig.ucTransmitterSel = 1; + num = 1; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + args.v3.acConfig.ucTransmitterSel = 2; + num = 2; + break; + } + + if (is_dp) + args.v3.acConfig.fCoherentMode = 1; /* DP requires coherent */ + else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { + if (dig->coherent_mode) + args.v3.acConfig.fCoherentMode = 1; + } + } else if (ASIC_IS_DCE32(rdev)) { + if (dig->dig_encoder == 1) args.v2.acConfig.ucEncoderSel = 1; if (dig_connector->linkb) args.v2.acConfig.ucLinkSel = 1; @@ -851,15 +908,13 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t } else { args.v1.ucConfig = ATOM_TRANSMITTER_CONFIG_CLKSRC_PPLL; + if (dig->dig_encoder) + args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER; + else + args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER; + switch (radeon_encoder->encoder_id) { case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: - /* XXX doesn't really matter which dig encoder we pick as long as it's - * not already in use - */ - if (dig_connector->linkb) - args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER; - else - args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER; if (rdev->flags & RADEON_IS_IGP) { if (radeon_encoder->pixel_clock > 165000) { if (dig_connector->igp_lane_info & 0x3) @@ -878,10 +933,6 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t } } break; - case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: - /* Only dig2 encoder can drive LVTMA */ - args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER; - break; } if (radeon_encoder->pixel_clock > 165000) @@ -1029,9 +1080,12 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode) atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } -union crtc_sourc_param { +union crtc_source_param { SELECT_CRTC_SOURCE_PS_ALLOCATION v1; SELECT_CRTC_SOURCE_PARAMETERS_V2 v2; }; @@ -1043,9 +1097,10 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder) struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); - union crtc_sourc_param args; + union crtc_source_param args; int index = GetIndexIntoMasterTable(COMMAND, SelectCRTC_Source); uint8_t frev, crev; + struct radeon_encoder_atom_dig *dig; memset(&args, 0, sizeof(args)); @@ -1109,40 +1164,32 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder) case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: - if (ASIC_IS_DCE32(rdev)) { - if (radeon_crtc->crtc_id) - args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID; - else - args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID; - } else { - struct drm_connector *connector; - struct radeon_connector *radeon_connector; - struct radeon_connector_atom_dig *dig_connector; - - connector = radeon_get_connector_for_encoder(encoder); - if (!connector) - return; - radeon_connector = to_radeon_connector(connector); - if (!radeon_connector->con_priv) - return; - dig_connector = radeon_connector->con_priv; - - /* XXX doesn't really matter which dig encoder we pick as long as it's - * not already in use - */ - if (dig_connector->linkb) - args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID; - else - args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID; + case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: + dig = radeon_encoder->enc_priv; + switch (dig->dig_encoder) { + case 0: + args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID; + break; + case 1: + args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID; + break; + case 2: + args.v2.ucEncoderID = ASIC_INT_DIG3_ENCODER_ID; + break; + case 3: + args.v2.ucEncoderID = ASIC_INT_DIG4_ENCODER_ID; + break; + case 4: + args.v2.ucEncoderID = ASIC_INT_DIG5_ENCODER_ID; + break; + case 5: + args.v2.ucEncoderID = ASIC_INT_DIG6_ENCODER_ID; + break; } break; case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1: args.v2.ucEncoderID = ASIC_INT_DVO_ENCODER_ID; break; - case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: - /* Only dig2 encoder can drive LVTMA */ - args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID; - break; case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1: if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID; @@ -1195,6 +1242,7 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder, } /* set scaler clears this on some chips */ + /* XXX check DCE4 */ if (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))) { if (ASIC_IS_AVIVO(rdev) && (mode->flags & DRM_MODE_FLAG_INTERLACE)) WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, @@ -1202,6 +1250,74 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder, } } +static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct drm_encoder *test_encoder; + struct radeon_encoder_atom_dig *dig; + uint32_t dig_enc_in_use = 0; + + if (ASIC_IS_DCE4(rdev)) { + struct radeon_connector_atom_dig *dig_connector = + radeon_get_atom_connector_priv_from_encoder(encoder); + + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: + if (dig_connector->linkb) + return 1; + else + return 0; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: + if (dig_connector->linkb) + return 3; + else + return 2; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + if (dig_connector->linkb) + return 5; + else + return 4; + break; + } + } + + /* on DCE32 and encoder can driver any block so just crtc id */ + if (ASIC_IS_DCE32(rdev)) { + return radeon_crtc->crtc_id; + } + + /* on DCE3 - LVTMA can only be driven by DIGB */ + list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) { + struct radeon_encoder *radeon_test_encoder; + + if (encoder == test_encoder) + continue; + + if (!radeon_encoder_is_digital(test_encoder)) + continue; + + radeon_test_encoder = to_radeon_encoder(test_encoder); + dig = radeon_test_encoder->enc_priv; + + if (dig->dig_encoder >= 0) + dig_enc_in_use |= (1 << dig->dig_encoder); + } + + if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA) { + if (dig_enc_in_use & 0x2) + DRM_ERROR("LVDS required digital encoder 2 but it was in use - stealing\n"); + return 1; + } + if (!(dig_enc_in_use & 1)) + return 0; + return 1; +} + static void radeon_atom_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, @@ -1214,12 +1330,9 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder, if (radeon_encoder->active_device & (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) { - if (radeon_encoder->enc_priv) { - struct radeon_encoder_atom_dig *dig; - - dig = radeon_encoder->enc_priv; - dig->dig_block = radeon_crtc->crtc_id; - } + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + if (dig) + dig->dig_encoder = radeon_atom_pick_dig_encoder(encoder); } radeon_encoder->pixel_clock = adjusted_mode->clock; @@ -1244,15 +1357,26 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder, case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: - /* disable the encoder and transmitter */ - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); - atombios_dig_encoder_setup(encoder, ATOM_DISABLE); - - /* setup and enable the encoder and transmitter */ - atombios_dig_encoder_setup(encoder, ATOM_ENABLE); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_SETUP, 0, 0); - atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); + if (ASIC_IS_DCE4(rdev)) { + /* disable the transmitter */ + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); + /* setup and enable the encoder */ + atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_SETUP); + + /* init and enable the transmitter */ + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0); + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); + } else { + /* disable the encoder and transmitter */ + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0); + atombios_dig_encoder_setup(encoder, ATOM_DISABLE); + + /* setup and enable the encoder and transmitter */ + atombios_dig_encoder_setup(encoder, ATOM_ENABLE); + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0); + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_SETUP, 0, 0); + atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0); + } break; case ENCODER_OBJECT_ID_INTERNAL_DDI: atombios_ddia_setup(encoder, ATOM_ENABLE); @@ -1272,7 +1396,9 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder, } atombios_apply_encoder_quirks(encoder, adjusted_mode); - r600_hdmi_setmode(encoder, adjusted_mode); + /* XXX */ + if (!ASIC_IS_DCE4(rdev)) + r600_hdmi_setmode(encoder, adjusted_mode); } static bool @@ -1379,7 +1505,13 @@ static void radeon_atom_encoder_commit(struct drm_encoder *encoder) static void radeon_atom_encoder_disable(struct drm_encoder *encoder) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct radeon_encoder_atom_dig *dig; radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF); + + if (radeon_encoder_is_digital(encoder)) { + dig = radeon_encoder->enc_priv; + dig->dig_encoder = -1; + } radeon_encoder->active_device = 0; } @@ -1436,6 +1568,7 @@ radeon_atombios_set_dig_info(struct radeon_encoder *radeon_encoder) /* coherent mode by default */ dig->coherent_mode = true; + dig->dig_encoder = -1; return dig; } @@ -1463,10 +1596,18 @@ radeon_add_atom_encoder(struct drm_device *dev, uint32_t encoder_id, uint32_t su return; encoder = &radeon_encoder->base; - if (rdev->flags & RADEON_SINGLE_CRTC) + switch (rdev->num_crtc) { + case 1: encoder->possible_crtcs = 0x1; - else + break; + case 2: + default: encoder->possible_crtcs = 0x3; + break; + case 6: + encoder->possible_crtcs = 0x3f; + break; + } radeon_encoder->enc_priv = NULL; diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h index 797972e344a..93c7d5d4191 100644 --- a/drivers/gpu/drm/radeon/radeon_family.h +++ b/drivers/gpu/drm/radeon/radeon_family.h @@ -75,6 +75,11 @@ enum radeon_family { CHIP_RV730, CHIP_RV710, CHIP_RV740, + CHIP_CEDAR, + CHIP_REDWOOD, + CHIP_JUNIPER, + CHIP_CYPRESS, + CHIP_HEMLOCK, CHIP_LAST, }; diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c index 3ba213d1b06..8fccbf29235 100644 --- a/drivers/gpu/drm/radeon/radeon_fb.c +++ b/drivers/gpu/drm/radeon/radeon_fb.c @@ -39,6 +39,8 @@ #include "drm_fb_helper.h" +#include <linux/vga_switcheroo.h> + struct radeon_fb_device { struct drm_fb_helper helper; struct radeon_framebuffer *rfb; @@ -148,7 +150,6 @@ int radeonfb_create(struct drm_device *dev, unsigned long tmp; bool fb_tiled = false; /* useful for testing */ u32 tiling_flags = 0; - int crtc_count; mode_cmd.width = surface_width; mode_cmd.height = surface_height; @@ -239,16 +240,12 @@ int radeonfb_create(struct drm_device *dev, rfbdev = info->par; rfbdev->helper.funcs = &radeon_fb_helper_funcs; rfbdev->helper.dev = dev; - if (rdev->flags & RADEON_SINGLE_CRTC) - crtc_count = 1; - else - crtc_count = 2; - ret = drm_fb_helper_init_crtc_count(&rfbdev->helper, crtc_count, + ret = drm_fb_helper_init_crtc_count(&rfbdev->helper, rdev->num_crtc, RADEONFB_CONN_LIMIT); if (ret) goto out_unref; - memset_io(fbptr, 0xff, aligned_size); + memset_io(fbptr, 0x0, aligned_size); strcpy(info->fix.id, "radeondrmfb"); @@ -257,7 +254,7 @@ int radeonfb_create(struct drm_device *dev, info->flags = FBINFO_DEFAULT; info->fbops = &radeonfb_ops; - tmp = fb_gpuaddr - rdev->mc.vram_location; + tmp = fb_gpuaddr - rdev->mc.vram_start; info->fix.smem_start = rdev->mc.aper_base + tmp; info->fix.smem_len = size; info->screen_base = fbptr; @@ -291,6 +288,7 @@ int radeonfb_create(struct drm_device *dev, rfbdev->rdev = rdev; mutex_unlock(&rdev->ddev->struct_mutex); + vga_switcheroo_client_fb_set(rdev->ddev->pdev, info); return 0; out_unref: diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c index e73d56e83fa..1770d3c07fd 100644 --- a/drivers/gpu/drm/radeon/radeon_gart.c +++ b/drivers/gpu/drm/radeon/radeon_gart.c @@ -139,6 +139,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset, unsigned t; unsigned p; int i, j; + u64 page_base; if (!rdev->gart.ready) { WARN(1, "trying to unbind memory to unitialized GART !\n"); @@ -151,9 +152,11 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset, pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p], PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); rdev->gart.pages[p] = NULL; - rdev->gart.pages_addr[p] = 0; + rdev->gart.pages_addr[p] = rdev->dummy_page.addr; + page_base = rdev->gart.pages_addr[p]; for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) { - radeon_gart_set_page(rdev, t, 0); + radeon_gart_set_page(rdev, t, page_base); + page_base += RADEON_GPU_PAGE_SIZE; } } } @@ -199,8 +202,26 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset, return 0; } +void radeon_gart_restore(struct radeon_device *rdev) +{ + int i, j, t; + u64 page_base; + + for (i = 0, t = 0; i < rdev->gart.num_cpu_pages; i++) { + page_base = rdev->gart.pages_addr[i]; + for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) { + radeon_gart_set_page(rdev, t, page_base); + page_base += RADEON_GPU_PAGE_SIZE; + } + } + mb(); + radeon_gart_tlb_flush(rdev); +} + int radeon_gart_init(struct radeon_device *rdev) { + int r, i; + if (rdev->gart.pages) { return 0; } @@ -209,6 +230,9 @@ int radeon_gart_init(struct radeon_device *rdev) DRM_ERROR("Page size is smaller than GPU page size!\n"); return -EINVAL; } + r = radeon_dummy_page_init(rdev); + if (r) + return r; /* Compute table size */ rdev->gart.num_cpu_pages = rdev->mc.gtt_size / PAGE_SIZE; rdev->gart.num_gpu_pages = rdev->mc.gtt_size / RADEON_GPU_PAGE_SIZE; @@ -227,6 +251,10 @@ int radeon_gart_init(struct radeon_device *rdev) radeon_gart_fini(rdev); return -ENOMEM; } + /* set GART entry to point to the dummy page by default */ + for (i = 0; i < rdev->gart.num_cpu_pages; i++) { + rdev->gart.pages_addr[i] = rdev->dummy_page.addr; + } return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index 0e1325e1853..ef92d147d8f 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -69,9 +69,7 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size, if (r != -ERESTARTSYS) DRM_ERROR("Failed to allocate GEM object (%d, %d, %u, %d)\n", size, initial_domain, alignment, r); - mutex_lock(&rdev->ddev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&rdev->ddev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } gobj->driver_private = robj; @@ -202,14 +200,10 @@ int radeon_gem_create_ioctl(struct drm_device *dev, void *data, } r = drm_gem_handle_create(filp, gobj, &handle); if (r) { - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } - mutex_lock(&dev->struct_mutex); - drm_gem_object_handle_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_handle_unreference_unlocked(gobj); args->handle = handle; return 0; } @@ -236,9 +230,7 @@ int radeon_gem_set_domain_ioctl(struct drm_device *dev, void *data, r = radeon_gem_set_domain(gobj, args->read_domains, args->write_domain); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } @@ -255,9 +247,7 @@ int radeon_gem_mmap_ioctl(struct drm_device *dev, void *data, } robj = gobj->driver_private; args->addr_ptr = radeon_bo_mmap_offset(robj); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return 0; } @@ -288,9 +278,7 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data, default: break; } - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } @@ -308,9 +296,10 @@ int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data, } robj = gobj->driver_private; r = radeon_bo_wait(robj, NULL, false); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + /* callback hw specific functions if any */ + if (robj->rdev->asic->ioctl_wait_idle) + robj->rdev->asic->ioctl_wait_idle(robj->rdev, robj); + drm_gem_object_unreference_unlocked(gobj); return r; } @@ -328,9 +317,7 @@ int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data, return -EINVAL; robj = gobj->driver_private; r = radeon_bo_set_tiling_flags(robj, args->tiling_flags, args->pitch); - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } @@ -353,8 +340,6 @@ int radeon_gem_get_tiling_ioctl(struct drm_device *dev, void *data, radeon_bo_get_tiling_flags(rbo, &args->tiling_flags, &args->pitch); radeon_bo_unreserve(rbo); out: - mutex_lock(&dev->struct_mutex); - drm_gem_object_unreference(gobj); - mutex_unlock(&dev->struct_mutex); + drm_gem_object_unreference_unlocked(gobj); return r; } diff --git a/drivers/gpu/drm/radeon/radeon_i2c.c b/drivers/gpu/drm/radeon/radeon_i2c.c index da3da1e89d0..4ae50c19589 100644 --- a/drivers/gpu/drm/radeon/radeon_i2c.c +++ b/drivers/gpu/drm/radeon/radeon_i2c.c @@ -26,6 +26,7 @@ #include "drmP.h" #include "radeon_drm.h" #include "radeon.h" +#include "atom.h" /** * radeon_ddc_probe @@ -59,7 +60,7 @@ bool radeon_ddc_probe(struct radeon_connector *radeon_connector) } -void radeon_i2c_do_lock(struct radeon_i2c_chan *i2c, int lock_state) +static void radeon_i2c_do_lock(struct radeon_i2c_chan *i2c, int lock_state) { struct radeon_device *rdev = i2c->dev->dev_private; struct radeon_i2c_bus_rec *rec = &i2c->rec; @@ -71,13 +72,25 @@ void radeon_i2c_do_lock(struct radeon_i2c_chan *i2c, int lock_state) */ if (rec->hw_capable) { if ((rdev->family >= CHIP_R200) && !ASIC_IS_AVIVO(rdev)) { - if (rec->a_clk_reg == RADEON_GPIO_MONID) { + u32 reg; + + if (rdev->family >= CHIP_RV350) + reg = RADEON_GPIO_MONID; + else if ((rdev->family == CHIP_R300) || + (rdev->family == CHIP_R350)) + reg = RADEON_GPIO_DVI_DDC; + else + reg = RADEON_GPIO_CRT2_DDC; + + mutex_lock(&rdev->dc_hw_i2c_mutex); + if (rec->a_clk_reg == reg) { WREG32(RADEON_DVI_I2C_CNTL_0, (RADEON_I2C_SOFT_RST | R200_DVI_I2C_PIN_SEL(R200_SEL_DDC1))); } else { WREG32(RADEON_DVI_I2C_CNTL_0, (RADEON_I2C_SOFT_RST | R200_DVI_I2C_PIN_SEL(R200_SEL_DDC3))); } + mutex_unlock(&rdev->dc_hw_i2c_mutex); } } @@ -168,6 +181,692 @@ static void set_data(void *i2c_priv, int data) WREG32(rec->en_data_reg, val); } +static u32 radeon_get_i2c_prescale(struct radeon_device *rdev) +{ + struct radeon_pll *spll = &rdev->clock.spll; + u32 sclk = radeon_get_engine_clock(rdev); + u32 prescale = 0; + u32 n, m; + u8 loop; + int i2c_clock; + + switch (rdev->family) { + case CHIP_R100: + case CHIP_RV100: + case CHIP_RS100: + case CHIP_RV200: + case CHIP_RS200: + case CHIP_R200: + case CHIP_RV250: + case CHIP_RS300: + case CHIP_RV280: + case CHIP_R300: + case CHIP_R350: + case CHIP_RV350: + n = (spll->reference_freq) / (4 * 6); + for (loop = 1; loop < 255; loop++) { + if ((loop * (loop - 1)) > n) + break; + } + m = loop - 1; + prescale = m | (loop << 8); + break; + case CHIP_RV380: + case CHIP_RS400: + case CHIP_RS480: + case CHIP_R420: + case CHIP_R423: + case CHIP_RV410: + sclk = radeon_get_engine_clock(rdev); + prescale = (((sclk * 10)/(4 * 128 * 100) + 1) << 8) + 128; + break; + case CHIP_RS600: + case CHIP_RS690: + case CHIP_RS740: + /* todo */ + break; + case CHIP_RV515: + case CHIP_R520: + case CHIP_RV530: + case CHIP_RV560: + case CHIP_RV570: + case CHIP_R580: + i2c_clock = 50; + sclk = radeon_get_engine_clock(rdev); + if (rdev->family == CHIP_R520) + prescale = (127 << 8) + ((sclk * 10) / (4 * 127 * i2c_clock)); + else + prescale = (((sclk * 10)/(4 * 128 * 100) + 1) << 8) + 128; + break; + case CHIP_R600: + case CHIP_RV610: + case CHIP_RV630: + case CHIP_RV670: + /* todo */ + break; + case CHIP_RV620: + case CHIP_RV635: + case CHIP_RS780: + case CHIP_RS880: + case CHIP_RV770: + case CHIP_RV730: + case CHIP_RV710: + case CHIP_RV740: + /* todo */ + break; + case CHIP_CEDAR: + case CHIP_REDWOOD: + case CHIP_JUNIPER: + case CHIP_CYPRESS: + case CHIP_HEMLOCK: + /* todo */ + break; + default: + DRM_ERROR("i2c: unhandled radeon chip\n"); + break; + } + return prescale; +} + + +/* hw i2c engine for r1xx-4xx hardware + * hw can buffer up to 15 bytes + */ +static int r100_hw_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num) +{ + struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); + struct radeon_device *rdev = i2c->dev->dev_private; + struct radeon_i2c_bus_rec *rec = &i2c->rec; + struct i2c_msg *p; + int i, j, k, ret = num; + u32 prescale; + u32 i2c_cntl_0, i2c_cntl_1, i2c_data; + u32 tmp, reg; + + mutex_lock(&rdev->dc_hw_i2c_mutex); + /* take the pm lock since we need a constant sclk */ + mutex_lock(&rdev->pm.mutex); + + prescale = radeon_get_i2c_prescale(rdev); + + reg = ((prescale << RADEON_I2C_PRESCALE_SHIFT) | + RADEON_I2C_START | + RADEON_I2C_STOP | + RADEON_I2C_GO); + + if (rdev->is_atom_bios) { + tmp = RREG32(RADEON_BIOS_6_SCRATCH); + WREG32(RADEON_BIOS_6_SCRATCH, tmp | ATOM_S6_HW_I2C_BUSY_STATE); + } + + if (rec->mm_i2c) { + i2c_cntl_0 = RADEON_I2C_CNTL_0; + i2c_cntl_1 = RADEON_I2C_CNTL_1; + i2c_data = RADEON_I2C_DATA; + } else { + i2c_cntl_0 = RADEON_DVI_I2C_CNTL_0; + i2c_cntl_1 = RADEON_DVI_I2C_CNTL_1; + i2c_data = RADEON_DVI_I2C_DATA; + + switch (rdev->family) { + case CHIP_R100: + case CHIP_RV100: + case CHIP_RS100: + case CHIP_RV200: + case CHIP_RS200: + case CHIP_RS300: + switch (rec->mask_clk_reg) { + case RADEON_GPIO_DVI_DDC: + /* no gpio select bit */ + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + break; + case CHIP_R200: + /* only bit 4 on r200 */ + switch (rec->mask_clk_reg) { + case RADEON_GPIO_DVI_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC1); + break; + case RADEON_GPIO_MONID: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC3); + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + break; + case CHIP_RV250: + case CHIP_RV280: + /* bits 3 and 4 */ + switch (rec->mask_clk_reg) { + case RADEON_GPIO_DVI_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC1); + break; + case RADEON_GPIO_VGA_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC2); + break; + case RADEON_GPIO_CRT2_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC3); + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + break; + case CHIP_R300: + case CHIP_R350: + /* only bit 4 on r300/r350 */ + switch (rec->mask_clk_reg) { + case RADEON_GPIO_VGA_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC1); + break; + case RADEON_GPIO_DVI_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC3); + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + break; + case CHIP_RV350: + case CHIP_RV380: + case CHIP_R420: + case CHIP_R423: + case CHIP_RV410: + case CHIP_RS400: + case CHIP_RS480: + /* bits 3 and 4 */ + switch (rec->mask_clk_reg) { + case RADEON_GPIO_VGA_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC1); + break; + case RADEON_GPIO_DVI_DDC: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC2); + break; + case RADEON_GPIO_MONID: + reg |= R200_DVI_I2C_PIN_SEL(R200_SEL_DDC3); + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + break; + default: + DRM_ERROR("unsupported asic\n"); + ret = -EINVAL; + goto done; + break; + } + } + + /* check for bus probe */ + p = &msgs[0]; + if ((num == 1) && (p->len == 0)) { + WREG32(i2c_cntl_0, (RADEON_I2C_DONE | + RADEON_I2C_NACK | + RADEON_I2C_HALT | + RADEON_I2C_SOFT_RST)); + WREG32(i2c_data, (p->addr << 1) & 0xff); + WREG32(i2c_data, 0); + WREG32(i2c_cntl_1, ((1 << RADEON_I2C_DATA_COUNT_SHIFT) | + (1 << RADEON_I2C_ADDR_COUNT_SHIFT) | + RADEON_I2C_EN | + (48 << RADEON_I2C_TIME_LIMIT_SHIFT))); + WREG32(i2c_cntl_0, reg); + for (k = 0; k < 32; k++) { + udelay(10); + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_GO) + continue; + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_DONE) + break; + else { + DRM_DEBUG("i2c write error 0x%08x\n", tmp); + WREG32(i2c_cntl_0, tmp | RADEON_I2C_ABORT); + ret = -EIO; + goto done; + } + } + goto done; + } + + for (i = 0; i < num; i++) { + p = &msgs[i]; + for (j = 0; j < p->len; j++) { + if (p->flags & I2C_M_RD) { + WREG32(i2c_cntl_0, (RADEON_I2C_DONE | + RADEON_I2C_NACK | + RADEON_I2C_HALT | + RADEON_I2C_SOFT_RST)); + WREG32(i2c_data, ((p->addr << 1) & 0xff) | 0x1); + WREG32(i2c_cntl_1, ((1 << RADEON_I2C_DATA_COUNT_SHIFT) | + (1 << RADEON_I2C_ADDR_COUNT_SHIFT) | + RADEON_I2C_EN | + (48 << RADEON_I2C_TIME_LIMIT_SHIFT))); + WREG32(i2c_cntl_0, reg | RADEON_I2C_RECEIVE); + for (k = 0; k < 32; k++) { + udelay(10); + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_GO) + continue; + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_DONE) + break; + else { + DRM_DEBUG("i2c read error 0x%08x\n", tmp); + WREG32(i2c_cntl_0, tmp | RADEON_I2C_ABORT); + ret = -EIO; + goto done; + } + } + p->buf[j] = RREG32(i2c_data) & 0xff; + } else { + WREG32(i2c_cntl_0, (RADEON_I2C_DONE | + RADEON_I2C_NACK | + RADEON_I2C_HALT | + RADEON_I2C_SOFT_RST)); + WREG32(i2c_data, (p->addr << 1) & 0xff); + WREG32(i2c_data, p->buf[j]); + WREG32(i2c_cntl_1, ((1 << RADEON_I2C_DATA_COUNT_SHIFT) | + (1 << RADEON_I2C_ADDR_COUNT_SHIFT) | + RADEON_I2C_EN | + (48 << RADEON_I2C_TIME_LIMIT_SHIFT))); + WREG32(i2c_cntl_0, reg); + for (k = 0; k < 32; k++) { + udelay(10); + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_GO) + continue; + tmp = RREG32(i2c_cntl_0); + if (tmp & RADEON_I2C_DONE) + break; + else { + DRM_DEBUG("i2c write error 0x%08x\n", tmp); + WREG32(i2c_cntl_0, tmp | RADEON_I2C_ABORT); + ret = -EIO; + goto done; + } + } + } + } + } + +done: + WREG32(i2c_cntl_0, 0); + WREG32(i2c_cntl_1, 0); + WREG32(i2c_cntl_0, (RADEON_I2C_DONE | + RADEON_I2C_NACK | + RADEON_I2C_HALT | + RADEON_I2C_SOFT_RST)); + + if (rdev->is_atom_bios) { + tmp = RREG32(RADEON_BIOS_6_SCRATCH); + tmp &= ~ATOM_S6_HW_I2C_BUSY_STATE; + WREG32(RADEON_BIOS_6_SCRATCH, tmp); + } + + mutex_unlock(&rdev->pm.mutex); + mutex_unlock(&rdev->dc_hw_i2c_mutex); + + return ret; +} + +/* hw i2c engine for r5xx hardware + * hw can buffer up to 15 bytes + */ +static int r500_hw_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num) +{ + struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); + struct radeon_device *rdev = i2c->dev->dev_private; + struct radeon_i2c_bus_rec *rec = &i2c->rec; + struct i2c_msg *p; + int i, j, remaining, current_count, buffer_offset, ret = num; + u32 prescale; + u32 tmp, reg; + u32 saved1, saved2; + + mutex_lock(&rdev->dc_hw_i2c_mutex); + /* take the pm lock since we need a constant sclk */ + mutex_lock(&rdev->pm.mutex); + + prescale = radeon_get_i2c_prescale(rdev); + + /* clear gpio mask bits */ + tmp = RREG32(rec->mask_clk_reg); + tmp &= ~rec->mask_clk_mask; + WREG32(rec->mask_clk_reg, tmp); + tmp = RREG32(rec->mask_clk_reg); + + tmp = RREG32(rec->mask_data_reg); + tmp &= ~rec->mask_data_mask; + WREG32(rec->mask_data_reg, tmp); + tmp = RREG32(rec->mask_data_reg); + + /* clear pin values */ + tmp = RREG32(rec->a_clk_reg); + tmp &= ~rec->a_clk_mask; + WREG32(rec->a_clk_reg, tmp); + tmp = RREG32(rec->a_clk_reg); + + tmp = RREG32(rec->a_data_reg); + tmp &= ~rec->a_data_mask; + WREG32(rec->a_data_reg, tmp); + tmp = RREG32(rec->a_data_reg); + + /* set the pins to input */ + tmp = RREG32(rec->en_clk_reg); + tmp &= ~rec->en_clk_mask; + WREG32(rec->en_clk_reg, tmp); + tmp = RREG32(rec->en_clk_reg); + + tmp = RREG32(rec->en_data_reg); + tmp &= ~rec->en_data_mask; + WREG32(rec->en_data_reg, tmp); + tmp = RREG32(rec->en_data_reg); + + /* */ + tmp = RREG32(RADEON_BIOS_6_SCRATCH); + WREG32(RADEON_BIOS_6_SCRATCH, tmp | ATOM_S6_HW_I2C_BUSY_STATE); + saved1 = RREG32(AVIVO_DC_I2C_CONTROL1); + saved2 = RREG32(0x494); + WREG32(0x494, saved2 | 0x1); + + WREG32(AVIVO_DC_I2C_ARBITRATION, AVIVO_DC_I2C_SW_WANTS_TO_USE_I2C); + for (i = 0; i < 50; i++) { + udelay(1); + if (RREG32(AVIVO_DC_I2C_ARBITRATION) & AVIVO_DC_I2C_SW_CAN_USE_I2C) + break; + } + if (i == 50) { + DRM_ERROR("failed to get i2c bus\n"); + ret = -EBUSY; + goto done; + } + + reg = AVIVO_DC_I2C_START | AVIVO_DC_I2C_STOP | AVIVO_DC_I2C_EN; + switch (rec->mask_clk_reg) { + case AVIVO_DC_GPIO_DDC1_MASK: + reg |= AVIVO_DC_I2C_PIN_SELECT(AVIVO_SEL_DDC1); + break; + case AVIVO_DC_GPIO_DDC2_MASK: + reg |= AVIVO_DC_I2C_PIN_SELECT(AVIVO_SEL_DDC2); + break; + case AVIVO_DC_GPIO_DDC3_MASK: + reg |= AVIVO_DC_I2C_PIN_SELECT(AVIVO_SEL_DDC3); + break; + default: + DRM_ERROR("gpio not supported with hw i2c\n"); + ret = -EINVAL; + goto done; + } + + /* check for bus probe */ + p = &msgs[0]; + if ((num == 1) && (p->len == 0)) { + WREG32(AVIVO_DC_I2C_STATUS1, (AVIVO_DC_I2C_DONE | + AVIVO_DC_I2C_NACK | + AVIVO_DC_I2C_HALT)); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_SOFT_RESET); + udelay(1); + WREG32(AVIVO_DC_I2C_RESET, 0); + + WREG32(AVIVO_DC_I2C_DATA, (p->addr << 1) & 0xff); + WREG32(AVIVO_DC_I2C_DATA, 0); + + WREG32(AVIVO_DC_I2C_CONTROL3, AVIVO_DC_I2C_TIME_LIMIT(48)); + WREG32(AVIVO_DC_I2C_CONTROL2, (AVIVO_DC_I2C_ADDR_COUNT(1) | + AVIVO_DC_I2C_DATA_COUNT(1) | + (prescale << 16))); + WREG32(AVIVO_DC_I2C_CONTROL1, reg); + WREG32(AVIVO_DC_I2C_STATUS1, AVIVO_DC_I2C_GO); + for (j = 0; j < 200; j++) { + udelay(50); + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_GO) + continue; + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_DONE) + break; + else { + DRM_DEBUG("i2c write error 0x%08x\n", tmp); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_ABORT); + ret = -EIO; + goto done; + } + } + goto done; + } + + for (i = 0; i < num; i++) { + p = &msgs[i]; + remaining = p->len; + buffer_offset = 0; + if (p->flags & I2C_M_RD) { + while (remaining) { + if (remaining > 15) + current_count = 15; + else + current_count = remaining; + WREG32(AVIVO_DC_I2C_STATUS1, (AVIVO_DC_I2C_DONE | + AVIVO_DC_I2C_NACK | + AVIVO_DC_I2C_HALT)); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_SOFT_RESET); + udelay(1); + WREG32(AVIVO_DC_I2C_RESET, 0); + + WREG32(AVIVO_DC_I2C_DATA, ((p->addr << 1) & 0xff) | 0x1); + WREG32(AVIVO_DC_I2C_CONTROL3, AVIVO_DC_I2C_TIME_LIMIT(48)); + WREG32(AVIVO_DC_I2C_CONTROL2, (AVIVO_DC_I2C_ADDR_COUNT(1) | + AVIVO_DC_I2C_DATA_COUNT(current_count) | + (prescale << 16))); + WREG32(AVIVO_DC_I2C_CONTROL1, reg | AVIVO_DC_I2C_RECEIVE); + WREG32(AVIVO_DC_I2C_STATUS1, AVIVO_DC_I2C_GO); + for (j = 0; j < 200; j++) { + udelay(50); + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_GO) + continue; + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_DONE) + break; + else { + DRM_DEBUG("i2c read error 0x%08x\n", tmp); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_ABORT); + ret = -EIO; + goto done; + } + } + for (j = 0; j < current_count; j++) + p->buf[buffer_offset + j] = RREG32(AVIVO_DC_I2C_DATA) & 0xff; + remaining -= current_count; + buffer_offset += current_count; + } + } else { + while (remaining) { + if (remaining > 15) + current_count = 15; + else + current_count = remaining; + WREG32(AVIVO_DC_I2C_STATUS1, (AVIVO_DC_I2C_DONE | + AVIVO_DC_I2C_NACK | + AVIVO_DC_I2C_HALT)); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_SOFT_RESET); + udelay(1); + WREG32(AVIVO_DC_I2C_RESET, 0); + + WREG32(AVIVO_DC_I2C_DATA, (p->addr << 1) & 0xff); + for (j = 0; j < current_count; j++) + WREG32(AVIVO_DC_I2C_DATA, p->buf[buffer_offset + j]); + + WREG32(AVIVO_DC_I2C_CONTROL3, AVIVO_DC_I2C_TIME_LIMIT(48)); + WREG32(AVIVO_DC_I2C_CONTROL2, (AVIVO_DC_I2C_ADDR_COUNT(1) | + AVIVO_DC_I2C_DATA_COUNT(current_count) | + (prescale << 16))); + WREG32(AVIVO_DC_I2C_CONTROL1, reg); + WREG32(AVIVO_DC_I2C_STATUS1, AVIVO_DC_I2C_GO); + for (j = 0; j < 200; j++) { + udelay(50); + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_GO) + continue; + tmp = RREG32(AVIVO_DC_I2C_STATUS1); + if (tmp & AVIVO_DC_I2C_DONE) + break; + else { + DRM_DEBUG("i2c write error 0x%08x\n", tmp); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_ABORT); + ret = -EIO; + goto done; + } + } + remaining -= current_count; + buffer_offset += current_count; + } + } + } + +done: + WREG32(AVIVO_DC_I2C_STATUS1, (AVIVO_DC_I2C_DONE | + AVIVO_DC_I2C_NACK | + AVIVO_DC_I2C_HALT)); + WREG32(AVIVO_DC_I2C_RESET, AVIVO_DC_I2C_SOFT_RESET); + udelay(1); + WREG32(AVIVO_DC_I2C_RESET, 0); + + WREG32(AVIVO_DC_I2C_ARBITRATION, AVIVO_DC_I2C_SW_DONE_USING_I2C); + WREG32(AVIVO_DC_I2C_CONTROL1, saved1); + WREG32(0x494, saved2); + tmp = RREG32(RADEON_BIOS_6_SCRATCH); + tmp &= ~ATOM_S6_HW_I2C_BUSY_STATE; + WREG32(RADEON_BIOS_6_SCRATCH, tmp); + + mutex_unlock(&rdev->pm.mutex); + mutex_unlock(&rdev->dc_hw_i2c_mutex); + + return ret; +} + +static int radeon_sw_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num) +{ + struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); + int ret; + + radeon_i2c_do_lock(i2c, 1); + ret = i2c_transfer(&i2c->algo.radeon.bit_adapter, msgs, num); + radeon_i2c_do_lock(i2c, 0); + + return ret; +} + +static int radeon_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num) +{ + struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); + struct radeon_device *rdev = i2c->dev->dev_private; + struct radeon_i2c_bus_rec *rec = &i2c->rec; + int ret; + + switch (rdev->family) { + case CHIP_R100: + case CHIP_RV100: + case CHIP_RS100: + case CHIP_RV200: + case CHIP_RS200: + case CHIP_R200: + case CHIP_RV250: + case CHIP_RS300: + case CHIP_RV280: + case CHIP_R300: + case CHIP_R350: + case CHIP_RV350: + case CHIP_RV380: + case CHIP_R420: + case CHIP_R423: + case CHIP_RV410: + case CHIP_RS400: + case CHIP_RS480: + if (rec->hw_capable) + ret = r100_hw_i2c_xfer(i2c_adap, msgs, num); + else + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + case CHIP_RS600: + case CHIP_RS690: + case CHIP_RS740: + /* XXX fill in hw i2c implementation */ + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + case CHIP_RV515: + case CHIP_R520: + case CHIP_RV530: + case CHIP_RV560: + case CHIP_RV570: + case CHIP_R580: + if (rec->hw_capable) { + if (rec->mm_i2c) + ret = r100_hw_i2c_xfer(i2c_adap, msgs, num); + else + ret = r500_hw_i2c_xfer(i2c_adap, msgs, num); + } else + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + case CHIP_R600: + case CHIP_RV610: + case CHIP_RV630: + case CHIP_RV670: + /* XXX fill in hw i2c implementation */ + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + case CHIP_RV620: + case CHIP_RV635: + case CHIP_RS780: + case CHIP_RS880: + case CHIP_RV770: + case CHIP_RV730: + case CHIP_RV710: + case CHIP_RV740: + /* XXX fill in hw i2c implementation */ + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + case CHIP_CEDAR: + case CHIP_REDWOOD: + case CHIP_JUNIPER: + case CHIP_CYPRESS: + case CHIP_HEMLOCK: + /* XXX fill in hw i2c implementation */ + ret = radeon_sw_i2c_xfer(i2c_adap, msgs, num); + break; + default: + DRM_ERROR("i2c: unhandled radeon chip\n"); + ret = -EIO; + break; + } + + return ret; +} + +static u32 radeon_i2c_func(struct i2c_adapter *adap) +{ + return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL; +} + +static const struct i2c_algorithm radeon_i2c_algo = { + .master_xfer = radeon_i2c_xfer, + .functionality = radeon_i2c_func, +}; + struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, struct radeon_i2c_bus_rec *rec, const char *name) @@ -179,23 +878,36 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, if (i2c == NULL) return NULL; - i2c->adapter.owner = THIS_MODULE; - i2c->dev = dev; - i2c_set_adapdata(&i2c->adapter, i2c); - i2c->adapter.algo_data = &i2c->algo.bit; - i2c->algo.bit.setsda = set_data; - i2c->algo.bit.setscl = set_clock; - i2c->algo.bit.getsda = get_data; - i2c->algo.bit.getscl = get_clock; - i2c->algo.bit.udelay = 20; + /* set the internal bit adapter */ + i2c->algo.radeon.bit_adapter.owner = THIS_MODULE; + i2c_set_adapdata(&i2c->algo.radeon.bit_adapter, i2c); + sprintf(i2c->algo.radeon.bit_adapter.name, "Radeon internal i2c bit bus %s", name); + i2c->algo.radeon.bit_adapter.algo_data = &i2c->algo.radeon.bit_data; + i2c->algo.radeon.bit_data.setsda = set_data; + i2c->algo.radeon.bit_data.setscl = set_clock; + i2c->algo.radeon.bit_data.getsda = get_data; + i2c->algo.radeon.bit_data.getscl = get_clock; + i2c->algo.radeon.bit_data.udelay = 20; /* vesa says 2.2 ms is enough, 1 jiffy doesn't seem to always * make this, 2 jiffies is a lot more reliable */ - i2c->algo.bit.timeout = 2; - i2c->algo.bit.data = i2c; + i2c->algo.radeon.bit_data.timeout = 2; + i2c->algo.radeon.bit_data.data = i2c; + ret = i2c_bit_add_bus(&i2c->algo.radeon.bit_adapter); + if (ret) { + DRM_ERROR("Failed to register internal bit i2c %s\n", name); + goto out_free; + } + /* set the radeon i2c adapter */ + i2c->dev = dev; i2c->rec = *rec; - ret = i2c_bit_add_bus(&i2c->adapter); + i2c->adapter.owner = THIS_MODULE; + i2c_set_adapdata(&i2c->adapter, i2c); + sprintf(i2c->adapter.name, "Radeon i2c %s", name); + i2c->adapter.algo_data = &i2c->algo.radeon; + i2c->adapter.algo = &radeon_i2c_algo; + ret = i2c_add_adapter(&i2c->adapter); if (ret) { - DRM_INFO("Failed to register i2c %s\n", name); + DRM_ERROR("Failed to register i2c %s\n", name); goto out_free; } @@ -237,11 +949,19 @@ out_free: } - void radeon_i2c_destroy(struct radeon_i2c_chan *i2c) { if (!i2c) return; + i2c_del_adapter(&i2c->algo.radeon.bit_adapter); + i2c_del_adapter(&i2c->adapter); + kfree(i2c); +} + +void radeon_i2c_destroy_dp(struct radeon_i2c_chan *i2c) +{ + if (!i2c) + return; i2c_del_adapter(&i2c->adapter); kfree(i2c); @@ -252,10 +972,10 @@ struct drm_encoder *radeon_best_encoder(struct drm_connector *connector) return NULL; } -void radeon_i2c_sw_get_byte(struct radeon_i2c_chan *i2c_bus, - u8 slave_addr, - u8 addr, - u8 *val) +void radeon_i2c_get_byte(struct radeon_i2c_chan *i2c_bus, + u8 slave_addr, + u8 addr, + u8 *val) { u8 out_buf[2]; u8 in_buf[2]; @@ -286,10 +1006,10 @@ void radeon_i2c_sw_get_byte(struct radeon_i2c_chan *i2c_bus, } } -void radeon_i2c_sw_put_byte(struct radeon_i2c_chan *i2c_bus, - u8 slave_addr, - u8 addr, - u8 val) +void radeon_i2c_put_byte(struct radeon_i2c_chan *i2c_bus, + u8 slave_addr, + u8 addr, + u8 val) { uint8_t out_buf[2]; struct i2c_msg msg = { diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index f23b05606eb..20ec276e759 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -30,6 +30,8 @@ #include "radeon.h" #include "radeon_drm.h" +#include <linux/vga_switcheroo.h> + int radeon_driver_unload_kms(struct drm_device *dev) { struct radeon_device *rdev = dev->dev_private; @@ -136,6 +138,7 @@ int radeon_driver_firstopen_kms(struct drm_device *dev) void radeon_driver_lastclose_kms(struct drm_device *dev) { + vga_switcheroo_process_delayed_switch(); } int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) @@ -276,17 +279,17 @@ struct drm_ioctl_desc radeon_ioctls_kms[] = { DRM_IOCTL_DEF(DRM_RADEON_SURF_ALLOC, radeon_surface_alloc_kms, DRM_AUTH), DRM_IOCTL_DEF(DRM_RADEON_SURF_FREE, radeon_surface_free_kms, DRM_AUTH), /* KMS */ - DRM_IOCTL_DEF(DRM_RADEON_GEM_INFO, radeon_gem_info_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_CREATE, radeon_gem_create_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_MMAP, radeon_gem_mmap_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_SET_DOMAIN, radeon_gem_set_domain_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_PREAD, radeon_gem_pread_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_PWRITE, radeon_gem_pwrite_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_WAIT_IDLE, radeon_gem_wait_idle_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_CS, radeon_cs_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_INFO, radeon_info_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH), - DRM_IOCTL_DEF(DRM_RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH), + DRM_IOCTL_DEF(DRM_RADEON_GEM_INFO, radeon_gem_info_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_CREATE, radeon_gem_create_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_MMAP, radeon_gem_mmap_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_SET_DOMAIN, radeon_gem_set_domain_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_PREAD, radeon_gem_pread_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_PWRITE, radeon_gem_pwrite_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_WAIT_IDLE, radeon_gem_wait_idle_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_CS, radeon_cs_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_INFO, radeon_info_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED), }; int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms); diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c index cc27485a07a..df23d6a01d0 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c @@ -339,69 +339,6 @@ void radeon_crtc_dpms(struct drm_crtc *crtc, int mode) } } -/* properly set crtc bpp when using atombios */ -void radeon_legacy_atom_set_surface(struct drm_crtc *crtc) -{ - struct drm_device *dev = crtc->dev; - struct radeon_device *rdev = dev->dev_private; - struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); - int format; - uint32_t crtc_gen_cntl; - uint32_t disp_merge_cntl; - uint32_t crtc_pitch; - - switch (crtc->fb->bits_per_pixel) { - case 8: - format = 2; - break; - case 15: /* 555 */ - format = 3; - break; - case 16: /* 565 */ - format = 4; - break; - case 24: /* RGB */ - format = 5; - break; - case 32: /* xRGB */ - format = 6; - break; - default: - return; - } - - crtc_pitch = ((((crtc->fb->pitch / (crtc->fb->bits_per_pixel / 8)) * crtc->fb->bits_per_pixel) + - ((crtc->fb->bits_per_pixel * 8) - 1)) / - (crtc->fb->bits_per_pixel * 8)); - crtc_pitch |= crtc_pitch << 16; - - WREG32(RADEON_CRTC_PITCH + radeon_crtc->crtc_offset, crtc_pitch); - - switch (radeon_crtc->crtc_id) { - case 0: - disp_merge_cntl = RREG32(RADEON_DISP_MERGE_CNTL); - disp_merge_cntl &= ~RADEON_DISP_RGB_OFFSET_EN; - WREG32(RADEON_DISP_MERGE_CNTL, disp_merge_cntl); - - crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL) & 0xfffff0ff; - crtc_gen_cntl |= (format << 8); - crtc_gen_cntl |= RADEON_CRTC_EXT_DISP_EN; - WREG32(RADEON_CRTC_GEN_CNTL, crtc_gen_cntl); - break; - case 1: - disp_merge_cntl = RREG32(RADEON_DISP2_MERGE_CNTL); - disp_merge_cntl &= ~RADEON_DISP2_RGB_OFFSET_EN; - WREG32(RADEON_DISP2_MERGE_CNTL, disp_merge_cntl); - - crtc_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL) & 0xfffff0ff; - crtc_gen_cntl |= (format << 8); - WREG32(RADEON_CRTC2_GEN_CNTL, crtc_gen_cntl); - WREG32(RADEON_FP_H2_SYNC_STRT_WID, RREG32(RADEON_CRTC2_H_SYNC_STRT_WID)); - WREG32(RADEON_FP_V2_SYNC_STRT_WID, RREG32(RADEON_CRTC2_V_SYNC_STRT_WID)); - break; - } -} - int radeon_crtc_set_base(struct drm_crtc *crtc, int x, int y, struct drm_framebuffer *old_fb) { @@ -466,7 +403,7 @@ int radeon_crtc_set_base(struct drm_crtc *crtc, int x, int y, /* if scanout was in GTT this really wouldn't work */ /* crtc offset is from display base addr not FB location */ - radeon_crtc->legacy_display_base_addr = rdev->mc.vram_location; + radeon_crtc->legacy_display_base_addr = rdev->mc.vram_start; base -= radeon_crtc->legacy_display_base_addr; @@ -645,29 +582,6 @@ static bool radeon_set_crtc_timing(struct drm_crtc *crtc, struct drm_display_mod ? RADEON_CRTC_V_SYNC_POL : 0)); - /* TODO -> Dell Server */ - if (0) { - uint32_t disp_hw_debug = RREG32(RADEON_DISP_HW_DEBUG); - uint32_t tv_dac_cntl = RREG32(RADEON_TV_DAC_CNTL); - uint32_t dac2_cntl = RREG32(RADEON_DAC_CNTL2); - uint32_t crtc2_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL); - - dac2_cntl &= ~RADEON_DAC2_DAC_CLK_SEL; - dac2_cntl |= RADEON_DAC2_DAC2_CLK_SEL; - - /* For CRT on DAC2, don't turn it on if BIOS didn't - enable it, even it's detected. - */ - disp_hw_debug |= RADEON_CRT2_DISP1_SEL; - tv_dac_cntl &= ~((1<<2) | (3<<8) | (7<<24) | (0xff<<16)); - tv_dac_cntl |= (0x03 | (2<<8) | (0x58<<16)); - - WREG32(RADEON_TV_DAC_CNTL, tv_dac_cntl); - WREG32(RADEON_DISP_HW_DEBUG, disp_hw_debug); - WREG32(RADEON_DAC_CNTL2, dac2_cntl); - WREG32(RADEON_CRTC2_GEN_CNTL, crtc2_gen_cntl); - } - if (radeon_crtc->crtc_id) { uint32_t crtc2_gen_cntl; uint32_t disp2_merge_cntl; @@ -755,7 +669,6 @@ static void radeon_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) uint32_t post_divider = 0; uint32_t freq = 0; uint8_t pll_gain; - int pll_flags = RADEON_PLL_LEGACY; bool use_bios_divs = false; /* PLL registers */ uint32_t pll_ref_div = 0; @@ -789,10 +702,16 @@ static void radeon_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) else pll = &rdev->clock.p1pll; + pll->flags = RADEON_PLL_LEGACY; + if (radeon_new_pll == 1) + pll->algo = PLL_ALGO_NEW; + else + pll->algo = PLL_ALGO_LEGACY; + if (mode->clock > 200000) /* range limits??? */ - pll_flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; + pll->flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else - pll_flags |= RADEON_PLL_PREFER_LOW_REF_DIV; + pll->flags |= RADEON_PLL_PREFER_LOW_REF_DIV; list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) { @@ -804,7 +723,7 @@ static void radeon_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) } if (encoder->encoder_type != DRM_MODE_ENCODER_DAC) - pll_flags |= RADEON_PLL_NO_ODD_POST_DIV; + pll->flags |= RADEON_PLL_NO_ODD_POST_DIV; if (encoder->encoder_type == DRM_MODE_ENCODER_LVDS) { if (!rdev->is_atom_bios) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); @@ -819,7 +738,7 @@ static void radeon_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) } } } - pll_flags |= RADEON_PLL_USE_REF_DIV; + pll->flags |= RADEON_PLL_USE_REF_DIV; } } } @@ -829,8 +748,7 @@ static void radeon_set_pll(struct drm_crtc *crtc, struct drm_display_mode *mode) if (!use_bios_divs) { radeon_compute_pll(pll, mode->clock, &freq, &feedback_div, &frac_fb_div, - &reference_div, &post_divider, - pll_flags); + &reference_div, &post_divider); for (post_div = &post_divs[0]; post_div->divider; ++post_div) { if (post_div->divider == post_divider) diff --git a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c index 38e45e231ef..cf389ce50a8 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c @@ -115,6 +115,9 @@ static void radeon_legacy_lvds_dpms(struct drm_encoder *encoder, int mode) radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); else radeon_combios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } static void radeon_legacy_lvds_prepare(struct drm_encoder *encoder) @@ -214,6 +217,11 @@ static bool radeon_legacy_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *adjusted_mode) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + + /* adjust pm to upcoming mode change */ + radeon_pm_compute_clocks(rdev); /* set the active encoder to connector routing */ radeon_encoder_set_active_device(encoder); @@ -285,6 +293,9 @@ static void radeon_legacy_primary_dac_dpms(struct drm_encoder *encoder, int mode radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); else radeon_combios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } static void radeon_legacy_primary_dac_prepare(struct drm_encoder *encoder) @@ -470,6 +481,9 @@ static void radeon_legacy_tmds_int_dpms(struct drm_encoder *encoder, int mode) radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); else radeon_combios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } static void radeon_legacy_tmds_int_prepare(struct drm_encoder *encoder) @@ -635,6 +649,9 @@ static void radeon_legacy_tmds_ext_dpms(struct drm_encoder *encoder, int mode) radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); else radeon_combios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } static void radeon_legacy_tmds_ext_prepare(struct drm_encoder *encoder) @@ -842,6 +859,9 @@ static void radeon_legacy_tv_dac_dpms(struct drm_encoder *encoder, int mode) radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); else radeon_combios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false); + + /* adjust pm to dpms change */ + radeon_pm_compute_clocks(rdev); } static void radeon_legacy_tv_dac_prepare(struct drm_encoder *encoder) diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index 91cb041cb40..1702b820aa4 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -83,6 +83,8 @@ struct radeon_i2c_bus_rec { bool valid; /* id used by atom */ uint8_t i2c_id; + /* id used by atom */ + uint8_t hpd_id; /* can be used with hw i2c engine */ bool hw_capable; /* uses multi-media i2c engine */ @@ -113,6 +115,7 @@ struct radeon_tmds_pll { #define RADEON_MAX_BIOS_CONNECTOR 16 +/* pll flags */ #define RADEON_PLL_USE_BIOS_DIVS (1 << 0) #define RADEON_PLL_NO_ODD_POST_DIV (1 << 1) #define RADEON_PLL_USE_REF_DIV (1 << 2) @@ -125,16 +128,30 @@ struct radeon_tmds_pll { #define RADEON_PLL_PREFER_HIGH_POST_DIV (1 << 9) #define RADEON_PLL_USE_FRAC_FB_DIV (1 << 10) #define RADEON_PLL_PREFER_CLOSEST_LOWER (1 << 11) +#define RADEON_PLL_USE_POST_DIV (1 << 12) + +/* pll algo */ +enum radeon_pll_algo { + PLL_ALGO_LEGACY, + PLL_ALGO_NEW +}; struct radeon_pll { - uint16_t reference_freq; - uint16_t reference_div; + /* reference frequency */ + uint32_t reference_freq; + + /* fixed dividers */ + uint32_t reference_div; + uint32_t post_div; + + /* pll in/out limits */ uint32_t pll_in_min; uint32_t pll_in_max; uint32_t pll_out_min; uint32_t pll_out_max; - uint16_t xclk; + uint32_t best_vco; + /* divider limits */ uint32_t min_ref_div; uint32_t max_ref_div; uint32_t min_post_div; @@ -143,7 +160,19 @@ struct radeon_pll { uint32_t max_feedback_div; uint32_t min_frac_feedback_div; uint32_t max_frac_feedback_div; - uint32_t best_vco; + + /* flags for the current clock */ + uint32_t flags; + + /* pll id */ + uint32_t id; + /* pll algo */ + enum radeon_pll_algo algo; +}; + +struct i2c_algo_radeon_data { + struct i2c_adapter bit_adapter; + struct i2c_algo_bit_data bit_data; }; struct radeon_i2c_chan { @@ -151,7 +180,7 @@ struct radeon_i2c_chan { struct drm_device *dev; union { struct i2c_algo_dp_aux_data dp; - struct i2c_algo_bit_data bit; + struct i2c_algo_radeon_data radeon; } algo; struct radeon_i2c_bus_rec rec; }; @@ -180,7 +209,7 @@ struct radeon_mode_info { struct card_info *atom_card_info; enum radeon_connector_table connector_table; bool mode_config_initialized; - struct radeon_crtc *crtcs[2]; + struct radeon_crtc *crtcs[6]; /* DVI-I properties */ struct drm_property *coherent_mode_property; /* DAC enable load detect */ @@ -189,7 +218,8 @@ struct radeon_mode_info { struct drm_property *tv_std_property; /* legacy TMDS PLL detect */ struct drm_property *tmds_pll_property; - + /* hardcoded DFP edid from BIOS */ + struct edid *bios_hardcoded_edid; }; #define MAX_H_CODE_TIMING_LEN 32 @@ -224,6 +254,7 @@ struct radeon_crtc { fixed20_12 vsc; fixed20_12 hsc; struct drm_display_mode native_mode; + int pll_id; }; struct radeon_encoder_primary_dac { @@ -286,10 +317,11 @@ struct radeon_atom_ss { struct radeon_encoder_atom_dig { /* atom dig */ bool coherent_mode; - int dig_block; + int dig_encoder; /* -1 disabled, 0 DIGA, 1 DIGB */ /* atom lvds */ uint32_t lvds_misc; uint16_t panel_pwr_delay; + enum radeon_pll_algo pll_algo; struct radeon_atom_ss *ss; /* panel mode */ struct drm_display_mode native_mode; @@ -385,6 +417,7 @@ extern void dp_link_train(struct drm_encoder *encoder, struct drm_connector *connector); extern u8 radeon_dp_getsinktype(struct radeon_connector *radeon_connector); extern bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector); +extern void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action); extern void atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t lane_num, uint8_t lane_set); @@ -398,14 +431,15 @@ extern struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, struct radeon_i2c_bus_rec *rec, const char *name); extern void radeon_i2c_destroy(struct radeon_i2c_chan *i2c); -extern void radeon_i2c_sw_get_byte(struct radeon_i2c_chan *i2c_bus, - u8 slave_addr, - u8 addr, - u8 *val); -extern void radeon_i2c_sw_put_byte(struct radeon_i2c_chan *i2c, - u8 slave_addr, - u8 addr, - u8 val); +extern void radeon_i2c_destroy_dp(struct radeon_i2c_chan *i2c); +extern void radeon_i2c_get_byte(struct radeon_i2c_chan *i2c_bus, + u8 slave_addr, + u8 addr, + u8 *val); +extern void radeon_i2c_put_byte(struct radeon_i2c_chan *i2c, + u8 slave_addr, + u8 addr, + u8 val); extern bool radeon_ddc_probe(struct radeon_connector *radeon_connector); extern int radeon_ddc_get_modes(struct radeon_connector *radeon_connector); @@ -417,17 +451,7 @@ extern void radeon_compute_pll(struct radeon_pll *pll, uint32_t *fb_div_p, uint32_t *frac_fb_div_p, uint32_t *ref_div_p, - uint32_t *post_div_p, - int flags); - -extern void radeon_compute_pll_avivo(struct radeon_pll *pll, - uint64_t freq, - uint32_t *dot_clock_p, - uint32_t *fb_div_p, - uint32_t *frac_fb_div_p, - uint32_t *ref_div_p, - uint32_t *post_div_p, - int flags); + uint32_t *post_div_p); extern void radeon_setup_encoder_clones(struct drm_device *dev); @@ -453,7 +477,6 @@ extern void atombios_crtc_dpms(struct drm_crtc *crtc, int mode); extern int radeon_crtc_set_base(struct drm_crtc *crtc, int x, int y, struct drm_framebuffer *old_fb); -extern void radeon_legacy_atom_set_surface(struct drm_crtc *crtc); extern int radeon_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv, @@ -463,6 +486,9 @@ extern int radeon_crtc_cursor_set(struct drm_crtc *crtc, extern int radeon_crtc_cursor_move(struct drm_crtc *crtc, int x, int y); +extern bool radeon_combios_check_hardcoded_edid(struct radeon_device *rdev); +extern struct edid * +radeon_combios_get_hardcoded_edid(struct radeon_device *rdev); extern bool radeon_atom_get_clock_info(struct drm_device *dev); extern bool radeon_combios_get_clock_info(struct drm_device *dev); extern struct radeon_encoder_atom_dig * @@ -521,7 +547,6 @@ void radeon_atombios_init_crtc(struct drm_device *dev, struct radeon_crtc *radeon_crtc); void radeon_legacy_init_crtc(struct drm_device *dev, struct radeon_crtc *radeon_crtc); -extern void radeon_i2c_do_lock(struct radeon_i2c_chan *i2c, int lock_state); void radeon_get_clock_info(struct drm_device *dev); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index 4e636de877b..fc9d00ac6b1 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -178,7 +178,6 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr) { int r, i; - radeon_ttm_placement_from_domain(bo, domain); if (bo->pin_count) { bo->pin_count++; if (gpu_addr) @@ -186,6 +185,8 @@ int radeon_bo_pin(struct radeon_bo *bo, u32 domain, u64 *gpu_addr) return 0; } radeon_ttm_placement_from_domain(bo, domain); + /* force to pin into visible video ram */ + bo->placement.lpfn = bo->rdev->mc.visible_vram_size >> PAGE_SHIFT; for (i = 0; i < bo->placement.num_placement; i++) bo->placements[i] |= TTM_PL_FLAG_NO_EVICT; r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false); @@ -220,7 +221,8 @@ int radeon_bo_unpin(struct radeon_bo *bo) int radeon_bo_evict_vram(struct radeon_device *rdev) { - if (rdev->flags & RADEON_IS_IGP) { + /* late 2.6.33 fix IGP hibernate - we need pm ops to do this correct */ + if (0 && (rdev->flags & RADEON_IS_IGP)) { if (rdev->mc.igp_sideport_enabled == false) /* Useless to evict on IGP chips */ return 0; @@ -305,11 +307,10 @@ void radeon_bo_list_unreserve(struct list_head *head) } } -int radeon_bo_list_validate(struct list_head *head, void *fence) +int radeon_bo_list_validate(struct list_head *head) { struct radeon_bo_list *lobj; struct radeon_bo *bo; - struct radeon_fence *old_fence = NULL; int r; r = radeon_bo_list_reserve(head); @@ -333,32 +334,27 @@ int radeon_bo_list_validate(struct list_head *head, void *fence) } lobj->gpu_offset = radeon_bo_gpu_offset(bo); lobj->tiling_flags = bo->tiling_flags; - if (fence) { - old_fence = (struct radeon_fence *)bo->tbo.sync_obj; - bo->tbo.sync_obj = radeon_fence_ref(fence); - bo->tbo.sync_obj_arg = NULL; - } - if (old_fence) { - radeon_fence_unref(&old_fence); - } } return 0; } -void radeon_bo_list_unvalidate(struct list_head *head, void *fence) +void radeon_bo_list_fence(struct list_head *head, void *fence) { struct radeon_bo_list *lobj; - struct radeon_fence *old_fence; - - if (fence) - list_for_each_entry(lobj, head, list) { - old_fence = to_radeon_fence(lobj->bo->tbo.sync_obj); - if (old_fence == fence) { - lobj->bo->tbo.sync_obj = NULL; - radeon_fence_unref(&old_fence); - } + struct radeon_bo *bo; + struct radeon_fence *old_fence = NULL; + + list_for_each_entry(lobj, head, list) { + bo = lobj->bo; + spin_lock(&bo->tbo.lock); + old_fence = (struct radeon_fence *)bo->tbo.sync_obj; + bo->tbo.sync_obj = radeon_fence_ref(fence); + bo->tbo.sync_obj_arg = NULL; + spin_unlock(&bo->tbo.lock); + if (old_fence) { + radeon_fence_unref(&old_fence); } - radeon_bo_list_unreserve(head); + } } int radeon_bo_fbdev_mmap(struct radeon_bo *bo, diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index a02f18011ad..7ab43de1e24 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -156,8 +156,8 @@ extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj, struct list_head *head); extern int radeon_bo_list_reserve(struct list_head *head); extern void radeon_bo_list_unreserve(struct list_head *head); -extern int radeon_bo_list_validate(struct list_head *head, void *fence); -extern void radeon_bo_list_unvalidate(struct list_head *head, void *fence); +extern int radeon_bo_list_validate(struct list_head *head); +extern void radeon_bo_list_fence(struct list_head *head, void *fence); extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo, struct vm_area_struct *vma); extern int radeon_bo_set_tiling_flags(struct radeon_bo *bo, diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index 8bce64cdc32..d4d1c39a0e9 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -18,21 +18,413 @@ * OTHER DEALINGS IN THE SOFTWARE. * * Authors: Rafał Miłecki <zajec5@gmail.com> + * Alex Deucher <alexdeucher@gmail.com> */ #include "drmP.h" #include "radeon.h" +#include "avivod.h" -int radeon_debugfs_pm_init(struct radeon_device *rdev); +#define RADEON_IDLE_LOOP_MS 100 +#define RADEON_RECLOCK_DELAY_MS 200 +#define RADEON_WAIT_VBLANK_TIMEOUT 200 + +static void radeon_pm_set_clocks_locked(struct radeon_device *rdev); +static void radeon_pm_set_clocks(struct radeon_device *rdev); +static void radeon_pm_idle_work_handler(struct work_struct *work); +static int radeon_debugfs_pm_init(struct radeon_device *rdev); + +static const char *pm_state_names[4] = { + "PM_STATE_DISABLED", + "PM_STATE_MINIMUM", + "PM_STATE_PAUSED", + "PM_STATE_ACTIVE" +}; + +static const char *pm_state_types[5] = { + "Default", + "Powersave", + "Battery", + "Balanced", + "Performance", +}; + +static void radeon_print_power_mode_info(struct radeon_device *rdev) +{ + int i, j; + bool is_default; + + DRM_INFO("%d Power State(s)\n", rdev->pm.num_power_states); + for (i = 0; i < rdev->pm.num_power_states; i++) { + if (rdev->pm.default_power_state == &rdev->pm.power_state[i]) + is_default = true; + else + is_default = false; + DRM_INFO("State %d %s %s\n", i, + pm_state_types[rdev->pm.power_state[i].type], + is_default ? "(default)" : ""); + if ((rdev->flags & RADEON_IS_PCIE) && !(rdev->flags & RADEON_IS_IGP)) + DRM_INFO("\t%d PCIE Lanes\n", rdev->pm.power_state[i].non_clock_info.pcie_lanes); + DRM_INFO("\t%d Clock Mode(s)\n", rdev->pm.power_state[i].num_clock_modes); + for (j = 0; j < rdev->pm.power_state[i].num_clock_modes; j++) { + if (rdev->flags & RADEON_IS_IGP) + DRM_INFO("\t\t%d engine: %d\n", + j, + rdev->pm.power_state[i].clock_info[j].sclk * 10); + else + DRM_INFO("\t\t%d engine/memory: %d/%d\n", + j, + rdev->pm.power_state[i].clock_info[j].sclk * 10, + rdev->pm.power_state[i].clock_info[j].mclk * 10); + } + } +} + +static struct radeon_power_state * radeon_pick_power_state(struct radeon_device *rdev, + enum radeon_pm_state_type type) +{ + int i, j; + enum radeon_pm_state_type wanted_types[2]; + int wanted_count; + + switch (type) { + case POWER_STATE_TYPE_DEFAULT: + default: + return rdev->pm.default_power_state; + case POWER_STATE_TYPE_POWERSAVE: + if (rdev->flags & RADEON_IS_MOBILITY) { + wanted_types[0] = POWER_STATE_TYPE_POWERSAVE; + wanted_types[1] = POWER_STATE_TYPE_BATTERY; + wanted_count = 2; + } else { + wanted_types[0] = POWER_STATE_TYPE_PERFORMANCE; + wanted_count = 1; + } + break; + case POWER_STATE_TYPE_BATTERY: + if (rdev->flags & RADEON_IS_MOBILITY) { + wanted_types[0] = POWER_STATE_TYPE_BATTERY; + wanted_types[1] = POWER_STATE_TYPE_POWERSAVE; + wanted_count = 2; + } else { + wanted_types[0] = POWER_STATE_TYPE_PERFORMANCE; + wanted_count = 1; + } + break; + case POWER_STATE_TYPE_BALANCED: + case POWER_STATE_TYPE_PERFORMANCE: + wanted_types[0] = type; + wanted_count = 1; + break; + } + + for (i = 0; i < wanted_count; i++) { + for (j = 0; j < rdev->pm.num_power_states; j++) { + if (rdev->pm.power_state[j].type == wanted_types[i]) + return &rdev->pm.power_state[j]; + } + } + + return rdev->pm.default_power_state; +} + +static struct radeon_pm_clock_info * radeon_pick_clock_mode(struct radeon_device *rdev, + struct radeon_power_state *power_state, + enum radeon_pm_clock_mode_type type) +{ + switch (type) { + case POWER_MODE_TYPE_DEFAULT: + default: + return power_state->default_clock_mode; + case POWER_MODE_TYPE_LOW: + return &power_state->clock_info[0]; + case POWER_MODE_TYPE_MID: + if (power_state->num_clock_modes > 2) + return &power_state->clock_info[1]; + else + return &power_state->clock_info[0]; + break; + case POWER_MODE_TYPE_HIGH: + return &power_state->clock_info[power_state->num_clock_modes - 1]; + } + +} + +static void radeon_get_power_state(struct radeon_device *rdev, + enum radeon_pm_action action) +{ + switch (action) { + case PM_ACTION_MINIMUM: + rdev->pm.requested_power_state = radeon_pick_power_state(rdev, POWER_STATE_TYPE_BATTERY); + rdev->pm.requested_clock_mode = + radeon_pick_clock_mode(rdev, rdev->pm.requested_power_state, POWER_MODE_TYPE_LOW); + break; + case PM_ACTION_DOWNCLOCK: + rdev->pm.requested_power_state = radeon_pick_power_state(rdev, POWER_STATE_TYPE_POWERSAVE); + rdev->pm.requested_clock_mode = + radeon_pick_clock_mode(rdev, rdev->pm.requested_power_state, POWER_MODE_TYPE_MID); + break; + case PM_ACTION_UPCLOCK: + rdev->pm.requested_power_state = radeon_pick_power_state(rdev, POWER_STATE_TYPE_DEFAULT); + rdev->pm.requested_clock_mode = + radeon_pick_clock_mode(rdev, rdev->pm.requested_power_state, POWER_MODE_TYPE_HIGH); + break; + case PM_ACTION_NONE: + default: + DRM_ERROR("Requested mode for not defined action\n"); + return; + } + DRM_INFO("Requested: e: %d m: %d p: %d\n", + rdev->pm.requested_clock_mode->sclk, + rdev->pm.requested_clock_mode->mclk, + rdev->pm.requested_power_state->non_clock_info.pcie_lanes); +} + +static void radeon_set_power_state(struct radeon_device *rdev) +{ + /* if *_clock_mode are the same, *_power_state are as well */ + if (rdev->pm.requested_clock_mode == rdev->pm.current_clock_mode) + return; + + DRM_INFO("Setting: e: %d m: %d p: %d\n", + rdev->pm.requested_clock_mode->sclk, + rdev->pm.requested_clock_mode->mclk, + rdev->pm.requested_power_state->non_clock_info.pcie_lanes); + /* set pcie lanes */ + /* set voltage */ + /* set engine clock */ + radeon_set_engine_clock(rdev, rdev->pm.requested_clock_mode->sclk); + /* set memory clock */ + + rdev->pm.current_power_state = rdev->pm.requested_power_state; + rdev->pm.current_clock_mode = rdev->pm.requested_clock_mode; +} int radeon_pm_init(struct radeon_device *rdev) { + rdev->pm.state = PM_STATE_DISABLED; + rdev->pm.planned_action = PM_ACTION_NONE; + rdev->pm.downclocked = false; + + if (rdev->bios) { + if (rdev->is_atom_bios) + radeon_atombios_get_power_modes(rdev); + else + radeon_combios_get_power_modes(rdev); + radeon_print_power_mode_info(rdev); + } + if (radeon_debugfs_pm_init(rdev)) { DRM_ERROR("Failed to register debugfs file for PM!\n"); } + INIT_DELAYED_WORK(&rdev->pm.idle_work, radeon_pm_idle_work_handler); + + if (radeon_dynpm != -1 && radeon_dynpm) { + rdev->pm.state = PM_STATE_PAUSED; + DRM_INFO("radeon: dynamic power management enabled\n"); + } + + DRM_INFO("radeon: power management initialized\n"); + return 0; } +void radeon_pm_compute_clocks(struct radeon_device *rdev) +{ + struct drm_device *ddev = rdev->ddev; + struct drm_connector *connector; + struct radeon_crtc *radeon_crtc; + int count = 0; + + if (rdev->pm.state == PM_STATE_DISABLED) + return; + + mutex_lock(&rdev->pm.mutex); + + rdev->pm.active_crtcs = 0; + list_for_each_entry(connector, + &ddev->mode_config.connector_list, head) { + if (connector->encoder && + connector->dpms != DRM_MODE_DPMS_OFF) { + radeon_crtc = to_radeon_crtc(connector->encoder->crtc); + rdev->pm.active_crtcs |= (1 << radeon_crtc->crtc_id); + ++count; + } + } + + if (count > 1) { + if (rdev->pm.state == PM_STATE_ACTIVE) { + cancel_delayed_work(&rdev->pm.idle_work); + + rdev->pm.state = PM_STATE_PAUSED; + rdev->pm.planned_action = PM_ACTION_UPCLOCK; + if (rdev->pm.downclocked) + radeon_pm_set_clocks(rdev); + + DRM_DEBUG("radeon: dynamic power management deactivated\n"); + } + } else if (count == 1) { + /* TODO: Increase clocks if needed for current mode */ + + if (rdev->pm.state == PM_STATE_MINIMUM) { + rdev->pm.state = PM_STATE_ACTIVE; + rdev->pm.planned_action = PM_ACTION_UPCLOCK; + radeon_pm_set_clocks(rdev); + + queue_delayed_work(rdev->wq, &rdev->pm.idle_work, + msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); + } + else if (rdev->pm.state == PM_STATE_PAUSED) { + rdev->pm.state = PM_STATE_ACTIVE; + queue_delayed_work(rdev->wq, &rdev->pm.idle_work, + msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); + DRM_DEBUG("radeon: dynamic power management activated\n"); + } + } + else { /* count == 0 */ + if (rdev->pm.state != PM_STATE_MINIMUM) { + cancel_delayed_work(&rdev->pm.idle_work); + + rdev->pm.state = PM_STATE_MINIMUM; + rdev->pm.planned_action = PM_ACTION_MINIMUM; + radeon_pm_set_clocks(rdev); + } + } + + mutex_unlock(&rdev->pm.mutex); +} + +static bool radeon_pm_debug_check_in_vbl(struct radeon_device *rdev, bool finish) +{ + u32 stat_crtc1 = 0, stat_crtc2 = 0; + bool in_vbl = true; + + if (ASIC_IS_AVIVO(rdev)) { + if (rdev->pm.active_crtcs & (1 << 0)) { + stat_crtc1 = RREG32(D1CRTC_STATUS); + if (!(stat_crtc1 & 1)) + in_vbl = false; + } + if (rdev->pm.active_crtcs & (1 << 1)) { + stat_crtc2 = RREG32(D2CRTC_STATUS); + if (!(stat_crtc2 & 1)) + in_vbl = false; + } + } + if (in_vbl == false) + DRM_INFO("not in vbl for pm change %08x %08x at %s\n", stat_crtc1, + stat_crtc2, finish ? "exit" : "entry"); + return in_vbl; +} +static void radeon_pm_set_clocks_locked(struct radeon_device *rdev) +{ + /*radeon_fence_wait_last(rdev);*/ + switch (rdev->pm.planned_action) { + case PM_ACTION_UPCLOCK: + rdev->pm.downclocked = false; + break; + case PM_ACTION_DOWNCLOCK: + rdev->pm.downclocked = true; + break; + case PM_ACTION_MINIMUM: + break; + case PM_ACTION_NONE: + DRM_ERROR("%s: PM_ACTION_NONE\n", __func__); + break; + } + + /* check if we are in vblank */ + radeon_pm_debug_check_in_vbl(rdev, false); + radeon_set_power_state(rdev); + radeon_pm_debug_check_in_vbl(rdev, true); + rdev->pm.planned_action = PM_ACTION_NONE; +} + +static void radeon_pm_set_clocks(struct radeon_device *rdev) +{ + radeon_get_power_state(rdev, rdev->pm.planned_action); + mutex_lock(&rdev->cp.mutex); + + if (rdev->pm.active_crtcs & (1 << 0)) { + rdev->pm.req_vblank |= (1 << 0); + drm_vblank_get(rdev->ddev, 0); + } + if (rdev->pm.active_crtcs & (1 << 1)) { + rdev->pm.req_vblank |= (1 << 1); + drm_vblank_get(rdev->ddev, 1); + } + if (rdev->pm.active_crtcs) + wait_event_interruptible_timeout( + rdev->irq.vblank_queue, 0, + msecs_to_jiffies(RADEON_WAIT_VBLANK_TIMEOUT)); + if (rdev->pm.req_vblank & (1 << 0)) { + rdev->pm.req_vblank &= ~(1 << 0); + drm_vblank_put(rdev->ddev, 0); + } + if (rdev->pm.req_vblank & (1 << 1)) { + rdev->pm.req_vblank &= ~(1 << 1); + drm_vblank_put(rdev->ddev, 1); + } + + radeon_pm_set_clocks_locked(rdev); + mutex_unlock(&rdev->cp.mutex); +} + +static void radeon_pm_idle_work_handler(struct work_struct *work) +{ + struct radeon_device *rdev; + rdev = container_of(work, struct radeon_device, + pm.idle_work.work); + + mutex_lock(&rdev->pm.mutex); + if (rdev->pm.state == PM_STATE_ACTIVE) { + unsigned long irq_flags; + int not_processed = 0; + + read_lock_irqsave(&rdev->fence_drv.lock, irq_flags); + if (!list_empty(&rdev->fence_drv.emited)) { + struct list_head *ptr; + list_for_each(ptr, &rdev->fence_drv.emited) { + /* count up to 3, that's enought info */ + if (++not_processed >= 3) + break; + } + } + read_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + + if (not_processed >= 3) { /* should upclock */ + if (rdev->pm.planned_action == PM_ACTION_DOWNCLOCK) { + rdev->pm.planned_action = PM_ACTION_NONE; + } else if (rdev->pm.planned_action == PM_ACTION_NONE && + rdev->pm.downclocked) { + rdev->pm.planned_action = + PM_ACTION_UPCLOCK; + rdev->pm.action_timeout = jiffies + + msecs_to_jiffies(RADEON_RECLOCK_DELAY_MS); + } + } else if (not_processed == 0) { /* should downclock */ + if (rdev->pm.planned_action == PM_ACTION_UPCLOCK) { + rdev->pm.planned_action = PM_ACTION_NONE; + } else if (rdev->pm.planned_action == PM_ACTION_NONE && + !rdev->pm.downclocked) { + rdev->pm.planned_action = + PM_ACTION_DOWNCLOCK; + rdev->pm.action_timeout = jiffies + + msecs_to_jiffies(RADEON_RECLOCK_DELAY_MS); + } + } + + if (rdev->pm.planned_action != PM_ACTION_NONE && + jiffies > rdev->pm.action_timeout) { + radeon_pm_set_clocks(rdev); + } + } + mutex_unlock(&rdev->pm.mutex); + + queue_delayed_work(rdev->wq, &rdev->pm.idle_work, + msecs_to_jiffies(RADEON_IDLE_LOOP_MS)); +} + /* * Debugfs info */ @@ -44,11 +436,14 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data) struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; + seq_printf(m, "state: %s\n", pm_state_names[rdev->pm.state]); seq_printf(m, "default engine clock: %u0 kHz\n", rdev->clock.default_sclk); seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev)); seq_printf(m, "default memory clock: %u0 kHz\n", rdev->clock.default_mclk); if (rdev->asic->get_memory_clock) seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev)); + if (rdev->asic->get_pcie_lanes) + seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev)); return 0; } @@ -58,7 +453,7 @@ static struct drm_info_list radeon_pm_info_list[] = { }; #endif -int radeon_debugfs_pm_init(struct radeon_device *rdev) +static int radeon_debugfs_pm_init(struct radeon_device *rdev) { #if defined(CONFIG_DEBUG_FS) return radeon_debugfs_add_files(rdev, radeon_pm_info_list, ARRAY_SIZE(radeon_pm_info_list)); diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h index 6d0a009dd4a..5c0dc082d33 100644 --- a/drivers/gpu/drm/radeon/radeon_reg.h +++ b/drivers/gpu/drm/radeon/radeon_reg.h @@ -54,7 +54,7 @@ #include "r300_reg.h" #include "r500_reg.h" #include "r600_reg.h" - +#include "evergreen_reg.h" #define RADEON_MC_AGP_LOCATION 0x014c #define RADEON_MC_AGP_START_MASK 0x0000FFFF @@ -1060,32 +1060,38 @@ /* Multimedia I2C bus */ #define RADEON_I2C_CNTL_0 0x0090 -#define RADEON_I2C_DONE (1 << 0) -#define RADEON_I2C_NACK (1 << 1) -#define RADEON_I2C_HALT (1 << 2) -#define RADEON_I2C_SOFT_RST (1 << 5) -#define RADEON_I2C_DRIVE_EN (1 << 6) -#define RADEON_I2C_DRIVE_SEL (1 << 7) -#define RADEON_I2C_START (1 << 8) -#define RADEON_I2C_STOP (1 << 9) -#define RADEON_I2C_RECEIVE (1 << 10) -#define RADEON_I2C_ABORT (1 << 11) -#define RADEON_I2C_GO (1 << 12) -#define RADEON_I2C_PRESCALE_SHIFT 16 +# define RADEON_I2C_DONE (1 << 0) +# define RADEON_I2C_NACK (1 << 1) +# define RADEON_I2C_HALT (1 << 2) +# define RADEON_I2C_SOFT_RST (1 << 5) +# define RADEON_I2C_DRIVE_EN (1 << 6) +# define RADEON_I2C_DRIVE_SEL (1 << 7) +# define RADEON_I2C_START (1 << 8) +# define RADEON_I2C_STOP (1 << 9) +# define RADEON_I2C_RECEIVE (1 << 10) +# define RADEON_I2C_ABORT (1 << 11) +# define RADEON_I2C_GO (1 << 12) +# define RADEON_I2C_PRESCALE_SHIFT 16 #define RADEON_I2C_CNTL_1 0x0094 -#define RADEON_I2C_DATA_COUNT_SHIFT 0 -#define RADEON_I2C_ADDR_COUNT_SHIFT 4 -#define RADEON_I2C_INTRA_BYTE_DELAY_SHIFT 8 -#define RADEON_I2C_SEL (1 << 16) -#define RADEON_I2C_EN (1 << 17) -#define RADEON_I2C_TIME_LIMIT_SHIFT 24 +# define RADEON_I2C_DATA_COUNT_SHIFT 0 +# define RADEON_I2C_ADDR_COUNT_SHIFT 4 +# define RADEON_I2C_INTRA_BYTE_DELAY_SHIFT 8 +# define RADEON_I2C_SEL (1 << 16) +# define RADEON_I2C_EN (1 << 17) +# define RADEON_I2C_TIME_LIMIT_SHIFT 24 #define RADEON_I2C_DATA 0x0098 #define RADEON_DVI_I2C_CNTL_0 0x02e0 # define R200_DVI_I2C_PIN_SEL(x) ((x) << 3) -# define R200_SEL_DDC1 0 /* 0x60 - VGA_DDC */ -# define R200_SEL_DDC2 1 /* 0x64 - DVI_DDC */ -# define R200_SEL_DDC3 2 /* 0x68 - MONID_DDC */ +# define R200_SEL_DDC1 0 /* depends on asic */ +# define R200_SEL_DDC2 1 /* depends on asic */ +# define R200_SEL_DDC3 2 /* depends on asic */ +# define RADEON_SW_WANTS_TO_USE_DVI_I2C (1 << 13) +# define RADEON_SW_CAN_USE_DVI_I2C (1 << 13) +# define RADEON_SW_DONE_USING_DVI_I2C (1 << 14) +# define RADEON_HW_NEEDS_DVI_I2C (1 << 14) +# define RADEON_ABORT_HW_DVI_I2C (1 << 15) +# define RADEON_HW_USING_DVI_I2C (1 << 15) #define RADEON_DVI_I2C_CNTL_1 0x02e4 #define RADEON_DVI_I2C_DATA 0x02e8 diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 4d12b2d17b4..e50513a6273 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -34,6 +34,36 @@ int radeon_debugfs_ib_init(struct radeon_device *rdev); +void radeon_ib_bogus_cleanup(struct radeon_device *rdev) +{ + struct radeon_ib *ib, *n; + + list_for_each_entry_safe(ib, n, &rdev->ib_pool.bogus_ib, list) { + list_del(&ib->list); + vfree(ib->ptr); + kfree(ib); + } +} + +void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ib *bib; + + bib = kmalloc(sizeof(*bib), GFP_KERNEL); + if (bib == NULL) + return; + bib->ptr = vmalloc(ib->length_dw * 4); + if (bib->ptr == NULL) { + kfree(bib); + return; + } + memcpy(bib->ptr, ib->ptr, ib->length_dw * 4); + bib->length_dw = ib->length_dw; + mutex_lock(&rdev->ib_pool.mutex); + list_add_tail(&bib->list, &rdev->ib_pool.bogus_ib); + mutex_unlock(&rdev->ib_pool.mutex); +} + /* * IB. */ @@ -41,68 +71,55 @@ int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib) { struct radeon_fence *fence; struct radeon_ib *nib; - unsigned long i; - int r = 0; + int r = 0, i, c; *ib = NULL; r = radeon_fence_create(rdev, &fence); if (r) { - DRM_ERROR("failed to create fence for new IB\n"); + dev_err(rdev->dev, "failed to create fence for new IB\n"); return r; } mutex_lock(&rdev->ib_pool.mutex); - i = find_first_zero_bit(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE); - if (i < RADEON_IB_POOL_SIZE) { - set_bit(i, rdev->ib_pool.alloc_bm); - rdev->ib_pool.ibs[i].length_dw = 0; - *ib = &rdev->ib_pool.ibs[i]; - mutex_unlock(&rdev->ib_pool.mutex); - goto out; + for (i = rdev->ib_pool.head_id, c = 0, nib = NULL; c < RADEON_IB_POOL_SIZE; c++, i++) { + i &= (RADEON_IB_POOL_SIZE - 1); + if (rdev->ib_pool.ibs[i].free) { + nib = &rdev->ib_pool.ibs[i]; + break; + } } - if (list_empty(&rdev->ib_pool.scheduled_ibs)) { - /* we go do nothings here */ + if (nib == NULL) { + /* This should never happen, it means we allocated all + * IB and haven't scheduled one yet, return EBUSY to + * userspace hoping that on ioctl recall we get better + * luck + */ + dev_err(rdev->dev, "no free indirect buffer !\n"); mutex_unlock(&rdev->ib_pool.mutex); - DRM_ERROR("all IB allocated none scheduled.\n"); - r = -EINVAL; - goto out; + radeon_fence_unref(&fence); + return -EBUSY; } - /* get the first ib on the scheduled list */ - nib = list_entry(rdev->ib_pool.scheduled_ibs.next, - struct radeon_ib, list); - if (nib->fence == NULL) { - /* we go do nothings here */ + rdev->ib_pool.head_id = (nib->idx + 1) & (RADEON_IB_POOL_SIZE - 1); + nib->free = false; + if (nib->fence) { mutex_unlock(&rdev->ib_pool.mutex); - DRM_ERROR("IB %lu scheduled without a fence.\n", nib->idx); - r = -EINVAL; - goto out; - } - mutex_unlock(&rdev->ib_pool.mutex); - - r = radeon_fence_wait(nib->fence, false); - if (r) { - DRM_ERROR("radeon: IB(%lu:0x%016lX:%u)\n", nib->idx, - (unsigned long)nib->gpu_addr, nib->length_dw); - DRM_ERROR("radeon: GPU lockup detected, fail to get a IB\n"); - goto out; + r = radeon_fence_wait(nib->fence, false); + if (r) { + dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n", + nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw); + mutex_lock(&rdev->ib_pool.mutex); + nib->free = true; + mutex_unlock(&rdev->ib_pool.mutex); + radeon_fence_unref(&fence); + return r; + } + mutex_lock(&rdev->ib_pool.mutex); } radeon_fence_unref(&nib->fence); - + nib->fence = fence; nib->length_dw = 0; - - /* scheduled list is accessed here */ - mutex_lock(&rdev->ib_pool.mutex); - list_del(&nib->list); - INIT_LIST_HEAD(&nib->list); mutex_unlock(&rdev->ib_pool.mutex); - *ib = nib; -out: - if (r) { - radeon_fence_unref(&fence); - } else { - (*ib)->fence = fence; - } - return r; + return 0; } void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib) @@ -113,19 +130,10 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib) if (tmp == NULL) { return; } - mutex_lock(&rdev->ib_pool.mutex); - if (!list_empty(&tmp->list) && !radeon_fence_signaled(tmp->fence)) { - /* IB is scheduled & not signaled don't do anythings */ - mutex_unlock(&rdev->ib_pool.mutex); - return; - } - list_del(&tmp->list); - INIT_LIST_HEAD(&tmp->list); - if (tmp->fence) + if (!tmp->fence->emited) radeon_fence_unref(&tmp->fence); - - tmp->length_dw = 0; - clear_bit(tmp->idx, rdev->ib_pool.alloc_bm); + mutex_lock(&rdev->ib_pool.mutex); + tmp->free = true; mutex_unlock(&rdev->ib_pool.mutex); } @@ -135,7 +143,7 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib) if (!ib->length_dw || !rdev->cp.ready) { /* TODO: Nothings in the ib we should report. */ - DRM_ERROR("radeon: couldn't schedule IB(%lu).\n", ib->idx); + DRM_ERROR("radeon: couldn't schedule IB(%u).\n", ib->idx); return -EINVAL; } @@ -148,7 +156,8 @@ int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_ib_execute(rdev, ib); radeon_fence_emit(rdev, ib->fence); mutex_lock(&rdev->ib_pool.mutex); - list_add_tail(&ib->list, &rdev->ib_pool.scheduled_ibs); + /* once scheduled IB is considered free and protected by the fence */ + ib->free = true; mutex_unlock(&rdev->ib_pool.mutex); radeon_ring_unlock_commit(rdev); return 0; @@ -163,8 +172,8 @@ int radeon_ib_pool_init(struct radeon_device *rdev) if (rdev->ib_pool.robj) return 0; + INIT_LIST_HEAD(&rdev->ib_pool.bogus_ib); /* Allocate 1M object buffer */ - INIT_LIST_HEAD(&rdev->ib_pool.scheduled_ibs); r = radeon_bo_create(rdev, NULL, RADEON_IB_POOL_SIZE*64*1024, true, RADEON_GEM_DOMAIN_GTT, &rdev->ib_pool.robj); @@ -195,9 +204,9 @@ int radeon_ib_pool_init(struct radeon_device *rdev) rdev->ib_pool.ibs[i].ptr = ptr + offset; rdev->ib_pool.ibs[i].idx = i; rdev->ib_pool.ibs[i].length_dw = 0; - INIT_LIST_HEAD(&rdev->ib_pool.ibs[i].list); + rdev->ib_pool.ibs[i].free = true; } - bitmap_zero(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE); + rdev->ib_pool.head_id = 0; rdev->ib_pool.ready = true; DRM_INFO("radeon: ib pool ready.\n"); if (radeon_debugfs_ib_init(rdev)) { @@ -214,7 +223,8 @@ void radeon_ib_pool_fini(struct radeon_device *rdev) return; } mutex_lock(&rdev->ib_pool.mutex); - bitmap_zero(rdev->ib_pool.alloc_bm, RADEON_IB_POOL_SIZE); + radeon_ib_bogus_cleanup(rdev); + if (rdev->ib_pool.robj) { r = radeon_bo_reserve(rdev->ib_pool.robj, false); if (likely(r == 0)) { @@ -363,7 +373,7 @@ static int radeon_debugfs_ib_info(struct seq_file *m, void *data) if (ib == NULL) { return 0; } - seq_printf(m, "IB %04lu\n", ib->idx); + seq_printf(m, "IB %04u\n", ib->idx); seq_printf(m, "IB fence %p\n", ib->fence); seq_printf(m, "IB size %05u dwords\n", ib->length_dw); for (i = 0; i < ib->length_dw; i++) { @@ -372,15 +382,49 @@ static int radeon_debugfs_ib_info(struct seq_file *m, void *data) return 0; } +static int radeon_debugfs_ib_bogus_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct radeon_device *rdev = node->info_ent->data; + struct radeon_ib *ib; + unsigned i; + + mutex_lock(&rdev->ib_pool.mutex); + if (list_empty(&rdev->ib_pool.bogus_ib)) { + mutex_unlock(&rdev->ib_pool.mutex); + seq_printf(m, "no bogus IB recorded\n"); + return 0; + } + ib = list_first_entry(&rdev->ib_pool.bogus_ib, struct radeon_ib, list); + list_del_init(&ib->list); + mutex_unlock(&rdev->ib_pool.mutex); + seq_printf(m, "IB size %05u dwords\n", ib->length_dw); + for (i = 0; i < ib->length_dw; i++) { + seq_printf(m, "[%05u]=0x%08X\n", i, ib->ptr[i]); + } + vfree(ib->ptr); + kfree(ib); + return 0; +} + static struct drm_info_list radeon_debugfs_ib_list[RADEON_IB_POOL_SIZE]; static char radeon_debugfs_ib_names[RADEON_IB_POOL_SIZE][32]; + +static struct drm_info_list radeon_debugfs_ib_bogus_info_list[] = { + {"radeon_ib_bogus", radeon_debugfs_ib_bogus_info, 0, NULL}, +}; #endif int radeon_debugfs_ib_init(struct radeon_device *rdev) { #if defined(CONFIG_DEBUG_FS) unsigned i; + int r; + radeon_debugfs_ib_bogus_info_list[0].data = rdev; + r = radeon_debugfs_add_files(rdev, radeon_debugfs_ib_bogus_info_list, 1); + if (r) + return r; for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { sprintf(radeon_debugfs_ib_names[i], "radeon_ib_%04u", i); radeon_debugfs_ib_list[i].name = radeon_debugfs_ib_names[i]; diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c index 067167cb39c..3c32f840dcd 100644 --- a/drivers/gpu/drm/radeon/radeon_state.c +++ b/drivers/gpu/drm/radeon/radeon_state.c @@ -29,6 +29,7 @@ #include "drmP.h" #include "drm.h" +#include "drm_buffer.h" #include "drm_sarea.h" #include "radeon_drm.h" #include "radeon_drv.h" @@ -91,21 +92,27 @@ static __inline__ int radeon_check_and_fixup_offset(drm_radeon_private_t * static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t * dev_priv, struct drm_file *file_priv, - int id, u32 *data) + int id, struct drm_buffer *buf) { + u32 *data; switch (id) { case RADEON_EMIT_PP_MISC: - if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[(RADEON_RB3D_DEPTHOFFSET - RADEON_PP_MISC) / 4])) { + data = drm_buffer_pointer_to_dword(buf, + (RADEON_RB3D_DEPTHOFFSET - RADEON_PP_MISC) / 4); + + if (radeon_check_and_fixup_offset(dev_priv, file_priv, data)) { DRM_ERROR("Invalid depth buffer offset\n"); return -EINVAL; } + dev_priv->have_z_offset = 1; break; case RADEON_EMIT_PP_CNTL: - if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[(RADEON_RB3D_COLOROFFSET - RADEON_PP_CNTL) / 4])) { + data = drm_buffer_pointer_to_dword(buf, + (RADEON_RB3D_COLOROFFSET - RADEON_PP_CNTL) / 4); + + if (radeon_check_and_fixup_offset(dev_priv, file_priv, data)) { DRM_ERROR("Invalid colour buffer offset\n"); return -EINVAL; } @@ -117,8 +124,8 @@ static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t * case R200_EMIT_PP_TXOFFSET_3: case R200_EMIT_PP_TXOFFSET_4: case R200_EMIT_PP_TXOFFSET_5: - if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[0])) { + data = drm_buffer_pointer_to_dword(buf, 0); + if (radeon_check_and_fixup_offset(dev_priv, file_priv, data)) { DRM_ERROR("Invalid R200 texture offset\n"); return -EINVAL; } @@ -127,8 +134,9 @@ static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t * case RADEON_EMIT_PP_TXFILTER_0: case RADEON_EMIT_PP_TXFILTER_1: case RADEON_EMIT_PP_TXFILTER_2: - if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[(RADEON_PP_TXOFFSET_0 - RADEON_PP_TXFILTER_0) / 4])) { + data = drm_buffer_pointer_to_dword(buf, + (RADEON_PP_TXOFFSET_0 - RADEON_PP_TXFILTER_0) / 4); + if (radeon_check_and_fixup_offset(dev_priv, file_priv, data)) { DRM_ERROR("Invalid R100 texture offset\n"); return -EINVAL; } @@ -142,9 +150,10 @@ static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t * case R200_EMIT_PP_CUBIC_OFFSETS_5:{ int i; for (i = 0; i < 5; i++) { + data = drm_buffer_pointer_to_dword(buf, i); if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[i])) { + data)) { DRM_ERROR ("Invalid R200 cubic texture offset\n"); return -EINVAL; @@ -158,9 +167,10 @@ static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t * case RADEON_EMIT_PP_CUBIC_OFFSETS_T2:{ int i; for (i = 0; i < 5; i++) { + data = drm_buffer_pointer_to_dword(buf, i); if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &data[i])) { + data)) { DRM_ERROR ("Invalid R100 cubic texture offset\n"); return -EINVAL; @@ -269,23 +279,24 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * cmdbuf, unsigned int *cmdsz) { - u32 *cmd = (u32 *) cmdbuf->buf; + u32 *cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0); u32 offset, narrays; int count, i, k; - *cmdsz = 2 + ((cmd[0] & RADEON_CP_PACKET_COUNT_MASK) >> 16); + count = ((*cmd & RADEON_CP_PACKET_COUNT_MASK) >> 16); + *cmdsz = 2 + count; - if ((cmd[0] & 0xc0000000) != RADEON_CP_PACKET3) { + if ((*cmd & 0xc0000000) != RADEON_CP_PACKET3) { DRM_ERROR("Not a type 3 packet\n"); return -EINVAL; } - if (4 * *cmdsz > cmdbuf->bufsz) { + if (4 * *cmdsz > drm_buffer_unprocessed(cmdbuf->buffer)) { DRM_ERROR("Packet size larger than size of data provided\n"); return -EINVAL; } - switch(cmd[0] & 0xff00) { + switch (*cmd & 0xff00) { /* XXX Are there old drivers needing other packets? */ case RADEON_3D_DRAW_IMMD: @@ -312,7 +323,6 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * break; case RADEON_3D_LOAD_VBPNTR: - count = (cmd[0] >> 16) & 0x3fff; if (count > 18) { /* 12 arrays max */ DRM_ERROR("Too large payload in 3D_LOAD_VBPNTR (count=%d)\n", @@ -321,13 +331,16 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * } /* carefully check packet contents */ - narrays = cmd[1] & ~0xc000; + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + + narrays = *cmd & ~0xc000; k = 0; i = 2; while ((k < narrays) && (i < (count + 2))) { i++; /* skip attribute field */ + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, i); if (radeon_check_and_fixup_offset(dev_priv, file_priv, - &cmd[i])) { + cmd)) { DRM_ERROR ("Invalid offset (k=%d i=%d) in 3D_LOAD_VBPNTR packet.\n", k, i); @@ -338,8 +351,10 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * if (k == narrays) break; /* have one more to process, they come in pairs */ + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, i); + if (radeon_check_and_fixup_offset(dev_priv, - file_priv, &cmd[i])) + file_priv, cmd)) { DRM_ERROR ("Invalid offset (k=%d i=%d) in 3D_LOAD_VBPNTR packet.\n", @@ -363,7 +378,9 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * DRM_ERROR("Invalid 3d packet for r200-class chip\n"); return -EINVAL; } - if (radeon_check_and_fixup_offset(dev_priv, file_priv, &cmd[1])) { + + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + if (radeon_check_and_fixup_offset(dev_priv, file_priv, cmd)) { DRM_ERROR("Invalid rndr_gen_indx offset\n"); return -EINVAL; } @@ -374,12 +391,15 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * DRM_ERROR("Invalid 3d packet for r100-class chip\n"); return -EINVAL; } - if ((cmd[1] & 0x8000ffff) != 0x80000810) { - DRM_ERROR("Invalid indx_buffer reg address %08X\n", cmd[1]); + + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + if ((*cmd & 0x8000ffff) != 0x80000810) { + DRM_ERROR("Invalid indx_buffer reg address %08X\n", *cmd); return -EINVAL; } - if (radeon_check_and_fixup_offset(dev_priv, file_priv, &cmd[2])) { - DRM_ERROR("Invalid indx_buffer offset is %08X\n", cmd[2]); + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 2); + if (radeon_check_and_fixup_offset(dev_priv, file_priv, cmd)) { + DRM_ERROR("Invalid indx_buffer offset is %08X\n", *cmd); return -EINVAL; } break; @@ -388,31 +408,34 @@ static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t * case RADEON_CNTL_PAINT_MULTI: case RADEON_CNTL_BITBLT_MULTI: /* MSB of opcode: next DWORD GUI_CNTL */ - if (cmd[1] & (RADEON_GMC_SRC_PITCH_OFFSET_CNTL + cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 1); + if (*cmd & (RADEON_GMC_SRC_PITCH_OFFSET_CNTL | RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { - offset = cmd[2] << 10; + u32 *cmd2 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 2); + offset = *cmd2 << 10; if (radeon_check_and_fixup_offset (dev_priv, file_priv, &offset)) { DRM_ERROR("Invalid first packet offset\n"); return -EINVAL; } - cmd[2] = (cmd[2] & 0xffc00000) | offset >> 10; + *cmd2 = (*cmd2 & 0xffc00000) | offset >> 10; } - if ((cmd[1] & RADEON_GMC_SRC_PITCH_OFFSET_CNTL) && - (cmd[1] & RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { - offset = cmd[3] << 10; + if ((*cmd & RADEON_GMC_SRC_PITCH_OFFSET_CNTL) && + (*cmd & RADEON_GMC_DST_PITCH_OFFSET_CNTL)) { + u32 *cmd3 = drm_buffer_pointer_to_dword(cmdbuf->buffer, 3); + offset = *cmd << 10; if (radeon_check_and_fixup_offset (dev_priv, file_priv, &offset)) { DRM_ERROR("Invalid second packet offset\n"); return -EINVAL; } - cmd[3] = (cmd[3] & 0xffc00000) | offset >> 10; + *cmd3 = (*cmd3 & 0xffc00000) | offset >> 10; } break; default: - DRM_ERROR("Invalid packet type %x\n", cmd[0] & 0xff00); + DRM_ERROR("Invalid packet type %x\n", *cmd & 0xff00); return -EINVAL; } @@ -876,6 +899,11 @@ static void radeon_cp_dispatch_clear(struct drm_device * dev, if (tmp & RADEON_BACK) flags |= RADEON_FRONT; } + if (flags & (RADEON_DEPTH|RADEON_STENCIL)) { + if (!dev_priv->have_z_offset) + printk_once(KERN_ERR "radeon: illegal depth clear request. Buggy mesa detected - please update.\n"); + flags &= ~(RADEON_DEPTH | RADEON_STENCIL); + } if (flags & (RADEON_FRONT | RADEON_BACK)) { @@ -2611,7 +2639,6 @@ static int radeon_emit_packets(drm_radeon_private_t * dev_priv, { int id = (int)header.packet.packet_id; int sz, reg; - int *data = (int *)cmdbuf->buf; RING_LOCALS; if (id >= RADEON_MAX_STATE_PACKETS) @@ -2620,23 +2647,22 @@ static int radeon_emit_packets(drm_radeon_private_t * dev_priv, sz = packet[id].len; reg = packet[id].start; - if (sz * sizeof(int) > cmdbuf->bufsz) { + if (sz * sizeof(u32) > drm_buffer_unprocessed(cmdbuf->buffer)) { DRM_ERROR("Packet size provided larger than data provided\n"); return -EINVAL; } - if (radeon_check_and_fixup_packets(dev_priv, file_priv, id, data)) { + if (radeon_check_and_fixup_packets(dev_priv, file_priv, id, + cmdbuf->buffer)) { DRM_ERROR("Packet verification failed\n"); return -EINVAL; } BEGIN_RING(sz + 1); OUT_RING(CP_PACKET0(reg, (sz - 1))); - OUT_RING_TABLE(data, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * sizeof(int); - cmdbuf->bufsz -= sz * sizeof(int); return 0; } @@ -2653,10 +2679,8 @@ static __inline__ int radeon_emit_scalars(drm_radeon_private_t *dev_priv, OUT_RING(CP_PACKET0(RADEON_SE_TCL_SCALAR_INDX_REG, 0)); OUT_RING(start | (stride << RADEON_SCAL_INDX_DWORD_STRIDE_SHIFT)); OUT_RING(CP_PACKET0_TABLE(RADEON_SE_TCL_SCALAR_DATA_REG, sz - 1)); - OUT_RING_TABLE(cmdbuf->buf, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * sizeof(int); - cmdbuf->bufsz -= sz * sizeof(int); return 0; } @@ -2675,10 +2699,8 @@ static __inline__ int radeon_emit_scalars2(drm_radeon_private_t *dev_priv, OUT_RING(CP_PACKET0(RADEON_SE_TCL_SCALAR_INDX_REG, 0)); OUT_RING(start | (stride << RADEON_SCAL_INDX_DWORD_STRIDE_SHIFT)); OUT_RING(CP_PACKET0_TABLE(RADEON_SE_TCL_SCALAR_DATA_REG, sz - 1)); - OUT_RING_TABLE(cmdbuf->buf, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * sizeof(int); - cmdbuf->bufsz -= sz * sizeof(int); return 0; } @@ -2696,11 +2718,9 @@ static __inline__ int radeon_emit_vectors(drm_radeon_private_t *dev_priv, OUT_RING(CP_PACKET0(RADEON_SE_TCL_VECTOR_INDX_REG, 0)); OUT_RING(start | (stride << RADEON_VEC_INDX_OCTWORD_STRIDE_SHIFT)); OUT_RING(CP_PACKET0_TABLE(RADEON_SE_TCL_VECTOR_DATA_REG, (sz - 1))); - OUT_RING_TABLE(cmdbuf->buf, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * sizeof(int); - cmdbuf->bufsz -= sz * sizeof(int); return 0; } @@ -2714,7 +2734,7 @@ static __inline__ int radeon_emit_veclinear(drm_radeon_private_t *dev_priv, if (!sz) return 0; - if (sz * 4 > cmdbuf->bufsz) + if (sz * 4 > drm_buffer_unprocessed(cmdbuf->buffer)) return -EINVAL; BEGIN_RING(5 + sz); @@ -2722,11 +2742,9 @@ static __inline__ int radeon_emit_veclinear(drm_radeon_private_t *dev_priv, OUT_RING(CP_PACKET0(RADEON_SE_TCL_VECTOR_INDX_REG, 0)); OUT_RING(start | (1 << RADEON_VEC_INDX_OCTWORD_STRIDE_SHIFT)); OUT_RING(CP_PACKET0_TABLE(RADEON_SE_TCL_VECTOR_DATA_REG, (sz - 1))); - OUT_RING_TABLE(cmdbuf->buf, sz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, sz); ADVANCE_RING(); - cmdbuf->buf += sz * sizeof(int); - cmdbuf->bufsz -= sz * sizeof(int); return 0; } @@ -2748,11 +2766,9 @@ static int radeon_emit_packet3(struct drm_device * dev, } BEGIN_RING(cmdsz); - OUT_RING_TABLE(cmdbuf->buf, cmdsz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, cmdsz); ADVANCE_RING(); - cmdbuf->buf += cmdsz * 4; - cmdbuf->bufsz -= cmdsz * 4; return 0; } @@ -2805,16 +2821,16 @@ static int radeon_emit_packet3_cliprect(struct drm_device *dev, } BEGIN_RING(cmdsz); - OUT_RING_TABLE(cmdbuf->buf, cmdsz); + OUT_RING_DRM_BUFFER(cmdbuf->buffer, cmdsz); ADVANCE_RING(); } while (++i < cmdbuf->nbox); if (cmdbuf->nbox == 1) cmdbuf->nbox = 0; + return 0; out: - cmdbuf->buf += cmdsz * 4; - cmdbuf->bufsz -= cmdsz * 4; + drm_buffer_advance(cmdbuf->buffer, cmdsz * 4); return 0; } @@ -2847,16 +2863,16 @@ static int radeon_emit_wait(struct drm_device * dev, int flags) return 0; } -static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_priv) +static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, + struct drm_file *file_priv) { drm_radeon_private_t *dev_priv = dev->dev_private; struct drm_device_dma *dma = dev->dma; struct drm_buf *buf = NULL; + drm_radeon_cmd_header_t stack_header; int idx; drm_radeon_kcmd_buffer_t *cmdbuf = data; - drm_radeon_cmd_header_t header; - int orig_nbox, orig_bufsz; - char *kbuf = NULL; + int orig_nbox; LOCK_TEST_WITH_RETURN(dev, file_priv); @@ -2871,17 +2887,16 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file * races between checking values and using those values in other code, * and simply to avoid a lot of function calls to copy in data. */ - orig_bufsz = cmdbuf->bufsz; - if (orig_bufsz != 0) { - kbuf = kmalloc(cmdbuf->bufsz, GFP_KERNEL); - if (kbuf == NULL) - return -ENOMEM; - if (DRM_COPY_FROM_USER(kbuf, (void __user *)cmdbuf->buf, - cmdbuf->bufsz)) { - kfree(kbuf); - return -EFAULT; - } - cmdbuf->buf = kbuf; + if (cmdbuf->bufsz != 0) { + int rv; + void __user *buffer = cmdbuf->buffer; + rv = drm_buffer_alloc(&cmdbuf->buffer, cmdbuf->bufsz); + if (rv) + return rv; + rv = drm_buffer_copy_from_user(cmdbuf->buffer, buffer, + cmdbuf->bufsz); + if (rv) + return rv; } orig_nbox = cmdbuf->nbox; @@ -2890,24 +2905,24 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file int temp; temp = r300_do_cp_cmdbuf(dev, file_priv, cmdbuf); - if (orig_bufsz != 0) - kfree(kbuf); + if (cmdbuf->bufsz != 0) + drm_buffer_free(cmdbuf->buffer); return temp; } /* microcode_version != r300 */ - while (cmdbuf->bufsz >= sizeof(header)) { + while (drm_buffer_unprocessed(cmdbuf->buffer) >= sizeof(stack_header)) { - header.i = *(int *)cmdbuf->buf; - cmdbuf->buf += sizeof(header); - cmdbuf->bufsz -= sizeof(header); + drm_radeon_cmd_header_t *header; + header = drm_buffer_read_object(cmdbuf->buffer, + sizeof(stack_header), &stack_header); - switch (header.header.cmd_type) { + switch (header->header.cmd_type) { case RADEON_CMD_PACKET: DRM_DEBUG("RADEON_CMD_PACKET\n"); if (radeon_emit_packets - (dev_priv, file_priv, header, cmdbuf)) { + (dev_priv, file_priv, *header, cmdbuf)) { DRM_ERROR("radeon_emit_packets failed\n"); goto err; } @@ -2915,7 +2930,7 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file case RADEON_CMD_SCALARS: DRM_DEBUG("RADEON_CMD_SCALARS\n"); - if (radeon_emit_scalars(dev_priv, header, cmdbuf)) { + if (radeon_emit_scalars(dev_priv, *header, cmdbuf)) { DRM_ERROR("radeon_emit_scalars failed\n"); goto err; } @@ -2923,7 +2938,7 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file case RADEON_CMD_VECTORS: DRM_DEBUG("RADEON_CMD_VECTORS\n"); - if (radeon_emit_vectors(dev_priv, header, cmdbuf)) { + if (radeon_emit_vectors(dev_priv, *header, cmdbuf)) { DRM_ERROR("radeon_emit_vectors failed\n"); goto err; } @@ -2931,7 +2946,7 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file case RADEON_CMD_DMA_DISCARD: DRM_DEBUG("RADEON_CMD_DMA_DISCARD\n"); - idx = header.dma.buf_idx; + idx = header->dma.buf_idx; if (idx < 0 || idx >= dma->buf_count) { DRM_ERROR("buffer index %d (of %d max)\n", idx, dma->buf_count - 1); @@ -2968,7 +2983,7 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file case RADEON_CMD_SCALARS2: DRM_DEBUG("RADEON_CMD_SCALARS2\n"); - if (radeon_emit_scalars2(dev_priv, header, cmdbuf)) { + if (radeon_emit_scalars2(dev_priv, *header, cmdbuf)) { DRM_ERROR("radeon_emit_scalars2 failed\n"); goto err; } @@ -2976,37 +2991,37 @@ static int radeon_cp_cmdbuf(struct drm_device *dev, void *data, struct drm_file case RADEON_CMD_WAIT: DRM_DEBUG("RADEON_CMD_WAIT\n"); - if (radeon_emit_wait(dev, header.wait.flags)) { + if (radeon_emit_wait(dev, header->wait.flags)) { DRM_ERROR("radeon_emit_wait failed\n"); goto err; } break; case RADEON_CMD_VECLINEAR: DRM_DEBUG("RADEON_CMD_VECLINEAR\n"); - if (radeon_emit_veclinear(dev_priv, header, cmdbuf)) { + if (radeon_emit_veclinear(dev_priv, *header, cmdbuf)) { DRM_ERROR("radeon_emit_veclinear failed\n"); goto err; } break; default: - DRM_ERROR("bad cmd_type %d at %p\n", - header.header.cmd_type, - cmdbuf->buf - sizeof(header)); + DRM_ERROR("bad cmd_type %d at byte %d\n", + header->header.cmd_type, + cmdbuf->buffer->iterator); goto err; } } - if (orig_bufsz != 0) - kfree(kbuf); + if (cmdbuf->bufsz != 0) + drm_buffer_free(cmdbuf->buffer); DRM_DEBUG("DONE\n"); COMMIT_RING(); return 0; err: - if (orig_bufsz != 0) - kfree(kbuf); + if (cmdbuf->bufsz != 0) + drm_buffer_free(cmdbuf->buffer); return -EINVAL; } diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c index 9f5e2f929da..313c96bc09d 100644 --- a/drivers/gpu/drm/radeon/radeon_test.c +++ b/drivers/gpu/drm/radeon/radeon_test.c @@ -186,7 +186,7 @@ void radeon_test_moves(struct radeon_device *rdev) radeon_bo_kunmap(gtt_obj[i]); DRM_INFO("Tested GTT->VRAM and VRAM->GTT copy for GTT offset 0x%llx\n", - gtt_addr - rdev->mc.gtt_location); + gtt_addr - rdev->mc.gtt_start); } out_cleanup: diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 58b5adf974c..43c5ab34b63 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -150,7 +150,7 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, man->default_caching = TTM_PL_FLAG_CACHED; break; case TTM_PL_TT: - man->gpu_offset = rdev->mc.gtt_location; + man->gpu_offset = rdev->mc.gtt_start; man->available_caching = TTM_PL_MASK_CACHING; man->default_caching = TTM_PL_FLAG_CACHED; man->flags = TTM_MEMTYPE_FLAG_MAPPABLE | TTM_MEMTYPE_FLAG_CMA; @@ -180,7 +180,7 @@ static int radeon_init_mem_type(struct ttm_bo_device *bdev, uint32_t type, break; case TTM_PL_VRAM: /* "On-card" video ram */ - man->gpu_offset = rdev->mc.vram_location; + man->gpu_offset = rdev->mc.vram_start; man->flags = TTM_MEMTYPE_FLAG_FIXED | TTM_MEMTYPE_FLAG_NEEDS_IOREMAP | TTM_MEMTYPE_FLAG_MAPPABLE; @@ -262,10 +262,10 @@ static int radeon_move_blit(struct ttm_buffer_object *bo, switch (old_mem->mem_type) { case TTM_PL_VRAM: - old_start += rdev->mc.vram_location; + old_start += rdev->mc.vram_start; break; case TTM_PL_TT: - old_start += rdev->mc.gtt_location; + old_start += rdev->mc.gtt_start; break; default: DRM_ERROR("Unknown placement %d\n", old_mem->mem_type); @@ -273,10 +273,10 @@ static int radeon_move_blit(struct ttm_buffer_object *bo, } switch (new_mem->mem_type) { case TTM_PL_VRAM: - new_start += rdev->mc.vram_location; + new_start += rdev->mc.vram_start; break; case TTM_PL_TT: - new_start += rdev->mc.gtt_location; + new_start += rdev->mc.gtt_start; break; default: DRM_ERROR("Unknown placement %d\n", old_mem->mem_type); diff --git a/drivers/gpu/drm/radeon/reg_srcs/r200 b/drivers/gpu/drm/radeon/reg_srcs/r200 index 6021c8849a1..c29ac434ac9 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r200 +++ b/drivers/gpu/drm/radeon/reg_srcs/r200 @@ -91,6 +91,8 @@ r200 0x3294 0x22b8 SE_TCL_TEX_CYL_WRAP_CTL 0x22c0 SE_TCL_UCP_VERT_BLEND_CNTL 0x22c4 SE_TCL_POINT_SPRITE_CNTL +0x22d0 SE_PVS_CNTL +0x22d4 SE_PVS_CONST_CNTL 0x2648 RE_POINTSIZE 0x26c0 RE_TOP_LEFT 0x26c4 RE_MISC diff --git a/drivers/gpu/drm/radeon/reg_srcs/r600 b/drivers/gpu/drm/radeon/reg_srcs/r600 new file mode 100644 index 00000000000..8f414a5f520 --- /dev/null +++ b/drivers/gpu/drm/radeon/reg_srcs/r600 @@ -0,0 +1,837 @@ +r600 0x9400 +0x000287A0 R7xx_CB_SHADER_CONTROL +0x00028230 R7xx_PA_SC_EDGERULE +0x000286C8 R7xx_SPI_THREAD_GROUPING +0x00008D8C R7xx_SQ_DYN_GPR_CNTL_PS_FLUSH_REQ +0x000088C4 VGT_CACHE_INVALIDATION +0x00028A50 VGT_ENHANCE +0x000088CC VGT_ES_PER_GS +0x00028A2C VGT_GROUP_DECR +0x00028A28 VGT_GROUP_FIRST_DECR +0x00028A24 VGT_GROUP_PRIM_TYPE +0x00028A30 VGT_GROUP_VECT_0_CNTL +0x00028A38 VGT_GROUP_VECT_0_FMT_CNTL +0x00028A34 VGT_GROUP_VECT_1_CNTL +0x00028A3C VGT_GROUP_VECT_1_FMT_CNTL +0x00028A40 VGT_GS_MODE +0x00028A6C VGT_GS_OUT_PRIM_TYPE +0x000088C8 VGT_GS_PER_ES +0x000088E8 VGT_GS_PER_VS +0x000088D4 VGT_GS_VERTEX_REUSE +0x00028A14 VGT_HOS_CNTL +0x00028A18 VGT_HOS_MAX_TESS_LEVEL +0x00028A1C VGT_HOS_MIN_TESS_LEVEL +0x00028A20 VGT_HOS_REUSE_DEPTH +0x0000895C VGT_INDEX_TYPE +0x00028408 VGT_INDX_OFFSET +0x00028AA0 VGT_INSTANCE_STEP_RATE_0 +0x00028AA4 VGT_INSTANCE_STEP_RATE_1 +0x000088C0 VGT_LAST_COPY_STATE +0x00028400 VGT_MAX_VTX_INDX +0x000088D8 VGT_MC_LAT_CNTL +0x00028404 VGT_MIN_VTX_INDX +0x00028A94 VGT_MULTI_PRIM_IB_RESET_EN +0x0002840C VGT_MULTI_PRIM_IB_RESET_INDX +0x00008970 VGT_NUM_INDICES +0x00008974 VGT_NUM_INSTANCES +0x00028A10 VGT_OUTPUT_PATH_CNTL +0x00028C5C VGT_OUT_DEALLOC_CNTL +0x00028A84 VGT_PRIMITIVEID_EN +0x00008958 VGT_PRIMITIVE_TYPE +0x00028AB4 VGT_REUSE_OFF +0x00028C58 VGT_VERTEX_REUSE_BLOCK_CNTL +0x00028AB8 VGT_VTX_CNT_EN +0x000088B0 VGT_VTX_VECT_EJECT_REG +0x00028810 PA_CL_CLIP_CNTL +0x00008A14 PA_CL_ENHANCE +0x00028C14 PA_CL_GB_HORZ_CLIP_ADJ +0x00028C18 PA_CL_GB_HORZ_DISC_ADJ +0x00028C0C PA_CL_GB_VERT_CLIP_ADJ +0x00028C10 PA_CL_GB_VERT_DISC_ADJ +0x00028820 PA_CL_NANINF_CNTL +0x00028E1C PA_CL_POINT_CULL_RAD +0x00028E18 PA_CL_POINT_SIZE +0x00028E10 PA_CL_POINT_X_RAD +0x00028E14 PA_CL_POINT_Y_RAD +0x00028E2C PA_CL_UCP_0_W +0x00028E3C PA_CL_UCP_1_W +0x00028E4C PA_CL_UCP_2_W +0x00028E5C PA_CL_UCP_3_W +0x00028E6C PA_CL_UCP_4_W +0x00028E7C PA_CL_UCP_5_W +0x00028E20 PA_CL_UCP_0_X +0x00028E30 PA_CL_UCP_1_X +0x00028E40 PA_CL_UCP_2_X +0x00028E50 PA_CL_UCP_3_X +0x00028E60 PA_CL_UCP_4_X +0x00028E70 PA_CL_UCP_5_X +0x00028E24 PA_CL_UCP_0_Y +0x00028E34 PA_CL_UCP_1_Y +0x00028E44 PA_CL_UCP_2_Y +0x00028E54 PA_CL_UCP_3_Y +0x00028E64 PA_CL_UCP_4_Y +0x00028E74 PA_CL_UCP_5_Y +0x00028E28 PA_CL_UCP_0_Z +0x00028E38 PA_CL_UCP_1_Z +0x00028E48 PA_CL_UCP_2_Z +0x00028E58 PA_CL_UCP_3_Z +0x00028E68 PA_CL_UCP_4_Z +0x00028E78 PA_CL_UCP_5_Z +0x00028440 PA_CL_VPORT_XOFFSET_0 +0x00028458 PA_CL_VPORT_XOFFSET_1 +0x00028470 PA_CL_VPORT_XOFFSET_2 +0x00028488 PA_CL_VPORT_XOFFSET_3 +0x000284A0 PA_CL_VPORT_XOFFSET_4 +0x000284B8 PA_CL_VPORT_XOFFSET_5 +0x000284D0 PA_CL_VPORT_XOFFSET_6 +0x000284E8 PA_CL_VPORT_XOFFSET_7 +0x00028500 PA_CL_VPORT_XOFFSET_8 +0x00028518 PA_CL_VPORT_XOFFSET_9 +0x00028530 PA_CL_VPORT_XOFFSET_10 +0x00028548 PA_CL_VPORT_XOFFSET_11 +0x00028560 PA_CL_VPORT_XOFFSET_12 +0x00028578 PA_CL_VPORT_XOFFSET_13 +0x00028590 PA_CL_VPORT_XOFFSET_14 +0x000285A8 PA_CL_VPORT_XOFFSET_15 +0x0002843C PA_CL_VPORT_XSCALE_0 +0x00028454 PA_CL_VPORT_XSCALE_1 +0x0002846C PA_CL_VPORT_XSCALE_2 +0x00028484 PA_CL_VPORT_XSCALE_3 +0x0002849C PA_CL_VPORT_XSCALE_4 +0x000284B4 PA_CL_VPORT_XSCALE_5 +0x000284CC PA_CL_VPORT_XSCALE_6 +0x000284E4 PA_CL_VPORT_XSCALE_7 +0x000284FC PA_CL_VPORT_XSCALE_8 +0x00028514 PA_CL_VPORT_XSCALE_9 +0x0002852C PA_CL_VPORT_XSCALE_10 +0x00028544 PA_CL_VPORT_XSCALE_11 +0x0002855C PA_CL_VPORT_XSCALE_12 +0x00028574 PA_CL_VPORT_XSCALE_13 +0x0002858C PA_CL_VPORT_XSCALE_14 +0x000285A4 PA_CL_VPORT_XSCALE_15 +0x00028448 PA_CL_VPORT_YOFFSET_0 +0x00028460 PA_CL_VPORT_YOFFSET_1 +0x00028478 PA_CL_VPORT_YOFFSET_2 +0x00028490 PA_CL_VPORT_YOFFSET_3 +0x000284A8 PA_CL_VPORT_YOFFSET_4 +0x000284C0 PA_CL_VPORT_YOFFSET_5 +0x000284D8 PA_CL_VPORT_YOFFSET_6 +0x000284F0 PA_CL_VPORT_YOFFSET_7 +0x00028508 PA_CL_VPORT_YOFFSET_8 +0x00028520 PA_CL_VPORT_YOFFSET_9 +0x00028538 PA_CL_VPORT_YOFFSET_10 +0x00028550 PA_CL_VPORT_YOFFSET_11 +0x00028568 PA_CL_VPORT_YOFFSET_12 +0x00028580 PA_CL_VPORT_YOFFSET_13 +0x00028598 PA_CL_VPORT_YOFFSET_14 +0x000285B0 PA_CL_VPORT_YOFFSET_15 +0x00028444 PA_CL_VPORT_YSCALE_0 +0x0002845C PA_CL_VPORT_YSCALE_1 +0x00028474 PA_CL_VPORT_YSCALE_2 +0x0002848C PA_CL_VPORT_YSCALE_3 +0x000284A4 PA_CL_VPORT_YSCALE_4 +0x000284BC PA_CL_VPORT_YSCALE_5 +0x000284D4 PA_CL_VPORT_YSCALE_6 +0x000284EC PA_CL_VPORT_YSCALE_7 +0x00028504 PA_CL_VPORT_YSCALE_8 +0x0002851C PA_CL_VPORT_YSCALE_9 +0x00028534 PA_CL_VPORT_YSCALE_10 +0x0002854C PA_CL_VPORT_YSCALE_11 +0x00028564 PA_CL_VPORT_YSCALE_12 +0x0002857C PA_CL_VPORT_YSCALE_13 +0x00028594 PA_CL_VPORT_YSCALE_14 +0x000285AC PA_CL_VPORT_YSCALE_15 +0x00028450 PA_CL_VPORT_ZOFFSET_0 +0x00028468 PA_CL_VPORT_ZOFFSET_1 +0x00028480 PA_CL_VPORT_ZOFFSET_2 +0x00028498 PA_CL_VPORT_ZOFFSET_3 +0x000284B0 PA_CL_VPORT_ZOFFSET_4 +0x000284C8 PA_CL_VPORT_ZOFFSET_5 +0x000284E0 PA_CL_VPORT_ZOFFSET_6 +0x000284F8 PA_CL_VPORT_ZOFFSET_7 +0x00028510 PA_CL_VPORT_ZOFFSET_8 +0x00028528 PA_CL_VPORT_ZOFFSET_9 +0x00028540 PA_CL_VPORT_ZOFFSET_10 +0x00028558 PA_CL_VPORT_ZOFFSET_11 +0x00028570 PA_CL_VPORT_ZOFFSET_12 +0x00028588 PA_CL_VPORT_ZOFFSET_13 +0x000285A0 PA_CL_VPORT_ZOFFSET_14 +0x000285B8 PA_CL_VPORT_ZOFFSET_15 +0x0002844C PA_CL_VPORT_ZSCALE_0 +0x00028464 PA_CL_VPORT_ZSCALE_1 +0x0002847C PA_CL_VPORT_ZSCALE_2 +0x00028494 PA_CL_VPORT_ZSCALE_3 +0x000284AC PA_CL_VPORT_ZSCALE_4 +0x000284C4 PA_CL_VPORT_ZSCALE_5 +0x000284DC PA_CL_VPORT_ZSCALE_6 +0x000284F4 PA_CL_VPORT_ZSCALE_7 +0x0002850C PA_CL_VPORT_ZSCALE_8 +0x00028524 PA_CL_VPORT_ZSCALE_9 +0x0002853C PA_CL_VPORT_ZSCALE_10 +0x00028554 PA_CL_VPORT_ZSCALE_11 +0x0002856C PA_CL_VPORT_ZSCALE_12 +0x00028584 PA_CL_VPORT_ZSCALE_13 +0x0002859C PA_CL_VPORT_ZSCALE_14 +0x000285B4 PA_CL_VPORT_ZSCALE_15 +0x0002881C PA_CL_VS_OUT_CNTL +0x00028818 PA_CL_VTE_CNTL +0x00028C48 PA_SC_AA_MASK +0x00008B40 PA_SC_AA_SAMPLE_LOCS_2S +0x00008B44 PA_SC_AA_SAMPLE_LOCS_4S +0x00008B48 PA_SC_AA_SAMPLE_LOCS_8S_WD0 +0x00008B4C PA_SC_AA_SAMPLE_LOCS_8S_WD1 +0x00028C20 PA_SC_AA_SAMPLE_LOCS_8S_WD1_MCTX +0x00028C1C PA_SC_AA_SAMPLE_LOCS_MCTX +0x00028214 PA_SC_CLIPRECT_0_BR +0x0002821C PA_SC_CLIPRECT_1_BR +0x00028224 PA_SC_CLIPRECT_2_BR +0x0002822C PA_SC_CLIPRECT_3_BR +0x00028210 PA_SC_CLIPRECT_0_TL +0x00028218 PA_SC_CLIPRECT_1_TL +0x00028220 PA_SC_CLIPRECT_2_TL +0x00028228 PA_SC_CLIPRECT_3_TL +0x0002820C PA_SC_CLIPRECT_RULE +0x00008BF0 PA_SC_ENHANCE +0x00028244 PA_SC_GENERIC_SCISSOR_BR +0x00028240 PA_SC_GENERIC_SCISSOR_TL +0x00028C00 PA_SC_LINE_CNTL +0x00028A0C PA_SC_LINE_STIPPLE +0x00008B10 PA_SC_LINE_STIPPLE_STATE +0x00028A4C PA_SC_MODE_CNTL +0x00028A48 PA_SC_MPASS_PS_CNTL +0x00008B20 PA_SC_MULTI_CHIP_CNTL +0x00028034 PA_SC_SCREEN_SCISSOR_BR +0x00028030 PA_SC_SCREEN_SCISSOR_TL +0x00028254 PA_SC_VPORT_SCISSOR_0_BR +0x0002825C PA_SC_VPORT_SCISSOR_1_BR +0x00028264 PA_SC_VPORT_SCISSOR_2_BR +0x0002826C PA_SC_VPORT_SCISSOR_3_BR +0x00028274 PA_SC_VPORT_SCISSOR_4_BR +0x0002827C PA_SC_VPORT_SCISSOR_5_BR +0x00028284 PA_SC_VPORT_SCISSOR_6_BR +0x0002828C PA_SC_VPORT_SCISSOR_7_BR +0x00028294 PA_SC_VPORT_SCISSOR_8_BR +0x0002829C PA_SC_VPORT_SCISSOR_9_BR +0x000282A4 PA_SC_VPORT_SCISSOR_10_BR +0x000282AC PA_SC_VPORT_SCISSOR_11_BR +0x000282B4 PA_SC_VPORT_SCISSOR_12_BR +0x000282BC PA_SC_VPORT_SCISSOR_13_BR +0x000282C4 PA_SC_VPORT_SCISSOR_14_BR +0x000282CC PA_SC_VPORT_SCISSOR_15_BR +0x00028250 PA_SC_VPORT_SCISSOR_0_TL +0x00028258 PA_SC_VPORT_SCISSOR_1_TL +0x00028260 PA_SC_VPORT_SCISSOR_2_TL +0x00028268 PA_SC_VPORT_SCISSOR_3_TL +0x00028270 PA_SC_VPORT_SCISSOR_4_TL +0x00028278 PA_SC_VPORT_SCISSOR_5_TL +0x00028280 PA_SC_VPORT_SCISSOR_6_TL +0x00028288 PA_SC_VPORT_SCISSOR_7_TL +0x00028290 PA_SC_VPORT_SCISSOR_8_TL +0x00028298 PA_SC_VPORT_SCISSOR_9_TL +0x000282A0 PA_SC_VPORT_SCISSOR_10_TL +0x000282A8 PA_SC_VPORT_SCISSOR_11_TL +0x000282B0 PA_SC_VPORT_SCISSOR_12_TL +0x000282B8 PA_SC_VPORT_SCISSOR_13_TL +0x000282C0 PA_SC_VPORT_SCISSOR_14_TL +0x000282C8 PA_SC_VPORT_SCISSOR_15_TL +0x000282D4 PA_SC_VPORT_ZMAX_0 +0x000282DC PA_SC_VPORT_ZMAX_1 +0x000282E4 PA_SC_VPORT_ZMAX_2 +0x000282EC PA_SC_VPORT_ZMAX_3 +0x000282F4 PA_SC_VPORT_ZMAX_4 +0x000282FC PA_SC_VPORT_ZMAX_5 +0x00028304 PA_SC_VPORT_ZMAX_6 +0x0002830C PA_SC_VPORT_ZMAX_7 +0x00028314 PA_SC_VPORT_ZMAX_8 +0x0002831C PA_SC_VPORT_ZMAX_9 +0x00028324 PA_SC_VPORT_ZMAX_10 +0x0002832C PA_SC_VPORT_ZMAX_11 +0x00028334 PA_SC_VPORT_ZMAX_12 +0x0002833C PA_SC_VPORT_ZMAX_13 +0x00028344 PA_SC_VPORT_ZMAX_14 +0x0002834C PA_SC_VPORT_ZMAX_15 +0x000282D0 PA_SC_VPORT_ZMIN_0 +0x000282D8 PA_SC_VPORT_ZMIN_1 +0x000282E0 PA_SC_VPORT_ZMIN_2 +0x000282E8 PA_SC_VPORT_ZMIN_3 +0x000282F0 PA_SC_VPORT_ZMIN_4 +0x000282F8 PA_SC_VPORT_ZMIN_5 +0x00028300 PA_SC_VPORT_ZMIN_6 +0x00028308 PA_SC_VPORT_ZMIN_7 +0x00028310 PA_SC_VPORT_ZMIN_8 +0x00028318 PA_SC_VPORT_ZMIN_9 +0x00028320 PA_SC_VPORT_ZMIN_10 +0x00028328 PA_SC_VPORT_ZMIN_11 +0x00028330 PA_SC_VPORT_ZMIN_12 +0x00028338 PA_SC_VPORT_ZMIN_13 +0x00028340 PA_SC_VPORT_ZMIN_14 +0x00028348 PA_SC_VPORT_ZMIN_15 +0x00028200 PA_SC_WINDOW_OFFSET +0x00028208 PA_SC_WINDOW_SCISSOR_BR +0x00028204 PA_SC_WINDOW_SCISSOR_TL +0x00028A08 PA_SU_LINE_CNTL +0x00028A04 PA_SU_POINT_MINMAX +0x00028A00 PA_SU_POINT_SIZE +0x00028E0C PA_SU_POLY_OFFSET_BACK_OFFSET +0x00028E08 PA_SU_POLY_OFFSET_BACK_SCALE +0x00028DFC PA_SU_POLY_OFFSET_CLAMP +0x00028DF8 PA_SU_POLY_OFFSET_DB_FMT_CNTL +0x00028E04 PA_SU_POLY_OFFSET_FRONT_OFFSET +0x00028E00 PA_SU_POLY_OFFSET_FRONT_SCALE +0x00028814 PA_SU_SC_MODE_CNTL +0x00028C08 PA_SU_VTX_CNTL +0x00008C00 SQ_CONFIG +0x00008C04 SQ_GPR_RESOURCE_MGMT_1 +0x00008C08 SQ_GPR_RESOURCE_MGMT_2 +0x00008C10 SQ_STACK_RESOURCE_MGMT_1 +0x00008C14 SQ_STACK_RESOURCE_MGMT_2 +0x00008C0C SQ_THREAD_RESOURCE_MGMT +0x00028380 SQ_VTX_SEMANTIC_0 +0x00028384 SQ_VTX_SEMANTIC_1 +0x00028388 SQ_VTX_SEMANTIC_2 +0x0002838C SQ_VTX_SEMANTIC_3 +0x00028390 SQ_VTX_SEMANTIC_4 +0x00028394 SQ_VTX_SEMANTIC_5 +0x00028398 SQ_VTX_SEMANTIC_6 +0x0002839C SQ_VTX_SEMANTIC_7 +0x000283A0 SQ_VTX_SEMANTIC_8 +0x000283A4 SQ_VTX_SEMANTIC_9 +0x000283A8 SQ_VTX_SEMANTIC_10 +0x000283AC SQ_VTX_SEMANTIC_11 +0x000283B0 SQ_VTX_SEMANTIC_12 +0x000283B4 SQ_VTX_SEMANTIC_13 +0x000283B8 SQ_VTX_SEMANTIC_14 +0x000283BC SQ_VTX_SEMANTIC_15 +0x000283C0 SQ_VTX_SEMANTIC_16 +0x000283C4 SQ_VTX_SEMANTIC_17 +0x000283C8 SQ_VTX_SEMANTIC_18 +0x000283CC SQ_VTX_SEMANTIC_19 +0x000283D0 SQ_VTX_SEMANTIC_20 +0x000283D4 SQ_VTX_SEMANTIC_21 +0x000283D8 SQ_VTX_SEMANTIC_22 +0x000283DC SQ_VTX_SEMANTIC_23 +0x000283E0 SQ_VTX_SEMANTIC_24 +0x000283E4 SQ_VTX_SEMANTIC_25 +0x000283E8 SQ_VTX_SEMANTIC_26 +0x000283EC SQ_VTX_SEMANTIC_27 +0x000283F0 SQ_VTX_SEMANTIC_28 +0x000283F4 SQ_VTX_SEMANTIC_29 +0x000283F8 SQ_VTX_SEMANTIC_30 +0x000283FC SQ_VTX_SEMANTIC_31 +0x000288E0 SQ_VTX_SEMANTIC_CLEAR +0x0003CFF4 SQ_VTX_START_INST_LOC +0x0003C000 SQ_TEX_SAMPLER_WORD0_0 +0x0003C004 SQ_TEX_SAMPLER_WORD1_0 +0x0003C008 SQ_TEX_SAMPLER_WORD2_0 +0x00030000 SQ_ALU_CONSTANT0_0 +0x00030004 SQ_ALU_CONSTANT1_0 +0x00030008 SQ_ALU_CONSTANT2_0 +0x0003000C SQ_ALU_CONSTANT3_0 +0x0003E380 SQ_BOOL_CONST_0 +0x0003E384 SQ_BOOL_CONST_1 +0x0003E388 SQ_BOOL_CONST_2 +0x0003E200 SQ_LOOP_CONST_0 +0x0003E200 SQ_LOOP_CONST_DX10_0 +0x000281C0 SQ_ALU_CONST_BUFFER_SIZE_GS_0 +0x000281C4 SQ_ALU_CONST_BUFFER_SIZE_GS_1 +0x000281C8 SQ_ALU_CONST_BUFFER_SIZE_GS_2 +0x000281CC SQ_ALU_CONST_BUFFER_SIZE_GS_3 +0x000281D0 SQ_ALU_CONST_BUFFER_SIZE_GS_4 +0x000281D4 SQ_ALU_CONST_BUFFER_SIZE_GS_5 +0x000281D8 SQ_ALU_CONST_BUFFER_SIZE_GS_6 +0x000281DC SQ_ALU_CONST_BUFFER_SIZE_GS_7 +0x000281E0 SQ_ALU_CONST_BUFFER_SIZE_GS_8 +0x000281E4 SQ_ALU_CONST_BUFFER_SIZE_GS_9 +0x000281E8 SQ_ALU_CONST_BUFFER_SIZE_GS_10 +0x000281EC SQ_ALU_CONST_BUFFER_SIZE_GS_11 +0x000281F0 SQ_ALU_CONST_BUFFER_SIZE_GS_12 +0x000281F4 SQ_ALU_CONST_BUFFER_SIZE_GS_13 +0x000281F8 SQ_ALU_CONST_BUFFER_SIZE_GS_14 +0x000281FC SQ_ALU_CONST_BUFFER_SIZE_GS_15 +0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0 +0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1 +0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2 +0x0002814C SQ_ALU_CONST_BUFFER_SIZE_PS_3 +0x00028150 SQ_ALU_CONST_BUFFER_SIZE_PS_4 +0x00028154 SQ_ALU_CONST_BUFFER_SIZE_PS_5 +0x00028158 SQ_ALU_CONST_BUFFER_SIZE_PS_6 +0x0002815C SQ_ALU_CONST_BUFFER_SIZE_PS_7 +0x00028160 SQ_ALU_CONST_BUFFER_SIZE_PS_8 +0x00028164 SQ_ALU_CONST_BUFFER_SIZE_PS_9 +0x00028168 SQ_ALU_CONST_BUFFER_SIZE_PS_10 +0x0002816C SQ_ALU_CONST_BUFFER_SIZE_PS_11 +0x00028170 SQ_ALU_CONST_BUFFER_SIZE_PS_12 +0x00028174 SQ_ALU_CONST_BUFFER_SIZE_PS_13 +0x00028178 SQ_ALU_CONST_BUFFER_SIZE_PS_14 +0x0002817C SQ_ALU_CONST_BUFFER_SIZE_PS_15 +0x00028180 SQ_ALU_CONST_BUFFER_SIZE_VS_0 +0x00028184 SQ_ALU_CONST_BUFFER_SIZE_VS_1 +0x00028188 SQ_ALU_CONST_BUFFER_SIZE_VS_2 +0x0002818C SQ_ALU_CONST_BUFFER_SIZE_VS_3 +0x00028190 SQ_ALU_CONST_BUFFER_SIZE_VS_4 +0x00028194 SQ_ALU_CONST_BUFFER_SIZE_VS_5 +0x00028198 SQ_ALU_CONST_BUFFER_SIZE_VS_6 +0x0002819C SQ_ALU_CONST_BUFFER_SIZE_VS_7 +0x000281A0 SQ_ALU_CONST_BUFFER_SIZE_VS_8 +0x000281A4 SQ_ALU_CONST_BUFFER_SIZE_VS_9 +0x000281A8 SQ_ALU_CONST_BUFFER_SIZE_VS_10 +0x000281AC SQ_ALU_CONST_BUFFER_SIZE_VS_11 +0x000281B0 SQ_ALU_CONST_BUFFER_SIZE_VS_12 +0x000281B4 SQ_ALU_CONST_BUFFER_SIZE_VS_13 +0x000281B8 SQ_ALU_CONST_BUFFER_SIZE_VS_14 +0x000281BC SQ_ALU_CONST_BUFFER_SIZE_VS_15 +0x000289C0 SQ_ALU_CONST_CACHE_GS_0 +0x000289C4 SQ_ALU_CONST_CACHE_GS_1 +0x000289C8 SQ_ALU_CONST_CACHE_GS_2 +0x000289CC SQ_ALU_CONST_CACHE_GS_3 +0x000289D0 SQ_ALU_CONST_CACHE_GS_4 +0x000289D4 SQ_ALU_CONST_CACHE_GS_5 +0x000289D8 SQ_ALU_CONST_CACHE_GS_6 +0x000289DC SQ_ALU_CONST_CACHE_GS_7 +0x000289E0 SQ_ALU_CONST_CACHE_GS_8 +0x000289E4 SQ_ALU_CONST_CACHE_GS_9 +0x000289E8 SQ_ALU_CONST_CACHE_GS_10 +0x000289EC SQ_ALU_CONST_CACHE_GS_11 +0x000289F0 SQ_ALU_CONST_CACHE_GS_12 +0x000289F4 SQ_ALU_CONST_CACHE_GS_13 +0x000289F8 SQ_ALU_CONST_CACHE_GS_14 +0x000289FC SQ_ALU_CONST_CACHE_GS_15 +0x00028940 SQ_ALU_CONST_CACHE_PS_0 +0x00028944 SQ_ALU_CONST_CACHE_PS_1 +0x00028948 SQ_ALU_CONST_CACHE_PS_2 +0x0002894C SQ_ALU_CONST_CACHE_PS_3 +0x00028950 SQ_ALU_CONST_CACHE_PS_4 +0x00028954 SQ_ALU_CONST_CACHE_PS_5 +0x00028958 SQ_ALU_CONST_CACHE_PS_6 +0x0002895C SQ_ALU_CONST_CACHE_PS_7 +0x00028960 SQ_ALU_CONST_CACHE_PS_8 +0x00028964 SQ_ALU_CONST_CACHE_PS_9 +0x00028968 SQ_ALU_CONST_CACHE_PS_10 +0x0002896C SQ_ALU_CONST_CACHE_PS_11 +0x00028970 SQ_ALU_CONST_CACHE_PS_12 +0x00028974 SQ_ALU_CONST_CACHE_PS_13 +0x00028978 SQ_ALU_CONST_CACHE_PS_14 +0x0002897C SQ_ALU_CONST_CACHE_PS_15 +0x00028980 SQ_ALU_CONST_CACHE_VS_0 +0x00028984 SQ_ALU_CONST_CACHE_VS_1 +0x00028988 SQ_ALU_CONST_CACHE_VS_2 +0x0002898C SQ_ALU_CONST_CACHE_VS_3 +0x00028990 SQ_ALU_CONST_CACHE_VS_4 +0x00028994 SQ_ALU_CONST_CACHE_VS_5 +0x00028998 SQ_ALU_CONST_CACHE_VS_6 +0x0002899C SQ_ALU_CONST_CACHE_VS_7 +0x000289A0 SQ_ALU_CONST_CACHE_VS_8 +0x000289A4 SQ_ALU_CONST_CACHE_VS_9 +0x000289A8 SQ_ALU_CONST_CACHE_VS_10 +0x000289AC SQ_ALU_CONST_CACHE_VS_11 +0x000289B0 SQ_ALU_CONST_CACHE_VS_12 +0x000289B4 SQ_ALU_CONST_CACHE_VS_13 +0x000289B8 SQ_ALU_CONST_CACHE_VS_14 +0x000289BC SQ_ALU_CONST_CACHE_VS_15 +0x000288D8 SQ_PGM_CF_OFFSET_ES +0x000288DC SQ_PGM_CF_OFFSET_FS +0x000288D4 SQ_PGM_CF_OFFSET_GS +0x000288CC SQ_PGM_CF_OFFSET_PS +0x000288D0 SQ_PGM_CF_OFFSET_VS +0x00028854 SQ_PGM_EXPORTS_PS +0x00028890 SQ_PGM_RESOURCES_ES +0x000288A4 SQ_PGM_RESOURCES_FS +0x0002887C SQ_PGM_RESOURCES_GS +0x00028850 SQ_PGM_RESOURCES_PS +0x00028868 SQ_PGM_RESOURCES_VS +0x00009100 SPI_CONFIG_CNTL +0x0000913C SPI_CONFIG_CNTL_1 +0x000286DC SPI_FOG_CNTL +0x000286E4 SPI_FOG_FUNC_BIAS +0x000286E0 SPI_FOG_FUNC_SCALE +0x000286D8 SPI_INPUT_Z +0x000286D4 SPI_INTERP_CONTROL_0 +0x00028644 SPI_PS_INPUT_CNTL_0 +0x00028648 SPI_PS_INPUT_CNTL_1 +0x0002864C SPI_PS_INPUT_CNTL_2 +0x00028650 SPI_PS_INPUT_CNTL_3 +0x00028654 SPI_PS_INPUT_CNTL_4 +0x00028658 SPI_PS_INPUT_CNTL_5 +0x0002865C SPI_PS_INPUT_CNTL_6 +0x00028660 SPI_PS_INPUT_CNTL_7 +0x00028664 SPI_PS_INPUT_CNTL_8 +0x00028668 SPI_PS_INPUT_CNTL_9 +0x0002866C SPI_PS_INPUT_CNTL_10 +0x00028670 SPI_PS_INPUT_CNTL_11 +0x00028674 SPI_PS_INPUT_CNTL_12 +0x00028678 SPI_PS_INPUT_CNTL_13 +0x0002867C SPI_PS_INPUT_CNTL_14 +0x00028680 SPI_PS_INPUT_CNTL_15 +0x00028684 SPI_PS_INPUT_CNTL_16 +0x00028688 SPI_PS_INPUT_CNTL_17 +0x0002868C SPI_PS_INPUT_CNTL_18 +0x00028690 SPI_PS_INPUT_CNTL_19 +0x00028694 SPI_PS_INPUT_CNTL_20 +0x00028698 SPI_PS_INPUT_CNTL_21 +0x0002869C SPI_PS_INPUT_CNTL_22 +0x000286A0 SPI_PS_INPUT_CNTL_23 +0x000286A4 SPI_PS_INPUT_CNTL_24 +0x000286A8 SPI_PS_INPUT_CNTL_25 +0x000286AC SPI_PS_INPUT_CNTL_26 +0x000286B0 SPI_PS_INPUT_CNTL_27 +0x000286B4 SPI_PS_INPUT_CNTL_28 +0x000286B8 SPI_PS_INPUT_CNTL_29 +0x000286BC SPI_PS_INPUT_CNTL_30 +0x000286C0 SPI_PS_INPUT_CNTL_31 +0x000286CC SPI_PS_IN_CONTROL_0 +0x000286D0 SPI_PS_IN_CONTROL_1 +0x000286C4 SPI_VS_OUT_CONFIG +0x00028614 SPI_VS_OUT_ID_0 +0x00028618 SPI_VS_OUT_ID_1 +0x0002861C SPI_VS_OUT_ID_2 +0x00028620 SPI_VS_OUT_ID_3 +0x00028624 SPI_VS_OUT_ID_4 +0x00028628 SPI_VS_OUT_ID_5 +0x0002862C SPI_VS_OUT_ID_6 +0x00028630 SPI_VS_OUT_ID_7 +0x00028634 SPI_VS_OUT_ID_8 +0x00028638 SPI_VS_OUT_ID_9 +0x00028438 SX_ALPHA_REF +0x00028410 SX_ALPHA_TEST_CONTROL +0x00028350 SX_MISC +0x0000A020 SMX_DC_CTL0 +0x0000A024 SMX_DC_CTL1 +0x0000A028 SMX_DC_CTL2 +0x00009608 TC_CNTL +0x00009604 TC_INVALIDATE +0x00009490 TD_CNTL +0x00009400 TD_FILTER4 +0x00009404 TD_FILTER4_1 +0x00009408 TD_FILTER4_2 +0x0000940C TD_FILTER4_3 +0x00009410 TD_FILTER4_4 +0x00009414 TD_FILTER4_5 +0x00009418 TD_FILTER4_6 +0x0000941C TD_FILTER4_7 +0x00009420 TD_FILTER4_8 +0x00009424 TD_FILTER4_9 +0x00009428 TD_FILTER4_10 +0x0000942C TD_FILTER4_11 +0x00009430 TD_FILTER4_12 +0x00009434 TD_FILTER4_13 +0x00009438 TD_FILTER4_14 +0x0000943C TD_FILTER4_15 +0x00009440 TD_FILTER4_16 +0x00009444 TD_FILTER4_17 +0x00009448 TD_FILTER4_18 +0x0000944C TD_FILTER4_19 +0x00009450 TD_FILTER4_20 +0x00009454 TD_FILTER4_21 +0x00009458 TD_FILTER4_22 +0x0000945C TD_FILTER4_23 +0x00009460 TD_FILTER4_24 +0x00009464 TD_FILTER4_25 +0x00009468 TD_FILTER4_26 +0x0000946C TD_FILTER4_27 +0x00009470 TD_FILTER4_28 +0x00009474 TD_FILTER4_29 +0x00009478 TD_FILTER4_30 +0x0000947C TD_FILTER4_31 +0x00009480 TD_FILTER4_32 +0x00009484 TD_FILTER4_33 +0x00009488 TD_FILTER4_34 +0x0000948C TD_FILTER4_35 +0x0000A80C TD_GS_SAMPLER0_BORDER_ALPHA +0x0000A81C TD_GS_SAMPLER1_BORDER_ALPHA +0x0000A82C TD_GS_SAMPLER2_BORDER_ALPHA +0x0000A83C TD_GS_SAMPLER3_BORDER_ALPHA +0x0000A84C TD_GS_SAMPLER4_BORDER_ALPHA +0x0000A85C TD_GS_SAMPLER5_BORDER_ALPHA +0x0000A86C TD_GS_SAMPLER6_BORDER_ALPHA +0x0000A87C TD_GS_SAMPLER7_BORDER_ALPHA +0x0000A88C TD_GS_SAMPLER8_BORDER_ALPHA +0x0000A89C TD_GS_SAMPLER9_BORDER_ALPHA +0x0000A8AC TD_GS_SAMPLER10_BORDER_ALPHA +0x0000A8BC TD_GS_SAMPLER11_BORDER_ALPHA +0x0000A8CC TD_GS_SAMPLER12_BORDER_ALPHA +0x0000A8DC TD_GS_SAMPLER13_BORDER_ALPHA +0x0000A8EC TD_GS_SAMPLER14_BORDER_ALPHA +0x0000A8FC TD_GS_SAMPLER15_BORDER_ALPHA +0x0000A90C TD_GS_SAMPLER16_BORDER_ALPHA +0x0000A91C TD_GS_SAMPLER17_BORDER_ALPHA +0x0000A808 TD_GS_SAMPLER0_BORDER_BLUE +0x0000A818 TD_GS_SAMPLER1_BORDER_BLUE +0x0000A828 TD_GS_SAMPLER2_BORDER_BLUE +0x0000A838 TD_GS_SAMPLER3_BORDER_BLUE +0x0000A848 TD_GS_SAMPLER4_BORDER_BLUE +0x0000A858 TD_GS_SAMPLER5_BORDER_BLUE +0x0000A868 TD_GS_SAMPLER6_BORDER_BLUE +0x0000A878 TD_GS_SAMPLER7_BORDER_BLUE +0x0000A888 TD_GS_SAMPLER8_BORDER_BLUE +0x0000A898 TD_GS_SAMPLER9_BORDER_BLUE +0x0000A8A8 TD_GS_SAMPLER10_BORDER_BLUE +0x0000A8B8 TD_GS_SAMPLER11_BORDER_BLUE +0x0000A8C8 TD_GS_SAMPLER12_BORDER_BLUE +0x0000A8D8 TD_GS_SAMPLER13_BORDER_BLUE +0x0000A8E8 TD_GS_SAMPLER14_BORDER_BLUE +0x0000A8F8 TD_GS_SAMPLER15_BORDER_BLUE +0x0000A908 TD_GS_SAMPLER16_BORDER_BLUE +0x0000A918 TD_GS_SAMPLER17_BORDER_BLUE +0x0000A804 TD_GS_SAMPLER0_BORDER_GREEN +0x0000A814 TD_GS_SAMPLER1_BORDER_GREEN +0x0000A824 TD_GS_SAMPLER2_BORDER_GREEN +0x0000A834 TD_GS_SAMPLER3_BORDER_GREEN +0x0000A844 TD_GS_SAMPLER4_BORDER_GREEN +0x0000A854 TD_GS_SAMPLER5_BORDER_GREEN +0x0000A864 TD_GS_SAMPLER6_BORDER_GREEN +0x0000A874 TD_GS_SAMPLER7_BORDER_GREEN +0x0000A884 TD_GS_SAMPLER8_BORDER_GREEN +0x0000A894 TD_GS_SAMPLER9_BORDER_GREEN +0x0000A8A4 TD_GS_SAMPLER10_BORDER_GREEN +0x0000A8B4 TD_GS_SAMPLER11_BORDER_GREEN +0x0000A8C4 TD_GS_SAMPLER12_BORDER_GREEN +0x0000A8D4 TD_GS_SAMPLER13_BORDER_GREEN +0x0000A8E4 TD_GS_SAMPLER14_BORDER_GREEN +0x0000A8F4 TD_GS_SAMPLER15_BORDER_GREEN +0x0000A904 TD_GS_SAMPLER16_BORDER_GREEN +0x0000A914 TD_GS_SAMPLER17_BORDER_GREEN +0x0000A800 TD_GS_SAMPLER0_BORDER_RED +0x0000A810 TD_GS_SAMPLER1_BORDER_RED +0x0000A820 TD_GS_SAMPLER2_BORDER_RED +0x0000A830 TD_GS_SAMPLER3_BORDER_RED +0x0000A840 TD_GS_SAMPLER4_BORDER_RED +0x0000A850 TD_GS_SAMPLER5_BORDER_RED +0x0000A860 TD_GS_SAMPLER6_BORDER_RED +0x0000A870 TD_GS_SAMPLER7_BORDER_RED +0x0000A880 TD_GS_SAMPLER8_BORDER_RED +0x0000A890 TD_GS_SAMPLER9_BORDER_RED +0x0000A8A0 TD_GS_SAMPLER10_BORDER_RED +0x0000A8B0 TD_GS_SAMPLER11_BORDER_RED +0x0000A8C0 TD_GS_SAMPLER12_BORDER_RED +0x0000A8D0 TD_GS_SAMPLER13_BORDER_RED +0x0000A8E0 TD_GS_SAMPLER14_BORDER_RED +0x0000A8F0 TD_GS_SAMPLER15_BORDER_RED +0x0000A900 TD_GS_SAMPLER16_BORDER_RED +0x0000A910 TD_GS_SAMPLER17_BORDER_RED +0x0000A40C TD_PS_SAMPLER0_BORDER_ALPHA +0x0000A41C TD_PS_SAMPLER1_BORDER_ALPHA +0x0000A42C TD_PS_SAMPLER2_BORDER_ALPHA +0x0000A43C TD_PS_SAMPLER3_BORDER_ALPHA +0x0000A44C TD_PS_SAMPLER4_BORDER_ALPHA +0x0000A45C TD_PS_SAMPLER5_BORDER_ALPHA +0x0000A46C TD_PS_SAMPLER6_BORDER_ALPHA +0x0000A47C TD_PS_SAMPLER7_BORDER_ALPHA +0x0000A48C TD_PS_SAMPLER8_BORDER_ALPHA +0x0000A49C TD_PS_SAMPLER9_BORDER_ALPHA +0x0000A4AC TD_PS_SAMPLER10_BORDER_ALPHA +0x0000A4BC TD_PS_SAMPLER11_BORDER_ALPHA +0x0000A4CC TD_PS_SAMPLER12_BORDER_ALPHA +0x0000A4DC TD_PS_SAMPLER13_BORDER_ALPHA +0x0000A4EC TD_PS_SAMPLER14_BORDER_ALPHA +0x0000A4FC TD_PS_SAMPLER15_BORDER_ALPHA +0x0000A50C TD_PS_SAMPLER16_BORDER_ALPHA +0x0000A51C TD_PS_SAMPLER17_BORDER_ALPHA +0x0000A408 TD_PS_SAMPLER0_BORDER_BLUE +0x0000A418 TD_PS_SAMPLER1_BORDER_BLUE +0x0000A428 TD_PS_SAMPLER2_BORDER_BLUE +0x0000A438 TD_PS_SAMPLER3_BORDER_BLUE +0x0000A448 TD_PS_SAMPLER4_BORDER_BLUE +0x0000A458 TD_PS_SAMPLER5_BORDER_BLUE +0x0000A468 TD_PS_SAMPLER6_BORDER_BLUE +0x0000A478 TD_PS_SAMPLER7_BORDER_BLUE +0x0000A488 TD_PS_SAMPLER8_BORDER_BLUE +0x0000A498 TD_PS_SAMPLER9_BORDER_BLUE +0x0000A4A8 TD_PS_SAMPLER10_BORDER_BLUE +0x0000A4B8 TD_PS_SAMPLER11_BORDER_BLUE +0x0000A4C8 TD_PS_SAMPLER12_BORDER_BLUE +0x0000A4D8 TD_PS_SAMPLER13_BORDER_BLUE +0x0000A4E8 TD_PS_SAMPLER14_BORDER_BLUE +0x0000A4F8 TD_PS_SAMPLER15_BORDER_BLUE +0x0000A508 TD_PS_SAMPLER16_BORDER_BLUE +0x0000A518 TD_PS_SAMPLER17_BORDER_BLUE +0x0000A404 TD_PS_SAMPLER0_BORDER_GREEN +0x0000A414 TD_PS_SAMPLER1_BORDER_GREEN +0x0000A424 TD_PS_SAMPLER2_BORDER_GREEN +0x0000A434 TD_PS_SAMPLER3_BORDER_GREEN +0x0000A444 TD_PS_SAMPLER4_BORDER_GREEN +0x0000A454 TD_PS_SAMPLER5_BORDER_GREEN +0x0000A464 TD_PS_SAMPLER6_BORDER_GREEN +0x0000A474 TD_PS_SAMPLER7_BORDER_GREEN +0x0000A484 TD_PS_SAMPLER8_BORDER_GREEN +0x0000A494 TD_PS_SAMPLER9_BORDER_GREEN +0x0000A4A4 TD_PS_SAMPLER10_BORDER_GREEN +0x0000A4B4 TD_PS_SAMPLER11_BORDER_GREEN +0x0000A4C4 TD_PS_SAMPLER12_BORDER_GREEN +0x0000A4D4 TD_PS_SAMPLER13_BORDER_GREEN +0x0000A4E4 TD_PS_SAMPLER14_BORDER_GREEN +0x0000A4F4 TD_PS_SAMPLER15_BORDER_GREEN +0x0000A504 TD_PS_SAMPLER16_BORDER_GREEN +0x0000A514 TD_PS_SAMPLER17_BORDER_GREEN +0x0000A400 TD_PS_SAMPLER0_BORDER_RED +0x0000A410 TD_PS_SAMPLER1_BORDER_RED +0x0000A420 TD_PS_SAMPLER2_BORDER_RED +0x0000A430 TD_PS_SAMPLER3_BORDER_RED +0x0000A440 TD_PS_SAMPLER4_BORDER_RED +0x0000A450 TD_PS_SAMPLER5_BORDER_RED +0x0000A460 TD_PS_SAMPLER6_BORDER_RED +0x0000A470 TD_PS_SAMPLER7_BORDER_RED +0x0000A480 TD_PS_SAMPLER8_BORDER_RED +0x0000A490 TD_PS_SAMPLER9_BORDER_RED +0x0000A4A0 TD_PS_SAMPLER10_BORDER_RED +0x0000A4B0 TD_PS_SAMPLER11_BORDER_RED +0x0000A4C0 TD_PS_SAMPLER12_BORDER_RED +0x0000A4D0 TD_PS_SAMPLER13_BORDER_RED +0x0000A4E0 TD_PS_SAMPLER14_BORDER_RED +0x0000A4F0 TD_PS_SAMPLER15_BORDER_RED +0x0000A500 TD_PS_SAMPLER16_BORDER_RED +0x0000A510 TD_PS_SAMPLER17_BORDER_RED +0x0000AA00 TD_PS_SAMPLER0_CLEARTYPE_KERNEL +0x0000AA04 TD_PS_SAMPLER1_CLEARTYPE_KERNEL +0x0000AA08 TD_PS_SAMPLER2_CLEARTYPE_KERNEL +0x0000AA0C TD_PS_SAMPLER3_CLEARTYPE_KERNEL +0x0000AA10 TD_PS_SAMPLER4_CLEARTYPE_KERNEL +0x0000AA14 TD_PS_SAMPLER5_CLEARTYPE_KERNEL +0x0000AA18 TD_PS_SAMPLER6_CLEARTYPE_KERNEL +0x0000AA1C TD_PS_SAMPLER7_CLEARTYPE_KERNEL +0x0000AA20 TD_PS_SAMPLER8_CLEARTYPE_KERNEL +0x0000AA24 TD_PS_SAMPLER9_CLEARTYPE_KERNEL +0x0000AA28 TD_PS_SAMPLER10_CLEARTYPE_KERNEL +0x0000AA2C TD_PS_SAMPLER11_CLEARTYPE_KERNEL +0x0000AA30 TD_PS_SAMPLER12_CLEARTYPE_KERNEL +0x0000AA34 TD_PS_SAMPLER13_CLEARTYPE_KERNEL +0x0000AA38 TD_PS_SAMPLER14_CLEARTYPE_KERNEL +0x0000AA3C TD_PS_SAMPLER15_CLEARTYPE_KERNEL +0x0000AA40 TD_PS_SAMPLER16_CLEARTYPE_KERNEL +0x0000AA44 TD_PS_SAMPLER17_CLEARTYPE_KERNEL +0x0000A60C TD_VS_SAMPLER0_BORDER_ALPHA +0x0000A61C TD_VS_SAMPLER1_BORDER_ALPHA +0x0000A62C TD_VS_SAMPLER2_BORDER_ALPHA +0x0000A63C TD_VS_SAMPLER3_BORDER_ALPHA +0x0000A64C TD_VS_SAMPLER4_BORDER_ALPHA +0x0000A65C TD_VS_SAMPLER5_BORDER_ALPHA +0x0000A66C TD_VS_SAMPLER6_BORDER_ALPHA +0x0000A67C TD_VS_SAMPLER7_BORDER_ALPHA +0x0000A68C TD_VS_SAMPLER8_BORDER_ALPHA +0x0000A69C TD_VS_SAMPLER9_BORDER_ALPHA +0x0000A6AC TD_VS_SAMPLER10_BORDER_ALPHA +0x0000A6BC TD_VS_SAMPLER11_BORDER_ALPHA +0x0000A6CC TD_VS_SAMPLER12_BORDER_ALPHA +0x0000A6DC TD_VS_SAMPLER13_BORDER_ALPHA +0x0000A6EC TD_VS_SAMPLER14_BORDER_ALPHA +0x0000A6FC TD_VS_SAMPLER15_BORDER_ALPHA +0x0000A70C TD_VS_SAMPLER16_BORDER_ALPHA +0x0000A71C TD_VS_SAMPLER17_BORDER_ALPHA +0x0000A608 TD_VS_SAMPLER0_BORDER_BLUE +0x0000A618 TD_VS_SAMPLER1_BORDER_BLUE +0x0000A628 TD_VS_SAMPLER2_BORDER_BLUE +0x0000A638 TD_VS_SAMPLER3_BORDER_BLUE +0x0000A648 TD_VS_SAMPLER4_BORDER_BLUE +0x0000A658 TD_VS_SAMPLER5_BORDER_BLUE +0x0000A668 TD_VS_SAMPLER6_BORDER_BLUE +0x0000A678 TD_VS_SAMPLER7_BORDER_BLUE +0x0000A688 TD_VS_SAMPLER8_BORDER_BLUE +0x0000A698 TD_VS_SAMPLER9_BORDER_BLUE +0x0000A6A8 TD_VS_SAMPLER10_BORDER_BLUE +0x0000A6B8 TD_VS_SAMPLER11_BORDER_BLUE +0x0000A6C8 TD_VS_SAMPLER12_BORDER_BLUE +0x0000A6D8 TD_VS_SAMPLER13_BORDER_BLUE +0x0000A6E8 TD_VS_SAMPLER14_BORDER_BLUE +0x0000A6F8 TD_VS_SAMPLER15_BORDER_BLUE +0x0000A708 TD_VS_SAMPLER16_BORDER_BLUE +0x0000A718 TD_VS_SAMPLER17_BORDER_BLUE +0x0000A604 TD_VS_SAMPLER0_BORDER_GREEN +0x0000A614 TD_VS_SAMPLER1_BORDER_GREEN +0x0000A624 TD_VS_SAMPLER2_BORDER_GREEN +0x0000A634 TD_VS_SAMPLER3_BORDER_GREEN +0x0000A644 TD_VS_SAMPLER4_BORDER_GREEN +0x0000A654 TD_VS_SAMPLER5_BORDER_GREEN +0x0000A664 TD_VS_SAMPLER6_BORDER_GREEN +0x0000A674 TD_VS_SAMPLER7_BORDER_GREEN +0x0000A684 TD_VS_SAMPLER8_BORDER_GREEN +0x0000A694 TD_VS_SAMPLER9_BORDER_GREEN +0x0000A6A4 TD_VS_SAMPLER10_BORDER_GREEN +0x0000A6B4 TD_VS_SAMPLER11_BORDER_GREEN +0x0000A6C4 TD_VS_SAMPLER12_BORDER_GREEN +0x0000A6D4 TD_VS_SAMPLER13_BORDER_GREEN +0x0000A6E4 TD_VS_SAMPLER14_BORDER_GREEN +0x0000A6F4 TD_VS_SAMPLER15_BORDER_GREEN +0x0000A704 TD_VS_SAMPLER16_BORDER_GREEN +0x0000A714 TD_VS_SAMPLER17_BORDER_GREEN +0x0000A600 TD_VS_SAMPLER0_BORDER_RED +0x0000A610 TD_VS_SAMPLER1_BORDER_RED +0x0000A620 TD_VS_SAMPLER2_BORDER_RED +0x0000A630 TD_VS_SAMPLER3_BORDER_RED +0x0000A640 TD_VS_SAMPLER4_BORDER_RED +0x0000A650 TD_VS_SAMPLER5_BORDER_RED +0x0000A660 TD_VS_SAMPLER6_BORDER_RED +0x0000A670 TD_VS_SAMPLER7_BORDER_RED +0x0000A680 TD_VS_SAMPLER8_BORDER_RED +0x0000A690 TD_VS_SAMPLER9_BORDER_RED +0x0000A6A0 TD_VS_SAMPLER10_BORDER_RED +0x0000A6B0 TD_VS_SAMPLER11_BORDER_RED +0x0000A6C0 TD_VS_SAMPLER12_BORDER_RED +0x0000A6D0 TD_VS_SAMPLER13_BORDER_RED +0x0000A6E0 TD_VS_SAMPLER14_BORDER_RED +0x0000A6F0 TD_VS_SAMPLER15_BORDER_RED +0x0000A700 TD_VS_SAMPLER16_BORDER_RED +0x0000A710 TD_VS_SAMPLER17_BORDER_RED +0x00009508 TA_CNTL_AUX +0x0002802C DB_DEPTH_CLEAR +0x00028D24 DB_HTILE_SURFACE +0x00028D34 DB_PREFETCH_LIMIT +0x00028D30 DB_PRELOAD_CONTROL +0x00028D0C DB_RENDER_CONTROL +0x00028D10 DB_RENDER_OVERRIDE +0x0002880C DB_SHADER_CONTROL +0x00028D2C DB_SRESULTS_COMPARE_STATE1 +0x00028430 DB_STENCILREFMASK +0x00028434 DB_STENCILREFMASK_BF +0x00028028 DB_STENCIL_CLEAR +0x00028780 CB_BLEND0_CONTROL +0x00028784 CB_BLEND1_CONTROL +0x00028788 CB_BLEND2_CONTROL +0x0002878C CB_BLEND3_CONTROL +0x00028790 CB_BLEND4_CONTROL +0x00028794 CB_BLEND5_CONTROL +0x00028798 CB_BLEND6_CONTROL +0x0002879C CB_BLEND7_CONTROL +0x00028804 CB_BLEND_CONTROL +0x00028420 CB_BLEND_ALPHA +0x0002841C CB_BLEND_BLUE +0x00028418 CB_BLEND_GREEN +0x00028414 CB_BLEND_RED +0x0002812C CB_CLEAR_ALPHA +0x00028128 CB_CLEAR_BLUE +0x00028124 CB_CLEAR_GREEN +0x00028120 CB_CLEAR_RED +0x00028C30 CB_CLRCMP_CONTROL +0x00028C38 CB_CLRCMP_DST +0x00028C3C CB_CLRCMP_MSK +0x00028C34 CB_CLRCMP_SRC +0x00028100 CB_COLOR0_MASK +0x00028104 CB_COLOR1_MASK +0x00028108 CB_COLOR2_MASK +0x0002810C CB_COLOR3_MASK +0x00028110 CB_COLOR4_MASK +0x00028114 CB_COLOR5_MASK +0x00028118 CB_COLOR6_MASK +0x0002811C CB_COLOR7_MASK +0x00028080 CB_COLOR0_VIEW +0x00028084 CB_COLOR1_VIEW +0x00028088 CB_COLOR2_VIEW +0x0002808C CB_COLOR3_VIEW +0x00028090 CB_COLOR4_VIEW +0x00028094 CB_COLOR5_VIEW +0x00028098 CB_COLOR6_VIEW +0x0002809C CB_COLOR7_VIEW +0x00028808 CB_COLOR_CONTROL +0x0002842C CB_FOG_BLUE +0x00028428 CB_FOG_GREEN +0x00028424 CB_FOG_RED +0x00008040 WAIT_UNTIL +0x00008950 CC_GC_SHADER_PIPE_CONFIG +0x00008954 GC_USER_SHADER_PIPE_CONFIG +0x00009714 VC_ENHANCE +0x00009830 DB_DEBUG +0x00009838 DB_WATERMARKS +0x00028D28 DB_SRESULTS_COMPARE_STATE0 +0x00028D44 DB_ALPHA_TO_MASK +0x00009504 TA_CNTL +0x00009700 VC_CNTL +0x00009718 VC_CONFIG +0x0000A02C SMX_DC_MC_INTF_CTL diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c index 9f5418983e2..626d51891ee 100644 --- a/drivers/gpu/drm/radeon/rs400.c +++ b/drivers/gpu/drm/radeon/rs400.c @@ -113,6 +113,7 @@ int rs400_gart_enable(struct radeon_device *rdev) uint32_t size_reg; uint32_t tmp; + radeon_gart_restore(rdev); tmp = RREG32_MC(RS690_AIC_CTRL_SCRATCH); tmp |= RS690_DIS_OUT_OF_PCI_GART_ACCESS; WREG32_MC(RS690_AIC_CTRL_SCRATCH, tmp); @@ -150,9 +151,8 @@ int rs400_gart_enable(struct radeon_device *rdev) WREG32(RADEON_AGP_BASE, 0xFFFFFFFF); WREG32(RS480_AGP_BASE_2, 0); } - tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; - tmp = REG_SET(RS690_MC_AGP_TOP, tmp >> 16); - tmp |= REG_SET(RS690_MC_AGP_START, rdev->mc.gtt_location >> 16); + tmp = REG_SET(RS690_MC_AGP_TOP, rdev->mc.gtt_end >> 16); + tmp |= REG_SET(RS690_MC_AGP_START, rdev->mc.gtt_start >> 16); if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740)) { WREG32_MC(RS690_MCCFG_AGP_LOCATION, tmp); tmp = RREG32(RADEON_BUS_CNTL) & ~RS600_BUS_MASTER_DIS; @@ -223,26 +223,47 @@ int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) return 0; } +int rs400_mc_wait_for_idle(struct radeon_device *rdev) +{ + unsigned i; + uint32_t tmp; + + for (i = 0; i < rdev->usec_timeout; i++) { + /* read MC_STATUS */ + tmp = RREG32(0x0150); + if (tmp & (1 << 2)) { + return 0; + } + DRM_UDELAY(1); + } + return -1; +} + void rs400_gpu_init(struct radeon_device *rdev) { /* FIXME: HDP same place on rs400 ? */ r100_hdp_reset(rdev); /* FIXME: is this correct ? */ r420_pipes_init(rdev); - if (r300_mc_wait_for_idle(rdev)) { - printk(KERN_WARNING "Failed to wait MC idle while " - "programming pipes. Bad things might happen.\n"); + if (rs400_mc_wait_for_idle(rdev)) { + printk(KERN_WARNING "rs400: Failed to wait MC idle while " + "programming pipes. Bad things might happen. %08x\n", RREG32(0x150)); } } -void rs400_vram_info(struct radeon_device *rdev) +void rs400_mc_init(struct radeon_device *rdev) { + u64 base; + rs400_gart_adjust_size(rdev); + rdev->mc.igp_sideport_enabled = radeon_combios_sideport_present(rdev); /* DDR for all card after R300 & IGP */ rdev->mc.vram_is_ddr = true; rdev->mc.vram_width = 128; - r100_vram_init_sizes(rdev); + base = (RREG32(RADEON_NB_TOM) & 0xffff) << 16; + radeon_vram_location(rdev, &rdev->mc, base); + radeon_gtt_location(rdev, &rdev->mc); } uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg) @@ -346,22 +367,6 @@ static int rs400_debugfs_pcie_gart_info_init(struct radeon_device *rdev) #endif } -static int rs400_mc_init(struct radeon_device *rdev) -{ - int r; - u32 tmp; - - /* Setup GPU memory space */ - tmp = RREG32(R_00015C_NB_TOM); - rdev->mc.vram_location = G_00015C_MC_FB_START(tmp) << 16; - rdev->mc.gtt_location = 0xFFFFFFFFUL; - r = radeon_mc_setup(rdev); - rdev->mc.igp_sideport_enabled = radeon_combios_sideport_present(rdev); - if (r) - return r; - return 0; -} - void rs400_mc_program(struct radeon_device *rdev) { struct r100_mc_save save; @@ -370,8 +375,8 @@ void rs400_mc_program(struct radeon_device *rdev) r100_mc_stop(rdev, &save); /* Wait for mc idle */ - if (r300_mc_wait_for_idle(rdev)) - dev_warn(rdev->dev, "Wait MC idle timeout before updating MC.\n"); + if (rs400_mc_wait_for_idle(rdev)) + dev_warn(rdev->dev, "rs400: Wait MC idle timeout before updating MC.\n"); WREG32(R_000148_MC_FB_LOCATION, S_000148_MC_FB_START(rdev->mc.vram_start >> 16) | S_000148_MC_FB_TOP(rdev->mc.vram_end >> 16)); @@ -448,7 +453,6 @@ int rs400_suspend(struct radeon_device *rdev) void rs400_fini(struct radeon_device *rdev) { - rs400_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -501,12 +505,8 @@ int rs400_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - rs400_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = rs400_mc_init(rdev); - if (r) - return r; + /* initialize memory controller */ + rs400_mc_init(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); if (r) @@ -527,7 +527,6 @@ int rs400_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - rs400_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c index d5255751e7b..47f046b78c6 100644 --- a/drivers/gpu/drm/radeon/rs600.c +++ b/drivers/gpu/drm/radeon/rs600.c @@ -45,23 +45,6 @@ void rs600_gpu_init(struct radeon_device *rdev); int rs600_mc_wait_for_idle(struct radeon_device *rdev); -int rs600_mc_init(struct radeon_device *rdev) -{ - /* read back the MC value from the hw */ - int r; - u32 tmp; - - /* Setup GPU memory space */ - tmp = RREG32_MC(R_000004_MC_FB_LOCATION); - rdev->mc.vram_location = G_000004_MC_FB_START(tmp) << 16; - rdev->mc.gtt_location = 0xffffffffUL; - r = radeon_mc_setup(rdev); - rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); - if (r) - return r; - return 0; -} - /* hpd for digital panel detect/disconnect */ bool rs600_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd) { @@ -213,6 +196,7 @@ int rs600_gart_enable(struct radeon_device *rdev) r = radeon_gart_table_vram_pin(rdev); if (r) return r; + radeon_gart_restore(rdev); /* Enable bus master */ tmp = RREG32(R_00004C_BUS_CNTL) & C_00004C_BUS_MASTER_DIS; WREG32(R_00004C_BUS_CNTL, tmp); @@ -406,10 +390,14 @@ int rs600_irq_process(struct radeon_device *rdev) if (G_000044_SW_INT(status)) radeon_fence_process(rdev); /* Vertical blank interrupts */ - if (G_007EDC_LB_D1_VBLANK_INTERRUPT(r500_disp_int)) + if (G_007EDC_LB_D1_VBLANK_INTERRUPT(r500_disp_int)) { drm_handle_vblank(rdev->ddev, 0); - if (G_007EDC_LB_D2_VBLANK_INTERRUPT(r500_disp_int)) + wake_up(&rdev->irq.vblank_queue); + } + if (G_007EDC_LB_D2_VBLANK_INTERRUPT(r500_disp_int)) { drm_handle_vblank(rdev->ddev, 1); + wake_up(&rdev->irq.vblank_queue); + } if (G_007EDC_DC_HOT_PLUG_DETECT1_INTERRUPT(r500_disp_int)) { queue_hotplug = true; DRM_DEBUG("HPD1\n"); @@ -470,22 +458,22 @@ void rs600_gpu_init(struct radeon_device *rdev) dev_warn(rdev->dev, "Wait MC idle timeout before updating MC.\n"); } -void rs600_vram_info(struct radeon_device *rdev) +void rs600_mc_init(struct radeon_device *rdev) { + u64 base; + + rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); + rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); rdev->mc.vram_is_ddr = true; rdev->mc.vram_width = 128; - rdev->mc.real_vram_size = RREG32(RADEON_CONFIG_MEMSIZE); rdev->mc.mc_vram_size = rdev->mc.real_vram_size; - - rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); - rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) - rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) - rdev->mc.real_vram_size = rdev->mc.aper_size; + rdev->mc.visible_vram_size = rdev->mc.aper_size; + rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); + base = RREG32_MC(R_000004_MC_FB_LOCATION); + base = G_000004_MC_FB_START(base) << 16; + radeon_vram_location(rdev, &rdev->mc, base); + radeon_gtt_location(rdev, &rdev->mc); } void rs600_bandwidth_update(struct radeon_device *rdev) @@ -610,7 +598,6 @@ int rs600_suspend(struct radeon_device *rdev) void rs600_fini(struct radeon_device *rdev) { - rs600_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -662,12 +649,8 @@ int rs600_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - rs600_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = rs600_mc_init(rdev); - if (r) - return r; + /* initialize memory controller */ + rs600_mc_init(rdev); rs600_debugfs(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); @@ -689,7 +672,6 @@ int rs600_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - rs600_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index cd31da91377..83b9174f76f 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -129,27 +129,21 @@ void rs690_pm_info(struct radeon_device *rdev) rdev->pm.sideport_bandwidth.full = rfixed_div(rdev->pm.sideport_bandwidth, tmp); } -void rs690_vram_info(struct radeon_device *rdev) +void rs690_mc_init(struct radeon_device *rdev) { fixed20_12 a; + u64 base; rs400_gart_adjust_size(rdev); - rdev->mc.vram_is_ddr = true; rdev->mc.vram_width = 128; - rdev->mc.real_vram_size = RREG32(RADEON_CONFIG_MEMSIZE); rdev->mc.mc_vram_size = rdev->mc.real_vram_size; - rdev->mc.aper_base = drm_get_resource_start(rdev->ddev, 0); rdev->mc.aper_size = drm_get_resource_len(rdev->ddev, 0); - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) - rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) - rdev->mc.real_vram_size = rdev->mc.aper_size; - + rdev->mc.visible_vram_size = rdev->mc.aper_size; + base = RREG32_MC(R_000100_MCCFG_FB_LOCATION); + base = G_000100_MC_FB_START(base) << 16; rs690_pm_info(rdev); /* FIXME: we should enforce default clock in case GPU is not in * default setup @@ -160,22 +154,9 @@ void rs690_vram_info(struct radeon_device *rdev) a.full = rfixed_const(16); /* core_bandwidth = sclk(Mhz) * 16 */ rdev->pm.core_bandwidth.full = rfixed_div(rdev->pm.sclk, a); -} - -static int rs690_mc_init(struct radeon_device *rdev) -{ - int r; - u32 tmp; - - /* Setup GPU memory space */ - tmp = RREG32_MC(R_000100_MCCFG_FB_LOCATION); - rdev->mc.vram_location = G_000100_MC_FB_START(tmp) << 16; - rdev->mc.gtt_location = 0xFFFFFFFFUL; - r = radeon_mc_setup(rdev); rdev->mc.igp_sideport_enabled = radeon_atombios_sideport_present(rdev); - if (r) - return r; - return 0; + radeon_vram_location(rdev, &rdev->mc, base); + radeon_gtt_location(rdev, &rdev->mc); } void rs690_line_buffer_adjust(struct radeon_device *rdev, @@ -676,7 +657,6 @@ int rs690_suspend(struct radeon_device *rdev) void rs690_fini(struct radeon_device *rdev) { - rs690_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -729,12 +709,8 @@ int rs690_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - rs690_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = rs690_mc_init(rdev); - if (r) - return r; + /* initialize memory controller */ + rs690_mc_init(rdev); rv515_debugfs(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); @@ -756,7 +732,6 @@ int rs690_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - rs690_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c index 62756717b04..bea747da123 100644 --- a/drivers/gpu/drm/radeon/rv515.c +++ b/drivers/gpu/drm/radeon/rv515.c @@ -277,13 +277,15 @@ static void rv515_vram_get_type(struct radeon_device *rdev) } } -void rv515_vram_info(struct radeon_device *rdev) +void rv515_mc_init(struct radeon_device *rdev) { fixed20_12 a; rv515_vram_get_type(rdev); - r100_vram_init_sizes(rdev); + radeon_vram_location(rdev, &rdev->mc, 0); + if (!(rdev->flags & RADEON_IS_AGP)) + radeon_gtt_location(rdev, &rdev->mc); /* FIXME: we should enforce default clock in case GPU is not in * default setup */ @@ -537,7 +539,6 @@ void rv515_set_safe_registers(struct radeon_device *rdev) void rv515_fini(struct radeon_device *rdev) { - rv515_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); @@ -588,12 +589,15 @@ int rv515_init(struct radeon_device *rdev) radeon_get_clock_info(rdev->ddev); /* Initialize power management */ radeon_pm_init(rdev); - /* Get vram informations */ - rv515_vram_info(rdev); - /* Initialize memory controller (also test AGP) */ - r = r420_mc_init(rdev); - if (r) - return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) { + radeon_agp_disable(rdev); + } + } + /* initialize memory controller */ + rv515_mc_init(rdev); rv515_debugfs(rdev); /* Fence driver */ r = radeon_fence_driver_init(rdev); @@ -615,13 +619,12 @@ int rv515_init(struct radeon_device *rdev) if (r) { /* Somethings want wront with the accel init stop accel */ dev_err(rdev->dev, "Disabling GPU acceleration\n"); - rv515_suspend(rdev); r100_cp_fini(rdev); r100_wb_fini(rdev); r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); rv370_pcie_gart_fini(rdev); radeon_agp_fini(rdev); - radeon_irq_kms_fini(rdev); rdev->accel_working = false; } return 0; diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index 59c71245fb9..37887dee12a 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -56,6 +56,7 @@ int rv770_pcie_gart_enable(struct radeon_device *rdev) r = radeon_gart_table_vram_pin(rdev); if (r) return r; + radeon_gart_restore(rdev); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | ENABLE_L2_FRAGMENT_PROCESSING | ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | @@ -273,9 +274,10 @@ static int rv770_cp_load_microcode(struct radeon_device *rdev) /* * Core functions */ -static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, - u32 num_backends, - u32 backend_disable_mask) +static u32 r700_get_tile_pipe_to_backend_map(struct radeon_device *rdev, + u32 num_tile_pipes, + u32 num_backends, + u32 backend_disable_mask) { u32 backend_map = 0; u32 enabled_backends_mask; @@ -284,6 +286,7 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, u32 swizzle_pipe[R7XX_MAX_PIPES]; u32 cur_backend; u32 i; + bool force_no_swizzle; if (num_tile_pipes > R7XX_MAX_PIPES) num_tile_pipes = R7XX_MAX_PIPES; @@ -313,6 +316,18 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, if (enabled_backends_count != num_backends) num_backends = enabled_backends_count; + switch (rdev->family) { + case CHIP_RV770: + case CHIP_RV730: + force_no_swizzle = false; + break; + case CHIP_RV710: + case CHIP_RV740: + default: + force_no_swizzle = true; + break; + } + memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * R7XX_MAX_PIPES); switch (num_tile_pipes) { case 1: @@ -323,49 +338,100 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, swizzle_pipe[1] = 1; break; case 3: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 1; + } break; case 4: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 3; - swizzle_pipe[3] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 3; + swizzle_pipe[3] = 1; + } break; case 5: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 1; - swizzle_pipe[4] = 3; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 1; + swizzle_pipe[4] = 3; + } break; case 6: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 5; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 5; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + } break; case 7: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 5; + } break; case 8: - swizzle_pipe[0] = 0; - swizzle_pipe[1] = 2; - swizzle_pipe[2] = 4; - swizzle_pipe[3] = 6; - swizzle_pipe[4] = 3; - swizzle_pipe[5] = 1; - swizzle_pipe[6] = 7; - swizzle_pipe[7] = 5; + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + swizzle_pipe[7] = 7; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 1; + swizzle_pipe[6] = 7; + swizzle_pipe[7] = 5; + } break; } @@ -385,8 +451,10 @@ static u32 r700_get_tile_pipe_to_backend_map(u32 num_tile_pipes, static void rv770_gpu_init(struct radeon_device *rdev) { int i, j, num_qd_pipes; + u32 ta_aux_cntl; u32 sx_debug_1; u32 smx_dc_ctl0; + u32 db_debug3; u32 num_gs_verts_per_thread; u32 vgt_gs_per_es; u32 gs_prim_buffer_depth = 0; @@ -515,6 +583,7 @@ static void rv770_gpu_init(struct radeon_device *rdev) switch (rdev->config.rv770.max_tile_pipes) { case 1: + default: gb_tiling_config |= PIPE_TILING(0); break; case 2: @@ -526,16 +595,17 @@ static void rv770_gpu_init(struct radeon_device *rdev) case 8: gb_tiling_config |= PIPE_TILING(3); break; - default: - break; } + rdev->config.rv770.tiling_npipes = rdev->config.rv770.max_tile_pipes; if (rdev->family == CHIP_RV770) gb_tiling_config |= BANK_TILING(1); else gb_tiling_config |= BANK_TILING((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT); + rdev->config.rv770.tiling_nbanks = 4 << ((gb_tiling_config >> 4) & 0x3); gb_tiling_config |= GROUP_SIZE(0); + rdev->config.rv770.tiling_group_size = 256; if (((mc_arb_ramcfg & NOOFROWS_MASK) >> NOOFROWS_SHIFT) > 3) { gb_tiling_config |= ROW_TILING(3); @@ -549,18 +619,27 @@ static void rv770_gpu_init(struct radeon_device *rdev) gb_tiling_config |= BANK_SWAPS(1); - backend_map = r700_get_tile_pipe_to_backend_map(rdev->config.rv770.max_tile_pipes, - rdev->config.rv770.max_backends, - (0xff << rdev->config.rv770.max_backends) & 0xff); - gb_tiling_config |= BACKEND_MAP(backend_map); + cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE) & 0x00ff0000; + cc_rb_backend_disable |= + BACKEND_DISABLE((R7XX_MAX_BACKENDS_MASK << rdev->config.rv770.max_backends) & R7XX_MAX_BACKENDS_MASK); - cc_gc_shader_pipe_config = + cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffffff00; + cc_gc_shader_pipe_config |= INACTIVE_QD_PIPES((R7XX_MAX_PIPES_MASK << rdev->config.rv770.max_pipes) & R7XX_MAX_PIPES_MASK); cc_gc_shader_pipe_config |= INACTIVE_SIMDS((R7XX_MAX_SIMDS_MASK << rdev->config.rv770.max_simds) & R7XX_MAX_SIMDS_MASK); - cc_rb_backend_disable = - BACKEND_DISABLE((R7XX_MAX_BACKENDS_MASK << rdev->config.rv770.max_backends) & R7XX_MAX_BACKENDS_MASK); + if (rdev->family == CHIP_RV740) + backend_map = 0x28; + else + backend_map = r700_get_tile_pipe_to_backend_map(rdev, + rdev->config.rv770.max_tile_pipes, + (R7XX_MAX_BACKENDS - + r600_count_pipe_bits((cc_rb_backend_disable & + R7XX_MAX_BACKENDS_MASK) >> 16)), + (cc_rb_backend_disable >> 16)); + gb_tiling_config |= BACKEND_MAP(backend_map); + WREG32(GB_TILING_CONFIG, gb_tiling_config); WREG32(DCP_TILING_CONFIG, (gb_tiling_config & 0xffff)); @@ -568,16 +647,13 @@ static void rv770_gpu_init(struct radeon_device *rdev) WREG32(CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); WREG32(CC_GC_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); - WREG32(GC_USER_SHADER_PIPE_CONFIG, cc_gc_shader_pipe_config); + WREG32(CC_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable); - WREG32(CC_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable); WREG32(CGTS_SYS_TCC_DISABLE, 0); WREG32(CGTS_TCC_DISABLE, 0); - WREG32(CGTS_USER_SYS_TCC_DISABLE, 0); - WREG32(CGTS_USER_TCC_DISABLE, 0); num_qd_pipes = - R7XX_MAX_BACKENDS - r600_count_pipe_bits(cc_gc_shader_pipe_config & INACTIVE_QD_PIPES_MASK); + R7XX_MAX_PIPES - r600_count_pipe_bits((cc_gc_shader_pipe_config & INACTIVE_QD_PIPES_MASK) >> 8); WREG32(VGT_OUT_DEALLOC_CNTL, (num_qd_pipes * 4) & DEALLOC_DIST_MASK); WREG32(VGT_VERTEX_REUSE_BLOCK_CNTL, ((num_qd_pipes * 4) - 2) & VTX_REUSE_DEPTH_MASK); @@ -587,10 +663,8 @@ static void rv770_gpu_init(struct radeon_device *rdev) WREG32(CP_MEQ_THRESHOLDS, STQ_SPLIT(0x30)); - WREG32(TA_CNTL_AUX, (DISABLE_CUBE_ANISO | - SYNC_GRADIENT | - SYNC_WALKER | - SYNC_ALIGNER)); + ta_aux_cntl = RREG32(TA_CNTL_AUX); + WREG32(TA_CNTL_AUX, ta_aux_cntl | DISABLE_CUBE_ANISO); sx_debug_1 = RREG32(SX_DEBUG_1); sx_debug_1 |= ENABLE_NEW_SMX_ADDRESS; @@ -601,14 +675,28 @@ static void rv770_gpu_init(struct radeon_device *rdev) smx_dc_ctl0 |= CACHE_DEPTH((rdev->config.rv770.sx_num_of_sets * 64) - 1); WREG32(SMX_DC_CTL0, smx_dc_ctl0); - WREG32(SMX_EVENT_CTL, (ES_FLUSH_CTL(4) | - GS_FLUSH_CTL(4) | - ACK_FLUSH_CTL(3) | - SYNC_FLUSH_CTL)); + if (rdev->family != CHIP_RV740) + WREG32(SMX_EVENT_CTL, (ES_FLUSH_CTL(4) | + GS_FLUSH_CTL(4) | + ACK_FLUSH_CTL(3) | + SYNC_FLUSH_CTL)); - if (rdev->family == CHIP_RV770) - WREG32(DB_DEBUG3, DB_CLK_OFF_DELAY(0x1f)); - else { + db_debug3 = RREG32(DB_DEBUG3); + db_debug3 &= ~DB_CLK_OFF_DELAY(0x1f); + switch (rdev->family) { + case CHIP_RV770: + case CHIP_RV740: + db_debug3 |= DB_CLK_OFF_DELAY(0x1f); + break; + case CHIP_RV710: + case CHIP_RV730: + default: + db_debug3 |= DB_CLK_OFF_DELAY(2); + break; + } + WREG32(DB_DEBUG3, db_debug3); + + if (rdev->family != CHIP_RV770) { db_debug4 = RREG32(DB_DEBUG4); db_debug4 |= DISABLE_TILE_COVERED_FOR_PS_ITER; WREG32(DB_DEBUG4, db_debug4); @@ -637,10 +725,10 @@ static void rv770_gpu_init(struct radeon_device *rdev) ALU_UPDATE_FIFO_HIWATER(0x8)); switch (rdev->family) { case CHIP_RV770: - sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x1); - break; case CHIP_RV730: case CHIP_RV710: + sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x1); + break; case CHIP_RV740: default: sq_ms_fifo_sizes |= FETCH_FIFO_HIWATER(0x4); @@ -779,7 +867,6 @@ int rv770_mc_init(struct radeon_device *rdev) fixed20_12 a; u32 tmp; int chansize, numchan; - int r; /* Get VRAM informations */ rdev->mc.vram_is_ddr = true; @@ -814,48 +901,13 @@ int rv770_mc_init(struct radeon_device *rdev) /* Setup GPU memory space */ rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE); rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE); - - if (rdev->mc.mc_vram_size > rdev->mc.aper_size) + rdev->mc.visible_vram_size = rdev->mc.aper_size; + /* FIXME remove this once we support unmappable VRAM */ + if (rdev->mc.mc_vram_size > rdev->mc.aper_size) { rdev->mc.mc_vram_size = rdev->mc.aper_size; - - if (rdev->mc.real_vram_size > rdev->mc.aper_size) rdev->mc.real_vram_size = rdev->mc.aper_size; - - if (rdev->flags & RADEON_IS_AGP) { - r = radeon_agp_init(rdev); - if (r) - return r; - /* gtt_size is setup by radeon_agp_init */ - rdev->mc.gtt_location = rdev->mc.agp_base; - tmp = 0xFFFFFFFFUL - rdev->mc.agp_base - rdev->mc.gtt_size; - /* Try to put vram before or after AGP because we - * we want SYSTEM_APERTURE to cover both VRAM and - * AGP so that GPU can catch out of VRAM/AGP access - */ - if (rdev->mc.gtt_location > rdev->mc.mc_vram_size) { - /* Enough place before */ - rdev->mc.vram_location = rdev->mc.gtt_location - - rdev->mc.mc_vram_size; - } else if (tmp > rdev->mc.mc_vram_size) { - /* Enough place after */ - rdev->mc.vram_location = rdev->mc.gtt_location + - rdev->mc.gtt_size; - } else { - /* Try to setup VRAM then AGP might not - * not work on some card - */ - rdev->mc.vram_location = 0x00000000UL; - rdev->mc.gtt_location = rdev->mc.mc_vram_size; - } - } else { - rdev->mc.vram_location = 0x00000000UL; - rdev->mc.gtt_location = rdev->mc.mc_vram_size; - rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; } - rdev->mc.vram_start = rdev->mc.vram_location; - rdev->mc.vram_end = rdev->mc.vram_location + rdev->mc.mc_vram_size - 1; - rdev->mc.gtt_start = rdev->mc.gtt_location; - rdev->mc.gtt_end = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; + r600_vram_gtt_location(rdev, &rdev->mc); /* FIXME: we should enforce default clock in case GPU is not in * default setup */ @@ -864,6 +916,7 @@ int rv770_mc_init(struct radeon_device *rdev) rdev->pm.sclk.full = rfixed_div(rdev->pm.sclk, a); return 0; } + int rv770_gpu_reset(struct radeon_device *rdev) { /* FIXME: implement any rv770 specific bits */ @@ -891,26 +944,25 @@ static int rv770_startup(struct radeon_device *rdev) return r; } rv770_gpu_init(rdev); - - if (!rdev->r600_blit.shader_obj) { - r = r600_blit_init(rdev); + r = r600_blit_init(rdev); + if (r) { + r600_blit_fini(rdev); + rdev->asic->copy = NULL; + dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); + } + /* pin copy shader into vram */ + if (rdev->r600_blit.shader_obj) { + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (unlikely(r != 0)) + return r; + r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->r600_blit.shader_gpu_addr); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); if (r) { - DRM_ERROR("radeon: failed blitter (%d).\n", r); + DRM_ERROR("failed to pin blit object %d\n", r); return r; } } - - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->r600_blit.shader_obj, RADEON_GEM_DOMAIN_VRAM, - &rdev->r600_blit.shader_gpu_addr); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); - if (r) { - DRM_ERROR("failed to pin blit object %d\n", r); - return r; - } - /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -972,13 +1024,16 @@ int rv770_suspend(struct radeon_device *rdev) /* FIXME: we should wait for ring to be empty */ r700_cp_stop(rdev); rdev->cp.ready = false; + r600_irq_suspend(rdev); r600_wb_disable(rdev); rv770_pcie_gart_disable(rdev); /* unpin shaders bo */ - r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); - if (likely(r == 0)) { - radeon_bo_unpin(rdev->r600_blit.shader_obj); - radeon_bo_unreserve(rdev->r600_blit.shader_obj); + if (rdev->r600_blit.shader_obj) { + r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false); + if (likely(r == 0)) { + radeon_bo_unpin(rdev->r600_blit.shader_obj); + radeon_bo_unreserve(rdev->r600_blit.shader_obj); + } } return 0; } @@ -1037,6 +1092,12 @@ int rv770_init(struct radeon_device *rdev) r = radeon_fence_driver_init(rdev); if (r) return r; + /* initialize AGP */ + if (rdev->flags & RADEON_IS_AGP) { + r = radeon_agp_init(rdev); + if (r) + radeon_agp_disable(rdev); + } r = rv770_mc_init(rdev); if (r) return r; @@ -1062,22 +1123,25 @@ int rv770_init(struct radeon_device *rdev) rdev->accel_working = true; r = rv770_startup(rdev); if (r) { - rv770_suspend(rdev); + dev_err(rdev->dev, "disabling GPU acceleration\n"); + r600_cp_fini(rdev); r600_wb_fini(rdev); - radeon_ring_fini(rdev); + r600_irq_fini(rdev); + radeon_irq_kms_fini(rdev); rv770_pcie_gart_fini(rdev); rdev->accel_working = false; } if (rdev->accel_working) { r = radeon_ib_pool_init(rdev); if (r) { - DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r); - rdev->accel_working = false; - } - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); rdev->accel_working = false; + } else { + r = r600_ib_test(rdev); + if (r) { + dev_err(rdev->dev, "IB test failed (%d).\n", r); + rdev->accel_working = false; + } } } return 0; @@ -1085,13 +1149,11 @@ int rv770_init(struct radeon_device *rdev) void rv770_fini(struct radeon_device *rdev) { - rv770_suspend(rdev); - r600_blit_fini(rdev); + r600_cp_fini(rdev); + r600_wb_fini(rdev); r600_irq_fini(rdev); radeon_irq_kms_fini(rdev); - radeon_ring_fini(rdev); - r600_wb_fini(rdev); rv770_pcie_gart_fini(rdev); radeon_gem_fini(rdev); radeon_fence_driver_fini(rdev); diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h index a1367ab6f26..9506f8cb99e 100644 --- a/drivers/gpu/drm/radeon/rv770d.h +++ b/drivers/gpu/drm/radeon/rv770d.h @@ -343,4 +343,6 @@ #define WAIT_UNTIL 0x8040 +#define SRBM_STATUS 0x0E50 + #endif diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 2920f9a279e..89c38c49066 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -128,7 +128,7 @@ static struct attribute *ttm_bo_global_attrs[] = { NULL }; -static struct sysfs_ops ttm_bo_global_ops = { +static const struct sysfs_ops ttm_bo_global_ops = { .show = &ttm_bo_global_show }; @@ -426,7 +426,8 @@ moved: bdev->man[bo->mem.mem_type].gpu_offset; bo->cur_placement = bo->mem.placement; spin_unlock(&bo->lock); - } + } else + bo->offset = 0; return 0; @@ -523,52 +524,44 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo, bool remove_all) static int ttm_bo_delayed_delete(struct ttm_bo_device *bdev, bool remove_all) { struct ttm_bo_global *glob = bdev->glob; - struct ttm_buffer_object *entry, *nentry; - struct list_head *list, *next; - int ret; + struct ttm_buffer_object *entry = NULL; + int ret = 0; spin_lock(&glob->lru_lock); - list_for_each_safe(list, next, &bdev->ddestroy) { - entry = list_entry(list, struct ttm_buffer_object, ddestroy); - nentry = NULL; + if (list_empty(&bdev->ddestroy)) + goto out_unlock; - /* - * Protect the next list entry from destruction while we - * unlock the lru_lock. - */ + entry = list_first_entry(&bdev->ddestroy, + struct ttm_buffer_object, ddestroy); + kref_get(&entry->list_kref); - if (next != &bdev->ddestroy) { - nentry = list_entry(next, struct ttm_buffer_object, - ddestroy); + for (;;) { + struct ttm_buffer_object *nentry = NULL; + + if (entry->ddestroy.next != &bdev->ddestroy) { + nentry = list_first_entry(&entry->ddestroy, + struct ttm_buffer_object, ddestroy); kref_get(&nentry->list_kref); } - kref_get(&entry->list_kref); spin_unlock(&glob->lru_lock); ret = ttm_bo_cleanup_refs(entry, remove_all); kref_put(&entry->list_kref, ttm_bo_release_list); + entry = nentry; + + if (ret || !entry) + goto out; spin_lock(&glob->lru_lock); - if (nentry) { - bool next_onlist = !list_empty(next); - spin_unlock(&glob->lru_lock); - kref_put(&nentry->list_kref, ttm_bo_release_list); - spin_lock(&glob->lru_lock); - /* - * Someone might have raced us and removed the - * next entry from the list. We don't bother restarting - * list traversal. - */ - - if (!next_onlist) - break; - } - if (ret) + if (list_empty(&entry->ddestroy)) break; } - ret = !list_empty(&bdev->ddestroy); - spin_unlock(&glob->lru_lock); +out_unlock: + spin_unlock(&glob->lru_lock); +out: + if (entry) + kref_put(&entry->list_kref, ttm_bo_release_list); return ret; } @@ -950,6 +943,14 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, ttm_flag_masked(&cur_flags, placement->busy_placement[i], ~TTM_PL_MASK_MEMTYPE); + + if (mem_type == TTM_PL_SYSTEM) { + mem->mem_type = mem_type; + mem->placement = cur_flags; + mem->mm_node = NULL; + return 0; + } + ret = ttm_bo_mem_force_space(bo, mem_type, placement, mem, interruptible, no_wait); if (ret == 0 && mem->mm_node) { @@ -1019,6 +1020,12 @@ static int ttm_bo_mem_compat(struct ttm_placement *placement, struct ttm_mem_reg *mem) { int i; + struct drm_mm_node *node = mem->mm_node; + + if (node && placement->lpfn != 0 && + (node->start < placement->fpfn || + node->start + node->size > placement->lpfn)) + return -1; for (i = 0; i < placement->num_placement; i++) { if ((placement->placement[i] & mem->placement & @@ -1844,6 +1851,9 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink) * anyone tries to access a ttm page. */ + if (bo->bdev->driver->swap_notify) + bo->bdev->driver->swap_notify(bo); + ret = ttm_tt_swapout(bo->ttm, bo->persistant_swap_storage); out: @@ -1864,3 +1874,4 @@ void ttm_bo_swapout_all(struct ttm_bo_device *bdev) while (ttm_bo_swapout(&bdev->glob->shrink) == 0) ; } +EXPORT_SYMBOL(ttm_bo_swapout_all); diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c index 2ecf7d0c64f..5ca37a58a98 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_util.c +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c @@ -53,7 +53,6 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo, { struct ttm_tt *ttm = bo->ttm; struct ttm_mem_reg *old_mem = &bo->mem; - uint32_t save_flags = old_mem->placement; int ret; if (old_mem->mem_type != TTM_PL_SYSTEM) { @@ -62,7 +61,6 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo, ttm_flag_masked(&old_mem->placement, TTM_PL_FLAG_SYSTEM, TTM_PL_MASK_MEM); old_mem->mem_type = TTM_PL_SYSTEM; - save_flags = old_mem->placement; } ret = ttm_tt_set_placement_caching(ttm, new_mem->placement); @@ -77,7 +75,7 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo, *old_mem = *new_mem; new_mem->mm_node = NULL; - ttm_flag_masked(&save_flags, new_mem->placement, TTM_PL_MASK_MEMTYPE); + return 0; } EXPORT_SYMBOL(ttm_bo_move_ttm); @@ -219,7 +217,6 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo, void *old_iomap; void *new_iomap; int ret; - uint32_t save_flags = old_mem->placement; unsigned long i; unsigned long page; unsigned long add = 0; @@ -270,7 +267,6 @@ out2: *old_mem = *new_mem; new_mem->mm_node = NULL; - ttm_flag_masked(&save_flags, new_mem->placement, TTM_PL_MASK_MEMTYPE); if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) && (ttm != NULL)) { ttm_tt_unbind(ttm); @@ -537,7 +533,6 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, struct ttm_mem_type_manager *man = &bdev->man[new_mem->mem_type]; struct ttm_mem_reg *old_mem = &bo->mem; int ret; - uint32_t save_flags = old_mem->placement; struct ttm_buffer_object *ghost_obj; void *tmp_obj = NULL; @@ -598,7 +593,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, *old_mem = *new_mem; new_mem->mm_node = NULL; - ttm_flag_masked(&save_flags, new_mem->placement, TTM_PL_MASK_MEMTYPE); + return 0; } EXPORT_SYMBOL(ttm_bo_move_accel_cleanup); diff --git a/drivers/gpu/drm/ttm/ttm_lock.c b/drivers/gpu/drm/ttm/ttm_lock.c index f619ebcaa4e..3d172ef04ee 100644 --- a/drivers/gpu/drm/ttm/ttm_lock.c +++ b/drivers/gpu/drm/ttm/ttm_lock.c @@ -288,6 +288,7 @@ void ttm_suspend_unlock(struct ttm_lock *lock) wake_up_all(&lock->queue); spin_unlock(&lock->lock); } +EXPORT_SYMBOL(ttm_suspend_unlock); static bool __ttm_suspend_lock(struct ttm_lock *lock) { @@ -309,3 +310,4 @@ void ttm_suspend_lock(struct ttm_lock *lock) { wait_event(lock->queue, __ttm_suspend_lock(lock)); } +EXPORT_SYMBOL(ttm_suspend_lock); diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c index f5245c02b8f..eb143e04d40 100644 --- a/drivers/gpu/drm/ttm/ttm_memory.c +++ b/drivers/gpu/drm/ttm/ttm_memory.c @@ -152,7 +152,7 @@ static struct attribute *ttm_mem_zone_attrs[] = { NULL }; -static struct sysfs_ops ttm_mem_zone_ops = { +static const struct sysfs_ops ttm_mem_zone_ops = { .show = &ttm_mem_zone_show, .store = &ttm_mem_zone_store }; diff --git a/drivers/gpu/drm/ttm/ttm_object.c b/drivers/gpu/drm/ttm/ttm_object.c index 1099abac824..75e9d6f86ba 100644 --- a/drivers/gpu/drm/ttm/ttm_object.c +++ b/drivers/gpu/drm/ttm/ttm_object.c @@ -109,8 +109,8 @@ struct ttm_ref_object { struct drm_hash_item hash; struct list_head head; struct kref kref; - struct ttm_base_object *obj; enum ttm_ref_type ref_type; + struct ttm_base_object *obj; struct ttm_object_file *tfile; }; diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index 9c2b1cc5dba..a759170763b 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -196,23 +196,34 @@ EXPORT_SYMBOL(ttm_tt_populate); #ifdef CONFIG_X86 static inline int ttm_tt_set_page_caching(struct page *p, - enum ttm_caching_state c_state) + enum ttm_caching_state c_old, + enum ttm_caching_state c_new) { + int ret = 0; + if (PageHighMem(p)) return 0; - switch (c_state) { - case tt_cached: - return set_pages_wb(p, 1); - case tt_wc: - return set_memory_wc((unsigned long) page_address(p), 1); - default: - return set_pages_uc(p, 1); + if (c_old != tt_cached) { + /* p isn't in the default caching state, set it to + * writeback first to free its current memtype. */ + + ret = set_pages_wb(p, 1); + if (ret) + return ret; } + + if (c_new == tt_wc) + ret = set_memory_wc((unsigned long) page_address(p), 1); + else if (c_new == tt_uncached) + ret = set_pages_uc(p, 1); + + return ret; } #else /* CONFIG_X86 */ static inline int ttm_tt_set_page_caching(struct page *p, - enum ttm_caching_state c_state) + enum ttm_caching_state c_old, + enum ttm_caching_state c_new) { return 0; } @@ -245,7 +256,9 @@ static int ttm_tt_set_caching(struct ttm_tt *ttm, for (i = 0; i < ttm->num_pages; ++i) { cur_page = ttm->pages[i]; if (likely(cur_page != NULL)) { - ret = ttm_tt_set_page_caching(cur_page, c_state); + ret = ttm_tt_set_page_caching(cur_page, + ttm->caching_state, + c_state); if (unlikely(ret != 0)) goto out_err; } @@ -259,7 +272,7 @@ out_err: for (j = 0; j < i; ++j) { cur_page = ttm->pages[j]; if (likely(cur_page != NULL)) { - (void)ttm_tt_set_page_caching(cur_page, + (void)ttm_tt_set_page_caching(cur_page, c_state, ttm->caching_state); } } @@ -467,7 +480,7 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) void *from_virtual; void *to_virtual; int i; - int ret; + int ret = -ENOMEM; if (ttm->page_flags & TTM_PAGE_FLAG_USER) { ret = ttm_tt_set_user(ttm, ttm->tsk, ttm->start, @@ -486,8 +499,10 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) for (i = 0; i < ttm->num_pages; ++i) { from_page = read_mapping_page(swap_space, i, NULL); - if (IS_ERR(from_page)) + if (IS_ERR(from_page)) { + ret = PTR_ERR(from_page); goto out_err; + } to_page = __ttm_tt_get_page(ttm, i); if (unlikely(to_page == NULL)) goto out_err; @@ -510,7 +525,7 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) return 0; out_err: ttm_tt_free_alloced_pages(ttm); - return -ENOMEM; + return ret; } int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistant_swap_storage) @@ -522,6 +537,7 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistant_swap_storage) void *from_virtual; void *to_virtual; int i; + int ret = -ENOMEM; BUG_ON(ttm->state != tt_unbound && ttm->state != tt_unpopulated); BUG_ON(ttm->caching_state != tt_cached); @@ -544,7 +560,7 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistant_swap_storage) 0); if (unlikely(IS_ERR(swap_storage))) { printk(KERN_ERR "Failed allocating swap storage.\n"); - return -ENOMEM; + return PTR_ERR(swap_storage); } } else swap_storage = persistant_swap_storage; @@ -556,9 +572,10 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistant_swap_storage) if (unlikely(from_page == NULL)) continue; to_page = read_mapping_page(swap_space, i, NULL); - if (unlikely(to_page == NULL)) + if (unlikely(IS_ERR(to_page))) { + ret = PTR_ERR(to_page); goto out_err; - + } preempt_disable(); from_virtual = kmap_atomic(from_page, KM_USER0); to_virtual = kmap_atomic(to_page, KM_USER1); @@ -582,5 +599,5 @@ out_err: if (!persistant_swap_storage) fput(swap_storage); - return -ENOMEM; + return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c index d6f2d2b882e..825ebe3d89d 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c @@ -48,6 +48,15 @@ struct ttm_placement vmw_vram_placement = { .busy_placement = &vram_placement_flags }; +struct ttm_placement vmw_vram_sys_placement = { + .fpfn = 0, + .lpfn = 0, + .num_placement = 1, + .placement = &vram_placement_flags, + .num_busy_placement = 1, + .busy_placement = &sys_placement_flags +}; + struct ttm_placement vmw_vram_ne_placement = { .fpfn = 0, .lpfn = 0, @@ -172,6 +181,18 @@ static int vmw_verify_access(struct ttm_buffer_object *bo, struct file *filp) return 0; } +static void vmw_move_notify(struct ttm_buffer_object *bo, + struct ttm_mem_reg *new_mem) +{ + if (new_mem->mem_type != TTM_PL_SYSTEM) + vmw_dmabuf_gmr_unbind(bo); +} + +static void vmw_swap_notify(struct ttm_buffer_object *bo) +{ + vmw_dmabuf_gmr_unbind(bo); +} + /** * FIXME: We're using the old vmware polling method to sync. * Do this with fences instead. @@ -225,5 +246,7 @@ struct ttm_bo_driver vmw_bo_driver = { .sync_obj_wait = vmw_sync_obj_wait, .sync_obj_flush = vmw_sync_obj_flush, .sync_obj_unref = vmw_sync_obj_unref, - .sync_obj_ref = vmw_sync_obj_ref + .sync_obj_ref = vmw_sync_obj_ref, + .move_notify = vmw_move_notify, + .swap_notify = vmw_swap_notify }; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 1db1ef30be2..0c9c0811f42 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -147,6 +147,8 @@ static char *vmw_devname = "vmwgfx"; static int vmw_probe(struct pci_dev *, const struct pci_device_id *); static void vmw_master_init(struct vmw_master *); +static int vmwgfx_pm_notifier(struct notifier_block *nb, unsigned long val, + void *ptr); static void vmw_print_capabilities(uint32_t capabilities) { @@ -207,6 +209,7 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) { struct vmw_private *dev_priv; int ret; + uint32_t svga_id; dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL); if (unlikely(dev_priv == NULL)) { @@ -217,6 +220,7 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) dev_priv->dev = dev; dev_priv->vmw_chipset = chipset; + dev_priv->last_read_sequence = (uint32_t) -100; mutex_init(&dev_priv->hw_mutex); mutex_init(&dev_priv->cmdbuf_mutex); rwlock_init(&dev_priv->resource_lock); @@ -236,6 +240,16 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) dev_priv->mmio_start = pci_resource_start(dev->pdev, 2); mutex_lock(&dev_priv->hw_mutex); + + vmw_write(dev_priv, SVGA_REG_ID, SVGA_ID_2); + svga_id = vmw_read(dev_priv, SVGA_REG_ID); + if (svga_id != SVGA_ID_2) { + ret = -ENOSYS; + DRM_ERROR("Unsuported SVGA ID 0x%x\n", svga_id); + mutex_unlock(&dev_priv->hw_mutex); + goto out_err0; + } + dev_priv->capabilities = vmw_read(dev_priv, SVGA_REG_CAPABILITIES); if (dev_priv->capabilities & SVGA_CAP_GMR) { @@ -334,22 +348,24 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) */ DRM_INFO("It appears like vesafb is loaded. " - "Ignore above error if any. Entering stealth mode.\n"); + "Ignore above error if any.\n"); ret = pci_request_region(dev->pdev, 2, "vmwgfx stealth probe"); if (unlikely(ret != 0)) { DRM_ERROR("Failed reserving the SVGA MMIO resource.\n"); goto out_no_device; } - vmw_kms_init(dev_priv); - vmw_overlay_init(dev_priv); - } else { - ret = vmw_request_device(dev_priv); - if (unlikely(ret != 0)) - goto out_no_device; - vmw_kms_init(dev_priv); - vmw_overlay_init(dev_priv); - vmw_fb_init(dev_priv); } + ret = vmw_request_device(dev_priv); + if (unlikely(ret != 0)) + goto out_no_device; + vmw_kms_init(dev_priv); + vmw_overlay_init(dev_priv); + vmw_fb_init(dev_priv); + + dev_priv->pm_nb.notifier_call = vmwgfx_pm_notifier; + register_pm_notifier(&dev_priv->pm_nb); + + DRM_INFO("%s", vmw_fifo_have_3d(dev_priv) ? "Have 3D\n" : "No 3D\n"); return 0; @@ -385,17 +401,17 @@ static int vmw_driver_unload(struct drm_device *dev) DRM_INFO(VMWGFX_DRIVER_NAME " unload.\n"); - if (!dev_priv->stealth) { - vmw_fb_close(dev_priv); - vmw_kms_close(dev_priv); - vmw_overlay_close(dev_priv); - vmw_release_device(dev_priv); - pci_release_regions(dev->pdev); - } else { - vmw_kms_close(dev_priv); - vmw_overlay_close(dev_priv); + unregister_pm_notifier(&dev_priv->pm_nb); + + vmw_fb_close(dev_priv); + vmw_kms_close(dev_priv); + vmw_overlay_close(dev_priv); + vmw_release_device(dev_priv); + if (dev_priv->stealth) pci_release_region(dev->pdev, 2); - } + else + pci_release_regions(dev->pdev); + if (dev_priv->capabilities & SVGA_CAP_IRQMASK) drm_irq_uninstall(dev_priv->dev); if (dev->devname == vmw_devname) @@ -564,11 +580,6 @@ static int vmw_master_set(struct drm_device *dev, int ret = 0; DRM_INFO("Master set.\n"); - if (dev_priv->stealth) { - ret = vmw_request_device(dev_priv); - if (unlikely(ret != 0)) - return ret; - } if (active) { BUG_ON(active != &dev_priv->fbdev_master); @@ -628,18 +639,11 @@ static void vmw_master_drop(struct drm_device *dev, ttm_lock_set_kill(&vmaster->lock, true, SIGTERM); - if (dev_priv->stealth) { - ret = ttm_bo_evict_mm(&dev_priv->bdev, TTM_PL_VRAM); - if (unlikely(ret != 0)) - DRM_ERROR("Unable to clean VRAM on master drop.\n"); - vmw_release_device(dev_priv); - } dev_priv->active_master = &dev_priv->fbdev_master; ttm_lock_set_kill(&dev_priv->fbdev_master.lock, false, SIGTERM); ttm_vt_unlock(&dev_priv->fbdev_master.lock); - if (!dev_priv->stealth) - vmw_fb_on(dev_priv); + vmw_fb_on(dev_priv); } @@ -650,6 +654,57 @@ static void vmw_remove(struct pci_dev *pdev) drm_put_dev(dev); } +static int vmwgfx_pm_notifier(struct notifier_block *nb, unsigned long val, + void *ptr) +{ + struct vmw_private *dev_priv = + container_of(nb, struct vmw_private, pm_nb); + struct vmw_master *vmaster = dev_priv->active_master; + + switch (val) { + case PM_HIBERNATION_PREPARE: + case PM_SUSPEND_PREPARE: + ttm_suspend_lock(&vmaster->lock); + + /** + * This empties VRAM and unbinds all GMR bindings. + * Buffer contents is moved to swappable memory. + */ + ttm_bo_swapout_all(&dev_priv->bdev); + break; + case PM_POST_HIBERNATION: + case PM_POST_SUSPEND: + ttm_suspend_unlock(&vmaster->lock); + break; + case PM_RESTORE_PREPARE: + break; + case PM_POST_RESTORE: + break; + default: + break; + } + return 0; +} + +/** + * These might not be needed with the virtual SVGA device. + */ + +int vmw_pci_suspend(struct pci_dev *pdev, pm_message_t state) +{ + pci_save_state(pdev); + pci_disable_device(pdev); + pci_set_power_state(pdev, PCI_D3hot); + return 0; +} + +int vmw_pci_resume(struct pci_dev *pdev) +{ + pci_set_power_state(pdev, PCI_D0); + pci_restore_state(pdev); + return pci_enable_device(pdev); +} + static struct drm_driver driver = { .driver_features = DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_MODESET, @@ -689,7 +744,9 @@ static struct drm_driver driver = { .name = VMWGFX_DRIVER_NAME, .id_table = vmw_pci_id_list, .probe = vmw_probe, - .remove = vmw_remove + .remove = vmw_remove, + .suspend = vmw_pci_suspend, + .resume = vmw_pci_resume }, .name = VMWGFX_DRIVER_NAME, .desc = VMWGFX_DRIVER_DESC, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index e61bd85b697..356dc935ec1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -32,16 +32,17 @@ #include "drmP.h" #include "vmwgfx_drm.h" #include "drm_hashtab.h" +#include "linux/suspend.h" #include "ttm/ttm_bo_driver.h" #include "ttm/ttm_object.h" #include "ttm/ttm_lock.h" #include "ttm/ttm_execbuf_util.h" #include "ttm/ttm_module.h" -#define VMWGFX_DRIVER_DATE "20090724" -#define VMWGFX_DRIVER_MAJOR 0 -#define VMWGFX_DRIVER_MINOR 1 -#define VMWGFX_DRIVER_PATCHLEVEL 2 +#define VMWGFX_DRIVER_DATE "20100209" +#define VMWGFX_DRIVER_MAJOR 1 +#define VMWGFX_DRIVER_MINOR 0 +#define VMWGFX_DRIVER_PATCHLEVEL 0 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024) #define VMWGFX_MAX_RELOCATIONS 2048 @@ -95,6 +96,8 @@ struct vmw_surface { struct drm_vmw_size *sizes; uint32_t num_sizes; + bool scanout; + /* TODO so far just a extra pointer */ struct vmw_cursor_snooper snooper; }; @@ -110,6 +113,7 @@ struct vmw_fifo_state { unsigned long static_buffer_size; bool using_bounce_buffer; uint32_t capabilities; + struct mutex fifo_mutex; struct rw_semaphore rwsem; }; @@ -210,7 +214,7 @@ struct vmw_private { * Fencing and IRQs. */ - uint32_t fence_seq; + atomic_t fence_seq; wait_queue_head_t fence_queue; wait_queue_head_t fifo_queue; atomic_t fence_queue_waiters; @@ -258,6 +262,7 @@ struct vmw_private { struct vmw_master *active_master; struct vmw_master fbdev_master; + struct notifier_block pm_nb; }; static inline struct vmw_private *vmw_priv(struct drm_device *dev) @@ -353,6 +358,7 @@ extern int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv, struct vmw_dma_buffer *bo); extern int vmw_dmabuf_from_vram(struct vmw_private *vmw_priv, struct vmw_dma_buffer *bo); +extern void vmw_dmabuf_gmr_unbind(struct ttm_buffer_object *bo); extern int vmw_stream_claim_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_stream_unref_ioctl(struct drm_device *dev, void *data, @@ -386,6 +392,7 @@ extern int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence); extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason); extern int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma); +extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv); /** * TTM glue - vmwgfx_ttm_glue.c @@ -401,6 +408,7 @@ extern int vmw_mmap(struct file *filp, struct vm_area_struct *vma); extern struct ttm_placement vmw_vram_placement; extern struct ttm_placement vmw_vram_ne_placement; +extern struct ttm_placement vmw_vram_sys_placement; extern struct ttm_placement vmw_sys_placement; extern struct ttm_bo_driver vmw_bo_driver; extern int vmw_dma_quiescent(struct drm_device *dev); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index 2e92da56740..0897359b3e4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -182,25 +182,19 @@ static int vmw_cmd_present_check(struct vmw_private *dev_priv, return vmw_cmd_sid_check(dev_priv, sw_context, &cmd->body.sid); } -static int vmw_cmd_dma(struct vmw_private *dev_priv, - struct vmw_sw_context *sw_context, - SVGA3dCmdHeader *header) +static int vmw_translate_guest_ptr(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGAGuestPtr *ptr, + struct vmw_dma_buffer **vmw_bo_p) { - uint32_t handle; struct vmw_dma_buffer *vmw_bo = NULL; struct ttm_buffer_object *bo; - struct vmw_surface *srf = NULL; - struct vmw_dma_cmd { - SVGA3dCmdHeader header; - SVGA3dCmdSurfaceDMA dma; - } *cmd; + uint32_t handle = ptr->gmrId; struct vmw_relocation *reloc; - int ret; uint32_t cur_validate_node; struct ttm_validate_buffer *val_buf; + int ret; - cmd = container_of(header, struct vmw_dma_cmd, header); - handle = cmd->dma.guest.ptr.gmrId; ret = vmw_user_dmabuf_lookup(sw_context->tfile, handle, &vmw_bo); if (unlikely(ret != 0)) { DRM_ERROR("Could not find or use GMR region.\n"); @@ -209,14 +203,14 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, bo = &vmw_bo->base; if (unlikely(sw_context->cur_reloc >= VMWGFX_MAX_RELOCATIONS)) { - DRM_ERROR("Max number of DMA commands per submission" + DRM_ERROR("Max number relocations per submission" " exceeded\n"); ret = -EINVAL; goto out_no_reloc; } reloc = &sw_context->relocs[sw_context->cur_reloc++]; - reloc->location = &cmd->dma.guest.ptr; + reloc->location = ptr; cur_validate_node = vmw_dmabuf_validate_node(bo, sw_context->cur_val_buf); if (unlikely(cur_validate_node >= VMWGFX_MAX_GMRS)) { @@ -234,7 +228,89 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, list_add_tail(&val_buf->head, &sw_context->validate_nodes); ++sw_context->cur_val_buf; } + *vmw_bo_p = vmw_bo; + return 0; + +out_no_reloc: + vmw_dmabuf_unreference(&vmw_bo); + vmw_bo_p = NULL; + return ret; +} + +static int vmw_cmd_end_query(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo; + struct vmw_query_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdEndQuery q; + } *cmd; + int ret; + cmd = container_of(header, struct vmw_query_cmd, header); + ret = vmw_cmd_cid_check(dev_priv, sw_context, header); + if (unlikely(ret != 0)) + return ret; + + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->q.guestResult, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + vmw_dmabuf_unreference(&vmw_bo); + return 0; +} + +static int vmw_cmd_wait_query(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo; + struct vmw_query_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdWaitForQuery q; + } *cmd; + int ret; + + cmd = container_of(header, struct vmw_query_cmd, header); + ret = vmw_cmd_cid_check(dev_priv, sw_context, header); + if (unlikely(ret != 0)) + return ret; + + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->q.guestResult, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + vmw_dmabuf_unreference(&vmw_bo); + return 0; +} + + +static int vmw_cmd_dma(struct vmw_private *dev_priv, + struct vmw_sw_context *sw_context, + SVGA3dCmdHeader *header) +{ + struct vmw_dma_buffer *vmw_bo = NULL; + struct ttm_buffer_object *bo; + struct vmw_surface *srf = NULL; + struct vmw_dma_cmd { + SVGA3dCmdHeader header; + SVGA3dCmdSurfaceDMA dma; + } *cmd; + int ret; + + cmd = container_of(header, struct vmw_dma_cmd, header); + ret = vmw_translate_guest_ptr(dev_priv, sw_context, + &cmd->dma.guest.ptr, + &vmw_bo); + if (unlikely(ret != 0)) + return ret; + + bo = &vmw_bo->base; ret = vmw_user_surface_lookup_handle(dev_priv, sw_context->tfile, cmd->dma.host.sid, &srf); if (ret) { @@ -379,8 +455,8 @@ static vmw_cmd_func vmw_cmd_funcs[SVGA_3D_CMD_MAX] = { VMW_CMD_DEF(SVGA_3D_CMD_DRAW_PRIMITIVES, &vmw_cmd_draw), VMW_CMD_DEF(SVGA_3D_CMD_SETSCISSORRECT, &vmw_cmd_cid_check), VMW_CMD_DEF(SVGA_3D_CMD_BEGIN_QUERY, &vmw_cmd_cid_check), - VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_cid_check), - VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_cid_check), + VMW_CMD_DEF(SVGA_3D_CMD_END_QUERY, &vmw_cmd_end_query), + VMW_CMD_DEF(SVGA_3D_CMD_WAIT_FOR_QUERY, &vmw_cmd_wait_query), VMW_CMD_DEF(SVGA_3D_CMD_PRESENT_READBACK, &vmw_cmd_ok), VMW_CMD_DEF(SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN, &vmw_cmd_blt_surf_screen_check) @@ -490,10 +566,29 @@ static int vmw_validate_single_buffer(struct vmw_private *dev_priv, if (vmw_dmabuf_gmr(bo) != SVGA_GMR_NULL) return 0; + /** + * Put BO in VRAM, only if there is space. + */ + + ret = ttm_bo_validate(bo, &vmw_vram_sys_placement, true, false); + if (unlikely(ret == -ERESTARTSYS)) + return ret; + + /** + * Otherwise, set it up as GMR. + */ + + if (vmw_dmabuf_gmr(bo) != SVGA_GMR_NULL) + return 0; + ret = vmw_gmr_bind(dev_priv, bo); if (likely(ret == 0 || ret == -ERESTARTSYS)) return ret; + /** + * If that failed, try VRAM again, this time evicting + * previous contents. + */ ret = ttm_bo_validate(bo, &vmw_vram_placement, true, false); return ret; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c index 641dde76ada..a93367041cd 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c @@ -559,6 +559,9 @@ int vmw_fb_init(struct vmw_private *vmw_priv) info->pixmap.scan_align = 1; #endif + info->aperture_base = vmw_priv->vram_start; + info->aperture_size = vmw_priv->vram_size; + /* * Dirty & Deferred IO */ @@ -649,14 +652,6 @@ int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv, if (unlikely(ret != 0)) goto err_unlock; - if (vmw_bo->gmr_bound) { - vmw_gmr_unbind(vmw_priv, vmw_bo->gmr_id); - spin_lock(&bo->glob->lru_lock); - ida_remove(&vmw_priv->gmr_ida, vmw_bo->gmr_id); - spin_unlock(&bo->glob->lru_lock); - vmw_bo->gmr_bound = NULL; - } - ret = ttm_bo_validate(bo, &ne_placement, false, false); ttm_bo_unreserve(bo); err_unlock: diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c index 01feb48af33..39d43a01d84 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c @@ -29,6 +29,25 @@ #include "drmP.h" #include "ttm/ttm_placement.h" +bool vmw_fifo_have_3d(struct vmw_private *dev_priv) +{ + __le32 __iomem *fifo_mem = dev_priv->mmio_virt; + uint32_t fifo_min, hwversion; + + fifo_min = ioread32(fifo_mem + SVGA_FIFO_MIN); + if (fifo_min <= SVGA_FIFO_3D_HWVERSION * sizeof(unsigned int)) + return false; + + hwversion = ioread32(fifo_mem + SVGA_FIFO_3D_HWVERSION); + if (hwversion == 0) + return false; + + if (hwversion < SVGA3D_HWVERSION_WS65_B1) + return false; + + return true; +} + int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) { __le32 __iomem *fifo_mem = dev_priv->mmio_virt; @@ -55,6 +74,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) fifo->reserved_size = 0; fifo->using_bounce_buffer = false; + mutex_init(&fifo->fifo_mutex); init_rwsem(&fifo->rwsem); /* @@ -98,8 +118,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo) (unsigned int) min, (unsigned int) fifo->capabilities); - dev_priv->fence_seq = (uint32_t) -100; - dev_priv->last_read_sequence = (uint32_t) -100; + atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence); iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE); return vmw_fifo_send_fence(dev_priv, &dummy); @@ -265,7 +284,7 @@ void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes) uint32_t reserveable = fifo_state->capabilities & SVGA_FIFO_CAP_RESERVE; int ret; - down_write(&fifo_state->rwsem); + mutex_lock(&fifo_state->fifo_mutex); max = ioread32(fifo_mem + SVGA_FIFO_MAX); min = ioread32(fifo_mem + SVGA_FIFO_MIN); next_cmd = ioread32(fifo_mem + SVGA_FIFO_NEXT_CMD); @@ -333,7 +352,7 @@ void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes) } out_err: fifo_state->reserved_size = 0; - up_write(&fifo_state->rwsem); + mutex_unlock(&fifo_state->fifo_mutex); return NULL; } @@ -408,6 +427,7 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes) } + down_write(&fifo_state->rwsem); if (fifo_state->using_bounce_buffer || reserveable) { next_cmd += bytes; if (next_cmd >= max) @@ -419,8 +439,9 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes) if (reserveable) iowrite32(0, fifo_mem + SVGA_FIFO_RESERVED); mb(); - vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC); up_write(&fifo_state->rwsem); + vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC); + mutex_unlock(&fifo_state->fifo_mutex); } int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence) @@ -433,9 +454,7 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence) fm = vmw_fifo_reserve(dev_priv, bytes); if (unlikely(fm == NULL)) { - down_write(&fifo_state->rwsem); - *sequence = dev_priv->fence_seq; - up_write(&fifo_state->rwsem); + *sequence = atomic_read(&dev_priv->fence_seq); ret = -ENOMEM; (void)vmw_fallback_wait(dev_priv, false, true, *sequence, false, 3*HZ); @@ -443,7 +462,7 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence) } do { - *sequence = dev_priv->fence_seq++; + *sequence = atomic_add_return(1, &dev_priv->fence_seq); } while (*sequence == 0); if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE)) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c index 5fa6a4ed238..1c7a316454d 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c @@ -43,11 +43,17 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data, param->value = vmw_overlay_num_free_overlays(dev_priv); break; case DRM_VMW_PARAM_3D: - param->value = dev_priv->capabilities & SVGA_CAP_3D ? 1 : 0; + param->value = vmw_fifo_have_3d(dev_priv) ? 1 : 0; break; case DRM_VMW_PARAM_FIFO_OFFSET: param->value = dev_priv->mmio_start; break; + case DRM_VMW_PARAM_HW_CAPS: + param->value = dev_priv->capabilities; + break; + case DRM_VMW_PARAM_FIFO_CAPS: + param->value = dev_priv->fifo.capabilities; + break; default: DRM_ERROR("Illegal vmwgfx get param request: %d\n", param->param); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c index d40086fc864..4d7cb539386 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c @@ -85,19 +85,12 @@ bool vmw_fence_signaled(struct vmw_private *dev_priv, return true; /** - * Below is to signal stale fences that have wrapped. - * First, block fence submission. - */ - - down_read(&fifo_state->rwsem); - - /** * Then check if the sequence is higher than what we've actually * emitted. Then the fence is stale and signaled. */ - ret = ((dev_priv->fence_seq - sequence) > VMW_FENCE_WRAP); - up_read(&fifo_state->rwsem); + ret = ((atomic_read(&dev_priv->fence_seq) - sequence) + > VMW_FENCE_WRAP); return ret; } @@ -127,7 +120,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv, if (fifo_idle) down_read(&fifo_state->rwsem); - signal_seq = dev_priv->fence_seq; + signal_seq = atomic_read(&dev_priv->fence_seq); ret = 0; for (;;) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index b1af76e371c..31f9afed0a6 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -553,9 +553,7 @@ int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer, } *cmd; int i, increment = 1; - if (!num_clips || - !(dev_priv->fifo.capabilities & - SVGA_FIFO_CAP_SCREEN_OBJECT)) { + if (!num_clips) { num_clips = 1; clips = &norect; norect.x1 = norect.y1 = 0; @@ -574,10 +572,10 @@ int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer, for (i = 0; i < num_clips; i++, clips += increment) { cmd[i].header = cpu_to_le32(SVGA_CMD_UPDATE); - cmd[i].body.x = cpu_to_le32(clips[i].x1); - cmd[i].body.y = cpu_to_le32(clips[i].y1); - cmd[i].body.width = cpu_to_le32(clips[i].x2 - clips[i].x1); - cmd[i].body.height = cpu_to_le32(clips[i].y2 - clips[i].y1); + cmd[i].body.x = cpu_to_le32(clips->x1); + cmd[i].body.y = cpu_to_le32(clips->y1); + cmd[i].body.width = cpu_to_le32(clips->x2 - clips->x1); + cmd[i].body.height = cpu_to_le32(clips->y2 - clips->y1); } vmw_fifo_commit(dev_priv, sizeof(*cmd) * num_clips); @@ -709,6 +707,9 @@ static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, if (ret) goto try_dmabuf; + if (!surface->scanout) + goto err_not_scanout; + ret = vmw_kms_new_framebuffer_surface(dev_priv, surface, &vfb, mode_cmd->width, mode_cmd->height); @@ -742,6 +743,13 @@ try_dmabuf: } return &vfb->base; + +err_not_scanout: + DRM_ERROR("surface not marked as scanout\n"); + /* vmw_user_surface_lookup takes one ref */ + vmw_surface_unreference(&surface); + + return NULL; } static int vmw_kms_fb_changed(struct drm_device *dev) @@ -761,10 +769,10 @@ int vmw_kms_init(struct vmw_private *dev_priv) drm_mode_config_init(dev); dev->mode_config.funcs = &vmw_kms_funcs; - dev->mode_config.min_width = 640; - dev->mode_config.min_height = 480; - dev->mode_config.max_width = 2048; - dev->mode_config.max_height = 2048; + dev->mode_config.min_width = 1; + dev->mode_config.min_height = 1; + dev->mode_config.max_width = dev_priv->fb_max_width; + dev->mode_config.max_height = dev_priv->fb_max_height; ret = vmw_kms_init_legacy_display_system(dev_priv); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c index bb6e6a096d2..5b6eabeb7f5 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c @@ -104,7 +104,6 @@ static int vmw_dmabuf_pin_in_vram(struct vmw_private *dev_priv, bool pin, bool interruptible) { struct ttm_buffer_object *bo = &buf->base; - struct ttm_bo_global *glob = bo->glob; struct ttm_placement *overlay_placement = &vmw_vram_placement; int ret; @@ -116,14 +115,6 @@ static int vmw_dmabuf_pin_in_vram(struct vmw_private *dev_priv, if (unlikely(ret != 0)) goto err; - if (buf->gmr_bound) { - vmw_gmr_unbind(dev_priv, buf->gmr_id); - spin_lock(&glob->lru_lock); - ida_remove(&dev_priv->gmr_ida, buf->gmr_id); - spin_unlock(&glob->lru_lock); - buf->gmr_bound = NULL; - } - if (pin) overlay_placement = &vmw_vram_ne_placement; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index c012d5927f6..f8fbbc67a40 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -574,6 +574,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, srf->flags = req->flags; srf->format = req->format; + srf->scanout = req->scanout; memcpy(srf->mip_levels, req->mip_levels, sizeof(srf->mip_levels)); srf->num_sizes = 0; for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i) @@ -599,6 +600,26 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, if (unlikely(ret != 0)) goto out_err1; + if (srf->scanout && + srf->num_sizes == 1 && + srf->sizes[0].width == 64 && + srf->sizes[0].height == 64 && + srf->format == SVGA3D_A8R8G8B8) { + + srf->snooper.image = kmalloc(64 * 64 * 4, GFP_KERNEL); + /* clear the image */ + if (srf->snooper.image) { + memset(srf->snooper.image, 0x00, 64 * 64 * 4); + } else { + DRM_ERROR("Failed to allocate cursor_image\n"); + ret = -ENOMEM; + goto out_err1; + } + } else { + srf->snooper.image = NULL; + } + srf->snooper.crtc = NULL; + user_srf->base.shareable = false; user_srf->base.tfile = NULL; @@ -622,24 +643,6 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data, return ret; } - if (srf->flags & (1 << 9) && - srf->num_sizes == 1 && - srf->sizes[0].width == 64 && - srf->sizes[0].height == 64 && - srf->format == SVGA3D_A8R8G8B8) { - - srf->snooper.image = kmalloc(64 * 64 * 4, GFP_KERNEL); - /* clear the image */ - if (srf->snooper.image) - memset(srf->snooper.image, 0x00, 64 * 64 * 4); - else - DRM_ERROR("Failed to allocate cursor_image\n"); - - } else { - srf->snooper.image = NULL; - } - srf->snooper.crtc = NULL; - rep->sid = user_srf->base.hash.key; if (rep->sid == SVGA3D_INVALID_ID) DRM_ERROR("Created bad Surface ID.\n"); @@ -754,20 +757,29 @@ static size_t vmw_dmabuf_acc_size(struct ttm_bo_global *glob, return bo_user_size + page_array_size; } -void vmw_dmabuf_bo_free(struct ttm_buffer_object *bo) +void vmw_dmabuf_gmr_unbind(struct ttm_buffer_object *bo) { struct vmw_dma_buffer *vmw_bo = vmw_dma_buffer(bo); struct ttm_bo_global *glob = bo->glob; struct vmw_private *dev_priv = container_of(bo->bdev, struct vmw_private, bdev); - ttm_mem_global_free(glob->mem_glob, bo->acc_size); if (vmw_bo->gmr_bound) { vmw_gmr_unbind(dev_priv, vmw_bo->gmr_id); spin_lock(&glob->lru_lock); ida_remove(&dev_priv->gmr_ida, vmw_bo->gmr_id); spin_unlock(&glob->lru_lock); + vmw_bo->gmr_bound = false; } +} + +void vmw_dmabuf_bo_free(struct ttm_buffer_object *bo) +{ + struct vmw_dma_buffer *vmw_bo = vmw_dma_buffer(bo); + struct ttm_bo_global *glob = bo->glob; + + vmw_dmabuf_gmr_unbind(bo); + ttm_mem_global_free(glob->mem_glob, bo->acc_size); kfree(vmw_bo); } @@ -813,18 +825,10 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv, static void vmw_user_dmabuf_destroy(struct ttm_buffer_object *bo) { struct vmw_user_dma_buffer *vmw_user_bo = vmw_user_dma_buffer(bo); - struct vmw_dma_buffer *vmw_bo = &vmw_user_bo->dma; struct ttm_bo_global *glob = bo->glob; - struct vmw_private *dev_priv = - container_of(bo->bdev, struct vmw_private, bdev); + vmw_dmabuf_gmr_unbind(bo); ttm_mem_global_free(glob->mem_glob, bo->acc_size); - if (vmw_bo->gmr_bound) { - vmw_gmr_unbind(dev_priv, vmw_bo->gmr_id); - spin_lock(&glob->lru_lock); - ida_remove(&dev_priv->gmr_ida, vmw_bo->gmr_id); - spin_unlock(&glob->lru_lock); - } kfree(vmw_user_bo); } @@ -868,7 +872,7 @@ int vmw_dmabuf_alloc_ioctl(struct drm_device *dev, void *data, } ret = vmw_dmabuf_init(dev_priv, &vmw_user_bo->dma, req->size, - &vmw_vram_placement, true, + &vmw_vram_sys_placement, true, &vmw_user_dmabuf_destroy); if (unlikely(ret != 0)) return ret; diff --git a/drivers/gpu/vga/Kconfig b/drivers/gpu/vga/Kconfig index 790e675b13e..61ab4daf0bb 100644 --- a/drivers/gpu/vga/Kconfig +++ b/drivers/gpu/vga/Kconfig @@ -8,3 +8,22 @@ config VGA_ARB are accessed at same time they need some kind of coordination. Please see Documentation/vgaarbiter.txt for more details. Select this to enable VGA arbiter. + +config VGA_ARB_MAX_GPUS + int "Maximum number of GPUs" + default 16 + depends on VGA_ARB + help + Reserves space in the kernel to maintain resource locking for + multiple GPUS. The overhead for each GPU is very small. + +config VGA_SWITCHEROO + bool "Laptop Hybrid Grapics - GPU switching support" + depends on X86 + depends on ACPI + help + Many laptops released in 2008/9/10 have two gpus with a multiplxer + to switch between them. This adds support for dynamic switching when + X isn't running and delayed switching until the next logoff. This + features is called hybrid graphics, ATI PowerXpress, and Nvidia + HybridPower. diff --git a/drivers/gpu/vga/Makefile b/drivers/gpu/vga/Makefile index 7cc8c1ed645..14ca30b75d0 100644 --- a/drivers/gpu/vga/Makefile +++ b/drivers/gpu/vga/Makefile @@ -1 +1,2 @@ obj-$(CONFIG_VGA_ARB) += vgaarb.o +obj-$(CONFIG_VGA_SWITCHEROO) += vga_switcheroo.o diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c new file mode 100644 index 00000000000..d6d1149d525 --- /dev/null +++ b/drivers/gpu/vga/vga_switcheroo.c @@ -0,0 +1,450 @@ +/* + * Copyright (c) 2010 Red Hat Inc. + * Author : Dave Airlie <airlied@redhat.com> + * + * + * Licensed under GPLv2 + * + * vga_switcheroo.c - Support for laptop with dual GPU using one set of outputs + + Switcher interface - methods require for ATPX and DCM + - switchto - this throws the output MUX switch + - discrete_set_power - sets the power state for the discrete card + + GPU driver interface + - set_gpu_state - this should do the equiv of s/r for the card + - this should *not* set the discrete power state + - switch_check - check if the device is in a position to switch now + */ + +#include <linux/module.h> +#include <linux/dmi.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/fs.h> +#include <linux/debugfs.h> +#include <linux/fb.h> + +#include <linux/pci.h> +#include <linux/vga_switcheroo.h> + +struct vga_switcheroo_client { + struct pci_dev *pdev; + struct fb_info *fb_info; + int pwr_state; + void (*set_gpu_state)(struct pci_dev *pdev, enum vga_switcheroo_state); + bool (*can_switch)(struct pci_dev *pdev); + int id; + bool active; +}; + +static DEFINE_MUTEX(vgasr_mutex); + +struct vgasr_priv { + + bool active; + bool delayed_switch_active; + enum vga_switcheroo_client_id delayed_client_id; + + struct dentry *debugfs_root; + struct dentry *switch_file; + + int registered_clients; + struct vga_switcheroo_client clients[VGA_SWITCHEROO_MAX_CLIENTS]; + + struct vga_switcheroo_handler *handler; +}; + +static int vga_switcheroo_debugfs_init(struct vgasr_priv *priv); +static void vga_switcheroo_debugfs_fini(struct vgasr_priv *priv); + +/* only one switcheroo per system */ +static struct vgasr_priv vgasr_priv; + +int vga_switcheroo_register_handler(struct vga_switcheroo_handler *handler) +{ + mutex_lock(&vgasr_mutex); + if (vgasr_priv.handler) { + mutex_unlock(&vgasr_mutex); + return -EINVAL; + } + + vgasr_priv.handler = handler; + mutex_unlock(&vgasr_mutex); + return 0; +} +EXPORT_SYMBOL(vga_switcheroo_register_handler); + +void vga_switcheroo_unregister_handler(void) +{ + mutex_lock(&vgasr_mutex); + vgasr_priv.handler = NULL; + mutex_unlock(&vgasr_mutex); +} +EXPORT_SYMBOL(vga_switcheroo_unregister_handler); + +static void vga_switcheroo_enable(void) +{ + int i; + int ret; + /* call the handler to init */ + vgasr_priv.handler->init(); + + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + ret = vgasr_priv.handler->get_client_id(vgasr_priv.clients[i].pdev); + if (ret < 0) + return; + + vgasr_priv.clients[i].id = ret; + } + vga_switcheroo_debugfs_init(&vgasr_priv); + vgasr_priv.active = true; +} + +int vga_switcheroo_register_client(struct pci_dev *pdev, + void (*set_gpu_state)(struct pci_dev *pdev, enum vga_switcheroo_state), + bool (*can_switch)(struct pci_dev *pdev)) +{ + int index; + + mutex_lock(&vgasr_mutex); + /* don't do IGD vs DIS here */ + if (vgasr_priv.registered_clients & 1) + index = 1; + else + index = 0; + + vgasr_priv.clients[index].pwr_state = VGA_SWITCHEROO_ON; + vgasr_priv.clients[index].pdev = pdev; + vgasr_priv.clients[index].set_gpu_state = set_gpu_state; + vgasr_priv.clients[index].can_switch = can_switch; + vgasr_priv.clients[index].id = -1; + if (pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW) + vgasr_priv.clients[index].active = true; + + vgasr_priv.registered_clients |= (1 << index); + + /* if we get two clients + handler */ + if (vgasr_priv.registered_clients == 0x3 && vgasr_priv.handler) { + printk(KERN_INFO "vga_switcheroo: enabled\n"); + vga_switcheroo_enable(); + } + mutex_unlock(&vgasr_mutex); + return 0; +} +EXPORT_SYMBOL(vga_switcheroo_register_client); + +void vga_switcheroo_unregister_client(struct pci_dev *pdev) +{ + int i; + + mutex_lock(&vgasr_mutex); + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].pdev == pdev) { + vgasr_priv.registered_clients &= ~(1 << i); + break; + } + } + + printk(KERN_INFO "vga_switcheroo: disabled\n"); + vga_switcheroo_debugfs_fini(&vgasr_priv); + vgasr_priv.active = false; + mutex_unlock(&vgasr_mutex); +} +EXPORT_SYMBOL(vga_switcheroo_unregister_client); + +void vga_switcheroo_client_fb_set(struct pci_dev *pdev, + struct fb_info *info) +{ + int i; + + mutex_lock(&vgasr_mutex); + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].pdev == pdev) { + vgasr_priv.clients[i].fb_info = info; + break; + } + } + mutex_unlock(&vgasr_mutex); +} +EXPORT_SYMBOL(vga_switcheroo_client_fb_set); + +static int vga_switcheroo_show(struct seq_file *m, void *v) +{ + int i; + mutex_lock(&vgasr_mutex); + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + seq_printf(m, "%d:%c:%s:%s\n", i, + vgasr_priv.clients[i].active ? '+' : ' ', + vgasr_priv.clients[i].pwr_state ? "Pwr" : "Off", + pci_name(vgasr_priv.clients[i].pdev)); + } + mutex_unlock(&vgasr_mutex); + return 0; +} + +static int vga_switcheroo_debugfs_open(struct inode *inode, struct file *file) +{ + return single_open(file, vga_switcheroo_show, NULL); +} + +static int vga_switchon(struct vga_switcheroo_client *client) +{ + int ret; + + ret = vgasr_priv.handler->power_state(client->id, VGA_SWITCHEROO_ON); + /* call the driver callback to turn on device */ + client->set_gpu_state(client->pdev, VGA_SWITCHEROO_ON); + client->pwr_state = VGA_SWITCHEROO_ON; + return 0; +} + +static int vga_switchoff(struct vga_switcheroo_client *client) +{ + /* call the driver callback to turn off device */ + client->set_gpu_state(client->pdev, VGA_SWITCHEROO_OFF); + vgasr_priv.handler->power_state(client->id, VGA_SWITCHEROO_OFF); + client->pwr_state = VGA_SWITCHEROO_OFF; + return 0; +} + +static int vga_switchto(struct vga_switcheroo_client *new_client) +{ + int ret; + int i; + struct vga_switcheroo_client *active = NULL; + + if (new_client->active == true) + return 0; + + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].active == true) { + active = &vgasr_priv.clients[i]; + break; + } + } + if (!active) + return 0; + + /* power up the first device */ + ret = pci_enable_device(new_client->pdev); + if (ret) + return ret; + + if (new_client->pwr_state == VGA_SWITCHEROO_OFF) + vga_switchon(new_client); + + /* swap shadow resource to denote boot VGA device has changed so X starts on new device */ + active->active = false; + + active->pdev->resource[PCI_ROM_RESOURCE].flags &= ~IORESOURCE_ROM_SHADOW; + new_client->pdev->resource[PCI_ROM_RESOURCE].flags |= IORESOURCE_ROM_SHADOW; + + if (new_client->fb_info) { + struct fb_event event; + event.info = new_client->fb_info; + fb_notifier_call_chain(FB_EVENT_REMAP_ALL_CONSOLE, &event); + } + + ret = vgasr_priv.handler->switchto(new_client->id); + if (ret) + return ret; + + if (active->pwr_state == VGA_SWITCHEROO_ON) + vga_switchoff(active); + + new_client->active = true; + return 0; +} + +static ssize_t +vga_switcheroo_debugfs_write(struct file *filp, const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + char usercmd[64]; + const char *pdev_name; + int i, ret; + bool delay = false, can_switch; + int client_id = -1; + struct vga_switcheroo_client *client = NULL; + + if (cnt > 63) + cnt = 63; + + if (copy_from_user(usercmd, ubuf, cnt)) + return -EFAULT; + + mutex_lock(&vgasr_mutex); + + if (!vgasr_priv.active) + return -EINVAL; + + /* pwr off the device not in use */ + if (strncmp(usercmd, "OFF", 3) == 0) { + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].active) + continue; + if (vgasr_priv.clients[i].pwr_state == VGA_SWITCHEROO_ON) + vga_switchoff(&vgasr_priv.clients[i]); + } + goto out; + } + /* pwr on the device not in use */ + if (strncmp(usercmd, "ON", 2) == 0) { + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].active) + continue; + if (vgasr_priv.clients[i].pwr_state == VGA_SWITCHEROO_OFF) + vga_switchon(&vgasr_priv.clients[i]); + } + goto out; + } + + /* request a delayed switch - test can we switch now */ + if (strncmp(usercmd, "DIGD", 4) == 0) { + client_id = VGA_SWITCHEROO_IGD; + delay = true; + } + + if (strncmp(usercmd, "DDIS", 4) == 0) { + client_id = VGA_SWITCHEROO_DIS; + delay = true; + } + + if (strncmp(usercmd, "IGD", 3) == 0) + client_id = VGA_SWITCHEROO_IGD; + + if (strncmp(usercmd, "DIS", 3) == 0) + client_id = VGA_SWITCHEROO_DIS; + + if (client_id == -1) + goto out; + + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].id == client_id) { + client = &vgasr_priv.clients[i]; + break; + } + } + + vgasr_priv.delayed_switch_active = false; + /* okay we want a switch - test if devices are willing to switch */ + can_switch = true; + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + can_switch = vgasr_priv.clients[i].can_switch(vgasr_priv.clients[i].pdev); + if (can_switch == false) { + printk(KERN_ERR "vga_switcheroo: client %d refused switch\n", i); + break; + } + } + + if (can_switch == false && delay == false) + goto out; + + if (can_switch == true) { + pdev_name = pci_name(client->pdev); + ret = vga_switchto(client); + if (ret) + printk(KERN_ERR "vga_switcheroo: switching failed %d\n", ret); + } else { + printk(KERN_INFO "vga_switcheroo: setting delayed switch to client %d\n", client->id); + vgasr_priv.delayed_switch_active = true; + vgasr_priv.delayed_client_id = client_id; + + /* we should at least power up the card to + make the switch faster */ + if (client->pwr_state == VGA_SWITCHEROO_OFF) + vga_switchon(client); + } + +out: + mutex_unlock(&vgasr_mutex); + return cnt; +} + +static const struct file_operations vga_switcheroo_debugfs_fops = { + .owner = THIS_MODULE, + .open = vga_switcheroo_debugfs_open, + .write = vga_switcheroo_debugfs_write, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static void vga_switcheroo_debugfs_fini(struct vgasr_priv *priv) +{ + if (priv->switch_file) { + debugfs_remove(priv->switch_file); + priv->switch_file = NULL; + } + if (priv->debugfs_root) { + debugfs_remove(priv->debugfs_root); + priv->debugfs_root = NULL; + } +} + +static int vga_switcheroo_debugfs_init(struct vgasr_priv *priv) +{ + /* already initialised */ + if (priv->debugfs_root) + return 0; + priv->debugfs_root = debugfs_create_dir("vgaswitcheroo", NULL); + + if (!priv->debugfs_root) { + printk(KERN_ERR "vga_switcheroo: Cannot create /sys/kernel/debug/vgaswitcheroo\n"); + goto fail; + } + + priv->switch_file = debugfs_create_file("switch", 0644, + priv->debugfs_root, NULL, &vga_switcheroo_debugfs_fops); + if (!priv->switch_file) { + printk(KERN_ERR "vga_switcheroo: cannot create /sys/kernel/debug/vgaswitcheroo/switch\n"); + goto fail; + } + return 0; +fail: + vga_switcheroo_debugfs_fini(priv); + return -1; +} + +int vga_switcheroo_process_delayed_switch(void) +{ + struct vga_switcheroo_client *client = NULL; + const char *pdev_name; + bool can_switch = true; + int i; + int ret; + int err = -EINVAL; + + mutex_lock(&vgasr_mutex); + if (!vgasr_priv.delayed_switch_active) + goto err; + + printk(KERN_INFO "vga_switcheroo: processing delayed switch to %d\n", vgasr_priv.delayed_client_id); + + for (i = 0; i < VGA_SWITCHEROO_MAX_CLIENTS; i++) { + if (vgasr_priv.clients[i].id == vgasr_priv.delayed_client_id) + client = &vgasr_priv.clients[i]; + can_switch = vgasr_priv.clients[i].can_switch(vgasr_priv.clients[i].pdev); + if (can_switch == false) { + printk(KERN_ERR "vga_switcheroo: client %d refused switch\n", i); + break; + } + } + + if (can_switch == false || client == NULL) + goto err; + + pdev_name = pci_name(client->pdev); + ret = vga_switchto(client); + if (ret) + printk(KERN_ERR "vga_switcheroo: delayed switching failed %d\n", ret); + + vgasr_priv.delayed_switch_active = false; + err = 0; +err: + mutex_unlock(&vgasr_mutex); + return err; +} +EXPORT_SYMBOL(vga_switcheroo_process_delayed_switch); + diff --git a/drivers/gpu/vga/vgaarb.c b/drivers/gpu/vga/vgaarb.c index 1ac0c93603c..8827814d073 100644 --- a/drivers/gpu/vga/vgaarb.c +++ b/drivers/gpu/vga/vgaarb.c @@ -688,7 +688,7 @@ EXPORT_SYMBOL(vga_client_register); * the arbiter. */ -#define MAX_USER_CARDS 16 +#define MAX_USER_CARDS CONFIG_VGA_ARB_MAX_GPUS #define PCI_INVALID_CARD ((struct pci_dev *)-1UL) /* @@ -954,6 +954,7 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, } } else if (strncmp(curr_pos, "target ", 7) == 0) { + struct pci_bus *pbus; unsigned int domain, bus, devfn; struct vga_device *vgadev; @@ -961,7 +962,7 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, remaining -= 7; pr_devel("client 0x%p called 'target'\n", priv); /* if target is default */ - if (!strncmp(buf, "default", 7)) + if (!strncmp(curr_pos, "default", 7)) pdev = pci_dev_get(vga_default_device()); else { if (!vga_pci_str_to_vars(curr_pos, remaining, @@ -969,18 +970,31 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, ret_val = -EPROTO; goto done; } - - pdev = pci_get_bus_and_slot(bus, devfn); + pr_devel("vgaarb: %s ==> %x:%x:%x.%x\n", curr_pos, + domain, bus, PCI_SLOT(devfn), PCI_FUNC(devfn)); + + pbus = pci_find_bus(domain, bus); + pr_devel("vgaarb: pbus %p\n", pbus); + if (pbus == NULL) { + pr_err("vgaarb: invalid PCI domain and/or bus address %x:%x\n", + domain, bus); + ret_val = -ENODEV; + goto done; + } + pdev = pci_get_slot(pbus, devfn); + pr_devel("vgaarb: pdev %p\n", pdev); if (!pdev) { - pr_info("vgaarb: invalid PCI address!\n"); + pr_err("vgaarb: invalid PCI address %x:%x\n", + bus, devfn); ret_val = -ENODEV; goto done; } } vgadev = vgadev_find(pdev); + pr_devel("vgaarb: vgadev %p\n", vgadev); if (vgadev == NULL) { - pr_info("vgaarb: this pci device is not a vga device\n"); + pr_err("vgaarb: this pci device is not a vga device\n"); pci_dev_put(pdev); ret_val = -ENODEV; goto done; @@ -998,7 +1012,8 @@ static ssize_t vga_arb_write(struct file *file, const char __user * buf, } } if (i == MAX_USER_CARDS) { - pr_err("vgaarb: maximum user cards number reached!\n"); + pr_err("vgaarb: maximum user cards (%d) number reached!\n", + MAX_USER_CARDS); pci_dev_put(pdev); /* XXX: which value to return? */ ret_val = -ENOMEM; |