From 771fe6b912fca54f03e8a72eb63058b582775362 Mon Sep 17 00:00:00 2001 From: Jerome Glisse Date: Fri, 5 Jun 2009 14:42:42 +0200 Subject: drm/radeon: introduce kernel modesetting for radeon hardware Add kernel modesetting support to radeon driver, use the ttm memory manager to manage memory and DRM/GEM to provide userspace API. In order to avoid backward compatibility issue and to allow clean design and code the radeon kernel modesetting use different code path than old radeon/drm driver. When kernel modesetting is enabled the IOCTL of radeon/drm driver are considered as invalid and an error message is printed in the log and they return failure. KMS enabled userspace will use new API to talk with the radeon/drm driver. The new API provide functions to create/destroy/share/mmap buffer object which are then managed by the kernel memory manager (here TTM). In order to submit command to the GPU the userspace provide a buffer holding the command stream, along this buffer userspace have to provide a list of buffer object used by the command stream. The kernel radeon driver will then place buffer in GPU accessible memory and will update command stream to reflect the position of the different buffers. The kernel will also perform security check on command stream provided by the user, we want to catch and forbid any illegal use of the GPU such as DMA into random system memory or into memory not owned by the process supplying the command stream. This part of the code is still incomplete and this why we propose that patch as a staging driver addition, future security might forbid current experimental userspace to run. This code support the following hardware : R1XX,R2XX,R3XX,R4XX,R5XX (radeon up to X1950). Works is underway to provide support for R6XX, R7XX and newer hardware (radeon from HD2XXX to HD4XXX). Authors: Jerome Glisse Dave Airlie Alex Deucher Signed-off-by: Jerome Glisse Signed-off-by: Dave Airlie Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 1524 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 1524 insertions(+) create mode 100644 drivers/gpu/drm/radeon/r100.c (limited to 'drivers/gpu/drm/radeon/r100.c') diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c new file mode 100644 index 00000000000..5225f5be7ea --- /dev/null +++ b/drivers/gpu/drm/radeon/r100.c @@ -0,0 +1,1524 @@ +/* + * Copyright 2008 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * 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: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ +#include +#include "drmP.h" +#include "drm.h" +#include "radeon_drm.h" +#include "radeon_microcode.h" +#include "radeon_reg.h" +#include "radeon.h" + +/* This files gather functions specifics to: + * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 + * + * Some of these functions might be used by newer ASICs. + */ +void r100_hdp_reset(struct radeon_device *rdev); +void r100_gpu_init(struct radeon_device *rdev); +int r100_gui_wait_for_idle(struct radeon_device *rdev); +int r100_mc_wait_for_idle(struct radeon_device *rdev); +void r100_gpu_wait_for_vsync(struct radeon_device *rdev); +void r100_gpu_wait_for_vsync2(struct radeon_device *rdev); +int r100_debugfs_mc_info_init(struct radeon_device *rdev); + + +/* + * PCI GART + */ +void r100_pci_gart_tlb_flush(struct radeon_device *rdev) +{ + /* TODO: can we do somethings here ? */ + /* It seems hw only cache one entry so we should discard this + * entry otherwise if first GPU GART read hit this entry it + * could end up in wrong address. */ +} + +int r100_pci_gart_enable(struct radeon_device *rdev) +{ + uint32_t tmp; + int r; + + /* Initialize common gart structure */ + r = radeon_gart_init(rdev); + if (r) { + return r; + } + if (rdev->gart.table.ram.ptr == NULL) { + rdev->gart.table_size = rdev->gart.num_gpu_pages * 4; + r = radeon_gart_table_ram_alloc(rdev); + if (r) { + return r; + } + } + /* 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); + /* Enable bus mastering */ + tmp = RREG32(RADEON_BUS_CNTL) & ~RADEON_BUS_MASTER_DIS; + WREG32(RADEON_BUS_CNTL, tmp); + /* 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; + WREG32(RADEON_AIC_CNTL, tmp); + r100_pci_gart_tlb_flush(rdev); + rdev->gart.ready = true; + return 0; +} + +void r100_pci_gart_disable(struct radeon_device *rdev) +{ + uint32_t tmp; + + /* discard memory request outside of configured range */ + tmp = RREG32(RADEON_AIC_CNTL) | RADEON_DIS_OUT_OF_PCI_GART_ACCESS; + WREG32(RADEON_AIC_CNTL, tmp & ~RADEON_PCIGART_TRANSLATE_EN); + WREG32(RADEON_AIC_LO_ADDR, 0); + WREG32(RADEON_AIC_HI_ADDR, 0); +} + +int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr) +{ + if (i < 0 || i > rdev->gart.num_gpu_pages) { + return -EINVAL; + } + rdev->gart.table.ram.ptr[i] = cpu_to_le32((uint32_t)addr); + return 0; +} + +int r100_gart_enable(struct radeon_device *rdev) +{ + if (rdev->flags & RADEON_IS_AGP) { + r100_pci_gart_disable(rdev); + return 0; + } + return r100_pci_gart_enable(rdev); +} + + +/* + * MC + */ +void r100_mc_disable_clients(struct radeon_device *rdev) +{ + uint32_t ov0_scale_cntl, crtc_ext_cntl, crtc_gen_cntl, crtc2_gen_cntl; + + /* FIXME: is this function correct for rs100,rs200,rs300 ? */ + if (r100_gui_wait_for_idle(rdev)) { + printk(KERN_WARNING "Failed to wait GUI idle while " + "programming pipes. Bad things might happen.\n"); + } + + /* stop display and memory access */ + ov0_scale_cntl = RREG32(RADEON_OV0_SCALE_CNTL); + WREG32(RADEON_OV0_SCALE_CNTL, ov0_scale_cntl & ~RADEON_SCALER_ENABLE); + crtc_ext_cntl = RREG32(RADEON_CRTC_EXT_CNTL); + WREG32(RADEON_CRTC_EXT_CNTL, crtc_ext_cntl | RADEON_CRTC_DISPLAY_DIS); + crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL); + + r100_gpu_wait_for_vsync(rdev); + + WREG32(RADEON_CRTC_GEN_CNTL, + (crtc_gen_cntl & ~(RADEON_CRTC_CUR_EN | RADEON_CRTC_ICON_EN)) | + RADEON_CRTC_DISP_REQ_EN_B | RADEON_CRTC_EXT_DISP_EN); + + if (!(rdev->flags & RADEON_SINGLE_CRTC)) { + crtc2_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL); + + r100_gpu_wait_for_vsync2(rdev); + WREG32(RADEON_CRTC2_GEN_CNTL, + (crtc2_gen_cntl & + ~(RADEON_CRTC2_CUR_EN | RADEON_CRTC2_ICON_EN)) | + RADEON_CRTC2_DISP_REQ_EN_B); + } + + udelay(500); +} + +void r100_mc_setup(struct radeon_device *rdev) +{ + uint32_t tmp; + int r; + + r = r100_debugfs_mc_info_init(rdev); + if (r) { + DRM_ERROR("Failed to register debugfs file for R100 MC !\n"); + } + /* Write VRAM size in case we are limiting it */ + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); + tmp = rdev->mc.vram_location + rdev->mc.vram_size - 1; + tmp = REG_SET(RADEON_MC_FB_TOP, tmp >> 16); + tmp |= REG_SET(RADEON_MC_FB_START, rdev->mc.vram_location >> 16); + WREG32(RADEON_MC_FB_LOCATION, tmp); + + /* Enable bus mastering */ + tmp = RREG32(RADEON_BUS_CNTL) & ~RADEON_BUS_MASTER_DIS; + WREG32(RADEON_BUS_CNTL, tmp); + + if (rdev->flags & RADEON_IS_AGP) { + tmp = rdev->mc.gtt_location + rdev->mc.gtt_size - 1; + tmp = REG_SET(RADEON_MC_AGP_TOP, tmp >> 16); + tmp |= REG_SET(RADEON_MC_AGP_START, rdev->mc.gtt_location >> 16); + WREG32(RADEON_MC_AGP_LOCATION, tmp); + WREG32(RADEON_AGP_BASE, rdev->mc.agp_base); + } else { + WREG32(RADEON_MC_AGP_LOCATION, 0x0FFFFFFF); + WREG32(RADEON_AGP_BASE, 0); + } + + tmp = RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL; + tmp |= (7 << 28); + WREG32(RADEON_HOST_PATH_CNTL, tmp | RADEON_HDP_SOFT_RESET | RADEON_HDP_READ_BUFFER_INVALIDATE); + (void)RREG32(RADEON_HOST_PATH_CNTL); + WREG32(RADEON_HOST_PATH_CNTL, tmp); + (void)RREG32(RADEON_HOST_PATH_CNTL); +} + +int r100_mc_init(struct radeon_device *rdev) +{ + int r; + + if (r100_debugfs_rbbm_init(rdev)) { + DRM_ERROR("Failed to register debugfs file for RBBM !\n"); + } + + r100_gpu_init(rdev); + /* Disable gart which also disable out of gart access */ + r100_pci_gart_disable(rdev); + + /* 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; + } + + r100_mc_disable_clients(rdev); + if (r100_mc_wait_for_idle(rdev)) { + printk(KERN_WARNING "Failed to wait MC idle while " + "programming pipes. Bad things might happen.\n"); + } + + r100_mc_setup(rdev); + return 0; +} + +void r100_mc_fini(struct radeon_device *rdev) +{ + r100_pci_gart_disable(rdev); + radeon_gart_table_ram_free(rdev); + radeon_gart_fini(rdev); +} + + +/* + * Fence emission + */ +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) */ + /* Wait until IDLE & CLEAN */ + radeon_ring_write(rdev, PACKET0(0x1720, 0)); + radeon_ring_write(rdev, (1 << 16) | (1 << 17)); + /* Emit fence sequence & fire IRQ */ + radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0)); + radeon_ring_write(rdev, fence->seq); + radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0)); + radeon_ring_write(rdev, RADEON_SW_INT_FIRE); +} + + +/* + * Writeback + */ +int r100_wb_init(struct radeon_device *rdev) +{ + int r; + + if (rdev->wb.wb_obj == NULL) { + r = radeon_object_create(rdev, NULL, 4096, + true, + RADEON_GEM_DOMAIN_GTT, + false, &rdev->wb.wb_obj); + if (r) { + DRM_ERROR("radeon: failed to create WB buffer (%d).\n", r); + return r; + } + r = radeon_object_pin(rdev->wb.wb_obj, + RADEON_GEM_DOMAIN_GTT, + &rdev->wb.gpu_addr); + if (r) { + DRM_ERROR("radeon: failed to pin WB buffer (%d).\n", r); + return r; + } + r = radeon_object_kmap(rdev->wb.wb_obj, (void **)&rdev->wb.wb); + if (r) { + DRM_ERROR("radeon: failed to map WB buffer (%d).\n", r); + return r; + } + } + WREG32(0x774, rdev->wb.gpu_addr); + WREG32(0x70C, rdev->wb.gpu_addr + 1024); + WREG32(0x770, 0xff); + return 0; +} + +void r100_wb_fini(struct radeon_device *rdev) +{ + if (rdev->wb.wb_obj) { + radeon_object_kunmap(rdev->wb.wb_obj); + radeon_object_unpin(rdev->wb.wb_obj); + radeon_object_unref(&rdev->wb.wb_obj); + rdev->wb.wb = NULL; + rdev->wb.wb_obj = NULL; + } +} + +int r100_copy_blit(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_pages, + struct radeon_fence *fence) +{ + uint32_t cur_pages; + uint32_t stride_bytes = PAGE_SIZE; + uint32_t pitch; + uint32_t stride_pixels; + unsigned ndw; + int num_loops; + int r = 0; + + /* radeon limited to 16k stride */ + stride_bytes &= 0x3fff; + /* radeon pitch is /64 */ + pitch = stride_bytes / 64; + stride_pixels = stride_bytes / 4; + num_loops = DIV_ROUND_UP(num_pages, 8191); + + /* Ask for enough room for blit + flush + fence */ + ndw = 64 + (10 * num_loops); + r = radeon_ring_lock(rdev, ndw); + if (r) { + DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw); + return -EINVAL; + } + while (num_pages > 0) { + cur_pages = num_pages; + if (cur_pages > 8191) { + cur_pages = 8191; + } + num_pages -= cur_pages; + + /* pages are in Y direction - height + page width in X direction - width */ + radeon_ring_write(rdev, PACKET3(PACKET3_BITBLT_MULTI, 8)); + radeon_ring_write(rdev, + RADEON_GMC_SRC_PITCH_OFFSET_CNTL | + RADEON_GMC_DST_PITCH_OFFSET_CNTL | + RADEON_GMC_SRC_CLIPPING | + RADEON_GMC_DST_CLIPPING | + RADEON_GMC_BRUSH_NONE | + (RADEON_COLOR_FORMAT_ARGB8888 << 8) | + RADEON_GMC_SRC_DATATYPE_COLOR | + RADEON_ROP3_S | + RADEON_DP_SRC_SOURCE_MEMORY | + RADEON_GMC_CLR_CMP_CNTL_DIS | + RADEON_GMC_WR_MSK_DIS); + radeon_ring_write(rdev, (pitch << 22) | (src_offset >> 10)); + radeon_ring_write(rdev, (pitch << 22) | (dst_offset >> 10)); + radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(rdev, 0); + radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(rdev, num_pages); + radeon_ring_write(rdev, num_pages); + radeon_ring_write(rdev, cur_pages | (stride_pixels << 16)); + } + radeon_ring_write(rdev, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(rdev, RADEON_RB2D_DC_FLUSH_ALL); + radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(rdev, + RADEON_WAIT_2D_IDLECLEAN | + RADEON_WAIT_HOST_IDLECLEAN | + RADEON_WAIT_DMA_GUI_IDLE); + if (fence) { + r = radeon_fence_emit(rdev, fence); + } + radeon_ring_unlock_commit(rdev); + return r; +} + + +/* + * CP + */ +void r100_ring_start(struct radeon_device *rdev) +{ + int r; + + r = radeon_ring_lock(rdev, 2); + if (r) { + return; + } + radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0)); + radeon_ring_write(rdev, + RADEON_ISYNC_ANY2D_IDLE3D | + RADEON_ISYNC_ANY3D_IDLE2D | + RADEON_ISYNC_WAIT_IDLEGUI | + RADEON_ISYNC_CPSCRATCH_IDLEGUI); + radeon_ring_unlock_commit(rdev); +} + +static void r100_cp_load_microcode(struct radeon_device *rdev) +{ + int i; + + if (r100_gui_wait_for_idle(rdev)) { + printk(KERN_WARNING "Failed to wait GUI idle while " + "programming pipes. Bad things might happen.\n"); + } + + WREG32(RADEON_CP_ME_RAM_ADDR, 0); + if ((rdev->family == CHIP_R100) || (rdev->family == CHIP_RV100) || + (rdev->family == CHIP_RV200) || (rdev->family == CHIP_RS100) || + (rdev->family == CHIP_RS200)) { + DRM_INFO("Loading R100 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, R100_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, R100_cp_microcode[i][0]); + } + } else if ((rdev->family == CHIP_R200) || + (rdev->family == CHIP_RV250) || + (rdev->family == CHIP_RV280) || + (rdev->family == CHIP_RS300)) { + DRM_INFO("Loading R200 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, R200_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, R200_cp_microcode[i][0]); + } + } else if ((rdev->family == CHIP_R300) || + (rdev->family == CHIP_R350) || + (rdev->family == CHIP_RV350) || + (rdev->family == CHIP_RV380) || + (rdev->family == CHIP_RS400) || + (rdev->family == CHIP_RS480)) { + DRM_INFO("Loading R300 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, R300_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, R300_cp_microcode[i][0]); + } + } else if ((rdev->family == CHIP_R420) || + (rdev->family == CHIP_R423) || + (rdev->family == CHIP_RV410)) { + DRM_INFO("Loading R400 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, R420_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, R420_cp_microcode[i][0]); + } + } else if ((rdev->family == CHIP_RS690) || + (rdev->family == CHIP_RS740)) { + DRM_INFO("Loading RS690/RS740 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, RS690_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, RS690_cp_microcode[i][0]); + } + } else if (rdev->family == CHIP_RS600) { + DRM_INFO("Loading RS600 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, RS600_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, RS600_cp_microcode[i][0]); + } + } else if ((rdev->family == CHIP_RV515) || + (rdev->family == CHIP_R520) || + (rdev->family == CHIP_RV530) || + (rdev->family == CHIP_R580) || + (rdev->family == CHIP_RV560) || + (rdev->family == CHIP_RV570)) { + DRM_INFO("Loading R500 Microcode\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_ME_RAM_DATAH, R520_cp_microcode[i][1]); + WREG32(RADEON_CP_ME_RAM_DATAL, R520_cp_microcode[i][0]); + } + } +} + +int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) +{ + unsigned rb_bufsz; + unsigned rb_blksz; + unsigned max_fetch; + unsigned pre_write_timer; + unsigned pre_write_limit; + unsigned indirect2_start; + unsigned indirect1_start; + uint32_t tmp; + int r; + + if (r100_debugfs_cp_init(rdev)) { + DRM_ERROR("Failed to register debugfs file for CP !\n"); + } + /* Reset CP */ + tmp = RREG32(RADEON_CP_CSQ_STAT); + if ((tmp & (1 << 31))) { + DRM_INFO("radeon: cp busy (0x%08X) resetting\n", tmp); + WREG32(RADEON_CP_CSQ_MODE, 0); + WREG32(RADEON_CP_CSQ_CNTL, 0); + WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_CP); + tmp = RREG32(RADEON_RBBM_SOFT_RESET); + mdelay(2); + WREG32(RADEON_RBBM_SOFT_RESET, 0); + tmp = RREG32(RADEON_RBBM_SOFT_RESET); + mdelay(2); + tmp = RREG32(RADEON_CP_CSQ_STAT); + if ((tmp & (1 << 31))) { + DRM_INFO("radeon: cp reset failed (0x%08X)\n", tmp); + } + } else { + DRM_INFO("radeon: cp idle (0x%08X)\n", tmp); + } + /* Align ring size */ + rb_bufsz = drm_order(ring_size / 8); + ring_size = (1 << (rb_bufsz + 1)) * 4; + r100_cp_load_microcode(rdev); + r = radeon_ring_init(rdev, ring_size); + if (r) { + return r; + } + /* Each time the cp read 1024 bytes (16 dword/quadword) update + * the rptr copy in system ram */ + rb_blksz = 9; + /* cp will read 128bytes at a time (4 dwords) */ + max_fetch = 1; + rdev->cp.align_mask = 16 - 1; + /* Write to CP_RB_WPTR will be delayed for pre_write_timer clocks */ + pre_write_timer = 64; + /* Force CP_RB_WPTR write if written more than one time before the + * delay expire + */ + pre_write_limit = 0; + /* Setup the cp cache like this (cache size is 96 dwords) : + * RING 0 to 15 + * INDIRECT1 16 to 79 + * INDIRECT2 80 to 95 + * So ring cache size is 16dwords (> (2 * max_fetch = 2 * 4dwords)) + * indirect1 cache size is 64dwords (> (2 * max_fetch = 2 * 4dwords)) + * indirect2 cache size is 16dwords (> (2 * max_fetch = 2 * 4dwords)) + * Idea being that most of the gpu cmd will be through indirect1 buffer + * so it gets the bigger cache. + */ + indirect2_start = 80; + indirect1_start = 16; + /* cp setup */ + WREG32(0x718, pre_write_timer | (pre_write_limit << 28)); + WREG32(RADEON_CP_RB_CNTL, + REG_SET(RADEON_RB_BUFSZ, rb_bufsz) | + REG_SET(RADEON_RB_BLKSZ, rb_blksz) | + REG_SET(RADEON_MAX_FETCH, max_fetch) | + RADEON_RB_NO_UPDATE); + /* Set ring address */ + DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)rdev->cp.gpu_addr); + WREG32(RADEON_CP_RB_BASE, rdev->cp.gpu_addr); + /* Force read & write ptr to 0 */ + tmp = RREG32(RADEON_CP_RB_CNTL); + WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA); + WREG32(RADEON_CP_RB_RPTR_WR, 0); + WREG32(RADEON_CP_RB_WPTR, 0); + WREG32(RADEON_CP_RB_CNTL, tmp); + udelay(10); + rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); + rdev->cp.wptr = RREG32(RADEON_CP_RB_WPTR); + /* Set cp mode to bus mastering & enable cp*/ + WREG32(RADEON_CP_CSQ_MODE, + REG_SET(RADEON_INDIRECT2_START, indirect2_start) | + REG_SET(RADEON_INDIRECT1_START, indirect1_start)); + WREG32(0x718, 0); + WREG32(0x744, 0x00004D4D); + WREG32(RADEON_CP_CSQ_CNTL, RADEON_CSQ_PRIBM_INDBM); + radeon_ring_start(rdev); + r = radeon_ring_test(rdev); + if (r) { + DRM_ERROR("radeon: cp isn't working (%d).\n", r); + return r; + } + rdev->cp.ready = true; + return 0; +} + +void r100_cp_fini(struct radeon_device *rdev) +{ + /* Disable ring */ + rdev->cp.ready = false; + WREG32(RADEON_CP_CSQ_CNTL, 0); + radeon_ring_fini(rdev); + DRM_INFO("radeon: cp finalized\n"); +} + +void r100_cp_disable(struct radeon_device *rdev) +{ + /* Disable ring */ + rdev->cp.ready = false; + WREG32(RADEON_CP_CSQ_MODE, 0); + WREG32(RADEON_CP_CSQ_CNTL, 0); + if (r100_gui_wait_for_idle(rdev)) { + printk(KERN_WARNING "Failed to wait GUI idle while " + "programming pipes. Bad things might happen.\n"); + } +} + +int r100_cp_reset(struct radeon_device *rdev) +{ + uint32_t tmp; + bool reinit_cp; + int i; + + reinit_cp = rdev->cp.ready; + rdev->cp.ready = false; + WREG32(RADEON_CP_CSQ_MODE, 0); + WREG32(RADEON_CP_CSQ_CNTL, 0); + WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_CP); + (void)RREG32(RADEON_RBBM_SOFT_RESET); + udelay(200); + WREG32(RADEON_RBBM_SOFT_RESET, 0); + /* Wait to prevent race in RBBM_STATUS */ + mdelay(1); + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_RBBM_STATUS); + if (!(tmp & (1 << 16))) { + DRM_INFO("CP reset succeed (RBBM_STATUS=0x%08X)\n", + tmp); + if (reinit_cp) { + return r100_cp_init(rdev, rdev->cp.ring_size); + } + return 0; + } + DRM_UDELAY(1); + } + tmp = RREG32(RADEON_RBBM_STATUS); + DRM_ERROR("Failed to reset CP (RBBM_STATUS=0x%08X)!\n", tmp); + return -1; +} + + +/* + * CS functions + */ +int r100_cs_parse_packet0(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned *auth, unsigned n, + radeon_packet0_check_t check) +{ + unsigned reg; + unsigned i, j, m; + unsigned idx; + int r; + + idx = pkt->idx + 1; + reg = pkt->reg; + if (pkt->one_reg_wr) { + if ((reg >> 7) > n) { + return -EINVAL; + } + } else { + if (((reg + (pkt->count << 2)) >> 7) > n) { + return -EINVAL; + } + } + for (i = 0; i <= pkt->count; i++, idx++) { + j = (reg >> 7); + m = 1 << ((reg >> 2) & 31); + if (auth[j] & m) { + r = check(p, pkt, idx, reg); + if (r) { + return r; + } + } + if (pkt->one_reg_wr) { + if (!(auth[j] & m)) { + break; + } + } else { + reg += 4; + } + } + return 0; +} + +int r100_cs_parse_packet3(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned *auth, unsigned n, + radeon_packet3_check_t check) +{ + unsigned i, m; + + if ((pkt->opcode >> 5) > n) { + return -EINVAL; + } + i = pkt->opcode >> 5; + m = 1 << (pkt->opcode & 31); + if (auth[i] & m) { + return check(p, pkt); + } + return 0; +} + +void r100_cs_dump_packet(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt) +{ + struct radeon_cs_chunk *ib_chunk; + volatile uint32_t *ib; + unsigned i; + unsigned idx; + + ib = p->ib->ptr; + ib_chunk = &p->chunks[p->chunk_ib_idx]; + idx = pkt->idx; + for (i = 0; i <= (pkt->count + 1); i++, idx++) { + DRM_INFO("ib[%d]=0x%08X\n", idx, ib[idx]); + } +} + +/** + * r100_cs_packet_parse() - parse cp packet and point ib index to next packet + * @parser: parser structure holding parsing context. + * @pkt: where to store packet informations + * + * Assume that chunk_ib_index is properly set. Will return -EINVAL + * if packet is bigger than remaining ib size. or if packets is unknown. + **/ +int r100_cs_packet_parse(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + unsigned idx) +{ + struct radeon_cs_chunk *ib_chunk = &p->chunks[p->chunk_ib_idx]; + uint32_t header = ib_chunk->kdata[idx]; + + if (idx >= ib_chunk->length_dw) { + DRM_ERROR("Can not parse packet at %d after CS end %d !\n", + idx, ib_chunk->length_dw); + return -EINVAL; + } + pkt->idx = idx; + pkt->type = CP_PACKET_GET_TYPE(header); + pkt->count = CP_PACKET_GET_COUNT(header); + switch (pkt->type) { + case PACKET_TYPE0: + pkt->reg = CP_PACKET0_GET_REG(header); + pkt->one_reg_wr = CP_PACKET0_GET_ONE_REG_WR(header); + break; + case PACKET_TYPE3: + pkt->opcode = CP_PACKET3_GET_OPCODE(header); + break; + case PACKET_TYPE2: + pkt->count = -1; + break; + default: + DRM_ERROR("Unknown packet type %d at %d !\n", pkt->type, idx); + return -EINVAL; + } + if ((pkt->count + 1 + pkt->idx) >= ib_chunk->length_dw) { + DRM_ERROR("Packet (%d:%d:%d) end after CS buffer (%d) !\n", + pkt->idx, pkt->type, pkt->count, ib_chunk->length_dw); + return -EINVAL; + } + return 0; +} + +/** + * r100_cs_packet_next_reloc() - parse next packet which should be reloc packet3 + * @parser: parser structure holding parsing context. + * @data: pointer to relocation data + * @offset_start: starting offset + * @offset_mask: offset mask (to align start offset on) + * @reloc: reloc informations + * + * Check next packet is relocation packet3, do bo validation and compute + * GPU offset using the provided start. + **/ +int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, + struct radeon_cs_reloc **cs_reloc) +{ + struct radeon_cs_chunk *ib_chunk; + struct radeon_cs_chunk *relocs_chunk; + struct radeon_cs_packet p3reloc; + unsigned idx; + int r; + + if (p->chunk_relocs_idx == -1) { + DRM_ERROR("No relocation chunk !\n"); + return -EINVAL; + } + *cs_reloc = NULL; + ib_chunk = &p->chunks[p->chunk_ib_idx]; + relocs_chunk = &p->chunks[p->chunk_relocs_idx]; + r = r100_cs_packet_parse(p, &p3reloc, p->idx); + if (r) { + return r; + } + p->idx += p3reloc.count + 2; + if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) { + DRM_ERROR("No packet3 for relocation for packet at %d.\n", + p3reloc.idx); + r100_cs_dump_packet(p, &p3reloc); + return -EINVAL; + } + idx = ib_chunk->kdata[p3reloc.idx + 1]; + if (idx >= relocs_chunk->length_dw) { + DRM_ERROR("Relocs at %d after relocations chunk end %d !\n", + idx, relocs_chunk->length_dw); + r100_cs_dump_packet(p, &p3reloc); + return -EINVAL; + } + /* FIXME: we assume reloc size is 4 dwords */ + *cs_reloc = p->relocs_ptr[(idx / 4)]; + return 0; +} + +static int r100_packet0_check(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt) +{ + struct radeon_cs_chunk *ib_chunk; + struct radeon_cs_reloc *reloc; + volatile uint32_t *ib; + uint32_t tmp; + unsigned reg; + unsigned i; + unsigned idx; + bool onereg; + int r; + + ib = p->ib->ptr; + ib_chunk = &p->chunks[p->chunk_ib_idx]; + idx = pkt->idx + 1; + reg = pkt->reg; + onereg = false; + if (CP_PACKET0_GET_ONE_REG_WR(ib_chunk->kdata[pkt->idx])) { + onereg = true; + } + for (i = 0; i <= pkt->count; i++, idx++, reg += 4) { + switch (reg) { + /* FIXME: only allow PACKET3 blit? easier to check for out of + * range access */ + case RADEON_DST_PITCH_OFFSET: + case RADEON_SRC_PITCH_OFFSET: + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + r100_cs_dump_packet(p, pkt); + return r; + } + tmp = ib_chunk->kdata[idx] & 0x003fffff; + tmp += (((u32)reloc->lobj.gpu_offset) >> 10); + ib[idx] = (ib_chunk->kdata[idx] & 0xffc00000) | tmp; + break; + case RADEON_RB3D_DEPTHOFFSET: + case RADEON_RB3D_COLOROFFSET: + case R300_RB3D_COLOROFFSET0: + case R300_ZB_DEPTHOFFSET: + case R200_PP_TXOFFSET_0: + case R200_PP_TXOFFSET_1: + case R200_PP_TXOFFSET_2: + case R200_PP_TXOFFSET_3: + case R200_PP_TXOFFSET_4: + case R200_PP_TXOFFSET_5: + case RADEON_PP_TXOFFSET_0: + case RADEON_PP_TXOFFSET_1: + case RADEON_PP_TXOFFSET_2: + case R300_TX_OFFSET_0: + case R300_TX_OFFSET_0+4: + case R300_TX_OFFSET_0+8: + case R300_TX_OFFSET_0+12: + case R300_TX_OFFSET_0+16: + case R300_TX_OFFSET_0+20: + case R300_TX_OFFSET_0+24: + case R300_TX_OFFSET_0+28: + case R300_TX_OFFSET_0+32: + case R300_TX_OFFSET_0+36: + case R300_TX_OFFSET_0+40: + case R300_TX_OFFSET_0+44: + case R300_TX_OFFSET_0+48: + case R300_TX_OFFSET_0+52: + case R300_TX_OFFSET_0+56: + case R300_TX_OFFSET_0+60: + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + break; + default: + /* FIXME: we don't want to allow anyothers packet */ + break; + } + if (onereg) { + /* FIXME: forbid onereg write to register on relocate */ + break; + } + } + return 0; +} + +static int r100_packet3_check(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt) +{ + struct radeon_cs_chunk *ib_chunk; + struct radeon_cs_reloc *reloc; + unsigned idx; + unsigned i, c; + volatile uint32_t *ib; + int r; + + ib = p->ib->ptr; + ib_chunk = &p->chunks[p->chunk_ib_idx]; + idx = pkt->idx + 1; + switch (pkt->opcode) { + case PACKET3_3D_LOAD_VBPNTR: + c = ib_chunk->kdata[idx++]; + for (i = 0; i < (c - 1); i += 2, idx += 3) { + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx+2] = ib_chunk->kdata[idx+2] + ((u32)reloc->lobj.gpu_offset); + } + if (c & 1) { + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", + pkt->opcode); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + } + break; + case PACKET3_INDX_BUFFER: + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + break; + case 0x23: + /* FIXME: cleanup */ + /* 3D_RNDR_GEN_INDX_PRIM on r100/r200 */ + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for packet3 %d\n", pkt->opcode); + r100_cs_dump_packet(p, pkt); + return r; + } + ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + break; + case PACKET3_3D_DRAW_IMMD: + /* triggers drawing using in-packet vertex data */ + case PACKET3_3D_DRAW_IMMD_2: + /* triggers drawing using in-packet vertex data */ + case PACKET3_3D_DRAW_VBUF_2: + /* triggers drawing of vertex buffers setup elsewhere */ + case PACKET3_3D_DRAW_INDX_2: + /* triggers drawing using indices to vertex buffer */ + case PACKET3_3D_DRAW_VBUF: + /* triggers drawing of vertex buffers setup elsewhere */ + case PACKET3_3D_DRAW_INDX: + /* triggers drawing using indices to vertex buffer */ + case PACKET3_NOP: + break; + default: + DRM_ERROR("Packet3 opcode %x not supported\n", pkt->opcode); + return -EINVAL; + } + return 0; +} + +int r100_cs_parse(struct radeon_cs_parser *p) +{ + struct radeon_cs_packet pkt; + int r; + + do { + r = r100_cs_packet_parse(p, &pkt, p->idx); + if (r) { + return r; + } + p->idx += pkt.count + 2; + switch (pkt.type) { + case PACKET_TYPE0: + r = r100_packet0_check(p, &pkt); + break; + case PACKET_TYPE2: + break; + case PACKET_TYPE3: + r = r100_packet3_check(p, &pkt); + break; + default: + DRM_ERROR("Unknown packet type %d !\n", + pkt.type); + return -EINVAL; + } + if (r) { + return r; + } + } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); + return 0; +} + + +/* + * Global GPU functions + */ +void r100_errata(struct radeon_device *rdev) +{ + rdev->pll_errata = 0; + + if (rdev->family == CHIP_RV200 || rdev->family == CHIP_RS200) { + rdev->pll_errata |= CHIP_ERRATA_PLL_DUMMYREADS; + } + + if (rdev->family == CHIP_RV100 || + rdev->family == CHIP_RS100 || + rdev->family == CHIP_RS200) { + rdev->pll_errata |= CHIP_ERRATA_PLL_DELAY; + } +} + +/* Wait for vertical sync on primary CRTC */ +void r100_gpu_wait_for_vsync(struct radeon_device *rdev) +{ + uint32_t crtc_gen_cntl, tmp; + int i; + + crtc_gen_cntl = RREG32(RADEON_CRTC_GEN_CNTL); + if ((crtc_gen_cntl & RADEON_CRTC_DISP_REQ_EN_B) || + !(crtc_gen_cntl & RADEON_CRTC_EN)) { + return; + } + /* Clear the CRTC_VBLANK_SAVE bit */ + WREG32(RADEON_CRTC_STATUS, RADEON_CRTC_VBLANK_SAVE_CLEAR); + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_CRTC_STATUS); + if (tmp & RADEON_CRTC_VBLANK_SAVE) { + return; + } + DRM_UDELAY(1); + } +} + +/* Wait for vertical sync on secondary CRTC */ +void r100_gpu_wait_for_vsync2(struct radeon_device *rdev) +{ + uint32_t crtc2_gen_cntl, tmp; + int i; + + crtc2_gen_cntl = RREG32(RADEON_CRTC2_GEN_CNTL); + if ((crtc2_gen_cntl & RADEON_CRTC2_DISP_REQ_EN_B) || + !(crtc2_gen_cntl & RADEON_CRTC2_EN)) + return; + + /* Clear the CRTC_VBLANK_SAVE bit */ + WREG32(RADEON_CRTC2_STATUS, RADEON_CRTC2_VBLANK_SAVE_CLEAR); + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_CRTC2_STATUS); + if (tmp & RADEON_CRTC2_VBLANK_SAVE) { + return; + } + DRM_UDELAY(1); + } +} + +int r100_rbbm_fifo_wait_for_entry(struct radeon_device *rdev, unsigned n) +{ + unsigned i; + uint32_t tmp; + + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_RBBM_STATUS) & RADEON_RBBM_FIFOCNT_MASK; + if (tmp >= n) { + return 0; + } + DRM_UDELAY(1); + } + return -1; +} + +int r100_gui_wait_for_idle(struct radeon_device *rdev) +{ + unsigned i; + uint32_t tmp; + + if (r100_rbbm_fifo_wait_for_entry(rdev, 64)) { + printk(KERN_WARNING "radeon: wait for empty RBBM fifo failed !" + " Bad things might happen.\n"); + } + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_RBBM_STATUS); + if (!(tmp & (1 << 31))) { + return 0; + } + DRM_UDELAY(1); + } + return -1; +} + +int r100_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 r100_gpu_init(struct radeon_device *rdev) +{ + /* TODO: anythings to do here ? pipes ? */ + r100_hdp_reset(rdev); +} + +void r100_hdp_reset(struct radeon_device *rdev) +{ + uint32_t tmp; + + tmp = RREG32(RADEON_HOST_PATH_CNTL) & RADEON_HDP_APER_CNTL; + tmp |= (7 << 28); + WREG32(RADEON_HOST_PATH_CNTL, tmp | RADEON_HDP_SOFT_RESET | RADEON_HDP_READ_BUFFER_INVALIDATE); + (void)RREG32(RADEON_HOST_PATH_CNTL); + udelay(200); + WREG32(RADEON_RBBM_SOFT_RESET, 0); + WREG32(RADEON_HOST_PATH_CNTL, tmp); + (void)RREG32(RADEON_HOST_PATH_CNTL); +} + +int r100_rb2d_reset(struct radeon_device *rdev) +{ + uint32_t tmp; + int i; + + WREG32(RADEON_RBBM_SOFT_RESET, RADEON_SOFT_RESET_E2); + (void)RREG32(RADEON_RBBM_SOFT_RESET); + udelay(200); + WREG32(RADEON_RBBM_SOFT_RESET, 0); + /* Wait to prevent race in RBBM_STATUS */ + mdelay(1); + for (i = 0; i < rdev->usec_timeout; i++) { + tmp = RREG32(RADEON_RBBM_STATUS); + if (!(tmp & (1 << 26))) { + DRM_INFO("RB2D reset succeed (RBBM_STATUS=0x%08X)\n", + tmp); + return 0; + } + DRM_UDELAY(1); + } + tmp = RREG32(RADEON_RBBM_STATUS); + DRM_ERROR("Failed to reset RB2D (RBBM_STATUS=0x%08X)!\n", tmp); + return -1; +} + +int r100_gpu_reset(struct radeon_device *rdev) +{ + uint32_t status; + + /* reset order likely matter */ + status = RREG32(RADEON_RBBM_STATUS); + /* reset HDP */ + r100_hdp_reset(rdev); + /* reset rb2d */ + if (status & ((1 << 17) | (1 << 18) | (1 << 27))) { + r100_rb2d_reset(rdev); + } + /* TODO: reset 3D engine */ + /* reset CP */ + status = RREG32(RADEON_RBBM_STATUS); + if (status & (1 << 16)) { + r100_cp_reset(rdev); + } + /* Check if GPU is idle */ + status = RREG32(RADEON_RBBM_STATUS); + if (status & (1 << 31)) { + DRM_ERROR("Failed to reset GPU (RBBM_STATUS=0x%08X)\n", status); + return -1; + } + DRM_INFO("GPU reset succeed (RBBM_STATUS=0x%08X)\n", status); + return 0; +} + + +/* + * VRAM info + */ +static void r100_vram_get_type(struct radeon_device *rdev) +{ + uint32_t tmp; + + rdev->mc.vram_is_ddr = false; + if (rdev->flags & RADEON_IS_IGP) + rdev->mc.vram_is_ddr = true; + else if (RREG32(RADEON_MEM_SDRAM_MODE_REG) & RADEON_MEM_CFG_TYPE_DDR) + rdev->mc.vram_is_ddr = true; + if ((rdev->family == CHIP_RV100) || + (rdev->family == CHIP_RS100) || + (rdev->family == CHIP_RS200)) { + tmp = RREG32(RADEON_MEM_CNTL); + if (tmp & RV100_HALF_MODE) { + rdev->mc.vram_width = 32; + } else { + rdev->mc.vram_width = 64; + } + if (rdev->flags & RADEON_SINGLE_CRTC) { + rdev->mc.vram_width /= 4; + rdev->mc.vram_is_ddr = true; + } + } else if (rdev->family <= CHIP_RV280) { + tmp = RREG32(RADEON_MEM_CNTL); + if (tmp & RADEON_MEM_NUM_CHANNELS_MASK) { + rdev->mc.vram_width = 128; + } else { + rdev->mc.vram_width = 64; + } + } else { + /* newer IGPs */ + rdev->mc.vram_width = 128; + } +} + +void r100_vram_info(struct radeon_device *rdev) +{ + r100_vram_get_type(rdev); + + 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.vram_size = (((tom >> 16) - (tom & 0xffff) + 1) << 16); + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); + } else { + rdev->mc.vram_size = RREG32(RADEON_CONFIG_MEMSIZE); + /* Some production boards of m6 will report 0 + * if it's 8 MB + */ + if (rdev->mc.vram_size == 0) { + rdev->mc.vram_size = 8192 * 1024; + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.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.aper_size > rdev->mc.vram_size) { + /* Why does some hw doesn't have CONFIG_MEMSIZE properly + * setup ? */ + rdev->mc.vram_size = rdev->mc.aper_size; + WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); + } +} + + +/* + * Indirect registers accessor + */ +void r100_pll_errata_after_index(struct radeon_device *rdev) +{ + if (!(rdev->pll_errata & CHIP_ERRATA_PLL_DUMMYREADS)) { + return; + } + (void)RREG32(RADEON_CLOCK_CNTL_DATA); + (void)RREG32(RADEON_CRTC_GEN_CNTL); +} + +static void r100_pll_errata_after_data(struct radeon_device *rdev) +{ + /* This workarounds is necessary on RV100, RS100 and RS200 chips + * or the chip could hang on a subsequent access + */ + if (rdev->pll_errata & CHIP_ERRATA_PLL_DELAY) { + udelay(5000); + } + + /* This function is required to workaround a hardware bug in some (all?) + * revisions of the R300. This workaround should be called after every + * CLOCK_CNTL_INDEX register access. If not, register reads afterward + * may not be correct. + */ + if (rdev->pll_errata & CHIP_ERRATA_R300_CG) { + uint32_t save, tmp; + + save = RREG32(RADEON_CLOCK_CNTL_INDEX); + tmp = save & ~(0x3f | RADEON_PLL_WR_EN); + WREG32(RADEON_CLOCK_CNTL_INDEX, tmp); + tmp = RREG32(RADEON_CLOCK_CNTL_DATA); + WREG32(RADEON_CLOCK_CNTL_INDEX, save); + } +} + +uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg) +{ + uint32_t data; + + WREG8(RADEON_CLOCK_CNTL_INDEX, reg & 0x3f); + r100_pll_errata_after_index(rdev); + data = RREG32(RADEON_CLOCK_CNTL_DATA); + r100_pll_errata_after_data(rdev); + return data; +} + +void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) +{ + WREG8(RADEON_CLOCK_CNTL_INDEX, ((reg & 0x3f) | RADEON_PLL_WR_EN)); + r100_pll_errata_after_index(rdev); + WREG32(RADEON_CLOCK_CNTL_DATA, v); + r100_pll_errata_after_data(rdev); +} + +uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg) +{ + if (reg < 0x10000) + return readl(((void __iomem *)rdev->rmmio) + reg); + else { + writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); + return readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA); + } +} + +void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) +{ + if (reg < 0x10000) + writel(v, ((void __iomem *)rdev->rmmio) + reg); + else { + writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX); + writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA); + } +} + +/* + * Debugfs info + */ +#if defined(CONFIG_DEBUG_FS) +static int r100_debugfs_rbbm_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + uint32_t reg, value; + unsigned i; + + seq_printf(m, "RBBM_STATUS 0x%08x\n", RREG32(RADEON_RBBM_STATUS)); + seq_printf(m, "RBBM_CMDFIFO_STAT 0x%08x\n", RREG32(0xE7C)); + seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT)); + for (i = 0; i < 64; i++) { + WREG32(RADEON_RBBM_CMDFIFO_ADDR, i | 0x100); + reg = (RREG32(RADEON_RBBM_CMDFIFO_DATA) - 1) >> 2; + WREG32(RADEON_RBBM_CMDFIFO_ADDR, i); + value = RREG32(RADEON_RBBM_CMDFIFO_DATA); + seq_printf(m, "[0x%03X] 0x%04X=0x%08X\n", i, reg, value); + } + return 0; +} + +static int r100_debugfs_cp_ring_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + uint32_t rdp, wdp; + unsigned count, i, j; + + radeon_ring_free_size(rdev); + rdp = RREG32(RADEON_CP_RB_RPTR); + wdp = RREG32(RADEON_CP_RB_WPTR); + count = (rdp + rdev->cp.ring_size - wdp) & rdev->cp.ptr_mask; + seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT)); + seq_printf(m, "CP_RB_WPTR 0x%08x\n", wdp); + seq_printf(m, "CP_RB_RPTR 0x%08x\n", rdp); + seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw); + seq_printf(m, "%u dwords in ring\n", count); + for (j = 0; j <= count; j++) { + i = (rdp + j) & rdev->cp.ptr_mask; + seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.ring[i]); + } + return 0; +} + + +static int r100_debugfs_cp_csq_fifo(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + uint32_t csq_stat, csq2_stat, tmp; + unsigned r_rptr, r_wptr, ib1_rptr, ib1_wptr, ib2_rptr, ib2_wptr; + unsigned i; + + seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT)); + seq_printf(m, "CP_CSQ_MODE 0x%08x\n", RREG32(RADEON_CP_CSQ_MODE)); + csq_stat = RREG32(RADEON_CP_CSQ_STAT); + csq2_stat = RREG32(RADEON_CP_CSQ2_STAT); + r_rptr = (csq_stat >> 0) & 0x3ff; + r_wptr = (csq_stat >> 10) & 0x3ff; + ib1_rptr = (csq_stat >> 20) & 0x3ff; + ib1_wptr = (csq2_stat >> 0) & 0x3ff; + ib2_rptr = (csq2_stat >> 10) & 0x3ff; + ib2_wptr = (csq2_stat >> 20) & 0x3ff; + seq_printf(m, "CP_CSQ_STAT 0x%08x\n", csq_stat); + seq_printf(m, "CP_CSQ2_STAT 0x%08x\n", csq2_stat); + seq_printf(m, "Ring rptr %u\n", r_rptr); + seq_printf(m, "Ring wptr %u\n", r_wptr); + seq_printf(m, "Indirect1 rptr %u\n", ib1_rptr); + seq_printf(m, "Indirect1 wptr %u\n", ib1_wptr); + seq_printf(m, "Indirect2 rptr %u\n", ib2_rptr); + seq_printf(m, "Indirect2 wptr %u\n", ib2_wptr); + /* FIXME: 0, 128, 640 depends on fifo setup see cp_init_kms + * 128 = indirect1_start * 8 & 640 = indirect2_start * 8 */ + seq_printf(m, "Ring fifo:\n"); + for (i = 0; i < 256; i++) { + WREG32(RADEON_CP_CSQ_ADDR, i << 2); + tmp = RREG32(RADEON_CP_CSQ_DATA); + seq_printf(m, "rfifo[%04d]=0x%08X\n", i, tmp); + } + seq_printf(m, "Indirect1 fifo:\n"); + for (i = 256; i <= 512; i++) { + WREG32(RADEON_CP_CSQ_ADDR, i << 2); + tmp = RREG32(RADEON_CP_CSQ_DATA); + seq_printf(m, "ib1fifo[%04d]=0x%08X\n", i, tmp); + } + seq_printf(m, "Indirect2 fifo:\n"); + for (i = 640; i < ib1_wptr; i++) { + WREG32(RADEON_CP_CSQ_ADDR, i << 2); + tmp = RREG32(RADEON_CP_CSQ_DATA); + seq_printf(m, "ib2fifo[%04d]=0x%08X\n", i, tmp); + } + return 0; +} + +static int r100_debugfs_mc_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + uint32_t tmp; + + tmp = RREG32(RADEON_CONFIG_MEMSIZE); + seq_printf(m, "CONFIG_MEMSIZE 0x%08x\n", tmp); + tmp = RREG32(RADEON_MC_FB_LOCATION); + seq_printf(m, "MC_FB_LOCATION 0x%08x\n", tmp); + tmp = RREG32(RADEON_BUS_CNTL); + seq_printf(m, "BUS_CNTL 0x%08x\n", tmp); + tmp = RREG32(RADEON_MC_AGP_LOCATION); + seq_printf(m, "MC_AGP_LOCATION 0x%08x\n", tmp); + tmp = RREG32(RADEON_AGP_BASE); + seq_printf(m, "AGP_BASE 0x%08x\n", tmp); + tmp = RREG32(RADEON_HOST_PATH_CNTL); + seq_printf(m, "HOST_PATH_CNTL 0x%08x\n", tmp); + tmp = RREG32(0x01D0); + seq_printf(m, "AIC_CTRL 0x%08x\n", tmp); + tmp = RREG32(RADEON_AIC_LO_ADDR); + seq_printf(m, "AIC_LO_ADDR 0x%08x\n", tmp); + tmp = RREG32(RADEON_AIC_HI_ADDR); + seq_printf(m, "AIC_HI_ADDR 0x%08x\n", tmp); + tmp = RREG32(0x01E4); + seq_printf(m, "AIC_TLB_ADDR 0x%08x\n", tmp); + return 0; +} + +static struct drm_info_list r100_debugfs_rbbm_list[] = { + {"r100_rbbm_info", r100_debugfs_rbbm_info, 0, NULL}, +}; + +static struct drm_info_list r100_debugfs_cp_list[] = { + {"r100_cp_ring_info", r100_debugfs_cp_ring_info, 0, NULL}, + {"r100_cp_csq_fifo", r100_debugfs_cp_csq_fifo, 0, NULL}, +}; + +static struct drm_info_list r100_debugfs_mc_info_list[] = { + {"r100_mc_info", r100_debugfs_mc_info, 0, NULL}, +}; +#endif + +int r100_debugfs_rbbm_init(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + return radeon_debugfs_add_files(rdev, r100_debugfs_rbbm_list, 1); +#else + return 0; +#endif +} + +int r100_debugfs_cp_init(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + return radeon_debugfs_add_files(rdev, r100_debugfs_cp_list, 2); +#else + return 0; +#endif +} + +int r100_debugfs_mc_info_init(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + return radeon_debugfs_add_files(rdev, r100_debugfs_mc_info_list, 1); +#else + return 0; +#endif +} -- cgit v1.2.3-70-g09d2 From 00fa28ae29f70c9f26023f9922c4d2e1ca1297e3 Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Thu, 18 Jun 2009 18:08:33 +1000 Subject: drm/radeon: this VRAM vs aperture test is wrong, just remove it. Its quite valid to have VRAM < aperture size. Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 6 ------ 1 file changed, 6 deletions(-) (limited to 'drivers/gpu/drm/radeon/r100.c') diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 5225f5be7ea..8f41f701f0b 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -1267,12 +1267,6 @@ void r100_vram_info(struct radeon_device *rdev) 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.aper_size > rdev->mc.vram_size) { - /* Why does some hw doesn't have CONFIG_MEMSIZE properly - * setup ? */ - rdev->mc.vram_size = rdev->mc.aper_size; - WREG32(RADEON_CONFIG_MEMSIZE, rdev->mc.vram_size); - } } -- cgit v1.2.3-70-g09d2 From 4e484e7dc5856ff5086b6329d82e36d4adaf1f02 Mon Sep 17 00:00:00 2001 From: Michel Dänzer Date: Tue, 16 Jun 2009 17:29:06 +0200 Subject: radeon: Fix CP byte order on big endian architectures with KMS. Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 3 +++ drivers/gpu/drm/radeon/radeon_reg.h | 1 + 2 files changed, 4 insertions(+) (limited to 'drivers/gpu/drm/radeon/r100.c') diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 8f41f701f0b..64a692c0c31 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -551,6 +551,9 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) /* cp setup */ WREG32(0x718, pre_write_timer | (pre_write_limit << 28)); WREG32(RADEON_CP_RB_CNTL, +#ifdef __BIG_ENDIAN + RADEON_BUF_SWAP_32BIT | +#endif REG_SET(RADEON_RB_BUFSZ, rb_bufsz) | REG_SET(RADEON_RB_BLKSZ, rb_blksz) | REG_SET(RADEON_MAX_FETCH, max_fetch) | diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h index 6d3d90406a2..e1b61857446 100644 --- a/drivers/gpu/drm/radeon/radeon_reg.h +++ b/drivers/gpu/drm/radeon/radeon_reg.h @@ -3184,6 +3184,7 @@ # define RADEON_RB_BUFSZ_MASK (0x3f << 0) # define RADEON_RB_BLKSZ_SHIFT 8 # define RADEON_RB_BLKSZ_MASK (0x3f << 8) +# define RADEON_BUF_SWAP_32BIT (1 << 17) # define RADEON_MAX_FETCH_SHIFT 18 # define RADEON_MAX_FETCH_MASK (0x3 << 18) # define RADEON_RB_NO_UPDATE (1 << 27) -- cgit v1.2.3-70-g09d2 From 068a117ca38f27c9641db7642f24fe9270d9424e Mon Sep 17 00:00:00 2001 From: Jerome Glisse Date: Wed, 17 Jun 2009 13:28:30 +0200 Subject: drm/radeon: command stream checker for r3xx-r5xx hardware For security purpose we want to make sure the userspace process doesn't access memory beyond buffer it owns. To achieve this we need to check states the userspace program. For color buffer and zbuffer we check that the clipping register will discard access beyond buffers set as color or zbuffer. For vertex buffer we check that no vertex fetch will happen beyond buffer end. For texture we check various texture states (number of mipmap level, texture size, texture depth, ...) to compute the amount of memory the texture fetcher might access. The command stream checking impact the performances so far quick benchmark shows an average of 3% decrease in fps of various applications. It can be optimized a bit more by caching result of checking and thus avoid a full recheck if no states changed since last check. Note that this patch is still incomplete on checking side as it doesn't check 2d rendering states. Signed-off-by: Jerome Glisse Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 76 +++--- drivers/gpu/drm/radeon/r300.c | 478 ++++++++++++++++++++++++++++++--- drivers/gpu/drm/radeon/r300.h | 36 +++ drivers/gpu/drm/radeon/radeon.h | 9 +- drivers/gpu/drm/radeon/radeon_asic.h | 15 +- drivers/gpu/drm/radeon/radeon_device.c | 4 + drivers/gpu/drm/radeon/rv515.c | 58 ++++ 7 files changed, 611 insertions(+), 65 deletions(-) create mode 100644 drivers/gpu/drm/radeon/r300.h (limited to 'drivers/gpu/drm/radeon/r100.c') diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 64a692c0c31..c550932a108 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -647,7 +647,7 @@ int r100_cp_reset(struct radeon_device *rdev) */ int r100_cs_parse_packet0(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, - unsigned *auth, unsigned n, + const unsigned *auth, unsigned n, radeon_packet0_check_t check) { unsigned reg; @@ -657,6 +657,10 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p, idx = pkt->idx + 1; reg = pkt->reg; + /* Check that register fall into register range + * determined by the number of entry (n) in the + * safe register bitmap. + */ if (pkt->one_reg_wr) { if ((reg >> 7) > n) { return -EINVAL; @@ -686,24 +690,6 @@ int r100_cs_parse_packet0(struct radeon_cs_parser *p, return 0; } -int r100_cs_parse_packet3(struct radeon_cs_parser *p, - struct radeon_cs_packet *pkt, - unsigned *auth, unsigned n, - radeon_packet3_check_t check) -{ - unsigned i, m; - - if ((pkt->opcode >> 5) > n) { - return -EINVAL; - } - i = pkt->opcode >> 5; - m = 1 << (pkt->opcode & 31); - if (auth[i] & m) { - return check(p, pkt); - } - return 0; -} - void r100_cs_dump_packet(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { @@ -904,6 +890,25 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return 0; } +int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + struct radeon_object *robj) +{ + struct radeon_cs_chunk *ib_chunk; + unsigned idx; + + ib_chunk = &p->chunks[p->chunk_ib_idx]; + idx = pkt->idx + 1; + if ((ib_chunk->kdata[idx+2] + 1) > radeon_object_size(robj)) { + DRM_ERROR("[drm] Buffer too small for PACKET3 INDX_BUFFER " + "(need %u have %lu) !\n", + ib_chunk->kdata[idx+2] + 1, + radeon_object_size(robj)); + return -EINVAL; + } + return 0; +} + static int r100_packet3_check(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { @@ -957,6 +962,10 @@ static int r100_packet3_check(struct radeon_cs_parser *p, return r; } ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); + if (r) { + return r; + } break; case 0x23: /* FIXME: cleanup */ @@ -1002,18 +1011,18 @@ int r100_cs_parse(struct radeon_cs_parser *p) } p->idx += pkt.count + 2; switch (pkt.type) { - case PACKET_TYPE0: - r = r100_packet0_check(p, &pkt); - break; - case PACKET_TYPE2: - break; - case PACKET_TYPE3: - r = r100_packet3_check(p, &pkt); - break; - default: - DRM_ERROR("Unknown packet type %d !\n", - pkt.type); - return -EINVAL; + case PACKET_TYPE0: + r = r100_packet0_check(p, &pkt); + break; + case PACKET_TYPE2: + break; + case PACKET_TYPE3: + r = r100_packet3_check(p, &pkt); + break; + default: + DRM_ERROR("Unknown packet type %d !\n", + pkt.type); + return -EINVAL; } if (r) { return r; @@ -1349,6 +1358,11 @@ void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v) } } +int r100_init(struct radeon_device *rdev) +{ + return 0; +} + /* * Debugfs info */ diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index f5870a099d4..e2ed5bc0817 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -48,14 +48,13 @@ int r100_cs_packet_next_reloc(struct radeon_cs_parser *p, struct radeon_cs_reloc **cs_reloc); int r100_cs_parse_packet0(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt, - unsigned *auth, unsigned n, + const unsigned *auth, unsigned n, radeon_packet0_check_t check); -int r100_cs_parse_packet3(struct radeon_cs_parser *p, - struct radeon_cs_packet *pkt, - unsigned *auth, unsigned n, - radeon_packet3_check_t check); void r100_cs_dump_packet(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt); +int r100_cs_track_check_pkt3_indx_buffer(struct radeon_cs_parser *p, + struct radeon_cs_packet *pkt, + struct radeon_object *robj); /* This files gather functions specifics to: * r300,r350,rv350,rv370,rv380 @@ -288,7 +287,7 @@ int r300_copy_dma(struct radeon_device *rdev, 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, PACKET0(RADEON_WAIT_UNTIL, 0 )); radeon_ring_write(rdev, (1 << 16)); for (i = 0; i < num_loops; i++) { cur_size = size; @@ -319,7 +318,7 @@ void r300_ring_start(struct radeon_device *rdev) /* Sub pixel 1/12 so we can have 4K rendering according to doc */ gb_tile_config = (R300_ENABLE_TILING | R300_TILE_SIZE_16); - switch (rdev->num_gb_pipes) { + switch(rdev->num_gb_pipes) { case 2: gb_tile_config |= R300_PIPE_COUNT_R300; break; @@ -452,8 +451,8 @@ void r300_gpu_init(struct radeon_device *rdev) case 4: gb_tile_config |= R300_PIPE_COUNT_R420; break; - case 1: default: + case 1: gb_tile_config |= R300_PIPE_COUNT_RV350; break; } @@ -725,18 +724,120 @@ struct r300_cs_track_cb { unsigned offset; }; +struct r300_cs_track_array { + struct radeon_object *robj; + unsigned esize; +}; + +struct r300_cs_track_texture { + struct radeon_object *robj; + unsigned pitch; + unsigned width; + unsigned height; + unsigned num_levels; + unsigned cpp; + unsigned tex_coord_type; + unsigned txdepth; + unsigned width_11; + unsigned height_11; + bool use_pitch; + bool enabled; + bool roundup_w; + bool roundup_h; +}; + struct r300_cs_track { - unsigned num_cb; - unsigned maxy; - struct r300_cs_track_cb cb[4]; - struct r300_cs_track_cb zb; - bool z_enabled; + unsigned num_cb; + unsigned maxy; + unsigned vtx_size; + unsigned vap_vf_cntl; + unsigned immd_dwords; + unsigned num_arrays; + unsigned max_indx; + struct r300_cs_track_array arrays[11]; + struct r300_cs_track_cb cb[4]; + struct r300_cs_track_cb zb; + struct r300_cs_track_texture textures[16]; + bool z_enabled; }; +static inline void r300_cs_track_texture_print(struct r300_cs_track_texture *t) +{ + DRM_ERROR("pitch %d\n", t->pitch); + DRM_ERROR("width %d\n", t->width); + DRM_ERROR("height %d\n", t->height); + DRM_ERROR("num levels %d\n", t->num_levels); + DRM_ERROR("depth %d\n", t->txdepth); + DRM_ERROR("bpp %d\n", t->cpp); + DRM_ERROR("coordinate type %d\n", t->tex_coord_type); + DRM_ERROR("width round to power of 2 %d\n", t->roundup_w); + DRM_ERROR("height round to power of 2 %d\n", t->roundup_h); +} + +static inline int r300_cs_track_texture_check(struct radeon_device *rdev, + struct r300_cs_track *track) +{ + struct radeon_object *robj; + unsigned long size; + unsigned u, i, w, h; + + for (u = 0; u < 16; u++) { + if (!track->textures[u].enabled) + continue; + robj = track->textures[u].robj; + if (robj == NULL) { + DRM_ERROR("No texture bound to unit %u\n", u); + return -EINVAL; + } + size = 0; + for (i = 0; i <= track->textures[u].num_levels; i++) { + if (track->textures[u].use_pitch) { + w = track->textures[u].pitch / (1 << i); + } else { + w = track->textures[u].width / (1 << i); + if (rdev->family >= CHIP_RV515) + w |= track->textures[u].width_11; + if (track->textures[u].roundup_w) + w = roundup_pow_of_two(w); + } + h = track->textures[u].height / (1 << i); + if (rdev->family >= CHIP_RV515) + h |= track->textures[u].height_11; + if (track->textures[u].roundup_h) + h = roundup_pow_of_two(h); + size += w * h; + } + size *= track->textures[u].cpp; + switch (track->textures[u].tex_coord_type) { + case 0: + break; + case 1: + size *= (1 << track->textures[u].txdepth); + break; + case 2: + size *= 6; + break; + default: + DRM_ERROR("Invalid texture coordinate type %u for unit " + "%u\n", track->textures[u].tex_coord_type, u); + return -EINVAL; + } + if (size > radeon_object_size(robj)) { + DRM_ERROR("Texture of unit %u needs %lu bytes but is " + "%lu\n", u, size, radeon_object_size(robj)); + r300_cs_track_texture_print(&track->textures[u]); + return -EINVAL; + } + } + return 0; +} + int r300_cs_track_check(struct radeon_device *rdev, struct r300_cs_track *track) { unsigned i; unsigned long size; + unsigned prim_walk; + unsigned nverts; for (i = 0; i < track->num_cb; i++) { if (track->cb[i].robj == NULL) { @@ -769,7 +870,59 @@ int r300_cs_track_check(struct radeon_device *rdev, struct r300_cs_track *track) return -EINVAL; } } - return 0; + prim_walk = (track->vap_vf_cntl >> 4) & 0x3; + nverts = (track->vap_vf_cntl >> 16) & 0xFFFF; + switch (prim_walk) { + case 1: + for (i = 0; i < track->num_arrays; i++) { + size = track->arrays[i].esize * track->max_indx * 4; + if (track->arrays[i].robj == NULL) { + DRM_ERROR("(PW %u) Vertex array %u no buffer " + "bound\n", prim_walk, i); + return -EINVAL; + } + if (size > radeon_object_size(track->arrays[i].robj)) { + DRM_ERROR("(PW %u) Vertex array %u need %lu dwords " + "have %lu dwords\n", prim_walk, i, + size >> 2, + radeon_object_size(track->arrays[i].robj) >> 2); + DRM_ERROR("Max indices %u\n", track->max_indx); + return -EINVAL; + } + } + break; + case 2: + for (i = 0; i < track->num_arrays; i++) { + size = track->arrays[i].esize * (nverts - 1) * 4; + if (track->arrays[i].robj == NULL) { + DRM_ERROR("(PW %u) Vertex array %u no buffer " + "bound\n", prim_walk, i); + return -EINVAL; + } + if (size > radeon_object_size(track->arrays[i].robj)) { + DRM_ERROR("(PW %u) Vertex array %u need %lu dwords " + "have %lu dwords\n", prim_walk, i, size >> 2, + radeon_object_size(track->arrays[i].robj) >> 2); + return -EINVAL; + } + } + break; + case 3: + size = track->vtx_size * nverts; + if (size != track->immd_dwords) { + DRM_ERROR("IMMD draw %u dwors but needs %lu dwords\n", + track->immd_dwords, size); + DRM_ERROR("VAP_VF_CNTL.NUM_VERTICES %u, VTX_SIZE %u\n", + nverts, track->vtx_size); + return -EINVAL; + } + break; + default: + DRM_ERROR("[drm] Invalid primitive walk %d for VAP_VF_CNTL\n", + prim_walk); + return -EINVAL; + } + return r300_cs_track_texture_check(rdev, track); } static inline void r300_cs_track_clear(struct r300_cs_track *track) @@ -789,9 +942,33 @@ static inline void r300_cs_track_clear(struct r300_cs_track *track) track->zb.pitch = 8192; track->zb.cpp = 4; track->zb.offset = 0; + track->vtx_size = 0x7F; + track->immd_dwords = 0xFFFFFFFFUL; + track->num_arrays = 11; + track->max_indx = 0x00FFFFFFUL; + for (i = 0; i < track->num_arrays; i++) { + track->arrays[i].robj = NULL; + track->arrays[i].esize = 0x7F; + } + for (i = 0; i < 16; i++) { + track->textures[i].pitch = 16536; + track->textures[i].width = 16536; + track->textures[i].height = 16536; + track->textures[i].width_11 = 1 << 11; + track->textures[i].height_11 = 1 << 11; + track->textures[i].num_levels = 12; + track->textures[i].txdepth = 16; + track->textures[i].cpp = 64; + track->textures[i].tex_coord_type = 1; + track->textures[i].robj = NULL; + /* CS IB emission code makes sure texture unit are disabled */ + track->textures[i].enabled = false; + track->textures[i].roundup_w = true; + track->textures[i].roundup_h = true; + } } -static unsigned r300_auth_reg[] = { +static const unsigned r300_reg_safe_bm[159] = { 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFBF, 0xFFFFFFFF, 0xFFFFFFBF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, @@ -808,7 +985,7 @@ static unsigned r300_auth_reg[] = { 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, - 0xFFFFFFFF, 0xFFFFCFCC, 0xF00E9FFF, 0x007C0000, + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000, 0xF0000078, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, @@ -824,9 +1001,9 @@ static unsigned r300_auth_reg[] = { 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, - 0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFC, 0xFFFFFFFF, + 0xFFFFFC78, 0xFFFFFFFF, 0xFFFFFFFE, 0xFFFFFFFF, 0x38FF8F50, 0xFFF88082, 0xF000000C, 0xFAE009FF, - 0x00000000, 0x00000000, 0xFFFF0000, 0x00000000, + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0x00000000, 0x0000C100, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0xFFFF0000, 0xFFFFFFFF, 0xFF80FFFF, @@ -848,8 +1025,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, ib = p->ib->ptr; ib_chunk = &p->chunks[p->chunk_ib_idx]; - track = (struct r300_cs_track *)p->track; - switch (reg) { + track = (struct r300_cs_track*)p->track; + switch(reg) { case RADEON_DST_PITCH_OFFSET: case RADEON_SRC_PITCH_OFFSET: r = r100_cs_packet_next_reloc(p, &reloc); @@ -907,6 +1084,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case R300_TX_OFFSET_0+52: case R300_TX_OFFSET_0+56: case R300_TX_OFFSET_0+60: + i = (reg - R300_TX_OFFSET_0) >> 2; r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", @@ -915,11 +1093,23 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } ib[idx] = ib_chunk->kdata[idx] + ((u32)reloc->lobj.gpu_offset); + track->textures[i].robj = reloc->robj; break; /* Tracked registers */ + case 0x2084: + /* VAP_VF_CNTL */ + track->vap_vf_cntl = ib_chunk->kdata[idx]; + break; + case 0x20B4: + /* VAP_VTX_SIZE */ + track->vtx_size = ib_chunk->kdata[idx] & 0x7F; + break; + case 0x2134: + /* VAP_VF_MAX_VTX_INDX */ + track->max_indx = ib_chunk->kdata[idx] & 0x00FFFFFFUL; + break; case 0x43E4: /* SC_SCISSOR1 */ - track->maxy = ((ib_chunk->kdata[idx] >> 13) & 0x1FFF) + 1; if (p->rdev->family < CHIP_RV515) { track->maxy -= 1440; @@ -994,8 +1184,166 @@ static int r300_packet0_check(struct radeon_cs_parser *p, /* ZB_DEPTHPITCH */ track->zb.pitch = ib_chunk->kdata[idx] & 0x3FFC; break; + case 0x4104: + for (i = 0; i < 16; i++) { + bool enabled; + + enabled = !!(ib_chunk->kdata[idx] & (1 << i)); + track->textures[i].enabled = enabled; + } + break; + case 0x44C0: + case 0x44C4: + case 0x44C8: + case 0x44CC: + case 0x44D0: + case 0x44D4: + case 0x44D8: + case 0x44DC: + case 0x44E0: + case 0x44E4: + case 0x44E8: + case 0x44EC: + case 0x44F0: + case 0x44F4: + case 0x44F8: + case 0x44FC: + /* TX_FORMAT1_[0-15] */ + i = (reg - 0x44C0) >> 2; + tmp = (ib_chunk->kdata[idx] >> 25) & 0x3; + track->textures[i].tex_coord_type = tmp; + switch ((ib_chunk->kdata[idx] & 0x1F)) { + case 0: + case 2: + case 5: + case 18: + case 20: + case 21: + track->textures[i].cpp = 1; + break; + case 1: + case 3: + case 6: + case 7: + case 10: + case 11: + case 19: + case 22: + case 24: + track->textures[i].cpp = 2; + break; + case 4: + case 8: + case 9: + case 12: + case 13: + case 23: + case 25: + case 27: + case 30: + track->textures[i].cpp = 4; + break; + case 14: + case 26: + case 28: + track->textures[i].cpp = 8; + break; + case 29: + track->textures[i].cpp = 16; + break; + default: + DRM_ERROR("Invalid texture format %u\n", + (ib_chunk->kdata[idx] & 0x1F)); + return -EINVAL; + break; + } + break; + case 0x4400: + case 0x4404: + case 0x4408: + case 0x440C: + case 0x4410: + case 0x4414: + case 0x4418: + case 0x441C: + case 0x4420: + case 0x4424: + case 0x4428: + case 0x442C: + case 0x4430: + case 0x4434: + case 0x4438: + case 0x443C: + /* TX_FILTER0_[0-15] */ + i = (reg - 0x4400) >> 2; + tmp = ib_chunk->kdata[idx] & 0x7;; + if (tmp == 2 || tmp == 4 || tmp == 6) { + track->textures[i].roundup_w = false; + } + tmp = (ib_chunk->kdata[idx] >> 3) & 0x7;; + if (tmp == 2 || tmp == 4 || tmp == 6) { + track->textures[i].roundup_h = false; + } + break; + case 0x4500: + case 0x4504: + case 0x4508: + case 0x450C: + case 0x4510: + case 0x4514: + case 0x4518: + case 0x451C: + case 0x4520: + case 0x4524: + case 0x4528: + case 0x452C: + case 0x4530: + case 0x4534: + case 0x4538: + case 0x453C: + /* TX_FORMAT2_[0-15] */ + i = (reg - 0x4500) >> 2; + tmp = ib_chunk->kdata[idx] & 0x3FFF; + track->textures[i].pitch = tmp + 1; + if (p->rdev->family >= CHIP_RV515) { + tmp = ((ib_chunk->kdata[idx] >> 15) & 1) << 11; + track->textures[i].width_11 = tmp; + tmp = ((ib_chunk->kdata[idx] >> 16) & 1) << 11; + track->textures[i].height_11 = tmp; + } + break; + case 0x4480: + case 0x4484: + case 0x4488: + case 0x448C: + case 0x4490: + case 0x4494: + case 0x4498: + case 0x449C: + case 0x44A0: + case 0x44A4: + case 0x44A8: + case 0x44AC: + case 0x44B0: + case 0x44B4: + case 0x44B8: + case 0x44BC: + /* TX_FORMAT0_[0-15] */ + i = (reg - 0x4480) >> 2; + tmp = ib_chunk->kdata[idx] & 0x7FF; + track->textures[i].width = tmp + 1; + tmp = (ib_chunk->kdata[idx] >> 11) & 0x7FF; + track->textures[i].height = tmp + 1; + tmp = (ib_chunk->kdata[idx] >> 26) & 0xF; + track->textures[i].num_levels = tmp; + tmp = ib_chunk->kdata[idx] & (1 << 31); + track->textures[i].use_pitch = !!tmp; + tmp = (ib_chunk->kdata[idx] >> 22) & 0xF; + track->textures[i].txdepth = tmp; + break; default: - printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", reg, idx); + printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", + reg, idx); return -EINVAL; } return 0; @@ -1015,11 +1363,12 @@ static int r300_packet3_check(struct radeon_cs_parser *p, ib = p->ib->ptr; ib_chunk = &p->chunks[p->chunk_ib_idx]; idx = pkt->idx + 1; - track = (struct r300_cs_track *)p->track; - switch (pkt->opcode) { + track = (struct r300_cs_track*)p->track; + switch(pkt->opcode) { case PACKET3_3D_LOAD_VBPNTR: - c = ib_chunk->kdata[idx++]; - for (i = 0; i < (c - 1); i += 2, idx += 3) { + c = ib_chunk->kdata[idx++] & 0x1F; + track->num_arrays = c; + for (i = 0; i < (c - 1); i+=2, idx+=3) { r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for packet3 %d\n", @@ -1028,6 +1377,9 @@ static int r300_packet3_check(struct radeon_cs_parser *p, return r; } ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + track->arrays[i + 0].robj = reloc->robj; + track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8; + track->arrays[i + 0].esize &= 0x7F; r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for packet3 %d\n", @@ -1036,6 +1388,9 @@ static int r300_packet3_check(struct radeon_cs_parser *p, return r; } ib[idx+2] = ib_chunk->kdata[idx+2] + ((u32)reloc->lobj.gpu_offset); + track->arrays[i + 1].robj = reloc->robj; + track->arrays[i + 1].esize = ib_chunk->kdata[idx] >> 24; + track->arrays[i + 1].esize &= 0x7F; } if (c & 1) { r = r100_cs_packet_next_reloc(p, &reloc); @@ -1046,6 +1401,9 @@ static int r300_packet3_check(struct radeon_cs_parser *p, return r; } ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + track->arrays[i + 0].robj = reloc->robj; + track->arrays[i + 0].esize = ib_chunk->kdata[idx] >> 8; + track->arrays[i + 0].esize &= 0x7F; } break; case PACKET3_INDX_BUFFER: @@ -1056,14 +1414,65 @@ static int r300_packet3_check(struct radeon_cs_parser *p, return r; } ib[idx+1] = ib_chunk->kdata[idx+1] + ((u32)reloc->lobj.gpu_offset); + r = r100_cs_track_check_pkt3_indx_buffer(p, pkt, reloc->robj); + if (r) { + return r; + } break; /* Draw packet */ - case PACKET3_3D_DRAW_VBUF: case PACKET3_3D_DRAW_IMMD: - case PACKET3_3D_DRAW_INDX: - case PACKET3_3D_DRAW_VBUF_2: + /* Number of dwords is vtx_size * (num_vertices - 1) + * PRIM_WALK must be equal to 3 vertex data in embedded + * in cmd stream */ + if (((ib_chunk->kdata[idx+1] >> 4) & 0x3) != 3) { + DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); + return -EINVAL; + } + track->vap_vf_cntl = ib_chunk->kdata[idx+1]; + track->immd_dwords = pkt->count - 1; + r = r300_cs_track_check(p->rdev, track); + if (r) { + return r; + } + break; case PACKET3_3D_DRAW_IMMD_2: + /* Number of dwords is vtx_size * (num_vertices - 1) + * PRIM_WALK must be equal to 3 vertex data in embedded + * in cmd stream */ + if (((ib_chunk->kdata[idx] >> 4) & 0x3) != 3) { + DRM_ERROR("PRIM_WALK must be 3 for IMMD draw\n"); + return -EINVAL; + } + track->vap_vf_cntl = ib_chunk->kdata[idx]; + track->immd_dwords = pkt->count; + r = r300_cs_track_check(p->rdev, track); + if (r) { + return r; + } + break; + case PACKET3_3D_DRAW_VBUF: + track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; + r = r300_cs_track_check(p->rdev, track); + if (r) { + return r; + } + break; + case PACKET3_3D_DRAW_VBUF_2: + track->vap_vf_cntl = ib_chunk->kdata[idx]; + r = r300_cs_track_check(p->rdev, track); + if (r) { + return r; + } + break; + case PACKET3_3D_DRAW_INDX: + track->vap_vf_cntl = ib_chunk->kdata[idx + 1]; + r = r300_cs_track_check(p->rdev, track); + if (r) { + return r; + } + break; case PACKET3_3D_DRAW_INDX_2: + track->vap_vf_cntl = ib_chunk->kdata[idx]; r = r300_cs_track_check(p->rdev, track); if (r) { return r; @@ -1095,8 +1504,8 @@ int r300_cs_parse(struct radeon_cs_parser *p) switch (pkt.type) { case PACKET_TYPE0: r = r100_cs_parse_packet0(p, &pkt, - r300_auth_reg, - ARRAY_SIZE(r300_auth_reg), + p->rdev->config.r300.reg_safe_bm, + p->rdev->config.r300.reg_safe_bm_size, &r300_packet0_check); break; case PACKET_TYPE2: @@ -1114,3 +1523,10 @@ int r300_cs_parse(struct radeon_cs_parser *p) } while (p->idx < p->chunks[p->chunk_ib_idx].length_dw); return 0; } + +int r300_init(struct radeon_device *rdev) +{ + rdev->config.r300.reg_safe_bm = r300_reg_safe_bm; + rdev->config.r300.reg_safe_bm_size = ARRAY_SIZE(r300_reg_safe_bm); + return 0; +} diff --git a/drivers/gpu/drm/radeon/r300.h b/drivers/gpu/drm/radeon/r300.h new file mode 100644 index 00000000000..8486b4da9d6 --- /dev/null +++ b/drivers/gpu/drm/radeon/r300.h @@ -0,0 +1,36 @@ +/* + * Copyright 2008 Advanced Micro Devices, Inc. + * Copyright 2008 Red Hat Inc. + * Copyright 2009 Jerome Glisse. + * + * 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: Dave Airlie + * Alex Deucher + * Jerome Glisse + */ +#ifndef R300_H +#define R300_H + +struct r300_asic { + const unsigned *reg_safe_bm; + unsigned reg_safe_bm_size; +}; + +#endif diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index c3f24cc5600..d61f2fc61df 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -51,7 +51,7 @@ #include "radeon_mode.h" #include "radeon_reg.h" - +#include "r300.h" /* * Modules parameters. @@ -496,6 +496,7 @@ int r100_debugfs_cp_init(struct radeon_device *rdev); * ASIC specific functions. */ struct radeon_asic { + int (*init)(struct radeon_device *rdev); void (*errata)(struct radeon_device *rdev); void (*vram_info)(struct radeon_device *rdev); int (*gpu_reset)(struct radeon_device *rdev); @@ -536,6 +537,10 @@ struct radeon_asic { void (*set_clock_gating)(struct radeon_device *rdev, int enable); }; +union radeon_asic_config { + struct r300_asic r300; +}; + /* * IOCTL. @@ -573,6 +578,7 @@ struct radeon_device { struct drm_device *ddev; struct pci_dev *pdev; /* ASIC */ + union radeon_asic_config config; enum radeon_family family; unsigned long flags; int usec_timeout; @@ -763,6 +769,7 @@ static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v) /* * ASICs macro. */ +#define radeon_init(rdev) (rdev)->asic->init((rdev)) #define radeon_cs_parse(p) rdev->asic->cs_parse((p)) #define radeon_errata(rdev) (rdev)->asic->errata((rdev)) #define radeon_vram_info(rdev) (rdev)->asic->vram_info((rdev)) diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index e57d8a784e9..e2e567395df 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -41,6 +41,7 @@ void radeon_atom_set_clock_gating(struct radeon_device *rdev, int enable); /* * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 */ +int r100_init(struct radeon_device *rdev); uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg); void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); void r100_errata(struct radeon_device *rdev); @@ -72,6 +73,7 @@ int r100_copy_blit(struct radeon_device *rdev, struct radeon_fence *fence); static struct radeon_asic r100_asic = { + .init = &r100_init, .errata = &r100_errata, .vram_info = &r100_vram_info, .gpu_reset = &r100_gpu_reset, @@ -104,6 +106,7 @@ static struct radeon_asic r100_asic = { /* * r300,r350,rv350,rv380 */ +int r300_init(struct radeon_device *rdev); void r300_errata(struct radeon_device *rdev); void r300_vram_info(struct radeon_device *rdev); int r300_gpu_reset(struct radeon_device *rdev); @@ -126,6 +129,7 @@ int r300_copy_dma(struct radeon_device *rdev, unsigned num_pages, struct radeon_fence *fence); static struct radeon_asic r300_asic = { + .init = &r300_init, .errata = &r300_errata, .vram_info = &r300_vram_info, .gpu_reset = &r300_gpu_reset, @@ -162,6 +166,7 @@ void r420_vram_info(struct radeon_device *rdev); int r420_mc_init(struct radeon_device *rdev); void r420_mc_fini(struct radeon_device *rdev); static struct radeon_asic r420_asic = { + .init = &r300_init, .errata = &r420_errata, .vram_info = &r420_vram_info, .gpu_reset = &r300_gpu_reset, @@ -205,6 +210,7 @@ int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); uint32_t rs400_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rs400_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); static struct radeon_asic rs400_asic = { + .init = &r300_init, .errata = &rs400_errata, .vram_info = &rs400_vram_info, .gpu_reset = &r300_gpu_reset, @@ -249,6 +255,7 @@ int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); uint32_t rs600_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rs600_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); static struct radeon_asic rs600_asic = { + .init = &r300_init, .errata = &rs600_errata, .vram_info = &rs600_vram_info, .gpu_reset = &r300_gpu_reset, @@ -288,6 +295,7 @@ void rs690_mc_fini(struct radeon_device *rdev); uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rs690_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); static struct radeon_asic rs690_asic = { + .init = &r300_init, .errata = &rs690_errata, .vram_info = &rs690_vram_info, .gpu_reset = &r300_gpu_reset, @@ -320,6 +328,7 @@ static struct radeon_asic rs690_asic = { /* * rv515 */ +int rv515_init(struct radeon_device *rdev); void rv515_errata(struct radeon_device *rdev); void rv515_vram_info(struct radeon_device *rdev); int rv515_gpu_reset(struct radeon_device *rdev); @@ -331,6 +340,7 @@ void rv515_ring_start(struct radeon_device *rdev); uint32_t rv515_pcie_rreg(struct radeon_device *rdev, uint32_t reg); void rv515_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); static struct radeon_asic rv515_asic = { + .init = &rv515_init, .errata = &rv515_errata, .vram_info = &rv515_vram_info, .gpu_reset = &rv515_gpu_reset, @@ -349,7 +359,7 @@ static struct radeon_asic rv515_asic = { .irq_set = &r100_irq_set, .irq_process = &r100_irq_process, .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r100_cs_parse, + .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, .copy_dma = &r300_copy_dma, .copy = &r100_copy_blit, @@ -368,6 +378,7 @@ void r520_vram_info(struct radeon_device *rdev); int r520_mc_init(struct radeon_device *rdev); void r520_mc_fini(struct radeon_device *rdev); static struct radeon_asic r520_asic = { + .init = &rv515_init, .errata = &r520_errata, .vram_info = &r520_vram_info, .gpu_reset = &rv515_gpu_reset, @@ -386,7 +397,7 @@ static struct radeon_asic r520_asic = { .irq_set = &r100_irq_set, .irq_process = &r100_irq_process, .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r100_cs_parse, + .cs_parse = &r300_cs_parse, .copy_blit = &r100_copy_blit, .copy_dma = &r300_copy_dma, .copy = &r100_copy_blit, diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 5fd2b639bf6..f30aa7274a5 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -470,6 +470,10 @@ int radeon_device_init(struct radeon_device *rdev, if (r) { return r; } + r = radeon_init(rdev); + if (r) { + return r; + } /* Report DMA addressing limitation */ r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(32)); diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c index 7eab95db58a..ffea37b1b3e 100644 --- a/drivers/gpu/drm/radeon/rv515.c +++ b/drivers/gpu/drm/radeon/rv515.c @@ -225,6 +225,8 @@ void rv515_ring_start(struct radeon_device *rdev) radeon_ring_write(rdev, R300_GEOMETRY_ROUND_NEAREST | R300_COLOR_ROUND_NEAREST); + radeon_ring_write(rdev, PACKET0(0x20C8, 0)); + radeon_ring_write(rdev, 0); radeon_ring_unlock_commit(rdev); } @@ -502,3 +504,59 @@ int rv515_debugfs_ga_info_init(struct radeon_device *rdev) return 0; #endif } + + +/* + * Asic initialization + */ +static const unsigned r500_reg_safe_bm[159] = { + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFBF, 0xFFFFFFFF, 0xFFFFFFBF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0x17FF1FFF, 0xFFFFFFFC, 0xFFFFFFFF, 0xFF30FFBF, + 0xFFFFFFF8, 0xC3E6FFFF, 0xFFFFF6DF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFF03F, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFEFCE, 0xF00EBFFF, 0x007C0000, + 0xF0000038, 0xFF000009, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFF7FF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0x1FFFFC78, 0xFFFFE000, 0xFFFFFFFE, 0xFFFFFFFF, + 0x38CF8F50, 0xFFF88082, 0xFF0000FC, 0xFAE009FF, + 0x0000FFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, + 0xFFFF8CFC, 0xFFFFC1FF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, + 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFF80FFFF, + 0x00000000, 0x00000000, 0x00000000, 0x00000000, + 0x0003FC01, 0x3FFFFCF8, 0xFE800B19, +}; + + + +int rv515_init(struct radeon_device *rdev) +{ + rdev->config.r300.reg_safe_bm = r500_reg_safe_bm; + rdev->config.r300.reg_safe_bm_size = ARRAY_SIZE(r500_reg_safe_bm); + return 0; +} -- cgit v1.2.3-70-g09d2