diff options
Diffstat (limited to 'drivers')
53 files changed, 2310 insertions, 365 deletions
diff --git a/drivers/block/viodasd.c b/drivers/block/viodasd.c index 709f809f79f..2d518aa2720 100644 --- a/drivers/block/viodasd.c +++ b/drivers/block/viodasd.c @@ -45,10 +45,10 @@ #include <asm/uaccess.h> #include <asm/vio.h> -#include <asm/iSeries/HvTypes.h> -#include <asm/iSeries/HvLpEvent.h> -#include <asm/iSeries/HvLpConfig.h> -#include <asm/iSeries/vio.h> +#include <asm/iseries/hv_types.h> +#include <asm/iseries/hv_lp_event.h> +#include <asm/iseries/hv_lp_config.h> +#include <asm/iseries/vio.h> MODULE_DESCRIPTION("iSeries Virtual DASD"); MODULE_AUTHOR("Dave Boutcher"); diff --git a/drivers/cdrom/viocd.c b/drivers/cdrom/viocd.c index 36f31d20222..b5191780ecc 100644 --- a/drivers/cdrom/viocd.c +++ b/drivers/cdrom/viocd.c @@ -46,9 +46,9 @@ #include <asm/vio.h> #include <asm/scatterlist.h> -#include <asm/iSeries/HvTypes.h> -#include <asm/iSeries/HvLpEvent.h> -#include <asm/iSeries/vio.h> +#include <asm/iseries/hv_types.h> +#include <asm/iseries/hv_lp_event.h> +#include <asm/iseries/vio.h> #define VIOCD_DEVICE "iseries/vcd" #define VIOCD_DEVICE_DEVFS "iseries/vcd" diff --git a/drivers/char/nvram.c b/drivers/char/nvram.c index 9e24bbd4090..1af733d0732 100644 --- a/drivers/char/nvram.c +++ b/drivers/char/nvram.c @@ -32,11 +32,9 @@ * added changelog * 1.2 Erik Gilling: Cobalt Networks support * Tim Hockin: general cleanup, Cobalt support - * 1.3 Jon Ringle: Comdial MP1000 support - * */ -#define NVRAM_VERSION "1.3" +#define NVRAM_VERSION "1.2" #include <linux/module.h> #include <linux/config.h> @@ -47,7 +45,6 @@ #define PC 1 #define ATARI 2 #define COBALT 3 -#define MP1000 4 /* select machine configuration */ #if defined(CONFIG_ATARI) @@ -57,9 +54,6 @@ # if defined(CONFIG_COBALT) # include <linux/cobalt-nvram.h> # define MACH COBALT -# elif defined(CONFIG_MACH_MP1000) -# undef MACH -# define MACH MP1000 # else # define MACH PC # endif @@ -118,23 +112,6 @@ #endif -#if MACH == MP1000 - -/* RTC in a MP1000 */ -#define CHECK_DRIVER_INIT() 1 - -#define MP1000_CKS_RANGE_START 0 -#define MP1000_CKS_RANGE_END 111 -#define MP1000_CKS_LOC 112 - -#define NVRAM_BYTES (128-NVRAM_FIRST_BYTE) - -#define mach_check_checksum mp1000_check_checksum -#define mach_set_checksum mp1000_set_checksum -#define mach_proc_infos mp1000_proc_infos - -#endif - /* Note that *all* calls to CMOS_READ and CMOS_WRITE must be done with * rtc_lock held. Due to the index-port/data-port design of the RTC, we * don't want two different things trying to get to it at once. (e.g. the @@ -938,91 +915,6 @@ atari_proc_infos(unsigned char *nvram, char *buffer, int *len, #endif /* MACH == ATARI */ -#if MACH == MP1000 - -static int -mp1000_check_checksum(void) -{ - int i; - unsigned short sum = 0; - unsigned short expect; - - for (i = MP1000_CKS_RANGE_START; i <= MP1000_CKS_RANGE_END; ++i) - sum += __nvram_read_byte(i); - - expect = __nvram_read_byte(MP1000_CKS_LOC+1)<<8 | - __nvram_read_byte(MP1000_CKS_LOC); - return ((sum & 0xffff) == expect); -} - -static void -mp1000_set_checksum(void) -{ - int i; - unsigned short sum = 0; - - for (i = MP1000_CKS_RANGE_START; i <= MP1000_CKS_RANGE_END; ++i) - sum += __nvram_read_byte(i); - __nvram_write_byte(sum >> 8, MP1000_CKS_LOC + 1); - __nvram_write_byte(sum & 0xff, MP1000_CKS_LOC); -} - -#ifdef CONFIG_PROC_FS - -#define SERVER_N_LEN 32 -#define PATH_N_LEN 32 -#define FILE_N_LEN 32 -#define NVRAM_MAGIC_SIG 0xdead - -typedef struct NvRamImage -{ - unsigned short int magic; - unsigned short int mode; - char fname[FILE_N_LEN]; - char path[PATH_N_LEN]; - char server[SERVER_N_LEN]; - char pad[12]; -} NvRam; - -static int -mp1000_proc_infos(unsigned char *nvram, char *buffer, int *len, - off_t *begin, off_t offset, int size) -{ - int checksum; - NvRam* nv = (NvRam*)nvram; - - spin_lock_irq(&rtc_lock); - checksum = __nvram_check_checksum(); - spin_unlock_irq(&rtc_lock); - - PRINT_PROC("Checksum status: %svalid\n", checksum ? "" : "not "); - - switch( nv->mode ) - { - case 0 : - PRINT_PROC( "\tMode 0, tftp prompt\n" ); - break; - case 1 : - PRINT_PROC( "\tMode 1, booting from disk\n" ); - break; - case 2 : - PRINT_PROC( "\tMode 2, Alternate boot from disk /boot/%s\n", nv->fname ); - break; - case 3 : - PRINT_PROC( "\tMode 3, Booting from net:\n" ); - PRINT_PROC( "\t\t%s:%s%s\n",nv->server, nv->path, nv->fname ); - break; - default: - PRINT_PROC( "\tInconsistant nvram?\n" ); - break; - } - - return 1; -} -#endif - -#endif /* MACH == MP1000 */ - MODULE_LICENSE("GPL"); EXPORT_SYMBOL(__nvram_read_byte); diff --git a/drivers/char/viocons.c b/drivers/char/viocons.c index 44f5fb4a46e..98601c7d04a 100644 --- a/drivers/char/viocons.c +++ b/drivers/char/viocons.c @@ -44,12 +44,12 @@ #include <linux/tty_flip.h> #include <linux/sysrq.h> -#include <asm/iSeries/vio.h> +#include <asm/iseries/vio.h> -#include <asm/iSeries/HvLpEvent.h> -#include <asm/iSeries/HvCallEvent.h> -#include <asm/iSeries/HvLpConfig.h> -#include <asm/iSeries/HvCall.h> +#include <asm/iseries/hv_lp_event.h> +#include <asm/iseries/hv_call_event.h> +#include <asm/iseries/hv_lp_config.h> +#include <asm/iseries/hv_call.h> #ifdef CONFIG_VT #error You must turn off CONFIG_VT to use CONFIG_VIOCONS diff --git a/drivers/char/viotape.c b/drivers/char/viotape.c index 51abd3defc1..867cc4e418c 100644 --- a/drivers/char/viotape.c +++ b/drivers/char/viotape.c @@ -29,7 +29,7 @@ * * All tape operations are performed by sending messages back and forth to * the OS/400 partition. The format of the messages is defined in - * iSeries/vio.h + * iseries/vio.h */ #include <linux/config.h> #include <linux/version.h> @@ -54,10 +54,10 @@ #include <asm/ioctls.h> #include <asm/vio.h> -#include <asm/iSeries/vio.h> -#include <asm/iSeries/HvLpEvent.h> -#include <asm/iSeries/HvCallEvent.h> -#include <asm/iSeries/HvLpConfig.h> +#include <asm/iseries/vio.h> +#include <asm/iseries/hv_lp_event.h> +#include <asm/iseries/hv_call_event.h> +#include <asm/iseries/hv_lp_config.h> #define VIOTAPE_VERSION "1.2" #define VIOTAPE_MAXREQ 1 diff --git a/drivers/infiniband/Kconfig b/drivers/infiniband/Kconfig index 325d502e25c..bdf0891a92d 100644 --- a/drivers/infiniband/Kconfig +++ b/drivers/infiniband/Kconfig @@ -33,4 +33,6 @@ source "drivers/infiniband/hw/mthca/Kconfig" source "drivers/infiniband/ulp/ipoib/Kconfig" +source "drivers/infiniband/ulp/srp/Kconfig" + endmenu diff --git a/drivers/infiniband/Makefile b/drivers/infiniband/Makefile index d256cf79821..a43fb34cca9 100644 --- a/drivers/infiniband/Makefile +++ b/drivers/infiniband/Makefile @@ -1,3 +1,4 @@ obj-$(CONFIG_INFINIBAND) += core/ obj-$(CONFIG_INFINIBAND_MTHCA) += hw/mthca/ obj-$(CONFIG_INFINIBAND_IPOIB) += ulp/ipoib/ +obj-$(CONFIG_INFINIBAND_SRP) += ulp/srp/ diff --git a/drivers/infiniband/core/agent.c b/drivers/infiniband/core/agent.c index 0c3c6952faa..7545775d38e 100644 --- a/drivers/infiniband/core/agent.c +++ b/drivers/infiniband/core/agent.c @@ -155,13 +155,12 @@ int ib_agent_port_open(struct ib_device *device, int port_num) int ret; /* Create new device info */ - port_priv = kmalloc(sizeof *port_priv, GFP_KERNEL); + port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL); if (!port_priv) { printk(KERN_ERR SPFX "No memory for ib_agent_port_private\n"); ret = -ENOMEM; goto error1; } - memset(port_priv, 0, sizeof *port_priv); /* Obtain send only MAD agent for SMI QP */ port_priv->agent[0] = ib_register_mad_agent(device, port_num, diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c index 580c3a2bb10..02110e00d14 100644 --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -544,11 +544,10 @@ struct ib_cm_id *ib_create_cm_id(struct ib_device *device, struct cm_id_private *cm_id_priv; int ret; - cm_id_priv = kmalloc(sizeof *cm_id_priv, GFP_KERNEL); + cm_id_priv = kzalloc(sizeof *cm_id_priv, GFP_KERNEL); if (!cm_id_priv) return ERR_PTR(-ENOMEM); - memset(cm_id_priv, 0, sizeof *cm_id_priv); cm_id_priv->id.state = IB_CM_IDLE; cm_id_priv->id.device = device; cm_id_priv->id.cm_handler = cm_handler; @@ -621,10 +620,9 @@ static struct cm_timewait_info * cm_create_timewait_info(__be32 local_id) { struct cm_timewait_info *timewait_info; - timewait_info = kmalloc(sizeof *timewait_info, GFP_KERNEL); + timewait_info = kzalloc(sizeof *timewait_info, GFP_KERNEL); if (!timewait_info) return ERR_PTR(-ENOMEM); - memset(timewait_info, 0, sizeof *timewait_info); timewait_info->work.local_id = local_id; INIT_WORK(&timewait_info->work.work, cm_work_handler, diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c index 5a6e4497640..e169e798354 100644 --- a/drivers/infiniband/core/device.c +++ b/drivers/infiniband/core/device.c @@ -161,17 +161,9 @@ static int alloc_name(char *name) */ struct ib_device *ib_alloc_device(size_t size) { - void *dev; - BUG_ON(size < sizeof (struct ib_device)); - dev = kmalloc(size, GFP_KERNEL); - if (!dev) - return NULL; - - memset(dev, 0, size); - - return dev; + return kzalloc(size, GFP_KERNEL); } EXPORT_SYMBOL(ib_alloc_device); diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c index 88f9f8c9eac..3d8175e5f05 100644 --- a/drivers/infiniband/core/mad.c +++ b/drivers/infiniband/core/mad.c @@ -255,12 +255,11 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device, } /* Allocate structures */ - mad_agent_priv = kmalloc(sizeof *mad_agent_priv, GFP_KERNEL); + mad_agent_priv = kzalloc(sizeof *mad_agent_priv, GFP_KERNEL); if (!mad_agent_priv) { ret = ERR_PTR(-ENOMEM); goto error1; } - memset(mad_agent_priv, 0, sizeof *mad_agent_priv); mad_agent_priv->agent.mr = ib_get_dma_mr(port_priv->qp_info[qpn].qp->pd, IB_ACCESS_LOCAL_WRITE); @@ -448,14 +447,13 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device, goto error1; } /* Allocate structures */ - mad_snoop_priv = kmalloc(sizeof *mad_snoop_priv, GFP_KERNEL); + mad_snoop_priv = kzalloc(sizeof *mad_snoop_priv, GFP_KERNEL); if (!mad_snoop_priv) { ret = ERR_PTR(-ENOMEM); goto error1; } /* Now, fill in the various structures */ - memset(mad_snoop_priv, 0, sizeof *mad_snoop_priv); mad_snoop_priv->qp_info = &port_priv->qp_info[qpn]; mad_snoop_priv->agent.device = device; mad_snoop_priv->agent.recv_handler = recv_handler; @@ -794,10 +792,9 @@ struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent, (!rmpp_active && buf_size > sizeof(struct ib_mad))) return ERR_PTR(-EINVAL); - buf = kmalloc(sizeof *mad_send_wr + buf_size, gfp_mask); + buf = kzalloc(sizeof *mad_send_wr + buf_size, gfp_mask); if (!buf) return ERR_PTR(-ENOMEM); - memset(buf, 0, sizeof *mad_send_wr + buf_size); mad_send_wr = buf + buf_size; mad_send_wr->send_buf.mad = buf; @@ -1039,14 +1036,12 @@ static int method_in_use(struct ib_mad_mgmt_method_table **method, static int allocate_method_table(struct ib_mad_mgmt_method_table **method) { /* Allocate management method table */ - *method = kmalloc(sizeof **method, GFP_ATOMIC); + *method = kzalloc(sizeof **method, GFP_ATOMIC); if (!*method) { printk(KERN_ERR PFX "No memory for " "ib_mad_mgmt_method_table\n"); return -ENOMEM; } - /* Clear management method table */ - memset(*method, 0, sizeof **method); return 0; } @@ -1137,15 +1132,14 @@ static int add_nonoui_reg_req(struct ib_mad_reg_req *mad_reg_req, class = &port_priv->version[mad_reg_req->mgmt_class_version].class; if (!*class) { /* Allocate management class table for "new" class version */ - *class = kmalloc(sizeof **class, GFP_ATOMIC); + *class = kzalloc(sizeof **class, GFP_ATOMIC); if (!*class) { printk(KERN_ERR PFX "No memory for " "ib_mad_mgmt_class_table\n"); ret = -ENOMEM; goto error1; } - /* Clear management class table */ - memset(*class, 0, sizeof(**class)); + /* Allocate method table for this management class */ method = &(*class)->method_table[mgmt_class]; if ((ret = allocate_method_table(method))) @@ -1209,25 +1203,24 @@ static int add_oui_reg_req(struct ib_mad_reg_req *mad_reg_req, mad_reg_req->mgmt_class_version].vendor; if (!*vendor_table) { /* Allocate mgmt vendor class table for "new" class version */ - vendor = kmalloc(sizeof *vendor, GFP_ATOMIC); + vendor = kzalloc(sizeof *vendor, GFP_ATOMIC); if (!vendor) { printk(KERN_ERR PFX "No memory for " "ib_mad_mgmt_vendor_class_table\n"); goto error1; } - /* Clear management vendor class table */ - memset(vendor, 0, sizeof(*vendor)); + *vendor_table = vendor; } if (!(*vendor_table)->vendor_class[vclass]) { /* Allocate table for this management vendor class */ - vendor_class = kmalloc(sizeof *vendor_class, GFP_ATOMIC); + vendor_class = kzalloc(sizeof *vendor_class, GFP_ATOMIC); if (!vendor_class) { printk(KERN_ERR PFX "No memory for " "ib_mad_mgmt_vendor_class\n"); goto error2; } - memset(vendor_class, 0, sizeof(*vendor_class)); + (*vendor_table)->vendor_class[vclass] = vendor_class; } for (i = 0; i < MAX_MGMT_OUI; i++) { @@ -2524,12 +2517,12 @@ static int ib_mad_port_open(struct ib_device *device, char name[sizeof "ib_mad123"]; /* Create new device info */ - port_priv = kmalloc(sizeof *port_priv, GFP_KERNEL); + port_priv = kzalloc(sizeof *port_priv, GFP_KERNEL); if (!port_priv) { printk(KERN_ERR PFX "No memory for ib_mad_port_private\n"); return -ENOMEM; } - memset(port_priv, 0, sizeof *port_priv); + port_priv->device = device; port_priv->port_num = port_num; spin_lock_init(&port_priv->reg_lock); diff --git a/drivers/infiniband/core/sysfs.c b/drivers/infiniband/core/sysfs.c index 7ce7a6c782f..b8120650e71 100644 --- a/drivers/infiniband/core/sysfs.c +++ b/drivers/infiniband/core/sysfs.c @@ -307,14 +307,13 @@ static ssize_t show_pma_counter(struct ib_port *p, struct port_attribute *attr, if (!p->ibdev->process_mad) return sprintf(buf, "N/A (no PMA)\n"); - in_mad = kmalloc(sizeof *in_mad, GFP_KERNEL); + in_mad = kzalloc(sizeof *in_mad, GFP_KERNEL); out_mad = kmalloc(sizeof *in_mad, GFP_KERNEL); if (!in_mad || !out_mad) { ret = -ENOMEM; goto out; } - memset(in_mad, 0, sizeof *in_mad); in_mad->mad_hdr.base_version = 1; in_mad->mad_hdr.mgmt_class = IB_MGMT_CLASS_PERF_MGMT; in_mad->mad_hdr.class_version = 1; @@ -508,10 +507,9 @@ static int add_port(struct ib_device *device, int port_num) if (ret) return ret; - p = kmalloc(sizeof *p, GFP_KERNEL); + p = kzalloc(sizeof *p, GFP_KERNEL); if (!p) return -ENOMEM; - memset(p, 0, sizeof *p); p->ibdev = device; p->port_num = port_num; diff --git a/drivers/infiniband/core/ucm.c b/drivers/infiniband/core/ucm.c index 28477565ecb..6e15787d1de 100644 --- a/drivers/infiniband/core/ucm.c +++ b/drivers/infiniband/core/ucm.c @@ -172,11 +172,10 @@ static struct ib_ucm_context *ib_ucm_ctx_alloc(struct ib_ucm_file *file) struct ib_ucm_context *ctx; int result; - ctx = kmalloc(sizeof(*ctx), GFP_KERNEL); + ctx = kzalloc(sizeof *ctx, GFP_KERNEL); if (!ctx) return NULL; - memset(ctx, 0, sizeof *ctx); atomic_set(&ctx->ref, 1); init_waitqueue_head(&ctx->wait); ctx->file = file; @@ -386,11 +385,10 @@ static int ib_ucm_event_handler(struct ib_cm_id *cm_id, ctx = cm_id->context; - uevent = kmalloc(sizeof(*uevent), GFP_KERNEL); + uevent = kzalloc(sizeof *uevent, GFP_KERNEL); if (!uevent) goto err1; - memset(uevent, 0, sizeof(*uevent)); uevent->ctx = ctx; uevent->cm_id = cm_id; uevent->resp.uid = ctx->uid; @@ -1345,11 +1343,10 @@ static void ib_ucm_add_one(struct ib_device *device) if (!device->alloc_ucontext) return; - ucm_dev = kmalloc(sizeof *ucm_dev, GFP_KERNEL); + ucm_dev = kzalloc(sizeof *ucm_dev, GFP_KERNEL); if (!ucm_dev) return; - memset(ucm_dev, 0, sizeof *ucm_dev); ucm_dev->ib_dev = device; ucm_dev->devnum = find_first_zero_bit(dev_map, IB_UCM_MAX_DEVICES); diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c index 97128e25f78..aed5ca23fb2 100644 --- a/drivers/infiniband/core/user_mad.c +++ b/drivers/infiniband/core/user_mad.c @@ -94,6 +94,9 @@ struct ib_umad_port { struct class_device *sm_class_dev; struct semaphore sm_sem; + struct rw_semaphore mutex; + struct list_head file_list; + struct ib_device *ib_dev; struct ib_umad_device *umad_dev; int dev_num; @@ -108,10 +111,10 @@ struct ib_umad_device { struct ib_umad_file { struct ib_umad_port *port; - spinlock_t recv_lock; struct list_head recv_list; + struct list_head port_list; + spinlock_t recv_lock; wait_queue_head_t recv_wait; - struct rw_semaphore agent_mutex; struct ib_mad_agent *agent[IB_UMAD_MAX_AGENTS]; struct ib_mr *mr[IB_UMAD_MAX_AGENTS]; }; @@ -148,7 +151,7 @@ static int queue_packet(struct ib_umad_file *file, { int ret = 1; - down_read(&file->agent_mutex); + down_read(&file->port->mutex); for (packet->mad.hdr.id = 0; packet->mad.hdr.id < IB_UMAD_MAX_AGENTS; packet->mad.hdr.id++) @@ -161,7 +164,7 @@ static int queue_packet(struct ib_umad_file *file, break; } - up_read(&file->agent_mutex); + up_read(&file->port->mutex); return ret; } @@ -322,7 +325,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, goto err; } - down_read(&file->agent_mutex); + down_read(&file->port->mutex); agent = file->agent[packet->mad.hdr.id]; if (!agent) { @@ -419,7 +422,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf, if (ret) goto err_msg; - up_read(&file->agent_mutex); + up_read(&file->port->mutex); return count; @@ -430,7 +433,7 @@ err_ah: ib_destroy_ah(ah); err_up: - up_read(&file->agent_mutex); + up_read(&file->port->mutex); err: kfree(packet); @@ -460,7 +463,12 @@ static int ib_umad_reg_agent(struct ib_umad_file *file, unsigned long arg) int agent_id; int ret; - down_write(&file->agent_mutex); + down_write(&file->port->mutex); + + if (!file->port->ib_dev) { + ret = -EPIPE; + goto out; + } if (copy_from_user(&ureq, (void __user *) arg, sizeof ureq)) { ret = -EFAULT; @@ -522,7 +530,7 @@ err: ib_unregister_mad_agent(agent); out: - up_write(&file->agent_mutex); + up_write(&file->port->mutex); return ret; } @@ -531,7 +539,7 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg) u32 id; int ret = 0; - down_write(&file->agent_mutex); + down_write(&file->port->mutex); if (get_user(id, (u32 __user *) arg)) { ret = -EFAULT; @@ -548,7 +556,7 @@ static int ib_umad_unreg_agent(struct ib_umad_file *file, unsigned long arg) file->agent[id] = NULL; out: - up_write(&file->agent_mutex); + up_write(&file->port->mutex); return ret; } @@ -569,6 +577,7 @@ static int ib_umad_open(struct inode *inode, struct file *filp) { struct ib_umad_port *port; struct ib_umad_file *file; + int ret = 0; spin_lock(&port_lock); port = umad_port[iminor(inode) - IB_UMAD_MINOR_BASE]; @@ -579,21 +588,32 @@ static int ib_umad_open(struct inode *inode, struct file *filp) if (!port) return -ENXIO; + down_write(&port->mutex); + + if (!port->ib_dev) { + ret = -ENXIO; + goto out; + } + file = kzalloc(sizeof *file, GFP_KERNEL); if (!file) { kref_put(&port->umad_dev->ref, ib_umad_release_dev); - return -ENOMEM; + ret = -ENOMEM; + goto out; } spin_lock_init(&file->recv_lock); - init_rwsem(&file->agent_mutex); INIT_LIST_HEAD(&file->recv_list); init_waitqueue_head(&file->recv_wait); file->port = port; filp->private_data = file; - return 0; + list_add_tail(&file->port_list, &port->file_list); + +out: + up_write(&port->mutex); + return ret; } static int ib_umad_close(struct inode *inode, struct file *filp) @@ -603,6 +623,7 @@ static int ib_umad_close(struct inode *inode, struct file *filp) struct ib_umad_packet *packet, *tmp; int i; + down_write(&file->port->mutex); for (i = 0; i < IB_UMAD_MAX_AGENTS; ++i) if (file->agent[i]) { ib_dereg_mr(file->mr[i]); @@ -612,6 +633,9 @@ static int ib_umad_close(struct inode *inode, struct file *filp) list_for_each_entry_safe(packet, tmp, &file->recv_list, list) kfree(packet); + list_del(&file->port_list); + up_write(&file->port->mutex); + kfree(file); kref_put(&dev->ref, ib_umad_release_dev); @@ -680,9 +704,13 @@ static int ib_umad_sm_close(struct inode *inode, struct file *filp) struct ib_port_modify props = { .clr_port_cap_mask = IB_PORT_SM }; - int ret; + int ret = 0; + + down_write(&port->mutex); + if (port->ib_dev) + ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props); + up_write(&port->mutex); - ret = ib_modify_port(port->ib_dev, port->port_num, 0, &props); up(&port->sm_sem); kref_put(&port->umad_dev->ref, ib_umad_release_dev); @@ -745,6 +773,8 @@ static int ib_umad_init_port(struct ib_device *device, int port_num, port->ib_dev = device; port->port_num = port_num; init_MUTEX(&port->sm_sem); + init_rwsem(&port->mutex); + INIT_LIST_HEAD(&port->file_list); port->dev = cdev_alloc(); if (!port->dev) @@ -813,6 +843,9 @@ err_cdev: static void ib_umad_kill_port(struct ib_umad_port *port) { + struct ib_umad_file *file; + int id; + class_set_devdata(port->class_dev, NULL); class_set_devdata(port->sm_class_dev, NULL); @@ -826,6 +859,21 @@ static void ib_umad_kill_port(struct ib_umad_port *port) umad_port[port->dev_num] = NULL; spin_unlock(&port_lock); + down_write(&port->mutex); + + port->ib_dev = NULL; + + list_for_each_entry(file, &port->file_list, port_list) + for (id = 0; id < IB_UMAD_MAX_AGENTS; ++id) { + if (!file->agent[id]) + continue; + ib_dereg_mr(file->mr[id]); + ib_unregister_mad_agent(file->agent[id]); + file->agent[id] = NULL; + } + + up_write(&port->mutex); + clear_bit(port->dev_num, dev_map); } diff --git a/drivers/infiniband/core/uverbs.h b/drivers/infiniband/core/uverbs.h index 031cdf3c066..ecb83012786 100644 --- a/drivers/infiniband/core/uverbs.h +++ b/drivers/infiniband/core/uverbs.h @@ -113,6 +113,7 @@ struct ib_uevent_object { struct ib_ucq_object { struct ib_uobject uobject; + struct ib_uverbs_file *uverbs_file; struct list_head comp_list; struct list_head async_list; u32 comp_events_reported; diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index 8c89abc8c76..63a74151c60 100644 --- a/drivers/infiniband/core/uverbs_cmd.c +++ b/drivers/infiniband/core/uverbs_cmd.c @@ -602,6 +602,7 @@ ssize_t ib_uverbs_create_cq(struct ib_uverbs_file *file, uobj->uobject.user_handle = cmd.user_handle; uobj->uobject.context = file->ucontext; + uobj->uverbs_file = file; uobj->comp_events_reported = 0; uobj->async_events_reported = 0; INIT_LIST_HEAD(&uobj->comp_list); diff --git a/drivers/infiniband/core/uverbs_main.c b/drivers/infiniband/core/uverbs_main.c index 0eb38f479b3..de6581d7cb8 100644 --- a/drivers/infiniband/core/uverbs_main.c +++ b/drivers/infiniband/core/uverbs_main.c @@ -442,13 +442,10 @@ static void ib_uverbs_async_handler(struct ib_uverbs_file *file, void ib_uverbs_cq_event_handler(struct ib_event *event, void *context_ptr) { - struct ib_uverbs_event_file *ev_file = context_ptr; - struct ib_ucq_object *uobj; + struct ib_ucq_object *uobj = container_of(event->element.cq->uobject, + struct ib_ucq_object, uobject); - uobj = container_of(event->element.cq->uobject, - struct ib_ucq_object, uobject); - - ib_uverbs_async_handler(ev_file->uverbs_file, uobj->uobject.user_handle, + ib_uverbs_async_handler(uobj->uverbs_file, uobj->uobject.user_handle, event->event, &uobj->async_list, &uobj->async_events_reported); @@ -728,12 +725,10 @@ static void ib_uverbs_add_one(struct ib_device *device) if (!device->alloc_ucontext) return; - uverbs_dev = kmalloc(sizeof *uverbs_dev, GFP_KERNEL); + uverbs_dev = kzalloc(sizeof *uverbs_dev, GFP_KERNEL); if (!uverbs_dev) return; - memset(uverbs_dev, 0, sizeof *uverbs_dev); - kref_init(&uverbs_dev->ref); spin_lock(&map_lock); diff --git a/drivers/infiniband/hw/mthca/mthca_cq.c b/drivers/infiniband/hw/mthca/mthca_cq.c index 8600b6c3e0c..f98e2355582 100644 --- a/drivers/infiniband/hw/mthca/mthca_cq.c +++ b/drivers/infiniband/hw/mthca/mthca_cq.c @@ -208,7 +208,7 @@ static inline void update_cons_index(struct mthca_dev *dev, struct mthca_cq *cq, } } -void mthca_cq_event(struct mthca_dev *dev, u32 cqn) +void mthca_cq_completion(struct mthca_dev *dev, u32 cqn) { struct mthca_cq *cq; @@ -224,6 +224,35 @@ void mthca_cq_event(struct mthca_dev *dev, u32 cqn) cq->ibcq.comp_handler(&cq->ibcq, cq->ibcq.cq_context); } +void mthca_cq_event(struct mthca_dev *dev, u32 cqn, + enum ib_event_type event_type) +{ + struct mthca_cq *cq; + struct ib_event event; + + spin_lock(&dev->cq_table.lock); + + cq = mthca_array_get(&dev->cq_table.cq, cqn & (dev->limits.num_cqs - 1)); + + if (cq) + atomic_inc(&cq->refcount); + spin_unlock(&dev->cq_table.lock); + + if (!cq) { + mthca_warn(dev, "Async event for bogus CQ %08x\n", cqn); + return; + } + + event.device = &dev->ib_dev; + event.event = event_type; + event.element.cq = &cq->ibcq; + if (cq->ibcq.event_handler) + cq->ibcq.event_handler(&event, cq->ibcq.cq_context); + + if (atomic_dec_and_test(&cq->refcount)) + wake_up(&cq->wait); +} + void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn, struct mthca_srq *srq) { diff --git a/drivers/infiniband/hw/mthca/mthca_dev.h b/drivers/infiniband/hw/mthca/mthca_dev.h index 7e68bd4a378..e7e5d3b4f00 100644 --- a/drivers/infiniband/hw/mthca/mthca_dev.h +++ b/drivers/infiniband/hw/mthca/mthca_dev.h @@ -460,7 +460,9 @@ int mthca_init_cq(struct mthca_dev *dev, int nent, struct mthca_cq *cq); void mthca_free_cq(struct mthca_dev *dev, struct mthca_cq *cq); -void mthca_cq_event(struct mthca_dev *dev, u32 cqn); +void mthca_cq_completion(struct mthca_dev *dev, u32 cqn); +void mthca_cq_event(struct mthca_dev *dev, u32 cqn, + enum ib_event_type event_type); void mthca_cq_clean(struct mthca_dev *dev, u32 cqn, u32 qpn, struct mthca_srq *srq); diff --git a/drivers/infiniband/hw/mthca/mthca_eq.c b/drivers/infiniband/hw/mthca/mthca_eq.c index e5a047a6dbe..34d68e5a72d 100644 --- a/drivers/infiniband/hw/mthca/mthca_eq.c +++ b/drivers/infiniband/hw/mthca/mthca_eq.c @@ -292,7 +292,7 @@ static int mthca_eq_int(struct mthca_dev *dev, struct mthca_eq *eq) case MTHCA_EVENT_TYPE_COMP: disarm_cqn = be32_to_cpu(eqe->event.comp.cqn) & 0xffffff; disarm_cq(dev, eq->eqn, disarm_cqn); - mthca_cq_event(dev, disarm_cqn); + mthca_cq_completion(dev, disarm_cqn); break; case MTHCA_EVENT_TYPE_PATH_MIG: @@ -364,6 +364,8 @@ static int mthca_eq_int(struct mthca_dev *dev, struct mthca_eq *eq) eqe->event.cq_err.syndrome == 1 ? "overrun" : "access violation", be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff); + mthca_cq_event(dev, be32_to_cpu(eqe->event.cq_err.cqn), + IB_EVENT_CQ_ERR); break; case MTHCA_EVENT_TYPE_EQ_OVERFLOW: diff --git a/drivers/infiniband/hw/mthca/mthca_main.c b/drivers/infiniband/hw/mthca/mthca_main.c index 883d1e5a79b..45c6328e780 100644 --- a/drivers/infiniband/hw/mthca/mthca_main.c +++ b/drivers/infiniband/hw/mthca/mthca_main.c @@ -1057,7 +1057,7 @@ static int __devinit mthca_init_one(struct pci_dev *pdev, goto err_cmd; if (mdev->fw_ver < mthca_hca_table[id->driver_data].latest_fw) { - mthca_warn(mdev, "HCA FW version %x.%x.%x is old (%x.%x.%x is current).\n", + mthca_warn(mdev, "HCA FW version %d.%d.%d is old (%d.%d.%d is current).\n", (int) (mdev->fw_ver >> 32), (int) (mdev->fw_ver >> 16) & 0xffff, (int) (mdev->fw_ver & 0xffff), (int) (mthca_hca_table[id->driver_data].latest_fw >> 32), diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c b/drivers/infiniband/hw/mthca/mthca_mr.c index 1f97a44477f..e995e2aa016 100644 --- a/drivers/infiniband/hw/mthca/mthca_mr.c +++ b/drivers/infiniband/hw/mthca/mthca_mr.c @@ -140,13 +140,11 @@ static int __devinit mthca_buddy_init(struct mthca_buddy *buddy, int max_order) buddy->max_order = max_order; spin_lock_init(&buddy->lock); - buddy->bits = kmalloc((buddy->max_order + 1) * sizeof (long *), + buddy->bits = kzalloc((buddy->max_order + 1) * sizeof (long *), GFP_KERNEL); if (!buddy->bits) goto err_out; - memset(buddy->bits, 0, (buddy->max_order + 1) * sizeof (long *)); - for (i = 0; i <= buddy->max_order; ++i) { s = BITS_TO_LONGS(1 << (buddy->max_order - i)); buddy->bits[i] = kmalloc(s * sizeof (long), GFP_KERNEL); diff --git a/drivers/infiniband/hw/mthca/mthca_profile.c b/drivers/infiniband/hw/mthca/mthca_profile.c index bd133868207..08a909371b0 100644 --- a/drivers/infiniband/hw/mthca/mthca_profile.c +++ b/drivers/infiniband/hw/mthca/mthca_profile.c @@ -82,12 +82,10 @@ u64 mthca_make_profile(struct mthca_dev *dev, struct mthca_resource tmp; int i, j; - profile = kmalloc(MTHCA_RES_NUM * sizeof *profile, GFP_KERNEL); + profile = kzalloc(MTHCA_RES_NUM * sizeof *profile, GFP_KERNEL); if (!profile) return -ENOMEM; - memset(profile, 0, MTHCA_RES_NUM * sizeof *profile); - profile[MTHCA_RES_QP].size = dev_lim->qpc_entry_sz; profile[MTHCA_RES_EEC].size = dev_lim->eec_entry_sz; profile[MTHCA_RES_SRQ].size = dev_lim->srq_entry_sz; diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c index 1b9477edbd7..6b016666826 100644 --- a/drivers/infiniband/hw/mthca/mthca_provider.c +++ b/drivers/infiniband/hw/mthca/mthca_provider.c @@ -1028,7 +1028,7 @@ static ssize_t show_rev(struct class_device *cdev, char *buf) static ssize_t show_fw_ver(struct class_device *cdev, char *buf) { struct mthca_dev *dev = container_of(cdev, struct mthca_dev, ib_dev.class_dev); - return sprintf(buf, "%x.%x.%x\n", (int) (dev->fw_ver >> 32), + return sprintf(buf, "%d.%d.%d\n", (int) (dev->fw_ver >> 32), (int) (dev->fw_ver >> 16) & 0xffff, (int) dev->fw_ver & 0xffff); } diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c b/drivers/infiniband/hw/mthca/mthca_qp.c index 7c9afde5ace..8852ea477c2 100644 --- a/drivers/infiniband/hw/mthca/mthca_qp.c +++ b/drivers/infiniband/hw/mthca/mthca_qp.c @@ -584,6 +584,13 @@ int mthca_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr, int attr_mask) return -EINVAL; } + if ((attr_mask & IB_QP_PKEY_INDEX) && + attr->pkey_index >= dev->limits.pkey_table_len) { + mthca_dbg(dev, "PKey index (%u) too large. max is %d\n", + attr->pkey_index,dev->limits.pkey_table_len-1); + return -EINVAL; + } + mailbox = mthca_alloc_mailbox(dev, GFP_KERNEL); if (IS_ERR(mailbox)) return PTR_ERR(mailbox); diff --git a/drivers/infiniband/hw/mthca/mthca_srq.c b/drivers/infiniband/hw/mthca/mthca_srq.c index 64f70aa1b3c..292f55be8cb 100644 --- a/drivers/infiniband/hw/mthca/mthca_srq.c +++ b/drivers/infiniband/hw/mthca/mthca_srq.c @@ -75,15 +75,16 @@ static void *get_wqe(struct mthca_srq *srq, int n) /* * Return a pointer to the location within a WQE that we're using as a - * link when the WQE is in the free list. We use an offset of 4 - * because in the Tavor case, posting a WQE may overwrite the first - * four bytes of the previous WQE. The offset avoids corrupting our - * free list if the WQE has already completed and been put on the free - * list when we post the next WQE. + * link when the WQE is in the free list. We use the imm field + * because in the Tavor case, posting a WQE may overwrite the next + * segment of the previous WQE, but a receive WQE will never touch the + * imm field. This avoids corrupting our free list if the previous + * WQE has already completed and been put on the free list when we + * post the next WQE. */ static inline int *wqe_to_link(void *wqe) { - return (int *) (wqe + 4); + return (int *) (wqe + offsetof(struct mthca_next_seg, imm)); } static void mthca_tavor_init_srq_context(struct mthca_dev *dev, diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h index c994a916a58..0095acc0fbb 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib.h +++ b/drivers/infiniband/ulp/ipoib/ipoib.h @@ -235,6 +235,7 @@ static inline void ipoib_put_ah(struct ipoib_ah *ah) kref_put(&ah->ref, ipoib_free_ah); } +int ipoib_open(struct net_device *dev); int ipoib_add_pkey_attr(struct net_device *dev); void ipoib_send(struct net_device *dev, struct sk_buff *skb, @@ -267,6 +268,7 @@ int ipoib_mcast_stop_thread(struct net_device *dev, int flush); void ipoib_mcast_dev_down(struct net_device *dev); void ipoib_mcast_dev_flush(struct net_device *dev); +#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev); void ipoib_mcast_iter_free(struct ipoib_mcast_iter *iter); int ipoib_mcast_iter_next(struct ipoib_mcast_iter *iter); @@ -276,6 +278,7 @@ void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter, unsigned int *queuelen, unsigned int *complete, unsigned int *send_only); +#endif int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid); diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c index 192fef884e2..54ef2fea530 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c @@ -486,15 +486,16 @@ int ipoib_ib_dev_stop(struct net_device *dev) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_attr qp_attr; - int attr_mask; unsigned long begin; struct ipoib_tx_buf *tx_req; int i; - /* Kill the existing QP and allocate a new one */ + /* + * Move our QP to the error state and then reinitialize in + * when all work requests have completed or have been flushed. + */ qp_attr.qp_state = IB_QPS_ERR; - attr_mask = IB_QP_STATE; - if (ib_modify_qp(priv->qp, &qp_attr, attr_mask)) + if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE)) ipoib_warn(priv, "Failed to modify QP to ERROR state\n"); /* Wait for all sends and receives to complete */ @@ -541,8 +542,7 @@ int ipoib_ib_dev_stop(struct net_device *dev) timeout: qp_attr.qp_state = IB_QPS_RESET; - attr_mask = IB_QP_STATE; - if (ib_modify_qp(priv->qp, &qp_attr, attr_mask)) + if (ib_modify_qp(priv->qp, &qp_attr, IB_QP_STATE)) ipoib_warn(priv, "Failed to modify QP to RESET state\n"); /* Wait for all AHs to be reaped */ @@ -636,7 +636,6 @@ void ipoib_ib_dev_cleanup(struct net_device *dev) * Bug #2507. This implementation will probably be removed when the P_Key * change async notification is available. */ -int ipoib_open(struct net_device *dev); static void ipoib_pkey_dev_check_presence(struct net_device *dev) { diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c index cd4f42328db..ce0296273e7 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_main.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c @@ -356,18 +356,15 @@ static struct ipoib_path *path_rec_create(struct net_device *dev, struct ipoib_dev_priv *priv = netdev_priv(dev); struct ipoib_path *path; - path = kmalloc(sizeof *path, GFP_ATOMIC); + path = kzalloc(sizeof *path, GFP_ATOMIC); if (!path) return NULL; - path->dev = dev; - path->pathrec.dlid = 0; - path->ah = NULL; + path->dev = dev; skb_queue_head_init(&path->queue); INIT_LIST_HEAD(&path->neigh_list); - path->query = NULL; init_completion(&path->done); memcpy(path->pathrec.dgid.raw, gid->raw, sizeof (union ib_gid)); @@ -551,11 +548,8 @@ static int ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) struct ipoib_neigh *neigh; unsigned long flags; - local_irq_save(flags); - if (!spin_trylock(&priv->tx_lock)) { - local_irq_restore(flags); + if (!spin_trylock_irqsave(&priv->tx_lock, flags)) return NETDEV_TX_LOCKED; - } /* * Check if our queue is stopped. Since we have the LLTX bit @@ -732,25 +726,21 @@ int ipoib_dev_init(struct net_device *dev, struct ib_device *ca, int port) /* Allocate RX/TX "rings" to hold queued skbs */ - priv->rx_ring = kmalloc(IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf), + priv->rx_ring = kzalloc(IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf), GFP_KERNEL); if (!priv->rx_ring) { printk(KERN_WARNING "%s: failed to allocate RX ring (%d entries)\n", ca->name, IPOIB_RX_RING_SIZE); goto out; } - memset(priv->rx_ring, 0, - IPOIB_RX_RING_SIZE * sizeof (struct ipoib_rx_buf)); - priv->tx_ring = kmalloc(IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf), + priv->tx_ring = kzalloc(IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf), GFP_KERNEL); if (!priv->tx_ring) { printk(KERN_WARNING "%s: failed to allocate TX ring (%d entries)\n", ca->name, IPOIB_TX_RING_SIZE); goto out_rx_ring_cleanup; } - memset(priv->tx_ring, 0, - IPOIB_TX_RING_SIZE * sizeof (struct ipoib_tx_buf)); /* priv->tx_head & tx_tail are already 0 */ @@ -807,10 +797,6 @@ static void ipoib_setup(struct net_device *dev) dev->watchdog_timeo = HZ; - dev->rebuild_header = NULL; - dev->set_mac_address = NULL; - dev->header_cache_update = NULL; - dev->flags |= IFF_BROADCAST | IFF_MULTICAST; /* diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c index 36ce29836bf..3ecf78a9493 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c @@ -135,12 +135,10 @@ static struct ipoib_mcast *ipoib_mcast_alloc(struct net_device *dev, { struct ipoib_mcast *mcast; - mcast = kmalloc(sizeof (*mcast), can_sleep ? GFP_KERNEL : GFP_ATOMIC); + mcast = kzalloc(sizeof *mcast, can_sleep ? GFP_KERNEL : GFP_ATOMIC); if (!mcast) return NULL; - memset(mcast, 0, sizeof (*mcast)); - init_completion(&mcast->done); mcast->dev = dev; @@ -919,6 +917,8 @@ void ipoib_mcast_restart_task(void *dev_ptr) ipoib_mcast_start_thread(dev); } +#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG + struct ipoib_mcast_iter *ipoib_mcast_iter_init(struct net_device *dev) { struct ipoib_mcast_iter *iter; @@ -991,3 +991,5 @@ void ipoib_mcast_iter_read(struct ipoib_mcast_iter *iter, *complete = iter->complete; *send_only = iter->send_only; } + +#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */ diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c index b5902a7ec24..e829e10400e 100644 --- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c +++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c @@ -41,7 +41,6 @@ int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid) { struct ipoib_dev_priv *priv = netdev_priv(dev); struct ib_qp_attr *qp_attr; - int attr_mask; int ret; u16 pkey_index; @@ -59,8 +58,7 @@ int ipoib_mcast_attach(struct net_device *dev, u16 mlid, union ib_gid *mgid) /* set correct QKey for QP */ qp_attr->qkey = priv->qkey; - attr_mask = IB_QP_QKEY; - ret = ib_modify_qp(priv->qp, qp_attr, attr_mask); + ret = ib_modify_qp(priv->qp, qp_attr, IB_QP_QKEY); if (ret) { ipoib_warn(priv, "failed to modify QP, ret = %d\n", ret); goto out; diff --git a/drivers/infiniband/ulp/srp/Kbuild b/drivers/infiniband/ulp/srp/Kbuild new file mode 100644 index 00000000000..a16c73c667c --- /dev/null +++ b/drivers/infiniband/ulp/srp/Kbuild @@ -0,0 +1 @@ +obj-$(CONFIG_INFINIBAND_SRP) += ib_srp.o diff --git a/drivers/infiniband/ulp/srp/Kconfig b/drivers/infiniband/ulp/srp/Kconfig new file mode 100644 index 00000000000..8fe3be4e991 --- /dev/null +++ b/drivers/infiniband/ulp/srp/Kconfig @@ -0,0 +1,11 @@ +config INFINIBAND_SRP + tristate "InfiniBand SCSI RDMA Protocol" + depends on INFINIBAND && SCSI + ---help--- + Support for the SCSI RDMA Protocol over InfiniBand. This + allows you to access storage devices that speak SRP over + InfiniBand. + + The SRP protocol is defined by the INCITS T10 technical + committee. See <http://www.t10.org/>. + diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c new file mode 100644 index 00000000000..2687e34aa5b --- /dev/null +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -0,0 +1,1700 @@ +/* + * Copyright (c) 2005 Cisco Systems. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * 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 AUTHORS OR COPYRIGHT HOLDERS + * 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. + * + * $Id: ib_srp.c 3932 2005-11-01 17:19:29Z roland $ + */ + +#include <linux/version.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/err.h> +#include <linux/string.h> +#include <linux/parser.h> +#include <linux/random.h> + +#include <asm/atomic.h> + +#include <scsi/scsi.h> +#include <scsi/scsi_device.h> +#include <scsi/scsi_dbg.h> +#include <scsi/srp.h> + +#include <rdma/ib_cache.h> + +#include "ib_srp.h" + +#define DRV_NAME "ib_srp" +#define PFX DRV_NAME ": " +#define DRV_VERSION "0.2" +#define DRV_RELDATE "November 1, 2005" + +MODULE_AUTHOR("Roland Dreier"); +MODULE_DESCRIPTION("InfiniBand SCSI RDMA Protocol initiator " + "v" DRV_VERSION " (" DRV_RELDATE ")"); +MODULE_LICENSE("Dual BSD/GPL"); + +static int topspin_workarounds = 1; + +module_param(topspin_workarounds, int, 0444); +MODULE_PARM_DESC(topspin_workarounds, + "Enable workarounds for Topspin/Cisco SRP target bugs if != 0"); + +static const u8 topspin_oui[3] = { 0x00, 0x05, 0xad }; + +static void srp_add_one(struct ib_device *device); +static void srp_remove_one(struct ib_device *device); +static void srp_completion(struct ib_cq *cq, void *target_ptr); +static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event); + +static struct ib_client srp_client = { + .name = "srp", + .add = srp_add_one, + .remove = srp_remove_one +}; + +static inline struct srp_target_port *host_to_target(struct Scsi_Host *host) +{ + return (struct srp_target_port *) host->hostdata; +} + +static const char *srp_target_info(struct Scsi_Host *host) +{ + return host_to_target(host)->target_name; +} + +static struct srp_iu *srp_alloc_iu(struct srp_host *host, size_t size, + gfp_t gfp_mask, + enum dma_data_direction direction) +{ + struct srp_iu *iu; + + iu = kmalloc(sizeof *iu, gfp_mask); + if (!iu) + goto out; + + iu->buf = kzalloc(size, gfp_mask); + if (!iu->buf) + goto out_free_iu; + + iu->dma = dma_map_single(host->dev->dma_device, iu->buf, size, direction); + if (dma_mapping_error(iu->dma)) + goto out_free_buf; + + iu->size = size; + iu->direction = direction; + + return iu; + +out_free_buf: + kfree(iu->buf); +out_free_iu: + kfree(iu); +out: + return NULL; +} + +static void srp_free_iu(struct srp_host *host, struct srp_iu *iu) +{ + if (!iu) + return; + + dma_unmap_single(host->dev->dma_device, iu->dma, iu->size, iu->direction); + kfree(iu->buf); + kfree(iu); +} + +static void srp_qp_event(struct ib_event *event, void *context) +{ + printk(KERN_ERR PFX "QP event %d\n", event->event); +} + +static int srp_init_qp(struct srp_target_port *target, + struct ib_qp *qp) +{ + struct ib_qp_attr *attr; + int ret; + + attr = kmalloc(sizeof *attr, GFP_KERNEL); + if (!attr) + return -ENOMEM; + + ret = ib_find_cached_pkey(target->srp_host->dev, + target->srp_host->port, + be16_to_cpu(target->path.pkey), + &attr->pkey_index); + if (ret) + goto out; + + attr->qp_state = IB_QPS_INIT; + attr->qp_access_flags = (IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE); + attr->port_num = target->srp_host->port; + + ret = ib_modify_qp(qp, attr, + IB_QP_STATE | + IB_QP_PKEY_INDEX | + IB_QP_ACCESS_FLAGS | + IB_QP_PORT); + +out: + kfree(attr); + return ret; +} + +static int srp_create_target_ib(struct srp_target_port *target) +{ + struct ib_qp_init_attr *init_attr; + int ret; + + init_attr = kzalloc(sizeof *init_attr, GFP_KERNEL); + if (!init_attr) + return -ENOMEM; + + target->cq = ib_create_cq(target->srp_host->dev, srp_completion, + NULL, target, SRP_CQ_SIZE); + if (IS_ERR(target->cq)) { + ret = PTR_ERR(target->cq); + goto out; + } + + ib_req_notify_cq(target->cq, IB_CQ_NEXT_COMP); + + init_attr->event_handler = srp_qp_event; + init_attr->cap.max_send_wr = SRP_SQ_SIZE; + init_attr->cap.max_recv_wr = SRP_RQ_SIZE; + init_attr->cap.max_recv_sge = 1; + init_attr->cap.max_send_sge = 1; + init_attr->sq_sig_type = IB_SIGNAL_ALL_WR; + init_attr->qp_type = IB_QPT_RC; + init_attr->send_cq = target->cq; + init_attr->recv_cq = target->cq; + + target->qp = ib_create_qp(target->srp_host->pd, init_attr); + if (IS_ERR(target->qp)) { + ret = PTR_ERR(target->qp); + ib_destroy_cq(target->cq); + goto out; + } + + ret = srp_init_qp(target, target->qp); + if (ret) { + ib_destroy_qp(target->qp); + ib_destroy_cq(target->cq); + goto out; + } + +out: + kfree(init_attr); + return ret; +} + +static void srp_free_target_ib(struct srp_target_port *target) +{ + int i; + + ib_destroy_qp(target->qp); + ib_destroy_cq(target->cq); + + for (i = 0; i < SRP_RQ_SIZE; ++i) + srp_free_iu(target->srp_host, target->rx_ring[i]); + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) + srp_free_iu(target->srp_host, target->tx_ring[i]); +} + +static void srp_path_rec_completion(int status, + struct ib_sa_path_rec *pathrec, + void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + target->status = status; + if (status) + printk(KERN_ERR PFX "Got failed path rec status %d\n", status); + else + target->path = *pathrec; + complete(&target->done); +} + +static int srp_lookup_path(struct srp_target_port *target) +{ + target->path.numb_path = 1; + + init_completion(&target->done); + + target->path_query_id = ib_sa_path_rec_get(target->srp_host->dev, + target->srp_host->port, + &target->path, + IB_SA_PATH_REC_DGID | + IB_SA_PATH_REC_SGID | + IB_SA_PATH_REC_NUMB_PATH | + IB_SA_PATH_REC_PKEY, + SRP_PATH_REC_TIMEOUT_MS, + GFP_KERNEL, + srp_path_rec_completion, + target, &target->path_query); + if (target->path_query_id < 0) + return target->path_query_id; + + wait_for_completion(&target->done); + + if (target->status < 0) + printk(KERN_WARNING PFX "Path record query failed\n"); + + return target->status; +} + +static int srp_send_req(struct srp_target_port *target) +{ + struct { + struct ib_cm_req_param param; + struct srp_login_req priv; + } *req = NULL; + int status; + + req = kzalloc(sizeof *req, GFP_KERNEL); + if (!req) + return -ENOMEM; + + req->param.primary_path = &target->path; + req->param.alternate_path = NULL; + req->param.service_id = target->service_id; + req->param.qp_num = target->qp->qp_num; + req->param.qp_type = target->qp->qp_type; + req->param.private_data = &req->priv; + req->param.private_data_len = sizeof req->priv; + req->param.flow_control = 1; + + get_random_bytes(&req->param.starting_psn, 4); + req->param.starting_psn &= 0xffffff; + + /* + * Pick some arbitrary defaults here; we could make these + * module parameters if anyone cared about setting them. + */ + req->param.responder_resources = 4; + req->param.remote_cm_response_timeout = 20; + req->param.local_cm_response_timeout = 20; + req->param.retry_count = 7; + req->param.rnr_retry_count = 7; + req->param.max_cm_retries = 15; + + req->priv.opcode = SRP_LOGIN_REQ; + req->priv.tag = 0; + req->priv.req_it_iu_len = cpu_to_be32(SRP_MAX_IU_LEN); + req->priv.req_buf_fmt = cpu_to_be16(SRP_BUF_FORMAT_DIRECT | + SRP_BUF_FORMAT_INDIRECT); + memcpy(req->priv.initiator_port_id, target->srp_host->initiator_port_id, 16); + /* + * Topspin/Cisco SRP targets will reject our login unless we + * zero out the first 8 bytes of our initiator port ID. The + * second 8 bytes must be our local node GUID, but we always + * use that anyway. + */ + if (topspin_workarounds && !memcmp(&target->ioc_guid, topspin_oui, 3)) { + printk(KERN_DEBUG PFX "Topspin/Cisco initiator port ID workaround " + "activated for target GUID %016llx\n", + (unsigned long long) be64_to_cpu(target->ioc_guid)); + memset(req->priv.initiator_port_id, 0, 8); + } + memcpy(req->priv.target_port_id, &target->id_ext, 8); + memcpy(req->priv.target_port_id + 8, &target->ioc_guid, 8); + + status = ib_send_cm_req(target->cm_id, &req->param); + + kfree(req); + + return status; +} + +static void srp_disconnect_target(struct srp_target_port *target) +{ + /* XXX should send SRP_I_LOGOUT request */ + + init_completion(&target->done); + ib_send_cm_dreq(target->cm_id, NULL, 0); + wait_for_completion(&target->done); +} + +static void srp_remove_work(void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state != SRP_TARGET_DEAD) { + spin_unlock_irq(target->scsi_host->host_lock); + scsi_host_put(target->scsi_host); + return; + } + target->state = SRP_TARGET_REMOVED; + spin_unlock_irq(target->scsi_host->host_lock); + + down(&target->srp_host->target_mutex); + list_del(&target->list); + up(&target->srp_host->target_mutex); + + scsi_remove_host(target->scsi_host); + ib_destroy_cm_id(target->cm_id); + srp_free_target_ib(target); + scsi_host_put(target->scsi_host); + /* And another put to really free the target port... */ + scsi_host_put(target->scsi_host); +} + +static int srp_connect_target(struct srp_target_port *target) +{ + int ret; + + ret = srp_lookup_path(target); + if (ret) + return ret; + + while (1) { + init_completion(&target->done); + ret = srp_send_req(target); + if (ret) + return ret; + wait_for_completion(&target->done); + + /* + * The CM event handling code will set status to + * SRP_PORT_REDIRECT if we get a port redirect REJ + * back, or SRP_DLID_REDIRECT if we get a lid/qp + * redirect REJ back. + */ + switch (target->status) { + case 0: + return 0; + + case SRP_PORT_REDIRECT: + ret = srp_lookup_path(target); + if (ret) + return ret; + break; + + case SRP_DLID_REDIRECT: + break; + + default: + return target->status; + } + } +} + +static int srp_reconnect_target(struct srp_target_port *target) +{ + struct ib_cm_id *new_cm_id; + struct ib_qp_attr qp_attr; + struct srp_request *req; + struct ib_wc wc; + int ret; + int i; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state != SRP_TARGET_LIVE) { + spin_unlock_irq(target->scsi_host->host_lock); + return -EAGAIN; + } + target->state = SRP_TARGET_CONNECTING; + spin_unlock_irq(target->scsi_host->host_lock); + + srp_disconnect_target(target); + /* + * Now get a new local CM ID so that we avoid confusing the + * target in case things are really fouled up. + */ + new_cm_id = ib_create_cm_id(target->srp_host->dev, + srp_cm_handler, target); + if (IS_ERR(new_cm_id)) { + ret = PTR_ERR(new_cm_id); + goto err; + } + ib_destroy_cm_id(target->cm_id); + target->cm_id = new_cm_id; + + qp_attr.qp_state = IB_QPS_RESET; + ret = ib_modify_qp(target->qp, &qp_attr, IB_QP_STATE); + if (ret) + goto err; + + ret = srp_init_qp(target, target->qp); + if (ret) + goto err; + + while (ib_poll_cq(target->cq, 1, &wc) > 0) + ; /* nothing */ + + list_for_each_entry(req, &target->req_queue, list) { + req->scmnd->result = DID_RESET << 16; + req->scmnd->scsi_done(req->scmnd); + } + + target->rx_head = 0; + target->tx_head = 0; + target->tx_tail = 0; + target->req_head = 0; + for (i = 0; i < SRP_SQ_SIZE - 1; ++i) + target->req_ring[i].next = i + 1; + target->req_ring[SRP_SQ_SIZE - 1].next = -1; + INIT_LIST_HEAD(&target->req_queue); + + ret = srp_connect_target(target); + if (ret) + goto err; + + spin_lock_irq(target->scsi_host->host_lock); + if (target->state == SRP_TARGET_CONNECTING) { + ret = 0; + target->state = SRP_TARGET_LIVE; + } else + ret = -EAGAIN; + spin_unlock_irq(target->scsi_host->host_lock); + + return ret; + +err: + printk(KERN_ERR PFX "reconnect failed (%d), removing target port.\n", ret); + + /* + * We couldn't reconnect, so kill our target port off. + * However, we have to defer the real removal because we might + * be in the context of the SCSI error handler now, which + * would deadlock if we call scsi_remove_host(). + */ + spin_lock_irq(target->scsi_host->host_lock); + if (target->state == SRP_TARGET_CONNECTING) { + target->state = SRP_TARGET_DEAD; + INIT_WORK(&target->work, srp_remove_work, target); + schedule_work(&target->work); + } + spin_unlock_irq(target->scsi_host->host_lock); + + return ret; +} + +static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target, + struct srp_request *req) +{ + struct srp_cmd *cmd = req->cmd->buf; + int len; + u8 fmt; + + if (!scmnd->request_buffer || scmnd->sc_data_direction == DMA_NONE) + return sizeof (struct srp_cmd); + + if (scmnd->sc_data_direction != DMA_FROM_DEVICE && + scmnd->sc_data_direction != DMA_TO_DEVICE) { + printk(KERN_WARNING PFX "Unhandled data direction %d\n", + scmnd->sc_data_direction); + return -EINVAL; + } + + if (scmnd->use_sg) { + struct scatterlist *scat = scmnd->request_buffer; + int n; + int i; + + n = dma_map_sg(target->srp_host->dev->dma_device, + scat, scmnd->use_sg, scmnd->sc_data_direction); + + if (n == 1) { + struct srp_direct_buf *buf = (void *) cmd->add_data; + + fmt = SRP_DATA_DESC_DIRECT; + + buf->va = cpu_to_be64(sg_dma_address(scat)); + buf->key = cpu_to_be32(target->srp_host->mr->rkey); + buf->len = cpu_to_be32(sg_dma_len(scat)); + + len = sizeof (struct srp_cmd) + + sizeof (struct srp_direct_buf); + } else { + struct srp_indirect_buf *buf = (void *) cmd->add_data; + u32 datalen = 0; + + fmt = SRP_DATA_DESC_INDIRECT; + + if (scmnd->sc_data_direction == DMA_TO_DEVICE) + cmd->data_out_desc_cnt = n; + else + cmd->data_in_desc_cnt = n; + + buf->table_desc.va = cpu_to_be64(req->cmd->dma + + sizeof *cmd + + sizeof *buf); + buf->table_desc.key = + cpu_to_be32(target->srp_host->mr->rkey); + buf->table_desc.len = + cpu_to_be32(n * sizeof (struct srp_direct_buf)); + + for (i = 0; i < n; ++i) { + buf->desc_list[i].va = cpu_to_be64(sg_dma_address(&scat[i])); + buf->desc_list[i].key = + cpu_to_be32(target->srp_host->mr->rkey); + buf->desc_list[i].len = cpu_to_be32(sg_dma_len(&scat[i])); + + datalen += sg_dma_len(&scat[i]); + } + + buf->len = cpu_to_be32(datalen); + + len = sizeof (struct srp_cmd) + + sizeof (struct srp_indirect_buf) + + n * sizeof (struct srp_direct_buf); + } + } else { + struct srp_direct_buf *buf = (void *) cmd->add_data; + dma_addr_t dma; + + dma = dma_map_single(target->srp_host->dev->dma_device, + scmnd->request_buffer, scmnd->request_bufflen, + scmnd->sc_data_direction); + if (dma_mapping_error(dma)) { + printk(KERN_WARNING PFX "unable to map %p/%d (dir %d)\n", + scmnd->request_buffer, (int) scmnd->request_bufflen, + scmnd->sc_data_direction); + return -EINVAL; + } + + pci_unmap_addr_set(req, direct_mapping, dma); + + buf->va = cpu_to_be64(dma); + buf->key = cpu_to_be32(target->srp_host->mr->rkey); + buf->len = cpu_to_be32(scmnd->request_bufflen); + + fmt = SRP_DATA_DESC_DIRECT; + + len = sizeof (struct srp_cmd) + sizeof (struct srp_direct_buf); + } + + if (scmnd->sc_data_direction == DMA_TO_DEVICE) + cmd->buf_fmt = fmt << 4; + else + cmd->buf_fmt = fmt; + + + return len; +} + +static void srp_unmap_data(struct scsi_cmnd *scmnd, + struct srp_target_port *target, + struct srp_request *req) +{ + if (!scmnd->request_buffer || + (scmnd->sc_data_direction != DMA_TO_DEVICE && + scmnd->sc_data_direction != DMA_FROM_DEVICE)) + return; + + if (scmnd->use_sg) + dma_unmap_sg(target->srp_host->dev->dma_device, + (struct scatterlist *) scmnd->request_buffer, + scmnd->use_sg, scmnd->sc_data_direction); + else + dma_unmap_single(target->srp_host->dev->dma_device, + pci_unmap_addr(req, direct_mapping), + scmnd->request_bufflen, + scmnd->sc_data_direction); +} + +static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp) +{ + struct srp_request *req; + struct scsi_cmnd *scmnd; + unsigned long flags; + s32 delta; + + delta = (s32) be32_to_cpu(rsp->req_lim_delta); + + spin_lock_irqsave(target->scsi_host->host_lock, flags); + + target->req_lim += delta; + + req = &target->req_ring[rsp->tag & ~SRP_TAG_TSK_MGMT]; + + if (unlikely(rsp->tag & SRP_TAG_TSK_MGMT)) { + if (be32_to_cpu(rsp->resp_data_len) < 4) + req->tsk_status = -1; + else + req->tsk_status = rsp->data[3]; + complete(&req->done); + } else { + scmnd = req->scmnd; + if (!scmnd) + printk(KERN_ERR "Null scmnd for RSP w/tag %016llx\n", + (unsigned long long) rsp->tag); + scmnd->result = rsp->status; + + if (rsp->flags & SRP_RSP_FLAG_SNSVALID) { + memcpy(scmnd->sense_buffer, rsp->data + + be32_to_cpu(rsp->resp_data_len), + min_t(int, be32_to_cpu(rsp->sense_data_len), + SCSI_SENSE_BUFFERSIZE)); + } + + if (rsp->flags & (SRP_RSP_FLAG_DOOVER | SRP_RSP_FLAG_DOUNDER)) + scmnd->resid = be32_to_cpu(rsp->data_out_res_cnt); + else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER)) + scmnd->resid = be32_to_cpu(rsp->data_in_res_cnt); + + srp_unmap_data(scmnd, target, req); + + if (!req->tsk_mgmt) { + req->scmnd = NULL; + scmnd->host_scribble = (void *) -1L; + scmnd->scsi_done(scmnd); + + list_del(&req->list); + req->next = target->req_head; + target->req_head = rsp->tag & ~SRP_TAG_TSK_MGMT; + } else + req->cmd_done = 1; + } + + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); +} + +static void srp_reconnect_work(void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + + srp_reconnect_target(target); +} + +static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc) +{ + struct srp_iu *iu; + u8 opcode; + + iu = target->rx_ring[wc->wr_id & ~SRP_OP_RECV]; + + dma_sync_single_for_cpu(target->srp_host->dev->dma_device, iu->dma, + target->max_ti_iu_len, DMA_FROM_DEVICE); + + opcode = *(u8 *) iu->buf; + + if (0) { + int i; + + printk(KERN_ERR PFX "recv completion, opcode 0x%02x\n", opcode); + + for (i = 0; i < wc->byte_len; ++i) { + if (i % 8 == 0) + printk(KERN_ERR " [%02x] ", i); + printk(" %02x", ((u8 *) iu->buf)[i]); + if ((i + 1) % 8 == 0) + printk("\n"); + } + + if (wc->byte_len % 8) + printk("\n"); + } + + switch (opcode) { + case SRP_RSP: + srp_process_rsp(target, iu->buf); + break; + + case SRP_T_LOGOUT: + /* XXX Handle target logout */ + printk(KERN_WARNING PFX "Got target logout request\n"); + break; + + default: + printk(KERN_WARNING PFX "Unhandled SRP opcode 0x%02x\n", opcode); + break; + } + + dma_sync_single_for_device(target->srp_host->dev->dma_device, iu->dma, + target->max_ti_iu_len, DMA_FROM_DEVICE); +} + +static void srp_completion(struct ib_cq *cq, void *target_ptr) +{ + struct srp_target_port *target = target_ptr; + struct ib_wc wc; + unsigned long flags; + + ib_req_notify_cq(cq, IB_CQ_NEXT_COMP); + while (ib_poll_cq(cq, 1, &wc) > 0) { + if (wc.status) { + printk(KERN_ERR PFX "failed %s status %d\n", + wc.wr_id & SRP_OP_RECV ? "receive" : "send", + wc.status); + spin_lock_irqsave(target->scsi_host->host_lock, flags); + if (target->state == SRP_TARGET_LIVE) + schedule_work(&target->work); + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); + break; + } + + if (wc.wr_id & SRP_OP_RECV) + srp_handle_recv(target, &wc); + else + ++target->tx_tail; + } +} + +static int __srp_post_recv(struct srp_target_port *target) +{ + struct srp_iu *iu; + struct ib_sge list; + struct ib_recv_wr wr, *bad_wr; + unsigned int next; + int ret; + + next = target->rx_head & (SRP_RQ_SIZE - 1); + wr.wr_id = next | SRP_OP_RECV; + iu = target->rx_ring[next]; + + list.addr = iu->dma; + list.length = iu->size; + list.lkey = target->srp_host->mr->lkey; + + wr.next = NULL; + wr.sg_list = &list; + wr.num_sge = 1; + + ret = ib_post_recv(target->qp, &wr, &bad_wr); + if (!ret) + ++target->rx_head; + + return ret; +} + +static int srp_post_recv(struct srp_target_port *target) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(target->scsi_host->host_lock, flags); + ret = __srp_post_recv(target); + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); + + return ret; +} + +/* + * Must be called with target->scsi_host->host_lock held to protect + * req_lim and tx_head. + */ +static struct srp_iu *__srp_get_tx_iu(struct srp_target_port *target) +{ + if (target->tx_head - target->tx_tail >= SRP_SQ_SIZE) + return NULL; + + return target->tx_ring[target->tx_head & SRP_SQ_SIZE]; +} + +/* + * Must be called with target->scsi_host->host_lock held to protect + * req_lim and tx_head. + */ +static int __srp_post_send(struct srp_target_port *target, + struct srp_iu *iu, int len) +{ + struct ib_sge list; + struct ib_send_wr wr, *bad_wr; + int ret = 0; + + if (target->req_lim < 1) { + printk(KERN_ERR PFX "Target has req_lim %d\n", target->req_lim); + return -EAGAIN; + } + + list.addr = iu->dma; + list.length = len; + list.lkey = target->srp_host->mr->lkey; + + wr.next = NULL; + wr.wr_id = target->tx_head & SRP_SQ_SIZE; + wr.sg_list = &list; + wr.num_sge = 1; + wr.opcode = IB_WR_SEND; + wr.send_flags = IB_SEND_SIGNALED; + + ret = ib_post_send(target->qp, &wr, &bad_wr); + + if (!ret) { + ++target->tx_head; + --target->req_lim; + } + + return ret; +} + +static int srp_queuecommand(struct scsi_cmnd *scmnd, + void (*done)(struct scsi_cmnd *)) +{ + struct srp_target_port *target = host_to_target(scmnd->device->host); + struct srp_request *req; + struct srp_iu *iu; + struct srp_cmd *cmd; + long req_index; + int len; + + if (target->state == SRP_TARGET_CONNECTING) + goto err; + + if (target->state == SRP_TARGET_DEAD || + target->state == SRP_TARGET_REMOVED) { + scmnd->result = DID_BAD_TARGET << 16; + done(scmnd); + return 0; + } + + iu = __srp_get_tx_iu(target); + if (!iu) + goto err; + + dma_sync_single_for_cpu(target->srp_host->dev->dma_device, iu->dma, + SRP_MAX_IU_LEN, DMA_TO_DEVICE); + + req_index = target->req_head; + + scmnd->scsi_done = done; + scmnd->result = 0; + scmnd->host_scribble = (void *) req_index; + + cmd = iu->buf; + memset(cmd, 0, sizeof *cmd); + + cmd->opcode = SRP_CMD; + cmd->lun = cpu_to_be64((u64) scmnd->device->lun << 48); + cmd->tag = req_index; + memcpy(cmd->cdb, scmnd->cmnd, scmnd->cmd_len); + + req = &target->req_ring[req_index]; + + req->scmnd = scmnd; + req->cmd = iu; + req->cmd_done = 0; + req->tsk_mgmt = NULL; + + len = srp_map_data(scmnd, target, req); + if (len < 0) { + printk(KERN_ERR PFX "Failed to map data\n"); + goto err; + } + + if (__srp_post_recv(target)) { + printk(KERN_ERR PFX "Recv failed\n"); + goto err_unmap; + } + + dma_sync_single_for_device(target->srp_host->dev->dma_device, iu->dma, + SRP_MAX_IU_LEN, DMA_TO_DEVICE); + + if (__srp_post_send(target, iu, len)) { + printk(KERN_ERR PFX "Send failed\n"); + goto err_unmap; + } + + target->req_head = req->next; + list_add_tail(&req->list, &target->req_queue); + + return 0; + +err_unmap: + srp_unmap_data(scmnd, target, req); + +err: + return SCSI_MLQUEUE_HOST_BUSY; +} + +static int srp_alloc_iu_bufs(struct srp_target_port *target) +{ + int i; + + for (i = 0; i < SRP_RQ_SIZE; ++i) { + target->rx_ring[i] = srp_alloc_iu(target->srp_host, + target->max_ti_iu_len, + GFP_KERNEL, DMA_FROM_DEVICE); + if (!target->rx_ring[i]) + goto err; + } + + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) { + target->tx_ring[i] = srp_alloc_iu(target->srp_host, + SRP_MAX_IU_LEN, + GFP_KERNEL, DMA_TO_DEVICE); + if (!target->tx_ring[i]) + goto err; + } + + return 0; + +err: + for (i = 0; i < SRP_RQ_SIZE; ++i) { + srp_free_iu(target->srp_host, target->rx_ring[i]); + target->rx_ring[i] = NULL; + } + + for (i = 0; i < SRP_SQ_SIZE + 1; ++i) { + srp_free_iu(target->srp_host, target->tx_ring[i]); + target->tx_ring[i] = NULL; + } + + return -ENOMEM; +} + +static void srp_cm_rej_handler(struct ib_cm_id *cm_id, + struct ib_cm_event *event, + struct srp_target_port *target) +{ + struct ib_class_port_info *cpi; + int opcode; + + switch (event->param.rej_rcvd.reason) { + case IB_CM_REJ_PORT_CM_REDIRECT: + cpi = event->param.rej_rcvd.ari; + target->path.dlid = cpi->redirect_lid; + target->path.pkey = cpi->redirect_pkey; + cm_id->remote_cm_qpn = be32_to_cpu(cpi->redirect_qp) & 0x00ffffff; + memcpy(target->path.dgid.raw, cpi->redirect_gid, 16); + + target->status = target->path.dlid ? + SRP_DLID_REDIRECT : SRP_PORT_REDIRECT; + break; + + case IB_CM_REJ_PORT_REDIRECT: + if (topspin_workarounds && + !memcmp(&target->ioc_guid, topspin_oui, 3)) { + /* + * Topspin/Cisco SRP gateways incorrectly send + * reject reason code 25 when they mean 24 + * (port redirect). + */ + memcpy(target->path.dgid.raw, + event->param.rej_rcvd.ari, 16); + + printk(KERN_DEBUG PFX "Topspin/Cisco redirect to target port GID %016llx%016llx\n", + (unsigned long long) be64_to_cpu(target->path.dgid.global.subnet_prefix), + (unsigned long long) be64_to_cpu(target->path.dgid.global.interface_id)); + + target->status = SRP_PORT_REDIRECT; + } else { + printk(KERN_WARNING " REJ reason: IB_CM_REJ_PORT_REDIRECT\n"); + target->status = -ECONNRESET; + } + break; + + case IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID: + printk(KERN_WARNING " REJ reason: IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID\n"); + target->status = -ECONNRESET; + break; + + case IB_CM_REJ_CONSUMER_DEFINED: + opcode = *(u8 *) event->private_data; + if (opcode == SRP_LOGIN_REJ) { + struct srp_login_rej *rej = event->private_data; + u32 reason = be32_to_cpu(rej->reason); + + if (reason == SRP_LOGIN_REJ_REQ_IT_IU_LENGTH_TOO_LARGE) + printk(KERN_WARNING PFX + "SRP_LOGIN_REJ: requested max_it_iu_len too large\n"); + else + printk(KERN_WARNING PFX + "SRP LOGIN REJECTED, reason 0x%08x\n", reason); + } else + printk(KERN_WARNING " REJ reason: IB_CM_REJ_CONSUMER_DEFINED," + " opcode 0x%02x\n", opcode); + target->status = -ECONNRESET; + break; + + default: + printk(KERN_WARNING " REJ reason 0x%x\n", + event->param.rej_rcvd.reason); + target->status = -ECONNRESET; + } +} + +static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event) +{ + struct srp_target_port *target = cm_id->context; + struct ib_qp_attr *qp_attr = NULL; + int attr_mask = 0; + int comp = 0; + int opcode = 0; + + switch (event->event) { + case IB_CM_REQ_ERROR: + printk(KERN_DEBUG PFX "Sending CM REQ failed\n"); + comp = 1; + target->status = -ECONNRESET; + break; + + case IB_CM_REP_RECEIVED: + comp = 1; + opcode = *(u8 *) event->private_data; + + if (opcode == SRP_LOGIN_RSP) { + struct srp_login_rsp *rsp = event->private_data; + + target->max_ti_iu_len = be32_to_cpu(rsp->max_ti_iu_len); + target->req_lim = be32_to_cpu(rsp->req_lim_delta); + + target->scsi_host->can_queue = min(target->req_lim, + target->scsi_host->can_queue); + } else { + printk(KERN_WARNING PFX "Unhandled RSP opcode %#x\n", opcode); + target->status = -ECONNRESET; + break; + } + + target->status = srp_alloc_iu_bufs(target); + if (target->status) + break; + + qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL); + if (!qp_attr) { + target->status = -ENOMEM; + break; + } + + qp_attr->qp_state = IB_QPS_RTR; + target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); + if (target->status) + break; + + target->status = ib_modify_qp(target->qp, qp_attr, attr_mask); + if (target->status) + break; + + target->status = srp_post_recv(target); + if (target->status) + break; + + qp_attr->qp_state = IB_QPS_RTS; + target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask); + if (target->status) + break; + + target->status = ib_modify_qp(target->qp, qp_attr, attr_mask); + if (target->status) + break; + + target->status = ib_send_cm_rtu(cm_id, NULL, 0); + if (target->status) + break; + + break; + + case IB_CM_REJ_RECEIVED: + printk(KERN_DEBUG PFX "REJ received\n"); + comp = 1; + + srp_cm_rej_handler(cm_id, event, target); + break; + + case IB_CM_MRA_RECEIVED: + printk(KERN_ERR PFX "MRA received\n"); + break; + + case IB_CM_DREP_RECEIVED: + break; + + case IB_CM_TIMEWAIT_EXIT: + printk(KERN_ERR PFX "connection closed\n"); + + comp = 1; + target->status = 0; + break; + + default: + printk(KERN_WARNING PFX "Unhandled CM event %d\n", event->event); + break; + } + + if (comp) + complete(&target->done); + + kfree(qp_attr); + + return 0; +} + +static int srp_send_tsk_mgmt(struct scsi_cmnd *scmnd, u8 func) +{ + struct srp_target_port *target = host_to_target(scmnd->device->host); + struct srp_request *req; + struct srp_iu *iu; + struct srp_tsk_mgmt *tsk_mgmt; + int req_index; + int ret = FAILED; + + spin_lock_irq(target->scsi_host->host_lock); + + if (scmnd->host_scribble == (void *) -1L) + goto out; + + req_index = (long) scmnd->host_scribble; + printk(KERN_ERR "Abort for req_index %d\n", req_index); + + req = &target->req_ring[req_index]; + init_completion(&req->done); + + iu = __srp_get_tx_iu(target); + if (!iu) + goto out; + + tsk_mgmt = iu->buf; + memset(tsk_mgmt, 0, sizeof *tsk_mgmt); + + tsk_mgmt->opcode = SRP_TSK_MGMT; + tsk_mgmt->lun = cpu_to_be64((u64) scmnd->device->lun << 48); + tsk_mgmt->tag = req_index | SRP_TAG_TSK_MGMT; + tsk_mgmt->tsk_mgmt_func = func; + tsk_mgmt->task_tag = req_index; + + if (__srp_post_send(target, iu, sizeof *tsk_mgmt)) + goto out; + + req->tsk_mgmt = iu; + + spin_unlock_irq(target->scsi_host->host_lock); + if (!wait_for_completion_timeout(&req->done, + msecs_to_jiffies(SRP_ABORT_TIMEOUT_MS))) + return FAILED; + spin_lock_irq(target->scsi_host->host_lock); + + if (req->cmd_done) { + list_del(&req->list); + req->next = target->req_head; + target->req_head = req_index; + + scmnd->scsi_done(scmnd); + } else if (!req->tsk_status) { + scmnd->result = DID_ABORT << 16; + ret = SUCCESS; + } + +out: + spin_unlock_irq(target->scsi_host->host_lock); + return ret; +} + +static int srp_abort(struct scsi_cmnd *scmnd) +{ + printk(KERN_ERR "SRP abort called\n"); + + return srp_send_tsk_mgmt(scmnd, SRP_TSK_ABORT_TASK); +} + +static int srp_reset_device(struct scsi_cmnd *scmnd) +{ + printk(KERN_ERR "SRP reset_device called\n"); + + return srp_send_tsk_mgmt(scmnd, SRP_TSK_LUN_RESET); +} + +static int srp_reset_host(struct scsi_cmnd *scmnd) +{ + struct srp_target_port *target = host_to_target(scmnd->device->host); + int ret = FAILED; + + printk(KERN_ERR PFX "SRP reset_host called\n"); + + if (!srp_reconnect_target(target)) + ret = SUCCESS; + + return ret; +} + +static struct scsi_host_template srp_template = { + .module = THIS_MODULE, + .name = DRV_NAME, + .info = srp_target_info, + .queuecommand = srp_queuecommand, + .eh_abort_handler = srp_abort, + .eh_device_reset_handler = srp_reset_device, + .eh_host_reset_handler = srp_reset_host, + .can_queue = SRP_SQ_SIZE, + .this_id = -1, + .sg_tablesize = SRP_MAX_INDIRECT, + .cmd_per_lun = SRP_SQ_SIZE, + .use_clustering = ENABLE_CLUSTERING +}; + +static int srp_add_target(struct srp_host *host, struct srp_target_port *target) +{ + sprintf(target->target_name, "SRP.T10:%016llX", + (unsigned long long) be64_to_cpu(target->id_ext)); + + if (scsi_add_host(target->scsi_host, host->dev->dma_device)) + return -ENODEV; + + down(&host->target_mutex); + list_add_tail(&target->list, &host->target_list); + up(&host->target_mutex); + + target->state = SRP_TARGET_LIVE; + + /* XXX: are we supposed to have a definition of SCAN_WILD_CARD ?? */ + scsi_scan_target(&target->scsi_host->shost_gendev, + 0, target->scsi_id, ~0, 0); + + return 0; +} + +static void srp_release_class_dev(struct class_device *class_dev) +{ + struct srp_host *host = + container_of(class_dev, struct srp_host, class_dev); + + complete(&host->released); +} + +static struct class srp_class = { + .name = "infiniband_srp", + .release = srp_release_class_dev +}; + +/* + * Target ports are added by writing + * + * id_ext=<SRP ID ext>,ioc_guid=<SRP IOC GUID>,dgid=<dest GID>, + * pkey=<P_Key>,service_id=<service ID> + * + * to the add_target sysfs attribute. + */ +enum { + SRP_OPT_ERR = 0, + SRP_OPT_ID_EXT = 1 << 0, + SRP_OPT_IOC_GUID = 1 << 1, + SRP_OPT_DGID = 1 << 2, + SRP_OPT_PKEY = 1 << 3, + SRP_OPT_SERVICE_ID = 1 << 4, + SRP_OPT_MAX_SECT = 1 << 5, + SRP_OPT_ALL = (SRP_OPT_ID_EXT | + SRP_OPT_IOC_GUID | + SRP_OPT_DGID | + SRP_OPT_PKEY | + SRP_OPT_SERVICE_ID), +}; + +static match_table_t srp_opt_tokens = { + { SRP_OPT_ID_EXT, "id_ext=%s" }, + { SRP_OPT_IOC_GUID, "ioc_guid=%s" }, + { SRP_OPT_DGID, "dgid=%s" }, + { SRP_OPT_PKEY, "pkey=%x" }, + { SRP_OPT_SERVICE_ID, "service_id=%s" }, + { SRP_OPT_MAX_SECT, "max_sect=%d" }, + { SRP_OPT_ERR, NULL } +}; + +static int srp_parse_options(const char *buf, struct srp_target_port *target) +{ + char *options, *sep_opt; + char *p; + char dgid[3]; + substring_t args[MAX_OPT_ARGS]; + int opt_mask = 0; + int token; + int ret = -EINVAL; + int i; + + options = kstrdup(buf, GFP_KERNEL); + if (!options) + return -ENOMEM; + + sep_opt = options; + while ((p = strsep(&sep_opt, ",")) != NULL) { + if (!*p) + continue; + + token = match_token(p, srp_opt_tokens, args); + opt_mask |= token; + + switch (token) { + case SRP_OPT_ID_EXT: + p = match_strdup(args); + target->id_ext = cpu_to_be64(simple_strtoull(p, NULL, 16)); + kfree(p); + break; + + case SRP_OPT_IOC_GUID: + p = match_strdup(args); + target->ioc_guid = cpu_to_be64(simple_strtoull(p, NULL, 16)); + kfree(p); + break; + + case SRP_OPT_DGID: + p = match_strdup(args); + if (strlen(p) != 32) { + printk(KERN_WARNING PFX "bad dest GID parameter '%s'\n", p); + goto out; + } + + for (i = 0; i < 16; ++i) { + strlcpy(dgid, p + i * 2, 3); + target->path.dgid.raw[i] = simple_strtoul(dgid, NULL, 16); + } + break; + + case SRP_OPT_PKEY: + if (match_hex(args, &token)) { + printk(KERN_WARNING PFX "bad P_Key parameter '%s'\n", p); + goto out; + } + target->path.pkey = cpu_to_be16(token); + break; + + case SRP_OPT_SERVICE_ID: + p = match_strdup(args); + target->service_id = cpu_to_be64(simple_strtoull(p, NULL, 16)); + kfree(p); + break; + + case SRP_OPT_MAX_SECT: + if (match_int(args, &token)) { + printk(KERN_WARNING PFX "bad max sect parameter '%s'\n", p); + goto out; + } + target->scsi_host->max_sectors = token; + break; + + default: + printk(KERN_WARNING PFX "unknown parameter or missing value " + "'%s' in target creation request\n", p); + goto out; + } + } + + if ((opt_mask & SRP_OPT_ALL) == SRP_OPT_ALL) + ret = 0; + else + for (i = 0; i < ARRAY_SIZE(srp_opt_tokens); ++i) + if ((srp_opt_tokens[i].token & SRP_OPT_ALL) && + !(srp_opt_tokens[i].token & opt_mask)) + printk(KERN_WARNING PFX "target creation request is " + "missing parameter '%s'\n", + srp_opt_tokens[i].pattern); + +out: + kfree(options); + return ret; +} + +static ssize_t srp_create_target(struct class_device *class_dev, + const char *buf, size_t count) +{ + struct srp_host *host = + container_of(class_dev, struct srp_host, class_dev); + struct Scsi_Host *target_host; + struct srp_target_port *target; + int ret; + int i; + + target_host = scsi_host_alloc(&srp_template, + sizeof (struct srp_target_port)); + if (!target_host) + return -ENOMEM; + + target = host_to_target(target_host); + memset(target, 0, sizeof *target); + + target->scsi_host = target_host; + target->srp_host = host; + + INIT_WORK(&target->work, srp_reconnect_work, target); + + for (i = 0; i < SRP_SQ_SIZE - 1; ++i) + target->req_ring[i].next = i + 1; + target->req_ring[SRP_SQ_SIZE - 1].next = -1; + INIT_LIST_HEAD(&target->req_queue); + + ret = srp_parse_options(buf, target); + if (ret) + goto err; + + ib_get_cached_gid(host->dev, host->port, 0, &target->path.sgid); + + printk(KERN_DEBUG PFX "new target: id_ext %016llx ioc_guid %016llx pkey %04x " + "service_id %016llx dgid %04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n", + (unsigned long long) be64_to_cpu(target->id_ext), + (unsigned long long) be64_to_cpu(target->ioc_guid), + be16_to_cpu(target->path.pkey), + (unsigned long long) be64_to_cpu(target->service_id), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[0]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[2]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[4]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[6]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[8]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[10]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[12]), + (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[14])); + + ret = srp_create_target_ib(target); + if (ret) + goto err; + + target->cm_id = ib_create_cm_id(host->dev, srp_cm_handler, target); + if (IS_ERR(target->cm_id)) { + ret = PTR_ERR(target->cm_id); + goto err_free; + } + + ret = srp_connect_target(target); + if (ret) { + printk(KERN_ERR PFX "Connection failed\n"); + goto err_cm_id; + } + + ret = srp_add_target(host, target); + if (ret) + goto err_disconnect; + + return count; + +err_disconnect: + srp_disconnect_target(target); + +err_cm_id: + ib_destroy_cm_id(target->cm_id); + +err_free: + srp_free_target_ib(target); + +err: + scsi_host_put(target_host); + + return ret; +} + +static CLASS_DEVICE_ATTR(add_target, S_IWUSR, NULL, srp_create_target); + +static ssize_t show_ibdev(struct class_device *class_dev, char *buf) +{ + struct srp_host *host = + container_of(class_dev, struct srp_host, class_dev); + + return sprintf(buf, "%s\n", host->dev->name); +} + +static CLASS_DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL); + +static ssize_t show_port(struct class_device *class_dev, char *buf) +{ + struct srp_host *host = + container_of(class_dev, struct srp_host, class_dev); + + return sprintf(buf, "%d\n", host->port); +} + +static CLASS_DEVICE_ATTR(port, S_IRUGO, show_port, NULL); + +static struct srp_host *srp_add_port(struct ib_device *device, + __be64 node_guid, u8 port) +{ + struct srp_host *host; + + host = kzalloc(sizeof *host, GFP_KERNEL); + if (!host) + return NULL; + + INIT_LIST_HEAD(&host->target_list); + init_MUTEX(&host->target_mutex); + init_completion(&host->released); + host->dev = device; + host->port = port; + + host->initiator_port_id[7] = port; + memcpy(host->initiator_port_id + 8, &node_guid, 8); + + host->pd = ib_alloc_pd(device); + if (IS_ERR(host->pd)) + goto err_free; + + host->mr = ib_get_dma_mr(host->pd, + IB_ACCESS_LOCAL_WRITE | + IB_ACCESS_REMOTE_READ | + IB_ACCESS_REMOTE_WRITE); + if (IS_ERR(host->mr)) + goto err_pd; + + host->class_dev.class = &srp_class; + host->class_dev.dev = device->dma_device; + snprintf(host->class_dev.class_id, BUS_ID_SIZE, "srp-%s-%d", + device->name, port); + + if (class_device_register(&host->class_dev)) + goto err_mr; + if (class_device_create_file(&host->class_dev, &class_device_attr_add_target)) + goto err_class; + if (class_device_create_file(&host->class_dev, &class_device_attr_ibdev)) + goto err_class; + if (class_device_create_file(&host->class_dev, &class_device_attr_port)) + goto err_class; + + return host; + +err_class: + class_device_unregister(&host->class_dev); + +err_mr: + ib_dereg_mr(host->mr); + +err_pd: + ib_dealloc_pd(host->pd); + +err_free: + kfree(host); + + return NULL; +} + +static void srp_add_one(struct ib_device *device) +{ + struct list_head *dev_list; + struct srp_host *host; + struct ib_device_attr *dev_attr; + int s, e, p; + + dev_attr = kmalloc(sizeof *dev_attr, GFP_KERNEL); + if (!dev_attr) + return; + + if (ib_query_device(device, dev_attr)) { + printk(KERN_WARNING PFX "Couldn't query node GUID for %s.\n", + device->name); + goto out; + } + + dev_list = kmalloc(sizeof *dev_list, GFP_KERNEL); + if (!dev_list) + goto out; + + INIT_LIST_HEAD(dev_list); + + if (device->node_type == IB_NODE_SWITCH) { + s = 0; + e = 0; + } else { + s = 1; + e = device->phys_port_cnt; + } + + for (p = s; p <= e; ++p) { + host = srp_add_port(device, dev_attr->node_guid, p); + if (host) + list_add_tail(&host->list, dev_list); + } + + ib_set_client_data(device, &srp_client, dev_list); + +out: + kfree(dev_attr); +} + +static void srp_remove_one(struct ib_device *device) +{ + struct list_head *dev_list; + struct srp_host *host, *tmp_host; + LIST_HEAD(target_list); + struct srp_target_port *target, *tmp_target; + unsigned long flags; + + dev_list = ib_get_client_data(device, &srp_client); + + list_for_each_entry_safe(host, tmp_host, dev_list, list) { + class_device_unregister(&host->class_dev); + /* + * Wait for the sysfs entry to go away, so that no new + * target ports can be created. + */ + wait_for_completion(&host->released); + + /* + * Mark all target ports as removed, so we stop queueing + * commands and don't try to reconnect. + */ + down(&host->target_mutex); + list_for_each_entry_safe(target, tmp_target, + &host->target_list, list) { + spin_lock_irqsave(target->scsi_host->host_lock, flags); + if (target->state != SRP_TARGET_REMOVED) + target->state = SRP_TARGET_REMOVED; + spin_unlock_irqrestore(target->scsi_host->host_lock, flags); + } + up(&host->target_mutex); + + /* + * Wait for any reconnection tasks that may have + * started before we marked our target ports as + * removed, and any target port removal tasks. + */ + flush_scheduled_work(); + + list_for_each_entry_safe(target, tmp_target, + &host->target_list, list) { + scsi_remove_host(target->scsi_host); + srp_disconnect_target(target); + ib_destroy_cm_id(target->cm_id); + srp_free_target_ib(target); + scsi_host_put(target->scsi_host); + } + + ib_dereg_mr(host->mr); + ib_dealloc_pd(host->pd); + kfree(host); + } + + kfree(dev_list); +} + +static int __init srp_init_module(void) +{ + int ret; + + ret = class_register(&srp_class); + if (ret) { + printk(KERN_ERR PFX "couldn't register class infiniband_srp\n"); + return ret; + } + + ret = ib_register_client(&srp_client); + if (ret) { + printk(KERN_ERR PFX "couldn't register IB client\n"); + class_unregister(&srp_class); + return ret; + } + + return 0; +} + +static void __exit srp_cleanup_module(void) +{ + ib_unregister_client(&srp_client); + class_unregister(&srp_class); +} + +module_init(srp_init_module); +module_exit(srp_cleanup_module); diff --git a/drivers/infiniband/ulp/srp/ib_srp.h b/drivers/infiniband/ulp/srp/ib_srp.h new file mode 100644 index 00000000000..4fec28a7136 --- /dev/null +++ b/drivers/infiniband/ulp/srp/ib_srp.h @@ -0,0 +1,150 @@ +/* + * Copyright (c) 2005 Cisco Systems. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * 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 AUTHORS OR COPYRIGHT HOLDERS + * 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. + * + * $Id: ib_srp.h 3932 2005-11-01 17:19:29Z roland $ + */ + +#ifndef IB_SRP_H +#define IB_SRP_H + +#include <linux/types.h> +#include <linux/list.h> + +#include <asm/semaphore.h> + +#include <scsi/scsi_host.h> +#include <scsi/scsi_cmnd.h> + +#include <rdma/ib_verbs.h> +#include <rdma/ib_sa.h> +#include <rdma/ib_cm.h> + +enum { + SRP_PATH_REC_TIMEOUT_MS = 1000, + SRP_ABORT_TIMEOUT_MS = 5000, + + SRP_PORT_REDIRECT = 1, + SRP_DLID_REDIRECT = 2, + + SRP_MAX_IU_LEN = 256, + + SRP_RQ_SHIFT = 6, + SRP_RQ_SIZE = 1 << SRP_RQ_SHIFT, + SRP_SQ_SIZE = SRP_RQ_SIZE - 1, + SRP_CQ_SIZE = SRP_SQ_SIZE + SRP_RQ_SIZE, + + SRP_TAG_TSK_MGMT = 1 << (SRP_RQ_SHIFT + 1) +}; + +#define SRP_OP_RECV (1 << 31) +#define SRP_MAX_INDIRECT ((SRP_MAX_IU_LEN - \ + sizeof (struct srp_cmd) - \ + sizeof (struct srp_indirect_buf)) / 16) + +enum srp_target_state { + SRP_TARGET_LIVE, + SRP_TARGET_CONNECTING, + SRP_TARGET_DEAD, + SRP_TARGET_REMOVED +}; + +struct srp_host { + u8 initiator_port_id[16]; + struct ib_device *dev; + u8 port; + struct ib_pd *pd; + struct ib_mr *mr; + struct class_device class_dev; + struct list_head target_list; + struct semaphore target_mutex; + struct completion released; + struct list_head list; +}; + +struct srp_request { + struct list_head list; + struct scsi_cmnd *scmnd; + struct srp_iu *cmd; + struct srp_iu *tsk_mgmt; + DECLARE_PCI_UNMAP_ADDR(direct_mapping) + struct completion done; + short next; + u8 cmd_done; + u8 tsk_status; +}; + +struct srp_target_port { + __be64 id_ext; + __be64 ioc_guid; + __be64 service_id; + struct srp_host *srp_host; + struct Scsi_Host *scsi_host; + char target_name[32]; + unsigned int scsi_id; + + struct ib_sa_path_rec path; + struct ib_sa_query *path_query; + int path_query_id; + + struct ib_cm_id *cm_id; + struct ib_cq *cq; + struct ib_qp *qp; + + int max_ti_iu_len; + s32 req_lim; + + unsigned rx_head; + struct srp_iu *rx_ring[SRP_RQ_SIZE]; + + unsigned tx_head; + unsigned tx_tail; + struct srp_iu *tx_ring[SRP_SQ_SIZE + 1]; + + int req_head; + struct list_head req_queue; + struct srp_request req_ring[SRP_SQ_SIZE]; + + struct work_struct work; + + struct list_head list; + struct completion done; + int status; + enum srp_target_state state; +}; + +struct srp_iu { + dma_addr_t dma; + void *buf; + size_t size; + enum dma_data_direction direction; +}; + +#endif /* IB_SRP_H */ diff --git a/drivers/macintosh/via-pmu.c b/drivers/macintosh/via-pmu.c index 91920a1140f..9bc6cc6e384 100644 --- a/drivers/macintosh/via-pmu.c +++ b/drivers/macintosh/via-pmu.c @@ -155,10 +155,10 @@ static spinlock_t pmu_lock; static u8 pmu_intr_mask; static int pmu_version; static int drop_interrupts; -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) static int option_lid_wakeup = 1; static int sleep_in_progress; -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ static unsigned long async_req_locks; static unsigned int pmu_irq_stats[11]; @@ -865,7 +865,7 @@ proc_read_options(char *page, char **start, off_t off, { char *p = page; -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) if (pmu_kind == PMU_KEYLARGO_BASED && pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,0,-1) >= 0) p += sprintf(p, "lid_wakeup=%d\n", option_lid_wakeup); @@ -906,7 +906,7 @@ proc_write_options(struct file *file, const char __user *buffer, *(val++) = 0; while(*val == ' ') val++; -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) if (pmu_kind == PMU_KEYLARGO_BASED && pmac_call_feature(PMAC_FTR_SLEEP_STATE,NULL,0,-1) >= 0) if (!strcmp(label, "lid_wakeup")) @@ -2063,6 +2063,9 @@ pmu_unregister_sleep_notifier(struct pmu_sleep_notifier* n) n->list.next = NULL; return 0; } +#endif /* CONFIG_PM */ + +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) /* Sleep is broadcast last-to-first */ static int @@ -2687,7 +2690,7 @@ powerbook_sleep_3400(void) return 0; } -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ /* * Support for /dev/pmu device @@ -2871,7 +2874,7 @@ pmu_ioctl(struct inode * inode, struct file *filp, int error = -EINVAL; switch (cmd) { -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) case PMU_IOC_SLEEP: if (!capable(CAP_SYS_ADMIN)) return -EACCES; @@ -2899,7 +2902,7 @@ pmu_ioctl(struct inode * inode, struct file *filp, return put_user(0, argp); else return put_user(1, argp); -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ #ifdef CONFIG_PMAC_BACKLIGHT /* Backlight should have its own device or go via @@ -3047,7 +3050,7 @@ pmu_polled_request(struct adb_request *req) * to do suspend-to-disk. */ -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) static int pmu_sys_suspended = 0; @@ -3082,7 +3085,7 @@ static int pmu_sys_resume(struct sys_device *sysdev) return 0; } -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ static struct sysdev_class pmu_sysclass = { set_kset_name("pmu"), @@ -3094,10 +3097,10 @@ static struct sys_device device_pmu = { }; static struct sysdev_driver driver_pmu = { -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) .suspend = &pmu_sys_suspend, .resume = &pmu_sys_resume, -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ }; static int __init init_pmu_sysfs(void) @@ -3135,12 +3138,12 @@ EXPORT_SYMBOL(pmu_i2c_combined_read); EXPORT_SYMBOL(pmu_i2c_stdsub_write); EXPORT_SYMBOL(pmu_i2c_simple_read); EXPORT_SYMBOL(pmu_i2c_simple_write); -#ifdef CONFIG_PM +#if defined(CONFIG_PM) && defined(CONFIG_PPC32) EXPORT_SYMBOL(pmu_register_sleep_notifier); EXPORT_SYMBOL(pmu_unregister_sleep_notifier); EXPORT_SYMBOL(pmu_enable_irled); EXPORT_SYMBOL(pmu_battery_count); EXPORT_SYMBOL(pmu_batteries); EXPORT_SYMBOL(pmu_power_flags); -#endif /* CONFIG_PM */ +#endif /* CONFIG_PM && CONFIG_PPC32 */ diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index 6d4f9ceb0a3..57edae4790e 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -1374,7 +1374,7 @@ config FORCEDETH config CS89x0 tristate "CS89x0 support" - depends on (NET_PCI && (ISA || ARCH_IXDP2X01)) || ARCH_PNX0105 || MACH_MP1000 + depends on (NET_PCI && (ISA || ARCH_IXDP2X01)) || ARCH_PNX0105 ---help--- Support for CS89x0 chipset based Ethernet cards. If you have a network (Ethernet) card of this type, say Y and read the diff --git a/drivers/net/cs89x0.c b/drivers/net/cs89x0.c index bfdae10036e..a6078ad9b65 100644 --- a/drivers/net/cs89x0.c +++ b/drivers/net/cs89x0.c @@ -182,10 +182,6 @@ static unsigned int cs8900_irq_map[] = {IRQ_IXDP2X01_CS8900, 0, 0, 0}; #define CIRRUS_DEFAULT_IRQ VH_INTC_INT_NUM_CASCADED_INTERRUPT_1 /* Event inputs bank 1 - ID 35/bit 3 */ static unsigned int netcard_portlist[] __initdata = {CIRRUS_DEFAULT_BASE, 0}; static unsigned int cs8900_irq_map[] = {CIRRUS_DEFAULT_IRQ, 0, 0, 0}; -#elif defined(CONFIG_MACH_MP1000) -#include <asm/arch/mp1000-seprom.h> -static unsigned int netcard_portlist[] __initdata = {MP1000_EIO_BASE+0x300, 0}; -static unsigned int cs8900_irq_map[] = {IRQ_EINT3,0,0,0}; #else static unsigned int netcard_portlist[] __initdata = { 0x300, 0x320, 0x340, 0x360, 0x200, 0x220, 0x240, 0x260, 0x280, 0x2a0, 0x2c0, 0x2e0, 0}; @@ -594,10 +590,6 @@ cs89x0_probe1(struct net_device *dev, int ioaddr, int modular) cnt -= j; } } else -#elif defined(CONFIG_MACH_MP1000) - if (1) { - memcpy(dev->dev_addr, get_eeprom_mac_address(), ETH_ALEN); - } else #endif if ((readreg(dev, PP_SelfST) & (EEPROM_OK | EEPROM_PRESENT)) == @@ -657,10 +649,6 @@ cs89x0_probe1(struct net_device *dev, int ioaddr, int modular) if (1) { printk(KERN_NOTICE "cs89x0: No EEPROM on HiCO.SH4\n"); } else -#elif defined(CONFIG_MACH_MP1000) - if (1) { - lp->force |= FORCE_RJ45; - } else #endif if ((readreg(dev, PP_SelfST) & EEPROM_PRESENT) == 0) printk(KERN_WARNING "cs89x0: No EEPROM, relying on command line....\n"); @@ -1243,7 +1231,7 @@ net_open(struct net_device *dev) else #endif { -#if !defined(CONFIG_ARCH_IXDP2X01) && !defined(CONFIG_ARCH_PNX0105) && !defined(CONFIG_MACH_MP1000) +#if !defined(CONFIG_ARCH_IXDP2X01) && !defined(CONFIG_ARCH_PNX0105) if (((1 << dev->irq) & lp->irq_map) == 0) { printk(KERN_ERR "%s: IRQ %d is not in our map of allowable IRQs, which is %x\n", dev->name, dev->irq, lp->irq_map); diff --git a/drivers/net/cs89x0.h b/drivers/net/cs89x0.h index f19d1ebe018..decea264f12 100644 --- a/drivers/net/cs89x0.h +++ b/drivers/net/cs89x0.h @@ -16,7 +16,7 @@ #include <linux/config.h> -#if defined(CONFIG_ARCH_IXDP2X01) || defined(CONFIG_ARCH_PNX0105) || defined (CONFIG_MACH_MP1000) +#if defined(CONFIG_ARCH_IXDP2X01) || defined(CONFIG_ARCH_PNX0105) /* IXDP2401/IXDP2801 uses dword-aligned register addressing */ #define CS89x0_PORT(reg) ((reg) * 2) #else diff --git a/drivers/net/iseries_veth.c b/drivers/net/iseries_veth.c index f5ea39ff101..d86d8f055a6 100644 --- a/drivers/net/iseries_veth.c +++ b/drivers/net/iseries_veth.c @@ -72,12 +72,12 @@ #include <linux/ethtool.h> #include <asm/abs_addr.h> -#include <asm/iSeries/mf.h> +#include <asm/iseries/mf.h> #include <asm/uaccess.h> -#include <asm/iSeries/HvLpConfig.h> -#include <asm/iSeries/HvTypes.h> -#include <asm/iSeries/HvLpEvent.h> +#include <asm/iseries/hv_lp_config.h> +#include <asm/iseries/hv_types.h> +#include <asm/iseries/hv_lp_event.h> #include <asm/iommu.h> #include <asm/vio.h> diff --git a/drivers/scsi/ahci.c b/drivers/scsi/ahci.c index e2a5657d5fd..4612312c0c2 100644 --- a/drivers/scsi/ahci.c +++ b/drivers/scsi/ahci.c @@ -307,14 +307,22 @@ static int ahci_port_start(struct ata_port *ap) void __iomem *port_mmio = ahci_port_base(mmio, ap->port_no); void *mem; dma_addr_t mem_dma; + int rc; pp = kmalloc(sizeof(*pp), GFP_KERNEL); if (!pp) return -ENOMEM; memset(pp, 0, sizeof(*pp)); + rc = ata_pad_alloc(ap, dev); + if (rc) { + kfree(pp); + return rc; + } + mem = dma_alloc_coherent(dev, AHCI_PORT_PRIV_DMA_SZ, &mem_dma, GFP_KERNEL); if (!mem) { + ata_pad_free(ap, dev); kfree(pp); return -ENOMEM; } @@ -390,6 +398,7 @@ static void ahci_port_stop(struct ata_port *ap) ap->private_data = NULL; dma_free_coherent(dev, AHCI_PORT_PRIV_DMA_SZ, pp->cmd_slot, pp->cmd_slot_dma); + ata_pad_free(ap, dev); kfree(pp); } @@ -468,23 +477,23 @@ static void ahci_tf_read(struct ata_port *ap, struct ata_taskfile *tf) static void ahci_fill_sg(struct ata_queued_cmd *qc) { struct ahci_port_priv *pp = qc->ap->private_data; - unsigned int i; + struct scatterlist *sg; + struct ahci_sg *ahci_sg; VPRINTK("ENTER\n"); /* * Next, the S/G list. */ - for (i = 0; i < qc->n_elem; i++) { - u32 sg_len; - dma_addr_t addr; - - addr = sg_dma_address(&qc->sg[i]); - sg_len = sg_dma_len(&qc->sg[i]); - - pp->cmd_tbl_sg[i].addr = cpu_to_le32(addr & 0xffffffff); - pp->cmd_tbl_sg[i].addr_hi = cpu_to_le32((addr >> 16) >> 16); - pp->cmd_tbl_sg[i].flags_size = cpu_to_le32(sg_len - 1); + ahci_sg = pp->cmd_tbl_sg; + ata_for_each_sg(sg, qc) { + dma_addr_t addr = sg_dma_address(sg); + u32 sg_len = sg_dma_len(sg); + + ahci_sg->addr = cpu_to_le32(addr & 0xffffffff); + ahci_sg->addr_hi = cpu_to_le32((addr >> 16) >> 16); + ahci_sg->flags_size = cpu_to_le32(sg_len - 1); + ahci_sg++; } } diff --git a/drivers/scsi/ibmvscsi/iseries_vscsi.c b/drivers/scsi/ibmvscsi/iseries_vscsi.c index e9202f2a827..1045872b017 100644 --- a/drivers/scsi/ibmvscsi/iseries_vscsi.c +++ b/drivers/scsi/ibmvscsi/iseries_vscsi.c @@ -28,10 +28,10 @@ * hypervisor system or a converged hypervisor system. */ -#include <asm/iSeries/vio.h> -#include <asm/iSeries/HvLpEvent.h> -#include <asm/iSeries/HvTypes.h> -#include <asm/iSeries/HvLpConfig.h> +#include <asm/iseries/vio.h> +#include <asm/iseries/hv_lp_event.h> +#include <asm/iseries/hv_types.h> +#include <asm/iseries/hv_lp_config.h> #include <asm/vio.h> #include <linux/device.h> #include "ibmvscsi.h" diff --git a/drivers/scsi/libata-core.c b/drivers/scsi/libata-core.c index ff18fa7044c..e1346cddd37 100644 --- a/drivers/scsi/libata-core.c +++ b/drivers/scsi/libata-core.c @@ -2390,8 +2390,9 @@ static void ata_dev_init_params(struct ata_port *ap, struct ata_device *dev) static void ata_sg_clean(struct ata_queued_cmd *qc) { struct ata_port *ap = qc->ap; - struct scatterlist *sg = qc->sg; + struct scatterlist *sg = qc->__sg; int dir = qc->dma_dir; + void *pad_buf = NULL; assert(qc->flags & ATA_QCFLAG_DMAMAP); assert(sg != NULL); @@ -2401,14 +2402,35 @@ static void ata_sg_clean(struct ata_queued_cmd *qc) DPRINTK("unmapping %u sg elements\n", qc->n_elem); - if (qc->flags & ATA_QCFLAG_SG) + /* if we padded the buffer out to 32-bit bound, and data + * xfer direction is from-device, we must copy from the + * pad buffer back into the supplied buffer + */ + if (qc->pad_len && !(qc->tf.flags & ATA_TFLAG_WRITE)) + pad_buf = ap->pad + (qc->tag * ATA_DMA_PAD_SZ); + + if (qc->flags & ATA_QCFLAG_SG) { dma_unmap_sg(ap->host_set->dev, sg, qc->n_elem, dir); - else + /* restore last sg */ + sg[qc->orig_n_elem - 1].length += qc->pad_len; + if (pad_buf) { + struct scatterlist *psg = &qc->pad_sgent; + void *addr = kmap_atomic(psg->page, KM_IRQ0); + memcpy(addr + psg->offset, pad_buf, qc->pad_len); + kunmap_atomic(psg->page, KM_IRQ0); + } + } else { dma_unmap_single(ap->host_set->dev, sg_dma_address(&sg[0]), sg_dma_len(&sg[0]), dir); + /* restore sg */ + sg->length += qc->pad_len; + if (pad_buf) + memcpy(qc->buf_virt + sg->length - qc->pad_len, + pad_buf, qc->pad_len); + } qc->flags &= ~ATA_QCFLAG_DMAMAP; - qc->sg = NULL; + qc->__sg = NULL; } /** @@ -2424,15 +2446,15 @@ static void ata_sg_clean(struct ata_queued_cmd *qc) */ static void ata_fill_sg(struct ata_queued_cmd *qc) { - struct scatterlist *sg = qc->sg; struct ata_port *ap = qc->ap; - unsigned int idx, nelem; + struct scatterlist *sg; + unsigned int idx; - assert(sg != NULL); + assert(qc->__sg != NULL); assert(qc->n_elem > 0); idx = 0; - for (nelem = qc->n_elem; nelem; nelem--,sg++) { + ata_for_each_sg(sg, qc) { u32 addr, offset; u32 sg_len, len; @@ -2518,12 +2540,18 @@ void ata_qc_prep(struct ata_queued_cmd *qc) void ata_sg_init_one(struct ata_queued_cmd *qc, void *buf, unsigned int buflen) { + struct scatterlist *sg; + qc->flags |= ATA_QCFLAG_SINGLE; - qc->sg = &qc->sgent; + memset(&qc->sgent, 0, sizeof(qc->sgent)); + qc->__sg = &qc->sgent; qc->n_elem = 1; + qc->orig_n_elem = 1; qc->buf_virt = buf; - sg_init_one(qc->sg, buf, buflen); + + sg = qc->__sg; + sg_init_one(sg, buf, buflen); } /** @@ -2544,8 +2572,9 @@ void ata_sg_init(struct ata_queued_cmd *qc, struct scatterlist *sg, unsigned int n_elem) { qc->flags |= ATA_QCFLAG_SG; - qc->sg = sg; + qc->__sg = sg; qc->n_elem = n_elem; + qc->orig_n_elem = n_elem; } /** @@ -2565,9 +2594,32 @@ static int ata_sg_setup_one(struct ata_queued_cmd *qc) { struct ata_port *ap = qc->ap; int dir = qc->dma_dir; - struct scatterlist *sg = qc->sg; + struct scatterlist *sg = qc->__sg; dma_addr_t dma_address; + /* we must lengthen transfers to end on a 32-bit boundary */ + qc->pad_len = sg->length & 3; + if (qc->pad_len) { + void *pad_buf = ap->pad + (qc->tag * ATA_DMA_PAD_SZ); + struct scatterlist *psg = &qc->pad_sgent; + + assert(qc->dev->class == ATA_DEV_ATAPI); + + memset(pad_buf, 0, ATA_DMA_PAD_SZ); + + if (qc->tf.flags & ATA_TFLAG_WRITE) + memcpy(pad_buf, qc->buf_virt + sg->length - qc->pad_len, + qc->pad_len); + + sg_dma_address(psg) = ap->pad_dma + (qc->tag * ATA_DMA_PAD_SZ); + sg_dma_len(psg) = ATA_DMA_PAD_SZ; + /* trim sg */ + sg->length -= qc->pad_len; + + DPRINTK("padding done, sg->length=%u pad_len=%u\n", + sg->length, qc->pad_len); + } + dma_address = dma_map_single(ap->host_set->dev, qc->buf_virt, sg->length, dir); if (dma_mapping_error(dma_address)) @@ -2599,12 +2651,47 @@ static int ata_sg_setup_one(struct ata_queued_cmd *qc) static int ata_sg_setup(struct ata_queued_cmd *qc) { struct ata_port *ap = qc->ap; - struct scatterlist *sg = qc->sg; + struct scatterlist *sg = qc->__sg; + struct scatterlist *lsg = &sg[qc->n_elem - 1]; int n_elem, dir; VPRINTK("ENTER, ata%u\n", ap->id); assert(qc->flags & ATA_QCFLAG_SG); + /* we must lengthen transfers to end on a 32-bit boundary */ + qc->pad_len = lsg->length & 3; + if (qc->pad_len) { + void *pad_buf = ap->pad + (qc->tag * ATA_DMA_PAD_SZ); + struct scatterlist *psg = &qc->pad_sgent; + unsigned int offset; + + assert(qc->dev->class == ATA_DEV_ATAPI); + + memset(pad_buf, 0, ATA_DMA_PAD_SZ); + + /* + * psg->page/offset are used to copy to-be-written + * data in this function or read data in ata_sg_clean. + */ + offset = lsg->offset + lsg->length - qc->pad_len; + psg->page = nth_page(lsg->page, offset >> PAGE_SHIFT); + psg->offset = offset_in_page(offset); + + if (qc->tf.flags & ATA_TFLAG_WRITE) { + void *addr = kmap_atomic(psg->page, KM_IRQ0); + memcpy(pad_buf, addr + psg->offset, qc->pad_len); + kunmap_atomic(psg->page, KM_IRQ0); + } + + sg_dma_address(psg) = ap->pad_dma + (qc->tag * ATA_DMA_PAD_SZ); + sg_dma_len(psg) = ATA_DMA_PAD_SZ; + /* trim last sg */ + lsg->length -= qc->pad_len; + + DPRINTK("padding done, sg[%d].length=%u pad_len=%u\n", + qc->n_elem - 1, lsg->length, qc->pad_len); + } + dir = qc->dma_dir; n_elem = dma_map_sg(ap->host_set->dev, sg, qc->n_elem, dir); if (n_elem < 1) @@ -2880,7 +2967,7 @@ static void ata_data_xfer(struct ata_port *ap, unsigned char *buf, static void ata_pio_sector(struct ata_queued_cmd *qc) { int do_write = (qc->tf.flags & ATA_TFLAG_WRITE); - struct scatterlist *sg = qc->sg; + struct scatterlist *sg = qc->__sg; struct ata_port *ap = qc->ap; struct page *page; unsigned int offset; @@ -2930,7 +3017,7 @@ static void ata_pio_sector(struct ata_queued_cmd *qc) static void __atapi_pio_bytes(struct ata_queued_cmd *qc, unsigned int bytes) { int do_write = (qc->tf.flags & ATA_TFLAG_WRITE); - struct scatterlist *sg = qc->sg; + struct scatterlist *sg = qc->__sg; struct ata_port *ap = qc->ap; struct page *page; unsigned char *buf; @@ -2963,7 +3050,7 @@ next_sg: return; } - sg = &qc->sg[qc->cursg]; + sg = &qc->__sg[qc->cursg]; page = sg->page; offset = sg->offset + qc->cursg_ofs; @@ -3320,7 +3407,7 @@ struct ata_queued_cmd *ata_qc_new_init(struct ata_port *ap, qc = ata_qc_new(ap); if (qc) { - qc->sg = NULL; + qc->__sg = NULL; qc->flags = 0; qc->scsicmd = NULL; qc->ap = ap; @@ -4004,11 +4091,18 @@ err_out: int ata_port_start (struct ata_port *ap) { struct device *dev = ap->host_set->dev; + int rc; ap->prd = dma_alloc_coherent(dev, ATA_PRD_TBL_SZ, &ap->prd_dma, GFP_KERNEL); if (!ap->prd) return -ENOMEM; + rc = ata_pad_alloc(ap, dev); + if (rc) { + dma_free_coherent(dev, ATA_PRD_TBL_SZ, ap->prd, ap->prd_dma); + return rc; + } + DPRINTK("prd alloc, virt %p, dma %llx\n", ap->prd, (unsigned long long) ap->prd_dma); return 0; @@ -4032,6 +4126,7 @@ void ata_port_stop (struct ata_port *ap) struct device *dev = ap->host_set->dev; dma_free_coherent(dev, ATA_PRD_TBL_SZ, ap->prd, ap->prd_dma); + ata_pad_free(ap, dev); } void ata_host_stop (struct ata_host_set *host_set) diff --git a/drivers/scsi/libata-scsi.c b/drivers/scsi/libata-scsi.c index 248baae9648..eb604b0a899 100644 --- a/drivers/scsi/libata-scsi.c +++ b/drivers/scsi/libata-scsi.c @@ -355,10 +355,10 @@ struct ata_queued_cmd *ata_scsi_qc_new(struct ata_port *ap, qc->scsidone = done; if (cmd->use_sg) { - qc->sg = (struct scatterlist *) cmd->request_buffer; + qc->__sg = (struct scatterlist *) cmd->request_buffer; qc->n_elem = cmd->use_sg; } else { - qc->sg = &qc->sgent; + qc->__sg = &qc->sgent; qc->n_elem = 1; } } else { @@ -702,6 +702,16 @@ int ata_scsi_slave_config(struct scsi_device *sdev) */ blk_queue_max_sectors(sdev->request_queue, 2048); } + + /* + * SATA DMA transfers must be multiples of 4 byte, so + * we need to pad ATAPI transfers using an extra sg. + * Decrement max hw segments accordingly. + */ + if (dev->class == ATA_DEV_ATAPI) { + request_queue_t *q = sdev->request_queue; + blk_queue_max_hw_segments(q, q->max_hw_segments - 1); + } } return 0; /* scsi layer doesn't check return value, sigh */ diff --git a/drivers/scsi/pdc_adma.c b/drivers/scsi/pdc_adma.c index 665017eda8a..a50588c60fa 100644 --- a/drivers/scsi/pdc_adma.c +++ b/drivers/scsi/pdc_adma.c @@ -293,14 +293,14 @@ static void adma_eng_timeout(struct ata_port *ap) static int adma_fill_sg(struct ata_queued_cmd *qc) { - struct scatterlist *sg = qc->sg; + struct scatterlist *sg; struct ata_port *ap = qc->ap; struct adma_port_priv *pp = ap->private_data; u8 *buf = pp->pkt; - int nelem, i = (2 + buf[3]) * 8; + int i = (2 + buf[3]) * 8; u8 pFLAGS = pORD | ((qc->tf.flags & ATA_TFLAG_WRITE) ? pDIRO : 0); - for (nelem = 0; nelem < qc->n_elem; nelem++,sg++) { + ata_for_each_sg(sg, qc) { u32 addr; u32 len; @@ -312,7 +312,7 @@ static int adma_fill_sg(struct ata_queued_cmd *qc) *(__le32 *)(buf + i) = cpu_to_le32(len); i += 4; - if ((nelem + 1) == qc->n_elem) + if (ata_sg_is_last(sg, qc)) pFLAGS |= pEND; buf[i++] = pFLAGS; buf[i++] = qc->dev->dma_mode & 0xf; diff --git a/drivers/scsi/sata_mv.c b/drivers/scsi/sata_mv.c index 46dbdee79f7..0f469e3dabe 100644 --- a/drivers/scsi/sata_mv.c +++ b/drivers/scsi/sata_mv.c @@ -670,6 +670,11 @@ static void mv_host_stop(struct ata_host_set *host_set) ata_host_stop(host_set); } +static inline void mv_priv_free(struct mv_port_priv *pp, struct device *dev) +{ + dma_free_coherent(dev, MV_PORT_PRIV_DMA_SZ, pp->crpb, pp->crpb_dma); +} + /** * mv_port_start - Port specific init/start routine. * @ap: ATA channel to manipulate @@ -687,21 +692,23 @@ static int mv_port_start(struct ata_port *ap) void __iomem *port_mmio = mv_ap_base(ap); void *mem; dma_addr_t mem_dma; + int rc = -ENOMEM; pp = kmalloc(sizeof(*pp), GFP_KERNEL); - if (!pp) { - return -ENOMEM; - } + if (!pp) + goto err_out; memset(pp, 0, sizeof(*pp)); mem = dma_alloc_coherent(dev, MV_PORT_PRIV_DMA_SZ, &mem_dma, GFP_KERNEL); - if (!mem) { - kfree(pp); - return -ENOMEM; - } + if (!mem) + goto err_out_pp; memset(mem, 0, MV_PORT_PRIV_DMA_SZ); + rc = ata_pad_alloc(ap, dev); + if (rc) + goto err_out_priv; + /* First item in chunk of DMA memory: * 32-slot command request table (CRQB), 32 bytes each in size */ @@ -746,6 +753,13 @@ static int mv_port_start(struct ata_port *ap) */ ap->private_data = pp; return 0; + +err_out_priv: + mv_priv_free(pp, dev); +err_out_pp: + kfree(pp); +err_out: + return rc; } /** @@ -768,7 +782,8 @@ static void mv_port_stop(struct ata_port *ap) spin_unlock_irqrestore(&ap->host_set->lock, flags); ap->private_data = NULL; - dma_free_coherent(dev, MV_PORT_PRIV_DMA_SZ, pp->crpb, pp->crpb_dma); + ata_pad_free(ap, dev); + mv_priv_free(pp, dev); kfree(pp); } @@ -784,23 +799,24 @@ static void mv_port_stop(struct ata_port *ap) static void mv_fill_sg(struct ata_queued_cmd *qc) { struct mv_port_priv *pp = qc->ap->private_data; - unsigned int i; + unsigned int i = 0; + struct scatterlist *sg; - for (i = 0; i < qc->n_elem; i++) { + ata_for_each_sg(sg, qc) { u32 sg_len; dma_addr_t addr; - addr = sg_dma_address(&qc->sg[i]); - sg_len = sg_dma_len(&qc->sg[i]); + addr = sg_dma_address(sg); + sg_len = sg_dma_len(sg); pp->sg_tbl[i].addr = cpu_to_le32(addr & 0xffffffff); pp->sg_tbl[i].addr_hi = cpu_to_le32((addr >> 16) >> 16); assert(0 == (sg_len & ~MV_DMA_BOUNDARY)); pp->sg_tbl[i].flags_size = cpu_to_le32(sg_len); - } - if (0 < qc->n_elem) { - pp->sg_tbl[qc->n_elem - 1].flags_size |= - cpu_to_le32(EPRD_FLAG_END_OF_TBL); + if (ata_sg_is_last(sg, qc)) + pp->sg_tbl[i].flags_size |= cpu_to_le32(EPRD_FLAG_END_OF_TBL); + + i++; } } diff --git a/drivers/scsi/sata_qstor.c b/drivers/scsi/sata_qstor.c index 9938dae782b..65502c157a5 100644 --- a/drivers/scsi/sata_qstor.c +++ b/drivers/scsi/sata_qstor.c @@ -271,16 +271,17 @@ static void qs_scr_write (struct ata_port *ap, unsigned int sc_reg, u32 val) static void qs_fill_sg(struct ata_queued_cmd *qc) { - struct scatterlist *sg = qc->sg; + struct scatterlist *sg; struct ata_port *ap = qc->ap; struct qs_port_priv *pp = ap->private_data; unsigned int nelem; u8 *prd = pp->pkt + QS_CPB_BYTES; - assert(sg != NULL); + assert(qc->__sg != NULL); assert(qc->n_elem > 0); - for (nelem = 0; nelem < qc->n_elem; nelem++,sg++) { + nelem = 0; + ata_for_each_sg(sg, qc) { u64 addr; u32 len; @@ -294,6 +295,7 @@ static void qs_fill_sg(struct ata_queued_cmd *qc) VPRINTK("PRD[%u] = (0x%llX, 0x%X)\n", nelem, (unsigned long long)addr, len); + nelem++; } } diff --git a/drivers/scsi/sata_sil24.c b/drivers/scsi/sata_sil24.c index c6654802565..e6c8e89c226 100644 --- a/drivers/scsi/sata_sil24.c +++ b/drivers/scsi/sata_sil24.c @@ -411,15 +411,20 @@ static void sil24_phy_reset(struct ata_port *ap) static inline void sil24_fill_sg(struct ata_queued_cmd *qc, struct sil24_cmd_block *cb) { - struct scatterlist *sg = qc->sg; struct sil24_sge *sge = cb->sge; - unsigned i; + struct scatterlist *sg; + unsigned int idx = 0; - for (i = 0; i < qc->n_elem; i++, sg++, sge++) { + ata_for_each_sg(sg, qc) { sge->addr = cpu_to_le64(sg_dma_address(sg)); sge->cnt = cpu_to_le32(sg_dma_len(sg)); - sge->flags = 0; - sge->flags = i < qc->n_elem - 1 ? 0 : cpu_to_le32(SGE_TRM); + if (ata_sg_is_last(sg, qc)) + sge->flags = cpu_to_le32(SGE_TRM); + else + sge->flags = 0; + + sge++; + idx++; } } @@ -630,6 +635,13 @@ static irqreturn_t sil24_interrupt(int irq, void *dev_instance, struct pt_regs * return IRQ_RETVAL(handled); } +static inline void sil24_cblk_free(struct sil24_port_priv *pp, struct device *dev) +{ + const size_t cb_size = sizeof(*pp->cmd_block); + + dma_free_coherent(dev, cb_size, pp->cmd_block, pp->cmd_block_dma); +} + static int sil24_port_start(struct ata_port *ap) { struct device *dev = ap->host_set->dev; @@ -637,36 +649,44 @@ static int sil24_port_start(struct ata_port *ap) struct sil24_cmd_block *cb; size_t cb_size = sizeof(*cb); dma_addr_t cb_dma; + int rc = -ENOMEM; - pp = kmalloc(sizeof(*pp), GFP_KERNEL); + pp = kzalloc(sizeof(*pp), GFP_KERNEL); if (!pp) - return -ENOMEM; - memset(pp, 0, sizeof(*pp)); + goto err_out; pp->tf.command = ATA_DRDY; cb = dma_alloc_coherent(dev, cb_size, &cb_dma, GFP_KERNEL); - if (!cb) { - kfree(pp); - return -ENOMEM; - } + if (!cb) + goto err_out_pp; memset(cb, 0, cb_size); + rc = ata_pad_alloc(ap, dev); + if (rc) + goto err_out_pad; + pp->cmd_block = cb; pp->cmd_block_dma = cb_dma; ap->private_data = pp; return 0; + +err_out_pad: + sil24_cblk_free(pp, dev); +err_out_pp: + kfree(pp); +err_out: + return rc; } static void sil24_port_stop(struct ata_port *ap) { struct device *dev = ap->host_set->dev; struct sil24_port_priv *pp = ap->private_data; - size_t cb_size = sizeof(*pp->cmd_block); - dma_free_coherent(dev, cb_size, pp->cmd_block, pp->cmd_block_dma); + sil24_cblk_free(pp, dev); kfree(pp); } diff --git a/drivers/scsi/sata_sx4.c b/drivers/scsi/sata_sx4.c index 0ec21e09f5d..f859bbd681e 100644 --- a/drivers/scsi/sata_sx4.c +++ b/drivers/scsi/sata_sx4.c @@ -450,14 +450,14 @@ static inline void pdc20621_host_pkt(struct ata_taskfile *tf, u8 *buf, static void pdc20621_dma_prep(struct ata_queued_cmd *qc) { - struct scatterlist *sg = qc->sg; + struct scatterlist *sg; struct ata_port *ap = qc->ap; struct pdc_port_priv *pp = ap->private_data; void __iomem *mmio = ap->host_set->mmio_base; struct pdc_host_priv *hpriv = ap->host_set->private_data; void __iomem *dimm_mmio = hpriv->dimm_mmio; unsigned int portno = ap->port_no; - unsigned int i, last, idx, total_len = 0, sgt_len; + unsigned int i, idx, total_len = 0, sgt_len; u32 *buf = (u32 *) &pp->dimm_buf[PDC_DIMM_HEADER_SZ]; assert(qc->flags & ATA_QCFLAG_DMAMAP); @@ -470,12 +470,11 @@ static void pdc20621_dma_prep(struct ata_queued_cmd *qc) /* * Build S/G table */ - last = qc->n_elem; idx = 0; - for (i = 0; i < last; i++) { - buf[idx++] = cpu_to_le32(sg_dma_address(&sg[i])); - buf[idx++] = cpu_to_le32(sg_dma_len(&sg[i])); - total_len += sg_dma_len(&sg[i]); + ata_for_each_sg(sg, qc) { + buf[idx++] = cpu_to_le32(sg_dma_address(sg)); + buf[idx++] = cpu_to_le32(sg_dma_len(sg)); + total_len += sg_dma_len(sg); } buf[idx - 1] |= cpu_to_le32(ATA_PRD_EOT); sgt_len = idx * 4; diff --git a/drivers/serial/8250_early.c b/drivers/serial/8250_early.c index b7a5dd71022..59ba5d993b4 100644 --- a/drivers/serial/8250_early.c +++ b/drivers/serial/8250_early.c @@ -164,7 +164,7 @@ static int __init parse_options(struct early_uart_device *device, char *options) if ((options = strchr(options, ','))) { options++; - device->baud = simple_strtoul(options, 0, 0); + device->baud = simple_strtoul(options, NULL, 0); length = min(strcspn(options, " "), sizeof(device->options)); strncpy(device->options, options, length); } else { diff --git a/drivers/serial/clps711x.c b/drivers/serial/clps711x.c index 6a67e8f585b..87ef368384f 100644 --- a/drivers/serial/clps711x.c +++ b/drivers/serial/clps711x.c @@ -408,11 +408,7 @@ static struct uart_port clps711x_ports[UART_NR] = { { .iobase = SYSCON1, .irq = IRQ_UTXINT1, /* IRQ_URXINT1, IRQ_UMSINT */ -#ifdef CONFIG_MP1000_90MHZ - .uartclk = 4515840, -#else .uartclk = 3686400, -#endif .fifosize = 16, .ops = &clps711x_pops, .line = 0, @@ -421,11 +417,7 @@ static struct uart_port clps711x_ports[UART_NR] = { { .iobase = SYSCON2, .irq = IRQ_UTXINT2, /* IRQ_URXINT2 */ -#ifdef CONFIG_MP1000_90MHZ - .uartclk = 4515840, -#else .uartclk = 3686400, -#endif .fifosize = 16, .ops = &clps711x_pops, .line = 1, @@ -559,7 +551,6 @@ console_initcall(clps711xuart_console_init); static struct uart_driver clps711x_reg = { .driver_name = "ttyCL", .dev_name = "ttyCL", - .devfs_name = "ttyCL", .major = SERIAL_CLPS711X_MAJOR, .minor = SERIAL_CLPS711X_MINOR, .nr = UART_NR, diff --git a/drivers/usb/host/pci-quirks.c b/drivers/usb/host/pci-quirks.c index b1aa350fd32..e46528c825b 100644 --- a/drivers/usb/host/pci-quirks.c +++ b/drivers/usb/host/pci-quirks.c @@ -316,4 +316,4 @@ static void __devinit quirk_usb_early_handoff(struct pci_dev *pdev) else if (pdev->class == PCI_CLASS_SERIAL_USB_EHCI) quirk_usb_disable_ehci(pdev); } -DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, quirk_usb_early_handoff); +DECLARE_PCI_FIXUP_FINAL(PCI_ANY_ID, PCI_ANY_ID, quirk_usb_early_handoff); diff --git a/drivers/video/nvidia/nvidia.c b/drivers/video/nvidia/nvidia.c index a7f020ada63..308defc389a 100644 --- a/drivers/video/nvidia/nvidia.c +++ b/drivers/video/nvidia/nvidia.c @@ -384,6 +384,14 @@ static struct pci_device_id nvidiafb_pci_tbl[] = { PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, {PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_GEFORCE_6800B_GT, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_GEFORCE_7800_GT, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_GEFORCE_7800_GTX, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_GEFORCE_GO_7800, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, + {PCI_VENDOR_ID_NVIDIA, PCI_DEVICE_ID_NVIDIA_GEFORCE_GO_7800_GTX, + PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, {PCI_VENDOR_ID_NVIDIA, 0x021d, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, {PCI_VENDOR_ID_NVIDIA, 0x021e, |