diff options
Diffstat (limited to 'drivers/scsi/scsi_lib.c')
| -rw-r--r-- | drivers/scsi/scsi_lib.c | 3118 |
1 files changed, 1596 insertions, 1522 deletions
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index f6097b89d5d3..51ad2ad07e43 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0-only /* * Copyright (C) 1999 Eric Youngdale * Copyright (C) 2014 Christoph Hellwig @@ -20,8 +21,9 @@ #include <linux/hardirq.h> #include <linux/scatterlist.h> #include <linux/blk-mq.h> +#include <linux/blk-integrity.h> #include <linux/ratelimit.h> -#include <asm/unaligned.h> +#include <linux/unaligned.h> #include <scsi/scsi.h> #include <scsi/scsi_cmnd.h> @@ -30,7 +32,7 @@ #include <scsi/scsi_driver.h> #include <scsi/scsi_eh.h> #include <scsi/scsi_host.h> -#include <scsi/scsi_transport.h> /* __scsi_init_queue() */ +#include <scsi/scsi_transport.h> /* scsi_init_limits() */ #include <scsi/scsi_dh.h> #include <trace/events/scsi.h> @@ -39,67 +41,40 @@ #include "scsi_priv.h" #include "scsi_logging.h" -static struct kmem_cache *scsi_sdb_cache; +/* + * Size of integrity metadata is usually small, 1 inline sg should + * cover normal cases. + */ +#ifdef CONFIG_ARCH_NO_SG_CHAIN +#define SCSI_INLINE_PROT_SG_CNT 0 +#define SCSI_INLINE_SG_CNT 0 +#else +#define SCSI_INLINE_PROT_SG_CNT 1 +#define SCSI_INLINE_SG_CNT 2 +#endif + static struct kmem_cache *scsi_sense_cache; -static struct kmem_cache *scsi_sense_isadma_cache; static DEFINE_MUTEX(scsi_sense_cache_mutex); -static inline struct kmem_cache * -scsi_select_sense_cache(bool unchecked_isa_dma) -{ - return unchecked_isa_dma ? scsi_sense_isadma_cache : scsi_sense_cache; -} - -static void scsi_free_sense_buffer(bool unchecked_isa_dma, - unsigned char *sense_buffer) -{ - kmem_cache_free(scsi_select_sense_cache(unchecked_isa_dma), - sense_buffer); -} - -static unsigned char *scsi_alloc_sense_buffer(bool unchecked_isa_dma, - gfp_t gfp_mask, int numa_node) -{ - return kmem_cache_alloc_node(scsi_select_sense_cache(unchecked_isa_dma), - gfp_mask, numa_node); -} +static void scsi_mq_uninit_cmd(struct scsi_cmnd *cmd); int scsi_init_sense_cache(struct Scsi_Host *shost) { - struct kmem_cache *cache; int ret = 0; - cache = scsi_select_sense_cache(shost->unchecked_isa_dma); - if (cache) - return 0; - mutex_lock(&scsi_sense_cache_mutex); - if (shost->unchecked_isa_dma) { - scsi_sense_isadma_cache = - kmem_cache_create("scsi_sense_cache(DMA)", - SCSI_SENSE_BUFFERSIZE, 0, - SLAB_HWCACHE_ALIGN | SLAB_CACHE_DMA, NULL); - if (!scsi_sense_isadma_cache) - ret = -ENOMEM; - } else { + if (!scsi_sense_cache) { scsi_sense_cache = - kmem_cache_create("scsi_sense_cache", - SCSI_SENSE_BUFFERSIZE, 0, SLAB_HWCACHE_ALIGN, NULL); + kmem_cache_create_usercopy("scsi_sense_cache", + SCSI_SENSE_BUFFERSIZE, 0, SLAB_HWCACHE_ALIGN, + 0, SCSI_SENSE_BUFFERSIZE, NULL); if (!scsi_sense_cache) ret = -ENOMEM; } - mutex_unlock(&scsi_sense_cache_mutex); return ret; } -/* - * When to reinvoke queueing after a resource shortage. It's 3 msecs to - * not change behaviour from the previous unplug mechanism, experimentation - * may prove this needs changing. - */ -#define SCSI_QUEUE_DELAY 3 - static void scsi_set_blocked(struct scsi_cmnd *cmd, int reason) { @@ -136,12 +111,20 @@ scsi_set_blocked(struct scsi_cmnd *cmd, int reason) } } -static void scsi_mq_requeue_cmd(struct scsi_cmnd *cmd) +static void scsi_mq_requeue_cmd(struct scsi_cmnd *cmd, unsigned long msecs) { - struct scsi_device *sdev = cmd->device; + struct request *rq = scsi_cmd_to_rq(cmd); + + if (rq->rq_flags & RQF_DONTPREP) { + rq->rq_flags &= ~RQF_DONTPREP; + scsi_mq_uninit_cmd(cmd); + } else { + WARN_ON_ONCE(true); + } - blk_mq_requeue_request(cmd->request, true); - put_device(&sdev->sdev_gendev); + blk_mq_requeue_request(rq, false); + if (!scsi_host_in_recovery(cmd->device->host)) + blk_mq_delay_kick_requeue_list(rq->q, msecs); } /** @@ -156,11 +139,9 @@ static void scsi_mq_requeue_cmd(struct scsi_cmnd *cmd) * for a requeue after completion, which should only occur in this * file. */ -static void __scsi_queue_insert(struct scsi_cmnd *cmd, int reason, int unbusy) +static void __scsi_queue_insert(struct scsi_cmnd *cmd, int reason, bool unbusy) { struct scsi_device *device = cmd->device; - struct request_queue *q = device->request_queue; - unsigned long flags; SCSI_LOG_MLQUEUE(1, scmd_printk(KERN_INFO, cmd, "Inserting command %p into mlqueue\n", cmd)); @@ -172,100 +153,187 @@ static void __scsi_queue_insert(struct scsi_cmnd *cmd, int reason, int unbusy) * active on the host/device. */ if (unbusy) - scsi_device_unbusy(device); + scsi_device_unbusy(device, cmd); /* * Requeue this command. It will go before all other commands * that are already in the queue. Schedule requeue work under * lock such that the kblockd_schedule_work() call happens - * before blk_cleanup_queue() finishes. + * before blk_mq_destroy_queue() finishes. */ cmd->result = 0; - if (q->mq_ops) { - scsi_mq_requeue_cmd(cmd); - return; - } - spin_lock_irqsave(q->queue_lock, flags); - blk_requeue_request(q, cmd->request); - kblockd_schedule_work(&device->requeue_work); - spin_unlock_irqrestore(q->queue_lock, flags); + + blk_mq_requeue_request(scsi_cmd_to_rq(cmd), + !scsi_host_in_recovery(cmd->device->host)); } -/* - * Function: scsi_queue_insert() - * - * Purpose: Insert a command in the midlevel queue. - * - * Arguments: cmd - command that we are adding to queue. - * reason - why we are inserting command to queue. - * - * Lock status: Assumed that lock is not held upon entry. +/** + * scsi_queue_insert - Reinsert a command in the queue. + * @cmd: command that we are adding to queue. + * @reason: why we are inserting command to queue. * - * Returns: Nothing. + * We do this for one of two cases. Either the host is busy and it cannot accept + * any more commands for the time being, or the device returned QUEUE_FULL and + * can accept no more commands. * - * Notes: We do this for one of two cases. Either the host is busy - * and it cannot accept any more commands for the time being, - * or the device returned QUEUE_FULL and can accept no more - * commands. - * Notes: This could be called either from an interrupt context or a - * normal process context. + * Context: This could be called either from an interrupt context or a normal + * process context. */ void scsi_queue_insert(struct scsi_cmnd *cmd, int reason) { - __scsi_queue_insert(cmd, reason, 1); + __scsi_queue_insert(cmd, reason, true); } +/** + * scsi_failures_reset_retries - reset all failures to zero + * @failures: &struct scsi_failures with specific failure modes set + */ +void scsi_failures_reset_retries(struct scsi_failures *failures) +{ + struct scsi_failure *failure; + + failures->total_retries = 0; + + for (failure = failures->failure_definitions; failure->result; + failure++) + failure->retries = 0; +} +EXPORT_SYMBOL_GPL(scsi_failures_reset_retries); /** - * scsi_execute - insert request and wait for the result - * @sdev: scsi device + * scsi_check_passthrough - Determine if passthrough scsi_cmnd needs a retry. + * @scmd: scsi_cmnd to check. + * @failures: scsi_failures struct that lists failures to check for. + * + * Returns -EAGAIN if the caller should retry else 0. + */ +static int scsi_check_passthrough(struct scsi_cmnd *scmd, + struct scsi_failures *failures) +{ + struct scsi_failure *failure; + struct scsi_sense_hdr sshdr; + enum sam_status status; + + if (!scmd->result) + return 0; + + if (!failures) + return 0; + + for (failure = failures->failure_definitions; failure->result; + failure++) { + if (failure->result == SCMD_FAILURE_RESULT_ANY) + goto maybe_retry; + + if (host_byte(scmd->result) && + host_byte(scmd->result) == host_byte(failure->result)) + goto maybe_retry; + + status = status_byte(scmd->result); + if (!status) + continue; + + if (failure->result == SCMD_FAILURE_STAT_ANY && + !scsi_status_is_good(scmd->result)) + goto maybe_retry; + + if (status != status_byte(failure->result)) + continue; + + if (status_byte(failure->result) != SAM_STAT_CHECK_CONDITION || + failure->sense == SCMD_FAILURE_SENSE_ANY) + goto maybe_retry; + + if (!scsi_command_normalize_sense(scmd, &sshdr)) + return 0; + + if (failure->sense != sshdr.sense_key) + continue; + + if (failure->asc == SCMD_FAILURE_ASC_ANY) + goto maybe_retry; + + if (failure->asc != sshdr.asc) + continue; + + if (failure->ascq == SCMD_FAILURE_ASCQ_ANY || + failure->ascq == sshdr.ascq) + goto maybe_retry; + } + + return 0; + +maybe_retry: + if (failure->allowed) { + if (failure->allowed == SCMD_FAILURE_NO_LIMIT || + ++failure->retries <= failure->allowed) + return -EAGAIN; + } else { + if (failures->total_allowed == SCMD_FAILURE_NO_LIMIT || + ++failures->total_retries <= failures->total_allowed) + return -EAGAIN; + } + + return 0; +} + +/** + * scsi_execute_cmd - insert request and wait for the result + * @sdev: scsi_device * @cmd: scsi command - * @data_direction: data direction + * @opf: block layer request cmd_flags * @buffer: data buffer * @bufflen: len of buffer - * @sense: optional sense buffer - * @sshdr: optional decoded sense header - * @timeout: request timeout in seconds - * @retries: number of times to retry request - * @flags: flags for ->cmd_flags - * @rq_flags: flags for ->rq_flags - * @resid: optional residual length + * @timeout: request timeout in HZ + * @ml_retries: number of times SCSI midlayer will retry request + * @args: Optional args. See struct definition for field descriptions * * Returns the scsi_cmnd result field if a command was executed, or a negative * Linux error code if we didn't get that far. */ -int scsi_execute(struct scsi_device *sdev, const unsigned char *cmd, - int data_direction, void *buffer, unsigned bufflen, - unsigned char *sense, struct scsi_sense_hdr *sshdr, - int timeout, int retries, u64 flags, req_flags_t rq_flags, - int *resid) +int scsi_execute_cmd(struct scsi_device *sdev, const unsigned char *cmd, + blk_opf_t opf, void *buffer, unsigned int bufflen, + int timeout, int ml_retries, + const struct scsi_exec_args *args) { + static const struct scsi_exec_args default_args; struct request *req; - struct scsi_request *rq; - int ret = DRIVER_ERROR << 24; + struct scsi_cmnd *scmd; + int ret; - req = blk_get_request(sdev->request_queue, - data_direction == DMA_TO_DEVICE ? - REQ_OP_SCSI_OUT : REQ_OP_SCSI_IN, __GFP_RECLAIM); - if (IS_ERR(req)) - return ret; - rq = scsi_req(req); + if (!args) + args = &default_args; + else if (WARN_ON_ONCE(args->sense && + args->sense_len != SCSI_SENSE_BUFFERSIZE)) + return -EINVAL; - if (bufflen && blk_rq_map_kern(sdev->request_queue, req, - buffer, bufflen, __GFP_RECLAIM)) - goto out; +retry: + req = scsi_alloc_request(sdev->request_queue, opf, args->req_flags); + if (IS_ERR(req)) + return PTR_ERR(req); - rq->cmd_len = COMMAND_SIZE(cmd[0]); - memcpy(rq->cmd, cmd, rq->cmd_len); - rq->retries = retries; + if (bufflen) { + ret = blk_rq_map_kern(req, buffer, bufflen, GFP_NOIO); + if (ret) + goto out; + } + scmd = blk_mq_rq_to_pdu(req); + scmd->cmd_len = COMMAND_SIZE(cmd[0]); + memcpy(scmd->cmnd, cmd, scmd->cmd_len); + scmd->allowed = ml_retries; + scmd->flags |= args->scmd_flags; req->timeout = timeout; - req->cmd_flags |= flags; - req->rq_flags |= rq_flags | RQF_QUIET | RQF_PREEMPT; + req->rq_flags |= RQF_QUIET; /* * head injection *required* here otherwise quiesce won't work */ - blk_execute_rq(req->q, NULL, req, 1); + blk_execute_rq(req, true); + + if (scsi_check_passthrough(scmd, args->failures) == -EAGAIN) { + blk_mq_free_request(req); + goto retry; + } /* * Some devices (USB mass-storage in particular) may transfer @@ -273,69 +341,76 @@ int scsi_execute(struct scsi_device *sdev, const unsigned char *cmd, * is invalid. Prevent the garbage from being misinterpreted * and prevent security leaks by zeroing out the excess data. */ - if (unlikely(rq->resid_len > 0 && rq->resid_len <= bufflen)) - memset(buffer + (bufflen - rq->resid_len), 0, rq->resid_len); - - if (resid) - *resid = rq->resid_len; - if (sense && rq->sense_len) - memcpy(sense, rq->sense, SCSI_SENSE_BUFFERSIZE); - if (sshdr) - scsi_normalize_sense(rq->sense, rq->sense_len, sshdr); - ret = rq->result; + if (unlikely(scmd->resid_len > 0 && scmd->resid_len <= bufflen)) + memset(buffer + bufflen - scmd->resid_len, 0, scmd->resid_len); + + if (args->resid) + *args->resid = scmd->resid_len; + if (args->sense) + memcpy(args->sense, scmd->sense_buffer, SCSI_SENSE_BUFFERSIZE); + if (args->sshdr) + scsi_normalize_sense(scmd->sense_buffer, scmd->sense_len, + args->sshdr); + + ret = scmd->result; out: - blk_put_request(req); + blk_mq_free_request(req); return ret; } -EXPORT_SYMBOL(scsi_execute); +EXPORT_SYMBOL(scsi_execute_cmd); /* - * Function: scsi_init_cmd_errh() - * - * Purpose: Initialize cmd fields related to error handling. - * - * Arguments: cmd - command that is ready to be queued. - * - * Notes: This function has the job of initializing a number of - * fields related to error handling. Typically this will - * be called once for each command, as required. + * Wake up the error handler if necessary. Avoid as follows that the error + * handler is not woken up if host in-flight requests number == + * shost->host_failed: use call_rcu() in scsi_eh_scmd_add() in combination + * with an RCU read lock in this function to ensure that this function in + * its entirety either finishes before scsi_eh_scmd_add() increases the + * host_failed counter or that it notices the shost state change made by + * scsi_eh_scmd_add(). */ -static void scsi_init_cmd_errh(struct scsi_cmnd *cmd) +static void scsi_dec_host_busy(struct Scsi_Host *shost, struct scsi_cmnd *cmd) { - cmd->serial_number = 0; - scsi_set_resid(cmd, 0); - memset(cmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE); - if (cmd->cmd_len == 0) - cmd->cmd_len = scsi_command_size(cmd->cmnd); + unsigned long flags; + + rcu_read_lock(); + __clear_bit(SCMD_STATE_INFLIGHT, &cmd->state); + if (unlikely(scsi_host_in_recovery(shost))) { + unsigned int busy = scsi_host_busy(shost); + + spin_lock_irqsave(shost->host_lock, flags); + if (shost->host_failed || shost->host_eh_scheduled) + scsi_eh_wakeup(shost, busy); + spin_unlock_irqrestore(shost->host_lock, flags); + } + rcu_read_unlock(); } -void scsi_device_unbusy(struct scsi_device *sdev) +void scsi_device_unbusy(struct scsi_device *sdev, struct scsi_cmnd *cmd) { struct Scsi_Host *shost = sdev->host; struct scsi_target *starget = scsi_target(sdev); - unsigned long flags; - atomic_dec(&shost->host_busy); + scsi_dec_host_busy(shost, cmd); + if (starget->can_queue > 0) atomic_dec(&starget->target_busy); - if (unlikely(scsi_host_in_recovery(shost) && - (shost->host_failed || shost->host_eh_scheduled))) { - spin_lock_irqsave(shost->host_lock, flags); - scsi_eh_wakeup(shost); - spin_unlock_irqrestore(shost->host_lock, flags); - } - - atomic_dec(&sdev->device_busy); + if (sdev->budget_map.map) + sbitmap_put(&sdev->budget_map, cmd->budget_token); + cmd->budget_token = -1; } -static void scsi_kick_queue(struct request_queue *q) +/* + * Kick the queue of SCSI device @sdev if @sdev != current_sdev. Called with + * interrupts disabled. + */ +static void scsi_kick_sdev_queue(struct scsi_device *sdev, void *data) { - if (q->mq_ops) - blk_mq_start_hw_queues(q); - else - blk_run_queue(q); + struct scsi_device *current_sdev = data; + + if (sdev != current_sdev) + blk_mq_run_hw_queues(sdev->request_queue, true); } /* @@ -348,7 +423,6 @@ static void scsi_kick_queue(struct request_queue *q) static void scsi_single_lun_run(struct scsi_device *current_sdev) { struct Scsi_Host *shost = current_sdev->host; - struct scsi_device *sdev, *tmp; struct scsi_target *starget = scsi_target(current_sdev); unsigned long flags; @@ -362,31 +436,19 @@ static void scsi_single_lun_run(struct scsi_device *current_sdev) * but in most cases, we will be first. Ideally, each LU on the * target would get some limited time or requests on the target. */ - scsi_kick_queue(current_sdev->request_queue); + blk_mq_run_hw_queues(current_sdev->request_queue, + shost->queuecommand_may_block); spin_lock_irqsave(shost->host_lock, flags); - if (starget->starget_sdev_user) - goto out; - list_for_each_entry_safe(sdev, tmp, &starget->devices, - same_target_siblings) { - if (sdev == current_sdev) - continue; - if (scsi_device_get(sdev)) - continue; - - spin_unlock_irqrestore(shost->host_lock, flags); - scsi_kick_queue(sdev->request_queue); - spin_lock_irqsave(shost->host_lock, flags); - - scsi_device_put(sdev); - } - out: + if (!starget->starget_sdev_user) + __starget_for_each_device(starget, current_sdev, + scsi_kick_sdev_queue); spin_unlock_irqrestore(shost->host_lock, flags); } static inline bool scsi_device_is_busy(struct scsi_device *sdev) { - if (atomic_read(&sdev->device_busy) >= sdev->queue_depth) + if (scsi_device_busy(sdev) >= sdev->queue_depth) return true; if (atomic_read(&sdev->device_blocked) > 0) return true; @@ -406,9 +468,6 @@ static inline bool scsi_target_is_busy(struct scsi_target *starget) static inline bool scsi_host_is_busy(struct Scsi_Host *shost) { - if (shost->can_queue > 0 && - atomic_read(&shost->host_busy) >= shost->can_queue) - return true; if (atomic_read(&shost->host_blocked) > 0) return true; if (shost->host_self_blocked) @@ -456,16 +515,16 @@ static void scsi_starved_list_run(struct Scsi_Host *shost) * it and the queue. Mitigate by taking a reference to the * queue and never touching the sdev again after we drop the * host lock. Note: if __scsi_remove_device() invokes - * blk_cleanup_queue() before the queue is run from this + * blk_mq_destroy_queue() before the queue is run from this * function then blk_run_queue() will return immediately since - * blk_cleanup_queue() marks the queue with QUEUE_FLAG_DYING. + * blk_mq_destroy_queue() marks the queue with QUEUE_FLAG_DYING. */ slq = sdev->request_queue; if (!blk_get_queue(slq)) continue; spin_unlock_irqrestore(shost->host_lock, flags); - scsi_kick_queue(slq); + blk_mq_run_hw_queues(slq, false); blk_put_queue(slq); spin_lock_irqsave(shost->host_lock, flags); @@ -475,17 +534,11 @@ static void scsi_starved_list_run(struct Scsi_Host *shost) spin_unlock_irqrestore(shost->host_lock, flags); } -/* - * Function: scsi_run_queue() - * - * Purpose: Select a proper request queue to serve next - * - * Arguments: q - last request's queue - * - * Returns: Nothing +/** + * scsi_run_queue - Select a proper request queue to serve next. + * @q: last request's queue * - * Notes: The previous command was completely finished, start - * a new one if possible. + * The previous command was completely finished, start a new one if possible. */ static void scsi_run_queue(struct request_queue *q) { @@ -496,10 +549,8 @@ static void scsi_run_queue(struct request_queue *q) if (!list_empty(&sdev->host->starved_list)) scsi_starved_list_run(sdev->host); - if (q->mq_ops) - blk_mq_run_hw_queues(q, false); - else - blk_run_queue(q); + /* Note: blk_mq_kick_requeue_list() runs the queue asynchronously. */ + blk_mq_kick_requeue_list(q); } void scsi_requeue_run_queue(struct work_struct *work) @@ -512,42 +563,6 @@ void scsi_requeue_run_queue(struct work_struct *work) scsi_run_queue(q); } -/* - * Function: scsi_requeue_command() - * - * Purpose: Handle post-processing of completed commands. - * - * Arguments: q - queue to operate on - * cmd - command that may need to be requeued. - * - * Returns: Nothing - * - * Notes: After command completion, there may be blocks left - * over which weren't finished by the previous command - * this can be for a number of reasons - the main one is - * I/O errors in the middle of the request, in which case - * we need to request the blocks that come after the bad - * sector. - * Notes: Upon return, cmd is a stale pointer. - */ -static void scsi_requeue_command(struct request_queue *q, struct scsi_cmnd *cmd) -{ - struct scsi_device *sdev = cmd->device; - struct request *req = cmd->request; - unsigned long flags; - - spin_lock_irqsave(q->queue_lock, flags); - blk_unprep_request(req); - req->special = NULL; - scsi_put_command(cmd); - blk_requeue_request(q, req); - spin_unlock_irqrestore(q->queue_lock, flags); - - scsi_run_queue(q); - - put_device(&sdev->sdev_gendev); -} - void scsi_run_host_queues(struct Scsi_Host *shost) { struct scsi_device *sdev; @@ -558,7 +573,7 @@ void scsi_run_host_queues(struct Scsi_Host *shost) static void scsi_uninit_cmd(struct scsi_cmnd *cmd) { - if (!blk_rq_is_passthrough(cmd->request)) { + if (!blk_rq_is_passthrough(scsi_cmd_to_rq(cmd))) { struct scsi_driver *drv = scsi_cmd_to_driver(cmd); if (drv->uninit_command) @@ -566,291 +581,223 @@ static void scsi_uninit_cmd(struct scsi_cmnd *cmd) } } -static void scsi_mq_free_sgtables(struct scsi_cmnd *cmd) +void scsi_free_sgtables(struct scsi_cmnd *cmd) { - struct scsi_data_buffer *sdb; - if (cmd->sdb.table.nents) - sg_free_table_chained(&cmd->sdb.table, true); - if (cmd->request->next_rq) { - sdb = cmd->request->next_rq->special; - if (sdb) - sg_free_table_chained(&sdb->table, true); - } + sg_free_table_chained(&cmd->sdb.table, + SCSI_INLINE_SG_CNT); if (scsi_prot_sg_count(cmd)) - sg_free_table_chained(&cmd->prot_sdb->table, true); + sg_free_table_chained(&cmd->prot_sdb->table, + SCSI_INLINE_PROT_SG_CNT); } +EXPORT_SYMBOL_GPL(scsi_free_sgtables); static void scsi_mq_uninit_cmd(struct scsi_cmnd *cmd) { - scsi_mq_free_sgtables(cmd); + scsi_free_sgtables(cmd); scsi_uninit_cmd(cmd); - scsi_del_cmd_from_list(cmd); } -/* - * Function: scsi_release_buffers() - * - * Purpose: Free resources allocate for a scsi_command. - * - * Arguments: cmd - command that we are bailing. - * - * Lock status: Assumed that no lock is held upon entry. - * - * Returns: Nothing - * - * Notes: In the event that an upper level driver rejects a - * command, we must release resources allocated during - * the __init_io() function. Primarily this would involve - * the scatter-gather table. - */ -static void scsi_release_buffers(struct scsi_cmnd *cmd) +static void scsi_run_queue_async(struct scsi_device *sdev) { - if (cmd->sdb.table.nents) - sg_free_table_chained(&cmd->sdb.table, false); - - memset(&cmd->sdb, 0, sizeof(cmd->sdb)); - - if (scsi_prot_sg_count(cmd)) - sg_free_table_chained(&cmd->prot_sdb->table, false); -} + if (scsi_host_in_recovery(sdev->host)) + return; -static void scsi_release_bidi_buffers(struct scsi_cmnd *cmd) -{ - struct scsi_data_buffer *bidi_sdb = cmd->request->next_rq->special; + if (scsi_target(sdev)->single_lun || + !list_empty(&sdev->host->starved_list)) { + kblockd_schedule_work(&sdev->requeue_work); + } else { + /* + * smp_mb() present in sbitmap_queue_clear() or implied in + * .end_io is for ordering writing .device_busy in + * scsi_device_unbusy() and reading sdev->restarts. + */ + int old = atomic_read(&sdev->restarts); - sg_free_table_chained(&bidi_sdb->table, false); - kmem_cache_free(scsi_sdb_cache, bidi_sdb); - cmd->request->next_rq->special = NULL; + /* + * ->restarts has to be kept as non-zero if new budget + * contention occurs. + * + * No need to run queue when either another re-run + * queue wins in updating ->restarts or a new budget + * contention occurs. + */ + if (old && atomic_cmpxchg(&sdev->restarts, old, 0) == old) + blk_mq_run_hw_queues(sdev->request_queue, true); + } } +/* Returns false when no more bytes to process, true if there are more */ static bool scsi_end_request(struct request *req, blk_status_t error, - unsigned int bytes, unsigned int bidi_bytes) + unsigned int bytes) { - struct scsi_cmnd *cmd = req->special; + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); struct scsi_device *sdev = cmd->device; struct request_queue *q = sdev->request_queue; if (blk_update_request(req, error, bytes)) return true; - /* Bidi request must be completed as a whole */ - if (unlikely(bidi_bytes) && - blk_update_request(req->next_rq, error, bidi_bytes)) - return true; - - if (blk_queue_add_random(q)) - add_disk_randomness(req->rq_disk); + if (q->limits.features & BLK_FEAT_ADD_RANDOM) + add_disk_randomness(req->q->disk); - if (req->mq_ctx) { - /* - * In the MQ case the command gets freed by __blk_mq_end_request, - * so we have to do all cleanup that depends on it earlier. - * - * We also can't kick the queues from irq context, so we - * will have to defer it to a workqueue. - */ - scsi_mq_uninit_cmd(cmd); + WARN_ON_ONCE(!blk_rq_is_passthrough(req) && + !(cmd->flags & SCMD_INITIALIZED)); + cmd->flags = 0; - __blk_mq_end_request(req, error); + /* + * Calling rcu_barrier() is not necessary here because the + * SCSI error handler guarantees that the function called by + * call_rcu() has been called before scsi_end_request() is + * called. + */ + destroy_rcu_head(&cmd->rcu); - if (scsi_target(sdev)->single_lun || - !list_empty(&sdev->host->starved_list)) - kblockd_schedule_work(&sdev->requeue_work); - else - blk_mq_run_hw_queues(q, true); - } else { - unsigned long flags; + /* + * In the MQ case the command gets freed by __blk_mq_end_request, + * so we have to do all cleanup that depends on it earlier. + * + * We also can't kick the queues from irq context, so we + * will have to defer it to a workqueue. + */ + scsi_mq_uninit_cmd(cmd); - if (bidi_bytes) - scsi_release_bidi_buffers(cmd); - scsi_release_buffers(cmd); - scsi_put_command(cmd); + /* + * queue is still alive, so grab the ref for preventing it + * from being cleaned up during running queue. + */ + percpu_ref_get(&q->q_usage_counter); - spin_lock_irqsave(q->queue_lock, flags); - blk_finish_request(req, error); - spin_unlock_irqrestore(q->queue_lock, flags); + __blk_mq_end_request(req, error); - scsi_run_queue(q); - } + scsi_run_queue_async(sdev); - put_device(&sdev->sdev_gendev); + percpu_ref_put(&q->q_usage_counter); return false; } /** - * __scsi_error_from_host_byte - translate SCSI error code into errno - * @cmd: SCSI command (unused) + * scsi_result_to_blk_status - translate a SCSI result code into blk_status_t * @result: scsi error code * - * Translate SCSI error code into block errors. + * Translate a SCSI result code into a blk_status_t value. */ -static blk_status_t __scsi_error_from_host_byte(struct scsi_cmnd *cmd, - int result) +static blk_status_t scsi_result_to_blk_status(int result) { + /* + * Check the scsi-ml byte first in case we converted a host or status + * byte. + */ + switch (scsi_ml_byte(result)) { + case SCSIML_STAT_OK: + break; + case SCSIML_STAT_RESV_CONFLICT: + return BLK_STS_RESV_CONFLICT; + case SCSIML_STAT_NOSPC: + return BLK_STS_NOSPC; + case SCSIML_STAT_MED_ERROR: + return BLK_STS_MEDIUM; + case SCSIML_STAT_TGT_FAILURE: + return BLK_STS_TARGET; + case SCSIML_STAT_DL_TIMEOUT: + return BLK_STS_DURATION_LIMIT; + } + switch (host_byte(result)) { + case DID_OK: + if (scsi_status_is_good(result)) + return BLK_STS_OK; + return BLK_STS_IOERR; case DID_TRANSPORT_FAILFAST: + case DID_TRANSPORT_MARGINAL: return BLK_STS_TRANSPORT; - case DID_TARGET_FAILURE: - set_host_byte(cmd, DID_OK); - return BLK_STS_TARGET; - case DID_NEXUS_FAILURE: - return BLK_STS_NEXUS; - case DID_ALLOC_FAILURE: - set_host_byte(cmd, DID_OK); - return BLK_STS_NOSPC; - case DID_MEDIUM_ERROR: - set_host_byte(cmd, DID_OK); - return BLK_STS_MEDIUM; default: return BLK_STS_IOERR; } } -/* - * Function: scsi_io_completion() - * - * Purpose: Completion processing for block device I/O requests. - * - * Arguments: cmd - command that is finished. - * - * Lock status: Assumed that no lock is held upon entry. - * - * Returns: Nothing - * - * Notes: We will finish off the specified number of sectors. If we - * are done, the command block will be released and the queue - * function will be goosed. If we are not done then we have to - * figure out what to do next: - * - * a) We can call scsi_requeue_command(). The request - * will be unprepared and put back on the queue. Then - * a new command will be created for it. This should - * be used if we made forward progress, or if we want - * to switch from READ(10) to READ(6) for example. +/** + * scsi_rq_err_bytes - determine number of bytes till the next failure boundary + * @rq: request to examine * - * b) We can call __scsi_queue_insert(). The request will - * be put back on the queue and retried using the same - * command as before, possibly after a delay. + * Description: + * A request could be merge of IOs which require different failure + * handling. This function determines the number of bytes which + * can be failed from the beginning of the request without + * crossing into area which need to be retried further. * - * c) We can call scsi_end_request() with -EIO to fail - * the remainder of the request. + * Return: + * The number of bytes to fail. */ -void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) +static unsigned int scsi_rq_err_bytes(const struct request *rq) { - int result = cmd->result; - struct request_queue *q = cmd->device->request_queue; - struct request *req = cmd->request; - blk_status_t error = BLK_STS_OK; - struct scsi_sense_hdr sshdr; - bool sense_valid = false; - int sense_deferred = 0, level = 0; - enum {ACTION_FAIL, ACTION_REPREP, ACTION_RETRY, - ACTION_DELAYED_RETRY} action; - unsigned long wait_for = (cmd->allowed + 1) * req->timeout; - - if (result) { - sense_valid = scsi_command_normalize_sense(cmd, &sshdr); - if (sense_valid) - sense_deferred = scsi_sense_is_deferred(&sshdr); - } + blk_opf_t ff = rq->cmd_flags & REQ_FAILFAST_MASK; + unsigned int bytes = 0; + struct bio *bio; - if (blk_rq_is_passthrough(req)) { - if (result) { - if (sense_valid) { - /* - * SG_IO wants current and deferred errors - */ - scsi_req(req)->sense_len = - min(8 + cmd->sense_buffer[7], - SCSI_SENSE_BUFFERSIZE); - } - if (!sense_deferred) - error = __scsi_error_from_host_byte(cmd, result); - } - /* - * __scsi_error_from_host_byte may have reset the host_byte - */ - scsi_req(req)->result = cmd->result; - scsi_req(req)->resid_len = scsi_get_resid(cmd); + if (!(rq->rq_flags & RQF_MIXED_MERGE)) + return blk_rq_bytes(rq); - if (scsi_bidi_cmnd(cmd)) { - /* - * Bidi commands Must be complete as a whole, - * both sides at once. - */ - scsi_req(req->next_rq)->resid_len = scsi_in(cmd)->resid; - if (scsi_end_request(req, BLK_STS_OK, blk_rq_bytes(req), - blk_rq_bytes(req->next_rq))) - BUG(); - return; - } - } else if (blk_rq_bytes(req) == 0 && result && !sense_deferred) { - /* - * Flush commands do not transfers any data, and thus cannot use - * good_bytes != blk_rq_bytes(req) as the signal for an error. - * This sets the error explicitly for the problem case. - */ - error = __scsi_error_from_host_byte(cmd, result); + /* + * Currently the only 'mixing' which can happen is between + * different fastfail types. We can safely fail portions + * which have all the failfast bits that the first one has - + * the ones which are at least as eager to fail as the first + * one. + */ + for (bio = rq->bio; bio; bio = bio->bi_next) { + if ((bio->bi_opf & ff) != ff) + break; + bytes += bio->bi_iter.bi_size; } - /* no bidi support for !blk_rq_is_passthrough yet */ - BUG_ON(blk_bidi_rq(req)); + /* this could lead to infinite loop */ + BUG_ON(blk_rq_bytes(rq) && !bytes); + return bytes; +} - /* - * Next deal with any sectors which we were able to correctly - * handle. - */ - SCSI_LOG_HLCOMPLETE(1, scmd_printk(KERN_INFO, cmd, - "%u sectors total, %d bytes done.\n", - blk_rq_sectors(req), good_bytes)); +static bool scsi_cmd_runtime_exceeced(struct scsi_cmnd *cmd) +{ + struct request *req = scsi_cmd_to_rq(cmd); + unsigned long wait_for; - /* - * Recovered errors need reporting, but they're always treated as - * success, so fiddle the result code here. For passthrough requests - * we already took a copy of the original into sreq->result which - * is what gets returned to the user - */ - if (sense_valid && (sshdr.sense_key == RECOVERED_ERROR)) { - /* if ATA PASS-THROUGH INFORMATION AVAILABLE skip - * print since caller wants ATA registers. Only occurs on - * SCSI ATA PASS_THROUGH commands when CK_COND=1 - */ - if ((sshdr.asc == 0x0) && (sshdr.ascq == 0x1d)) - ; - else if (!(req->rq_flags & RQF_QUIET)) - scsi_print_sense(cmd); - result = 0; - /* for passthrough error may be set */ - error = BLK_STS_OK; + if (cmd->allowed == SCSI_CMD_RETRIES_NO_LIMIT) + return false; + + wait_for = (cmd->allowed + 1) * req->timeout; + if (time_before(cmd->jiffies_at_alloc + wait_for, jiffies)) { + scmd_printk(KERN_ERR, cmd, "timing out command, waited %lus\n", + wait_for/HZ); + return true; } + return false; +} - /* - * special case: failed zero length commands always need to - * drop down into the retry code. Otherwise, if we finished - * all bytes in the request we are done now. - */ - if (!(blk_rq_bytes(req) == 0 && error) && - !scsi_end_request(req, error, good_bytes, 0)) - return; +/* + * When ALUA transition state is returned, reprep the cmd to + * use the ALUA handler's transition timeout. Delay the reprep + * 1 sec to avoid aggressive retries of the target in that + * state. + */ +#define ALUA_TRANSITION_REPREP_DELAY 1000 - /* - * Kill remainder if no retrys. - */ - if (error && scsi_noretry_cmd(cmd)) { - if (scsi_end_request(req, error, blk_rq_bytes(req), 0)) - BUG(); - return; - } +/* Helper for scsi_io_completion() when special action required. */ +static void scsi_io_completion_action(struct scsi_cmnd *cmd, int result) +{ + struct request *req = scsi_cmd_to_rq(cmd); + int level = 0; + enum {ACTION_FAIL, ACTION_REPREP, ACTION_DELAYED_REPREP, + ACTION_RETRY, ACTION_DELAYED_RETRY} action; + struct scsi_sense_hdr sshdr; + bool sense_valid; + bool sense_current = true; /* false implies "deferred sense" */ + blk_status_t blk_stat; - /* - * If there had been no error, but we have leftover bytes in the - * requeues just queue the command up again. - */ - if (result == 0) - goto requeue; + sense_valid = scsi_command_normalize_sense(cmd, &sshdr); + if (sense_valid) + sense_current = !scsi_sense_is_deferred(&sshdr); - error = __scsi_error_from_host_byte(cmd, result); + blk_stat = scsi_result_to_blk_status(result); if (host_byte(result) == DID_RESET) { /* Third party bus reset or reset for error recovery @@ -858,7 +805,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) * happens. */ action = ACTION_RETRY; - } else if (sense_valid && !sense_deferred) { + } else if (sense_valid && sense_current) { switch (sshdr.sense_key) { case UNIT_ATTENTION: if (cmd->device->removable) { @@ -894,18 +841,18 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) action = ACTION_REPREP; } else if (sshdr.asc == 0x10) /* DIX */ { action = ACTION_FAIL; - error = BLK_STS_PROTECTION; + blk_stat = BLK_STS_PROTECTION; /* INVALID COMMAND OPCODE or INVALID FIELD IN CDB */ } else if (sshdr.asc == 0x20 || sshdr.asc == 0x24) { action = ACTION_FAIL; - error = BLK_STS_TARGET; + blk_stat = BLK_STS_TARGET; } else action = ACTION_FAIL; break; case ABORTED_COMMAND: action = ACTION_FAIL; if (sshdr.asc == 0x10) /* DIF */ - error = BLK_STS_PROTECTION; + blk_stat = BLK_STS_PROTECTION; break; case NOT_READY: /* If the device is in the process of becoming @@ -920,9 +867,23 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) case 0x07: /* operation in progress */ case 0x08: /* Long write in progress */ case 0x09: /* self test in progress */ + case 0x11: /* notify (enable spinup) required */ case 0x14: /* space allocation in progress */ + case 0x1a: /* start stop unit in progress */ + case 0x1b: /* sanitize in progress */ + case 0x1d: /* configuration in progress */ action = ACTION_DELAYED_RETRY; break; + case 0x0a: /* ALUA state transition */ + action = ACTION_DELAYED_REPREP; + break; + /* + * Depopulation might take many hours, + * thus it is not worthwhile to retry. + */ + case 0x24: /* depopulation in progress */ + case 0x25: /* depopulation restore in progress */ + fallthrough; default: action = ACTION_FAIL; break; @@ -934,6 +895,17 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) /* See SSC3rXX or current. */ action = ACTION_FAIL; break; + case DATA_PROTECT: + action = ACTION_FAIL; + if ((sshdr.asc == 0x0C && sshdr.ascq == 0x12) || + (sshdr.asc == 0x55 && + (sshdr.ascq == 0x0E || sshdr.ascq == 0x0F))) { + /* Insufficient zone resources */ + blk_stat = BLK_STS_ZONE_OPEN_RESOURCE; + } + break; + case COMPLETED: + fallthrough; default: action = ACTION_FAIL; break; @@ -941,8 +913,7 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) } else action = ACTION_FAIL; - if (action != ACTION_FAIL && - time_before(cmd->jiffies_at_alloc + wait_for, jiffies)) + if (action != ACTION_FAIL && scsi_cmd_runtime_exceeced(cmd)) action = ACTION_FAIL; switch (action) { @@ -954,8 +925,9 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) DEFAULT_RATELIMIT_BURST); if (unlikely(scsi_logging_level)) - level = SCSI_LOG_LEVEL(SCSI_LOG_MLCOMPLETE_SHIFT, - SCSI_LOG_MLCOMPLETE_BITS); + level = + SCSI_LOG_LEVEL(SCSI_LOG_MLCOMPLETE_SHIFT, + SCSI_LOG_MLCOMPLETE_BITS); /* * if logging is enabled the failure will be printed @@ -963,215 +935,353 @@ void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) */ if (!level && __ratelimit(&_rs)) { scsi_print_result(cmd, NULL, FAILED); - if (driver_byte(result) & DRIVER_SENSE) + if (sense_valid) scsi_print_sense(cmd); scsi_print_command(cmd); } } - if (!scsi_end_request(req, error, blk_rq_err_bytes(req), 0)) + if (!scsi_end_request(req, blk_stat, scsi_rq_err_bytes(req))) return; - /*FALLTHRU*/ + fallthrough; case ACTION_REPREP: - requeue: - /* Unprep the request and put it back at the head of the queue. - * A new command will be prepared and issued. - */ - if (q->mq_ops) { - cmd->request->rq_flags &= ~RQF_DONTPREP; - scsi_mq_uninit_cmd(cmd); - scsi_mq_requeue_cmd(cmd); - } else { - scsi_release_buffers(cmd); - scsi_requeue_command(q, cmd); - } + scsi_mq_requeue_cmd(cmd, 0); + break; + case ACTION_DELAYED_REPREP: + scsi_mq_requeue_cmd(cmd, ALUA_TRANSITION_REPREP_DELAY); break; case ACTION_RETRY: /* Retry the same command immediately */ - __scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY, 0); + __scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY, false); break; case ACTION_DELAYED_RETRY: /* Retry the same command after a delay */ - __scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY, 0); + __scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY, false); break; } } -static int scsi_init_sgtable(struct request *req, struct scsi_data_buffer *sdb) +/* + * Helper for scsi_io_completion() when cmd->result is non-zero. Returns a + * new result that may suppress further error checking. Also modifies + * *blk_statp in some cases. + */ +static int scsi_io_completion_nz_result(struct scsi_cmnd *cmd, int result, + blk_status_t *blk_statp) { - int count; + bool sense_valid; + bool sense_current = true; /* false implies "deferred sense" */ + struct request *req = scsi_cmd_to_rq(cmd); + struct scsi_sense_hdr sshdr; + sense_valid = scsi_command_normalize_sense(cmd, &sshdr); + if (sense_valid) + sense_current = !scsi_sense_is_deferred(&sshdr); + + if (blk_rq_is_passthrough(req)) { + if (sense_valid) { + /* + * SG_IO wants current and deferred errors + */ + cmd->sense_len = min(8 + cmd->sense_buffer[7], + SCSI_SENSE_BUFFERSIZE); + } + if (sense_current) + *blk_statp = scsi_result_to_blk_status(result); + } else if (blk_rq_bytes(req) == 0 && sense_current) { + /* + * Flush commands do not transfers any data, and thus cannot use + * good_bytes != blk_rq_bytes(req) as the signal for an error. + * This sets *blk_statp explicitly for the problem case. + */ + *blk_statp = scsi_result_to_blk_status(result); + } /* - * If sg table allocation fails, requeue request later. + * Recovered errors need reporting, but they're always treated as + * success, so fiddle the result code here. For passthrough requests + * we already took a copy of the original into sreq->result which + * is what gets returned to the user + */ + if (sense_valid && (sshdr.sense_key == RECOVERED_ERROR)) { + bool do_print = true; + /* + * if ATA PASS-THROUGH INFORMATION AVAILABLE [0x0, 0x1d] + * skip print since caller wants ATA registers. Only occurs + * on SCSI ATA PASS_THROUGH commands when CK_COND=1 + */ + if ((sshdr.asc == 0x0) && (sshdr.ascq == 0x1d)) + do_print = false; + else if (req->rq_flags & RQF_QUIET) + do_print = false; + if (do_print) + scsi_print_sense(cmd); + result = 0; + /* for passthrough, *blk_statp may be set */ + *blk_statp = BLK_STS_OK; + } + /* + * Another corner case: the SCSI status byte is non-zero but 'good'. + * Example: PRE-FETCH command returns SAM_STAT_CONDITION_MET when + * it is able to fit nominated LBs in its cache (and SAM_STAT_GOOD + * if it can't fit). Treat SAM_STAT_CONDITION_MET and the related + * intermediate statuses (both obsolete in SAM-4) as good. */ - if (unlikely(sg_alloc_table_chained(&sdb->table, - blk_rq_nr_phys_segments(req), sdb->table.sgl))) - return BLKPREP_DEFER; + if ((result & 0xff) && scsi_status_is_good(result)) { + result = 0; + *blk_statp = BLK_STS_OK; + } + return result; +} - /* - * Next, walk the list, and fill in the addresses and sizes of - * each segment. +/** + * scsi_io_completion - Completion processing for SCSI commands. + * @cmd: command that is finished. + * @good_bytes: number of processed bytes. + * + * We will finish off the specified number of sectors. If we are done, the + * command block will be released and the queue function will be goosed. If we + * are not done then we have to figure out what to do next: + * + * a) We can call scsi_mq_requeue_cmd(). The request will be + * unprepared and put back on the queue. Then a new command will + * be created for it. This should be used if we made forward + * progress, or if we want to switch from READ(10) to READ(6) for + * example. + * + * b) We can call scsi_io_completion_action(). The request will be + * put back on the queue and retried using the same command as + * before, possibly after a delay. + * + * c) We can call scsi_end_request() with blk_stat other than + * BLK_STS_OK, to fail the remainder of the request. + */ +void scsi_io_completion(struct scsi_cmnd *cmd, unsigned int good_bytes) +{ + int result = cmd->result; + struct request *req = scsi_cmd_to_rq(cmd); + blk_status_t blk_stat = BLK_STS_OK; + + if (unlikely(result)) /* a nz result may or may not be an error */ + result = scsi_io_completion_nz_result(cmd, result, &blk_stat); + + /* + * Next deal with any sectors which we were able to correctly + * handle. + */ + SCSI_LOG_HLCOMPLETE(1, scmd_printk(KERN_INFO, cmd, + "%u sectors total, %d bytes done.\n", + blk_rq_sectors(req), good_bytes)); + + /* + * Failed, zero length commands always need to drop down + * to retry code. Fast path should return in this block. */ - count = blk_rq_map_sg(req->q, req, sdb->table.sgl); - BUG_ON(count > sdb->table.nents); - sdb->table.nents = count; - sdb->length = blk_rq_payload_bytes(req); - return BLKPREP_OK; + if (likely(blk_rq_bytes(req) > 0 || blk_stat == BLK_STS_OK)) { + if (likely(!scsi_end_request(req, blk_stat, good_bytes))) + return; /* no bytes remaining */ + } + + /* Kill remainder if no retries. */ + if (unlikely(blk_stat && scsi_noretry_cmd(cmd))) { + if (scsi_end_request(req, blk_stat, blk_rq_bytes(req))) + WARN_ONCE(true, + "Bytes remaining after failed, no-retry command"); + return; + } + + /* + * If there had been no error, but we have leftover bytes in the + * request just queue the command up again. + */ + if (likely(result == 0)) + scsi_mq_requeue_cmd(cmd, 0); + else + scsi_io_completion_action(cmd, result); } -/* - * Function: scsi_init_io() - * - * Purpose: SCSI I/O initialize function. +static inline bool scsi_cmd_needs_dma_drain(struct scsi_device *sdev, + struct request *rq) +{ + return sdev->dma_drain_len && blk_rq_is_passthrough(rq) && + !op_is_write(req_op(rq)) && + sdev->host->hostt->dma_need_drain(rq); +} + +/** + * scsi_alloc_sgtables - Allocate and initialize data and integrity scatterlists + * @cmd: SCSI command data structure to initialize. * - * Arguments: cmd - Command descriptor we wish to initialize + * Initializes @cmd->sdb and also @cmd->prot_sdb if data integrity is enabled + * for @cmd. * - * Returns: 0 on success - * BLKPREP_DEFER if the failure is retryable - * BLKPREP_KILL if the failure is fatal + * Returns: + * * BLK_STS_OK - on success + * * BLK_STS_RESOURCE - if the failure is retryable + * * BLK_STS_IOERR - if the failure is fatal */ -int scsi_init_io(struct scsi_cmnd *cmd) +blk_status_t scsi_alloc_sgtables(struct scsi_cmnd *cmd) { struct scsi_device *sdev = cmd->device; - struct request *rq = cmd->request; - bool is_mq = (rq->mq_ctx != NULL); - int error = BLKPREP_KILL; - - if (WARN_ON_ONCE(!blk_rq_nr_phys_segments(rq))) - goto err_exit; - - error = scsi_init_sgtable(rq, &cmd->sdb); - if (error) - goto err_exit; - - if (blk_bidi_rq(rq)) { - if (!rq->q->mq_ops) { - struct scsi_data_buffer *bidi_sdb = - kmem_cache_zalloc(scsi_sdb_cache, GFP_ATOMIC); - if (!bidi_sdb) { - error = BLKPREP_DEFER; - goto err_exit; - } + struct request *rq = scsi_cmd_to_rq(cmd); + unsigned short nr_segs = blk_rq_nr_phys_segments(rq); + struct scatterlist *last_sg = NULL; + blk_status_t ret; + bool need_drain = scsi_cmd_needs_dma_drain(sdev, rq); + int count; - rq->next_rq->special = bidi_sdb; - } + if (WARN_ON_ONCE(!nr_segs)) + return BLK_STS_IOERR; + + /* + * Make sure there is space for the drain. The driver must adjust + * max_hw_segments to be prepared for this. + */ + if (need_drain) + nr_segs++; + + /* + * If sg table allocation fails, requeue request later. + */ + if (unlikely(sg_alloc_table_chained(&cmd->sdb.table, nr_segs, + cmd->sdb.table.sgl, SCSI_INLINE_SG_CNT))) + return BLK_STS_RESOURCE; + + /* + * Next, walk the list, and fill in the addresses and sizes of + * each segment. + */ + count = __blk_rq_map_sg(rq, cmd->sdb.table.sgl, &last_sg); + + if (blk_rq_bytes(rq) & rq->q->limits.dma_pad_mask) { + unsigned int pad_len = + (rq->q->limits.dma_pad_mask & ~blk_rq_bytes(rq)) + 1; + + last_sg->length += pad_len; + cmd->extra_len += pad_len; + } - error = scsi_init_sgtable(rq->next_rq, rq->next_rq->special); - if (error) - goto err_exit; + if (need_drain) { + sg_unmark_end(last_sg); + last_sg = sg_next(last_sg); + sg_set_buf(last_sg, sdev->dma_drain_buf, sdev->dma_drain_len); + sg_mark_end(last_sg); + + cmd->extra_len += sdev->dma_drain_len; + count++; } + BUG_ON(count > cmd->sdb.table.nents); + cmd->sdb.table.nents = count; + cmd->sdb.length = blk_rq_payload_bytes(rq); + if (blk_integrity_rq(rq)) { struct scsi_data_buffer *prot_sdb = cmd->prot_sdb; - int ivecs, count; - if (prot_sdb == NULL) { + if (WARN_ON_ONCE(!prot_sdb)) { /* * This can happen if someone (e.g. multipath) * queues a command to a device on an adapter * that does not support DIX. */ - WARN_ON_ONCE(1); - error = BLKPREP_KILL; - goto err_exit; + ret = BLK_STS_IOERR; + goto out_free_sgtables; } - ivecs = blk_rq_count_integrity_sg(rq->q, rq->bio); - - if (sg_alloc_table_chained(&prot_sdb->table, ivecs, - prot_sdb->table.sgl)) { - error = BLKPREP_DEFER; - goto err_exit; + if (sg_alloc_table_chained(&prot_sdb->table, + rq->nr_integrity_segments, + prot_sdb->table.sgl, + SCSI_INLINE_PROT_SG_CNT)) { + ret = BLK_STS_RESOURCE; + goto out_free_sgtables; } - count = blk_rq_map_integrity_sg(rq->q, rq->bio, - prot_sdb->table.sgl); - BUG_ON(unlikely(count > ivecs)); - BUG_ON(unlikely(count > queue_max_integrity_segments(rq->q))); - + count = blk_rq_map_integrity_sg(rq, prot_sdb->table.sgl); cmd->prot_sdb = prot_sdb; cmd->prot_sdb->table.nents = count; } - return BLKPREP_OK; -err_exit: - if (is_mq) { - scsi_mq_free_sgtables(cmd); - } else { - scsi_release_buffers(cmd); - cmd->request->special = NULL; - scsi_put_command(cmd); - put_device(&sdev->sdev_gendev); - } - return error; + return BLK_STS_OK; +out_free_sgtables: + scsi_free_sgtables(cmd); + return ret; } -EXPORT_SYMBOL(scsi_init_io); +EXPORT_SYMBOL(scsi_alloc_sgtables); /** - * scsi_initialize_rq - initialize struct scsi_cmnd.req + * scsi_initialize_rq - initialize struct scsi_cmnd partially + * @rq: Request associated with the SCSI command to be initialized. * - * Called from inside blk_get_request(). + * This function initializes the members of struct scsi_cmnd that must be + * initialized before request processing starts and that won't be + * reinitialized if a SCSI command is requeued. */ -void scsi_initialize_rq(struct request *rq) +static void scsi_initialize_rq(struct request *rq) { struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); - scsi_req_init(&cmd->req); + memset(cmd->cmnd, 0, sizeof(cmd->cmnd)); + cmd->cmd_len = MAX_COMMAND_SIZE; + cmd->sense_len = 0; + init_rcu_head(&cmd->rcu); + cmd->jiffies_at_alloc = jiffies; + cmd->retries = 0; } -EXPORT_SYMBOL(scsi_initialize_rq); -/* Add a command to the list used by the aacraid and dpt_i2o drivers */ -void scsi_add_cmd_to_list(struct scsi_cmnd *cmd) +/** + * scsi_alloc_request - allocate a block request and partially + * initialize its &scsi_cmnd + * @q: the device's request queue + * @opf: the request operation code + * @flags: block layer allocation flags + * + * Return: &struct request pointer on success or %NULL on failure + */ +struct request *scsi_alloc_request(struct request_queue *q, blk_opf_t opf, + blk_mq_req_flags_t flags) { - struct scsi_device *sdev = cmd->device; - struct Scsi_Host *shost = sdev->host; - unsigned long flags; + struct request *rq; - if (shost->use_cmd_list) { - spin_lock_irqsave(&sdev->list_lock, flags); - list_add_tail(&cmd->list, &sdev->cmd_list); - spin_unlock_irqrestore(&sdev->list_lock, flags); - } + rq = blk_mq_alloc_request(q, opf, flags); + if (!IS_ERR(rq)) + scsi_initialize_rq(rq); + return rq; } +EXPORT_SYMBOL_GPL(scsi_alloc_request); -/* Remove a command from the list used by the aacraid and dpt_i2o drivers */ -void scsi_del_cmd_from_list(struct scsi_cmnd *cmd) +/* + * Only called when the request isn't completed by SCSI, and not freed by + * SCSI + */ +static void scsi_cleanup_rq(struct request *rq) { - struct scsi_device *sdev = cmd->device; - struct Scsi_Host *shost = sdev->host; - unsigned long flags; + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); - if (shost->use_cmd_list) { - spin_lock_irqsave(&sdev->list_lock, flags); - BUG_ON(list_empty(&cmd->list)); - list_del_init(&cmd->list); - spin_unlock_irqrestore(&sdev->list_lock, flags); + cmd->flags = 0; + + if (rq->rq_flags & RQF_DONTPREP) { + scsi_mq_uninit_cmd(cmd); + rq->rq_flags &= ~RQF_DONTPREP; } } -/* Called after a request has been started. */ +/* Called before a request is prepared. See also scsi_mq_prep_fn(). */ void scsi_init_command(struct scsi_device *dev, struct scsi_cmnd *cmd) { - void *buf = cmd->sense_buffer; - void *prot = cmd->prot_sdb; - unsigned int unchecked_isa_dma = cmd->flags & SCMD_UNCHECKED_ISA_DMA; + struct request *rq = scsi_cmd_to_rq(cmd); - /* zero out the cmd, except for the embedded scsi_request */ - memset((char *)cmd + sizeof(cmd->req), 0, - sizeof(*cmd) - sizeof(cmd->req) + dev->host->hostt->cmd_size); + if (!blk_rq_is_passthrough(rq) && !(cmd->flags & SCMD_INITIALIZED)) { + cmd->flags |= SCMD_INITIALIZED; + scsi_initialize_rq(rq); + } cmd->device = dev; - cmd->sense_buffer = buf; - cmd->prot_sdb = prot; - cmd->flags = unchecked_isa_dma; + INIT_LIST_HEAD(&cmd->eh_entry); INIT_DELAYED_WORK(&cmd->abort_work, scmd_eh_abort_handler); - cmd->jiffies_at_alloc = jiffies; - - scsi_add_cmd_to_list(cmd); } -static int scsi_setup_scsi_cmnd(struct scsi_device *sdev, struct request *req) +static blk_status_t scsi_setup_scsi_cmnd(struct scsi_device *sdev, + struct request *req) { - struct scsi_cmnd *cmd = req->special; + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); /* * Passthrough requests may transfer data, in which case they must @@ -1180,8 +1290,8 @@ static int scsi_setup_scsi_cmnd(struct scsi_device *sdev, struct request *req) * submit a request without an attached bio. */ if (req->bio) { - int ret = scsi_init_io(cmd); - if (unlikely(ret)) + blk_status_t ret = scsi_alloc_sgtables(cmd); + if (unlikely(ret != BLK_STS_OK)) return ret; } else { BUG_ON(blk_rq_bytes(req)); @@ -1189,212 +1299,92 @@ static int scsi_setup_scsi_cmnd(struct scsi_device *sdev, struct request *req) memset(&cmd->sdb, 0, sizeof(cmd->sdb)); } - cmd->cmd_len = scsi_req(req)->cmd_len; - cmd->cmnd = scsi_req(req)->cmd; cmd->transfersize = blk_rq_bytes(req); - cmd->allowed = scsi_req(req)->retries; - return BLKPREP_OK; -} - -/* - * Setup a normal block command. These are simple request from filesystems - * that still need to be translated to SCSI CDBs from the ULD. - */ -static int scsi_setup_fs_cmnd(struct scsi_device *sdev, struct request *req) -{ - struct scsi_cmnd *cmd = req->special; - - if (unlikely(sdev->handler && sdev->handler->prep_fn)) { - int ret = sdev->handler->prep_fn(sdev, req); - if (ret != BLKPREP_OK) - return ret; - } - - cmd->cmnd = scsi_req(req)->cmd = scsi_req(req)->__cmd; - memset(cmd->cmnd, 0, BLK_MAX_CDB); - return scsi_cmd_to_driver(cmd)->init_command(cmd); -} - -static int scsi_setup_cmnd(struct scsi_device *sdev, struct request *req) -{ - struct scsi_cmnd *cmd = req->special; - - if (!blk_rq_bytes(req)) - cmd->sc_data_direction = DMA_NONE; - else if (rq_data_dir(req) == WRITE) - cmd->sc_data_direction = DMA_TO_DEVICE; - else - cmd->sc_data_direction = DMA_FROM_DEVICE; - - if (blk_rq_is_scsi(req)) - return scsi_setup_scsi_cmnd(sdev, req); - else - return scsi_setup_fs_cmnd(sdev, req); + return BLK_STS_OK; } -static int -scsi_prep_state_check(struct scsi_device *sdev, struct request *req) +static blk_status_t +scsi_device_state_check(struct scsi_device *sdev, struct request *req) { - int ret = BLKPREP_OK; - - /* - * If the device is not in running state we will reject some - * or all commands. - */ - if (unlikely(sdev->sdev_state != SDEV_RUNNING)) { - switch (sdev->sdev_state) { - case SDEV_OFFLINE: - case SDEV_TRANSPORT_OFFLINE: - /* - * If the device is offline we refuse to process any - * commands. The device must be brought online - * before trying any recovery commands. - */ + switch (sdev->sdev_state) { + case SDEV_CREATED: + return BLK_STS_OK; + case SDEV_OFFLINE: + case SDEV_TRANSPORT_OFFLINE: + /* + * If the device is offline we refuse to process any + * commands. The device must be brought online + * before trying any recovery commands. + */ + if (!sdev->offline_already) { + sdev->offline_already = true; sdev_printk(KERN_ERR, sdev, "rejecting I/O to offline device\n"); - ret = BLKPREP_KILL; - break; - case SDEV_DEL: - /* - * If the device is fully deleted, we refuse to - * process any commands as well. - */ - sdev_printk(KERN_ERR, sdev, - "rejecting I/O to dead device\n"); - ret = BLKPREP_KILL; - break; - case SDEV_BLOCK: - case SDEV_CREATED_BLOCK: - ret = BLKPREP_DEFER; - break; - case SDEV_QUIESCE: - /* - * If the devices is blocked we defer normal commands. - */ - if (!(req->rq_flags & RQF_PREEMPT)) - ret = BLKPREP_DEFER; - break; - default: - /* - * For any other not fully online state we only allow - * special commands. In particular any user initiated - * command is not allowed. - */ - if (!(req->rq_flags & RQF_PREEMPT)) - ret = BLKPREP_KILL; - break; - } - } - return ret; -} - -static int -scsi_prep_return(struct request_queue *q, struct request *req, int ret) -{ - struct scsi_device *sdev = q->queuedata; - - switch (ret) { - case BLKPREP_KILL: - case BLKPREP_INVALID: - scsi_req(req)->result = DID_NO_CONNECT << 16; - /* release the command and kill it */ - if (req->special) { - struct scsi_cmnd *cmd = req->special; - scsi_release_buffers(cmd); - scsi_put_command(cmd); - put_device(&sdev->sdev_gendev); - req->special = NULL; } - break; - case BLKPREP_DEFER: + return BLK_STS_IOERR; + case SDEV_DEL: /* - * If we defer, the blk_peek_request() returns NULL, but the - * queue must be restarted, so we schedule a callback to happen - * shortly. + * If the device is fully deleted, we refuse to + * process any commands as well. */ - if (atomic_read(&sdev->device_busy) == 0) - blk_delay_queue(q, SCSI_QUEUE_DELAY); - break; + sdev_printk(KERN_ERR, sdev, + "rejecting I/O to dead device\n"); + return BLK_STS_IOERR; + case SDEV_BLOCK: + case SDEV_CREATED_BLOCK: + return BLK_STS_RESOURCE; + case SDEV_QUIESCE: + /* + * If the device is blocked we only accept power management + * commands. + */ + if (req && WARN_ON_ONCE(!(req->rq_flags & RQF_PM))) + return BLK_STS_RESOURCE; + return BLK_STS_OK; default: - req->rq_flags |= RQF_DONTPREP; - } - - return ret; -} - -static int scsi_prep_fn(struct request_queue *q, struct request *req) -{ - struct scsi_device *sdev = q->queuedata; - struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); - int ret; - - ret = scsi_prep_state_check(sdev, req); - if (ret != BLKPREP_OK) - goto out; - - if (!req->special) { - /* Bail if we can't get a reference to the device */ - if (unlikely(!get_device(&sdev->sdev_gendev))) { - ret = BLKPREP_DEFER; - goto out; - } - - scsi_init_command(sdev, cmd); - req->special = cmd; + /* + * For any other not fully online state we only allow + * power management commands. + */ + if (req && !(req->rq_flags & RQF_PM)) + return BLK_STS_OFFLINE; + return BLK_STS_OK; } - - cmd->tag = req->tag; - cmd->request = req; - cmd->prot_op = SCSI_PROT_NORMAL; - - ret = scsi_setup_cmnd(sdev, req); -out: - return scsi_prep_return(q, req, ret); -} - -static void scsi_unprep_fn(struct request_queue *q, struct request *req) -{ - scsi_uninit_cmd(req->special); } /* - * scsi_dev_queue_ready: if we can send requests to sdev, return 1 else - * return 0. - * - * Called with the queue_lock held. + * scsi_dev_queue_ready: if we can send requests to sdev, assign one token + * and return the token else return -1. */ static inline int scsi_dev_queue_ready(struct request_queue *q, struct scsi_device *sdev) { - unsigned int busy; + int token; - busy = atomic_inc_return(&sdev->device_busy) - 1; - if (atomic_read(&sdev->device_blocked)) { - if (busy) - goto out_dec; + if (!sdev->budget_map.map) + return INT_MAX; - /* - * unblock after device_blocked iterates to zero - */ - if (atomic_dec_return(&sdev->device_blocked) > 0) { - /* - * For the MQ case we take care of this in the caller. - */ - if (!q->mq_ops) - blk_delay_queue(q, SCSI_QUEUE_DELAY); - goto out_dec; - } - SCSI_LOG_MLQUEUE(3, sdev_printk(KERN_INFO, sdev, - "unblocking device at zero depth\n")); + token = sbitmap_get(&sdev->budget_map); + if (token < 0) + return -1; + + if (!atomic_read(&sdev->device_blocked)) + return token; + + /* + * Only unblock if no other commands are pending and + * if device_blocked has decreased to zero + */ + if (scsi_device_busy(sdev) > 1 || + atomic_dec_return(&sdev->device_blocked) > 0) { + sbitmap_put(&sdev->budget_map, token); + return -1; } - if (busy >= sdev->queue_depth) - goto out_dec; + SCSI_LOG_MLQUEUE(3, sdev_printk(KERN_INFO, sdev, + "unblocking device at zero depth\n")); - return 1; -out_dec: - atomic_dec(&sdev->device_busy); - return 0; + return token; } /* @@ -1458,16 +1448,11 @@ out_dec: */ static inline int scsi_host_queue_ready(struct request_queue *q, struct Scsi_Host *shost, - struct scsi_device *sdev) + struct scsi_device *sdev, + struct scsi_cmnd *cmd) { - unsigned int busy; - - if (scsi_host_in_recovery(shost)) - return 0; - - busy = atomic_inc_return(&shost->host_busy) - 1; if (atomic_read(&shost->host_blocked) > 0) { - if (busy) + if (scsi_host_busy(shost) > 0) goto starved; /* @@ -1481,8 +1466,6 @@ static inline int scsi_host_queue_ready(struct request_queue *q, "unblocking host at zero depth\n")); } - if (shost->can_queue > 0 && busy >= shost->can_queue) - goto starved; if (shost->host_self_blocked) goto starved; @@ -1494,6 +1477,8 @@ static inline int scsi_host_queue_ready(struct request_queue *q, spin_unlock_irq(shost->host_lock); } + __set_bit(SCMD_STATE_INFLIGHT, &cmd->state); + return 1; starved: @@ -1502,7 +1487,7 @@ starved: list_add_tail(&sdev->starved_entry, &shost->starved_list); spin_unlock_irq(shost->host_lock); out_dec: - atomic_dec(&shost->host_busy); + scsi_dec_host_busy(shost, cmd); return 0; } @@ -1518,13 +1503,13 @@ out_dec: * needs to return 'not busy'. Otherwise, request stacking drivers * may hold requests forever. */ -static int scsi_lld_busy(struct request_queue *q) +static bool scsi_mq_lld_busy(struct request_queue *q) { struct scsi_device *sdev = q->queuedata; struct Scsi_Host *shost; if (blk_queue_dying(q)) - return 0; + return false; shost = sdev->host; @@ -1535,50 +1520,27 @@ static int scsi_lld_busy(struct request_queue *q) * in SCSI layer. */ if (scsi_host_in_recovery(shost) || scsi_device_is_busy(sdev)) - return 1; + return true; - return 0; + return false; } /* - * Kill a request for a dead device + * Block layer request completion callback. May be called from interrupt + * context. */ -static void scsi_kill_request(struct request *req, struct request_queue *q) +static void scsi_complete(struct request *rq) { - struct scsi_cmnd *cmd = req->special; - struct scsi_device *sdev; - struct scsi_target *starget; - struct Scsi_Host *shost; - - blk_start_request(req); - - scmd_printk(KERN_INFO, cmd, "killing request\n"); - - sdev = cmd->device; - starget = scsi_target(sdev); - shost = sdev->host; - scsi_init_cmd_errh(cmd); - cmd->result = DID_NO_CONNECT << 16; - atomic_inc(&cmd->device->iorequest_cnt); - - /* - * SCSI request completion path will do scsi_device_unbusy(), - * bump busy counts. To bump the counters, we need to dance - * with the locks as normal issue path does. - */ - atomic_inc(&sdev->device_busy); - atomic_inc(&shost->host_busy); - if (starget->can_queue > 0) - atomic_inc(&starget->target_busy); - - blk_complete_request(req); -} + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); + enum scsi_disposition disposition; -static void scsi_softirq_done(struct request *rq) -{ - struct scsi_cmnd *cmd = rq->special; - unsigned long wait_for = (cmd->allowed + 1) * rq->timeout; - int disposition; + if (blk_mq_is_reserved_rq(rq)) { + /* Only pass-through requests are supported in this code path. */ + WARN_ON_ONCE(!blk_rq_is_passthrough(scsi_cmd_to_rq(cmd))); + scsi_mq_uninit_cmd(cmd); + __blk_mq_end_request(rq, scsi_result_to_blk_status(cmd->result)); + return; + } INIT_LIST_HEAD(&cmd->eh_entry); @@ -1587,34 +1549,29 @@ static void scsi_softirq_done(struct request *rq) atomic_inc(&cmd->device->ioerr_cnt); disposition = scsi_decide_disposition(cmd); - if (disposition != SUCCESS && - time_before(cmd->jiffies_at_alloc + wait_for, jiffies)) { - sdev_printk(KERN_ERR, cmd->device, - "timing out command, waited %lus\n", - wait_for/HZ); + if (disposition != SUCCESS && scsi_cmd_runtime_exceeced(cmd)) disposition = SUCCESS; - } scsi_log_completion(cmd, disposition); switch (disposition) { - case SUCCESS: - scsi_finish_command(cmd); - break; - case NEEDS_RETRY: - scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY); - break; - case ADD_TO_MLQUEUE: - scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY); - break; - default: - scsi_eh_scmd_add(cmd); - break; + case SUCCESS: + scsi_finish_command(cmd); + break; + case NEEDS_RETRY: + scsi_queue_insert(cmd, SCSI_MLQUEUE_EH_RETRY); + break; + case ADD_TO_MLQUEUE: + scsi_queue_insert(cmd, SCSI_MLQUEUE_DEVICE_BUSY); + break; + default: + scsi_eh_scmd_add(cmd); + break; } } /** - * scsi_dispatch_command - Dispatch a command to the low-level driver. + * scsi_dispatch_cmd - Dispatch a command to the low-level driver. * @cmd: command block we are dispatching. * * Return: nonzero return request was rejected and device's queue needs to be @@ -1647,6 +1604,7 @@ static int scsi_dispatch_cmd(struct scsi_cmnd *cmd) */ SCSI_LOG_MLQUEUE(3, scmd_printk(KERN_INFO, cmd, "queuecommand : device blocked\n")); + atomic_dec(&cmd->device->iorequest_cnt); return SCSI_MLQUEUE_DEVICE_BUSY; } @@ -1679,6 +1637,7 @@ static int scsi_dispatch_cmd(struct scsi_cmnd *cmd) trace_scsi_dispatch_cmd_start(cmd); rtn = host->hostt->queuecommand(host, cmd); if (rtn) { + atomic_dec(&cmd->device->iorequest_cnt); trace_scsi_dispatch_cmd_error(cmd, rtn); if (rtn != SCSI_MLQUEUE_DEVICE_BUSY && rtn != SCSI_MLQUEUE_TARGET_BUSY) @@ -1690,224 +1649,172 @@ static int scsi_dispatch_cmd(struct scsi_cmnd *cmd) return rtn; done: - cmd->scsi_done(cmd); + scsi_done(cmd); return 0; } -/** - * scsi_done - Invoke completion on finished SCSI command. - * @cmd: The SCSI Command for which a low-level device driver (LLDD) gives - * ownership back to SCSI Core -- i.e. the LLDD has finished with it. - * - * Description: This function is the mid-level's (SCSI Core) interrupt routine, - * which regains ownership of the SCSI command (de facto) from a LLDD, and - * calls blk_complete_request() for further processing. - * - * This function is interrupt context safe. - */ -static void scsi_done(struct scsi_cmnd *cmd) +/* Size in bytes of the sg-list stored in the scsi-mq command-private data. */ +static unsigned int scsi_mq_inline_sgl_size(struct Scsi_Host *shost) { - trace_scsi_dispatch_cmd_done(cmd); - blk_complete_request(cmd->request); + return min_t(unsigned int, shost->sg_tablesize, SCSI_INLINE_SG_CNT) * + sizeof(struct scatterlist); } -/* - * Function: scsi_request_fn() - * - * Purpose: Main strategy routine for SCSI. - * - * Arguments: q - Pointer to actual queue. - * - * Returns: Nothing - * - * Lock status: IO request lock assumed to be held when called. - */ -static void scsi_request_fn(struct request_queue *q) - __releases(q->queue_lock) - __acquires(q->queue_lock) +static blk_status_t scsi_prepare_cmd(struct request *req) { - struct scsi_device *sdev = q->queuedata; - struct Scsi_Host *shost; - struct scsi_cmnd *cmd; - struct request *req; + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); + struct scsi_device *sdev = req->q->queuedata; + struct Scsi_Host *shost = sdev->host; + bool in_flight = test_bit(SCMD_STATE_INFLIGHT, &cmd->state); + struct scatterlist *sg; - /* - * To start with, we keep looping until the queue is empty, or until - * the host is no longer able to accept any more requests. - */ - shost = sdev->host; - for (;;) { - int rtn; - /* - * get next queueable request. We do this early to make sure - * that the request is fully prepared even if we cannot - * accept it. - */ - req = blk_peek_request(q); - if (!req) - break; + scsi_init_command(sdev, cmd); - if (unlikely(!scsi_device_online(sdev))) { - sdev_printk(KERN_ERR, sdev, - "rejecting I/O to offline device\n"); - scsi_kill_request(req, q); - continue; - } + cmd->eh_eflags = 0; + cmd->prot_type = 0; + cmd->prot_flags = 0; + cmd->submitter = 0; + memset(&cmd->sdb, 0, sizeof(cmd->sdb)); + cmd->underflow = 0; + cmd->transfersize = 0; + cmd->host_scribble = NULL; + cmd->result = 0; + cmd->extra_len = 0; + cmd->state = 0; + if (in_flight) + __set_bit(SCMD_STATE_INFLIGHT, &cmd->state); - if (!scsi_dev_queue_ready(q, sdev)) - break; + cmd->prot_op = SCSI_PROT_NORMAL; + if (blk_rq_bytes(req)) + cmd->sc_data_direction = rq_dma_dir(req); + else + cmd->sc_data_direction = DMA_NONE; - /* - * Remove the request from the request list. - */ - if (!(blk_queue_tagged(q) && !blk_queue_start_tag(q, req))) - blk_start_request(req); - - spin_unlock_irq(q->queue_lock); - cmd = req->special; - if (unlikely(cmd == NULL)) { - printk(KERN_CRIT "impossible request in %s.\n" - "please mail a stack trace to " - "linux-scsi@vger.kernel.org\n", - __func__); - blk_dump_rq_flags(req, "foo"); - BUG(); - } + sg = (void *)cmd + sizeof(struct scsi_cmnd) + shost->hostt->cmd_size; + cmd->sdb.table.sgl = sg; - /* - * We hit this when the driver is using a host wide - * tag map. For device level tag maps the queue_depth check - * in the device ready fn would prevent us from trying - * to allocate a tag. Since the map is a shared host resource - * we add the dev to the starved list so it eventually gets - * a run when a tag is freed. - */ - if (blk_queue_tagged(q) && !(req->rq_flags & RQF_QUEUED)) { - spin_lock_irq(shost->host_lock); - if (list_empty(&sdev->starved_entry)) - list_add_tail(&sdev->starved_entry, - &shost->starved_list); - spin_unlock_irq(shost->host_lock); - goto not_ready; - } + if (scsi_host_get_prot(shost)) { + memset(cmd->prot_sdb, 0, sizeof(struct scsi_data_buffer)); - if (!scsi_target_queue_ready(shost, sdev)) - goto not_ready; + cmd->prot_sdb->table.sgl = + (struct scatterlist *)(cmd->prot_sdb + 1); + } - if (!scsi_host_queue_ready(q, shost, sdev)) - goto host_not_ready; - - if (sdev->simple_tags) - cmd->flags |= SCMD_TAGGED; - else - cmd->flags &= ~SCMD_TAGGED; + /* + * Special handling for passthrough commands, which don't go to the ULP + * at all: + */ + if (blk_rq_is_passthrough(req)) + return scsi_setup_scsi_cmnd(sdev, req); - /* - * Finally, initialize any error handling parameters, and set up - * the timers for timeouts. - */ - scsi_init_cmd_errh(cmd); + if (sdev->handler && sdev->handler->prep_fn) { + blk_status_t ret = sdev->handler->prep_fn(sdev, req); - /* - * Dispatch the command to the low-level driver. - */ - cmd->scsi_done = scsi_done; - rtn = scsi_dispatch_cmd(cmd); - if (rtn) { - scsi_queue_insert(cmd, rtn); - spin_lock_irq(q->queue_lock); - goto out_delay; - } - spin_lock_irq(q->queue_lock); + if (ret != BLK_STS_OK) + return ret; } - return; - - host_not_ready: - if (scsi_target(sdev)->can_queue > 0) - atomic_dec(&scsi_target(sdev)->target_busy); - not_ready: - /* - * lock q, handle tag, requeue req, and decrement device_busy. We - * must return with queue_lock held. - * - * Decrementing device_busy without checking it is OK, as all such - * cases (host limits or settings) should run the queue at some - * later time. - */ - spin_lock_irq(q->queue_lock); - blk_requeue_request(q, req); - atomic_dec(&sdev->device_busy); -out_delay: - if (!atomic_read(&sdev->device_busy) && !scsi_device_blocked(sdev)) - blk_delay_queue(q, SCSI_QUEUE_DELAY); + /* Usually overridden by the ULP */ + cmd->allowed = 0; + memset(cmd->cmnd, 0, sizeof(cmd->cmnd)); + return scsi_cmd_to_driver(cmd)->init_command(cmd); } -static inline blk_status_t prep_to_mq(int ret) +static void scsi_done_internal(struct scsi_cmnd *cmd, bool complete_directly) { - switch (ret) { - case BLKPREP_OK: - return BLK_STS_OK; - case BLKPREP_DEFER: - return BLK_STS_RESOURCE; - default: - return BLK_STS_IOERR; + struct request *req = scsi_cmd_to_rq(cmd); + + switch (cmd->submitter) { + case SUBMITTED_BY_BLOCK_LAYER: + break; + case SUBMITTED_BY_SCSI_ERROR_HANDLER: + return scsi_eh_done(cmd); + case SUBMITTED_BY_SCSI_RESET_IOCTL: + return; } + + if (unlikely(blk_should_fake_timeout(scsi_cmd_to_rq(cmd)->q))) + return; + if (unlikely(test_and_set_bit(SCMD_STATE_COMPLETE, &cmd->state))) + return; + trace_scsi_dispatch_cmd_done(cmd); + + if (complete_directly) + blk_mq_complete_request_direct(req, scsi_complete); + else + blk_mq_complete_request(req); } -/* Size in bytes of the sg-list stored in the scsi-mq command-private data. */ -static unsigned int scsi_mq_sgl_size(struct Scsi_Host *shost) +void scsi_done(struct scsi_cmnd *cmd) { - return min_t(unsigned int, shost->sg_tablesize, SG_CHUNK_SIZE) * - sizeof(struct scatterlist); + scsi_done_internal(cmd, false); } +EXPORT_SYMBOL(scsi_done); -static int scsi_mq_prep_fn(struct request *req) +void scsi_done_direct(struct scsi_cmnd *cmd) { - struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); - struct scsi_device *sdev = req->q->queuedata; - struct Scsi_Host *shost = sdev->host; - struct scatterlist *sg; - - scsi_init_command(sdev, cmd); - - req->special = cmd; + scsi_done_internal(cmd, true); +} +EXPORT_SYMBOL(scsi_done_direct); - cmd->request = req; +static void scsi_mq_put_budget(struct request_queue *q, int budget_token) +{ + struct scsi_device *sdev = q->queuedata; - cmd->tag = req->tag; - cmd->prot_op = SCSI_PROT_NORMAL; + if (sdev->budget_map.map) + sbitmap_put(&sdev->budget_map, budget_token); +} - sg = (void *)cmd + sizeof(struct scsi_cmnd) + shost->hostt->cmd_size; - cmd->sdb.table.sgl = sg; +/* + * When to reinvoke queueing after a resource shortage. It's 3 msecs to + * not change behaviour from the previous unplug mechanism, experimentation + * may prove this needs changing. + */ +#define SCSI_QUEUE_DELAY 3 - if (scsi_host_get_prot(shost)) { - memset(cmd->prot_sdb, 0, sizeof(struct scsi_data_buffer)); +static int scsi_mq_get_budget(struct request_queue *q) +{ + struct scsi_device *sdev = q->queuedata; + int token = scsi_dev_queue_ready(q, sdev); - cmd->prot_sdb->table.sgl = - (struct scatterlist *)(cmd->prot_sdb + 1); - } + if (token >= 0) + return token; - if (blk_bidi_rq(req)) { - struct request *next_rq = req->next_rq; - struct scsi_data_buffer *bidi_sdb = blk_mq_rq_to_pdu(next_rq); + atomic_inc(&sdev->restarts); - memset(bidi_sdb, 0, sizeof(struct scsi_data_buffer)); - bidi_sdb->table.sgl = - (struct scatterlist *)(bidi_sdb + 1); + /* + * Orders atomic_inc(&sdev->restarts) and atomic_read(&sdev->device_busy). + * .restarts must be incremented before .device_busy is read because the + * code in scsi_run_queue_async() depends on the order of these operations. + */ + smp_mb__after_atomic(); - next_rq->special = bidi_sdb; - } + /* + * If all in-flight requests originated from this LUN are completed + * before reading .device_busy, sdev->device_busy will be observed as + * zero, then blk_mq_delay_run_hw_queues() will dispatch this request + * soon. Otherwise, completion of one of these requests will observe + * the .restarts flag, and the request queue will be run for handling + * this request, see scsi_end_request(). + */ + if (unlikely(scsi_device_busy(sdev) == 0 && + !scsi_device_blocked(sdev))) + blk_mq_delay_run_hw_queues(sdev->request_queue, SCSI_QUEUE_DELAY); + return -1; +} - blk_mq_start_request(req); +static void scsi_mq_set_rq_budget_token(struct request *req, int token) +{ + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); - return scsi_setup_cmnd(sdev, req); + cmd->budget_token = token; } -static void scsi_mq_done(struct scsi_cmnd *cmd) +static int scsi_mq_get_rq_budget_token(struct request *req) { - trace_scsi_dispatch_cmd_done(cmd); - blk_mq_complete_request(cmd->request); + struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(req); + + return cmd->budget_token; } static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx, @@ -1921,38 +1828,70 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx, blk_status_t ret; int reason; - ret = prep_to_mq(scsi_prep_state_check(sdev, req)); - if (ret != BLK_STS_OK) - goto out; + WARN_ON_ONCE(cmd->budget_token < 0); + + /* + * Bypass the SCSI device, SCSI target and SCSI host checks for + * reserved commands. + */ + if (!blk_mq_is_reserved_rq(req)) { + /* + * If the device is not in running state we will reject some or + * all commands. + */ + if (unlikely(sdev->sdev_state != SDEV_RUNNING)) { + ret = scsi_device_state_check(sdev, req); + if (ret != BLK_STS_OK) + goto out_put_budget; + } - ret = BLK_STS_RESOURCE; - if (!get_device(&sdev->sdev_gendev)) - goto out; + ret = BLK_STS_RESOURCE; + if (!scsi_target_queue_ready(shost, sdev)) + goto out_put_budget; + if (unlikely(scsi_host_in_recovery(shost))) { + if (cmd->flags & SCMD_FAIL_IF_RECOVERING) + ret = BLK_STS_OFFLINE; + goto out_dec_target_busy; + } + if (!scsi_host_queue_ready(q, shost, sdev, cmd)) + goto out_dec_target_busy; + } - if (!scsi_dev_queue_ready(q, sdev)) - goto out_put_device; - if (!scsi_target_queue_ready(shost, sdev)) - goto out_dec_device_busy; - if (!scsi_host_queue_ready(q, shost, sdev)) - goto out_dec_target_busy; + /* + * Only clear the driver-private command data if the LLD does not supply + * a function to initialize that data. + */ + if (shost->hostt->cmd_size && !shost->hostt->init_cmd_priv) + memset(scsi_cmd_priv(cmd), 0, shost->hostt->cmd_size); if (!(req->rq_flags & RQF_DONTPREP)) { - ret = prep_to_mq(scsi_mq_prep_fn(req)); + ret = scsi_prepare_cmd(req); if (ret != BLK_STS_OK) goto out_dec_host_busy; req->rq_flags |= RQF_DONTPREP; } else { - blk_mq_start_request(req); + clear_bit(SCMD_STATE_COMPLETE, &cmd->state); } + cmd->flags &= SCMD_PRESERVED_FLAGS; if (sdev->simple_tags) cmd->flags |= SCMD_TAGGED; - else - cmd->flags &= ~SCMD_TAGGED; + if (bd->last) + cmd->flags |= SCMD_LAST; - scsi_init_cmd_errh(cmd); - cmd->scsi_done = scsi_mq_done; + scsi_set_resid(cmd, 0); + memset(cmd->sense_buffer, 0, SCSI_SENSE_BUFFERSIZE); + cmd->submitter = SUBMITTED_BY_BLOCK_LAYER; + blk_mq_start_request(req); + if (blk_mq_is_reserved_rq(req)) { + reason = shost->hostt->queue_reserved_command(shost, cmd); + if (reason) { + ret = BLK_STS_RESOURCE; + goto out_put_budget; + } + return BLK_STS_OK; + } reason = scsi_dispatch_cmd(cmd); if (reason) { scsi_set_blocked(cmd, reason); @@ -1963,269 +1902,277 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx, return BLK_STS_OK; out_dec_host_busy: - atomic_dec(&shost->host_busy); + scsi_dec_host_busy(shost, cmd); out_dec_target_busy: if (scsi_target(sdev)->can_queue > 0) atomic_dec(&scsi_target(sdev)->target_busy); -out_dec_device_busy: - atomic_dec(&sdev->device_busy); -out_put_device: - put_device(&sdev->sdev_gendev); -out: +out_put_budget: + scsi_mq_put_budget(q, cmd->budget_token); + cmd->budget_token = -1; switch (ret) { case BLK_STS_OK: break; case BLK_STS_RESOURCE: - if (atomic_read(&sdev->device_busy) == 0 && - !scsi_device_blocked(sdev)) - blk_mq_delay_run_hw_queue(hctx, SCSI_QUEUE_DELAY); + if (scsi_device_blocked(sdev)) + ret = BLK_STS_DEV_RESOURCE; + break; + case BLK_STS_AGAIN: + cmd->result = DID_BUS_BUSY << 16; + if (req->rq_flags & RQF_DONTPREP) + scsi_mq_uninit_cmd(cmd); break; default: + if (unlikely(!scsi_device_online(sdev))) + cmd->result = DID_NO_CONNECT << 16; + else + cmd->result = DID_ERROR << 16; /* - * Make sure to release all allocated ressources when + * Make sure to release all allocated resources when * we hit an error, as we will never see this command * again. */ if (req->rq_flags & RQF_DONTPREP) scsi_mq_uninit_cmd(cmd); + scsi_run_queue_async(sdev); break; } return ret; } -static enum blk_eh_timer_return scsi_timeout(struct request *req, - bool reserved) -{ - if (reserved) - return BLK_EH_RESET_TIMER; - return scsi_times_out(req); -} - -static int scsi_init_request(struct blk_mq_tag_set *set, struct request *rq, - unsigned int hctx_idx, unsigned int numa_node) +static int scsi_mq_init_request(struct blk_mq_tag_set *set, struct request *rq, + unsigned int hctx_idx, unsigned int numa_node) { struct Scsi_Host *shost = set->driver_data; - const bool unchecked_isa_dma = shost->unchecked_isa_dma; struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); struct scatterlist *sg; + int ret = 0; - if (unchecked_isa_dma) - cmd->flags |= SCMD_UNCHECKED_ISA_DMA; - cmd->sense_buffer = scsi_alloc_sense_buffer(unchecked_isa_dma, - GFP_KERNEL, numa_node); + cmd->sense_buffer = + kmem_cache_alloc_node(scsi_sense_cache, GFP_KERNEL, numa_node); if (!cmd->sense_buffer) return -ENOMEM; - cmd->req.sense = cmd->sense_buffer; if (scsi_host_get_prot(shost)) { sg = (void *)cmd + sizeof(struct scsi_cmnd) + shost->hostt->cmd_size; - cmd->prot_sdb = (void *)sg + scsi_mq_sgl_size(shost); + cmd->prot_sdb = (void *)sg + scsi_mq_inline_sgl_size(shost); } - return 0; + if (shost->hostt->init_cmd_priv) { + ret = shost->hostt->init_cmd_priv(shost, cmd); + if (ret < 0) + kmem_cache_free(scsi_sense_cache, cmd->sense_buffer); + } + + return ret; } -static void scsi_exit_request(struct blk_mq_tag_set *set, struct request *rq, - unsigned int hctx_idx) +static void scsi_mq_exit_request(struct blk_mq_tag_set *set, struct request *rq, + unsigned int hctx_idx) { + struct Scsi_Host *shost = set->driver_data; struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); - scsi_free_sense_buffer(cmd->flags & SCMD_UNCHECKED_ISA_DMA, - cmd->sense_buffer); + if (shost->hostt->exit_cmd_priv) + shost->hostt->exit_cmd_priv(shost, cmd); + kmem_cache_free(scsi_sense_cache, cmd->sense_buffer); } -static int scsi_map_queues(struct blk_mq_tag_set *set) + +static int scsi_mq_poll(struct blk_mq_hw_ctx *hctx, struct io_comp_batch *iob) { - struct Scsi_Host *shost = container_of(set, struct Scsi_Host, tag_set); + struct Scsi_Host *shost = hctx->driver_data; - if (shost->hostt->map_queues) - return shost->hostt->map_queues(shost); - return blk_mq_map_queues(set); + if (shost->hostt->mq_poll) + return shost->hostt->mq_poll(shost, hctx->queue_num); + + return 0; } -static u64 scsi_calculate_bounce_limit(struct Scsi_Host *shost) +static int scsi_init_hctx(struct blk_mq_hw_ctx *hctx, void *data, + unsigned int hctx_idx) { - struct device *host_dev; - u64 bounce_limit = 0xffffffff; + struct Scsi_Host *shost = data; - if (shost->unchecked_isa_dma) - return BLK_BOUNCE_ISA; - /* - * Platforms with virtual-DMA translation - * hardware have no practical limit. - */ - if (!PCI_DMA_BUS_IS_PHYS) - return BLK_BOUNCE_ANY; + hctx->driver_data = shost; + return 0; +} - host_dev = scsi_get_device(shost); - if (host_dev && host_dev->dma_mask) - bounce_limit = (u64)dma_max_pfn(host_dev) << PAGE_SHIFT; +static void scsi_map_queues(struct blk_mq_tag_set *set) +{ + struct Scsi_Host *shost = container_of(set, struct Scsi_Host, tag_set); - return bounce_limit; + if (shost->hostt->map_queues) + return shost->hostt->map_queues(shost); + blk_mq_map_queues(&set->map[HCTX_TYPE_DEFAULT]); } -void __scsi_init_queue(struct Scsi_Host *shost, struct request_queue *q) +void scsi_init_limits(struct Scsi_Host *shost, struct queue_limits *lim) { struct device *dev = shost->dma_dev; - queue_flag_set_unlocked(QUEUE_FLAG_SCSI_PASSTHROUGH, q); - - /* - * this limit is imposed by hardware restrictions - */ - blk_queue_max_segments(q, min_t(unsigned short, shost->sg_tablesize, - SG_MAX_SEGMENTS)); + memset(lim, 0, sizeof(*lim)); + lim->max_segments = + min_t(unsigned short, shost->sg_tablesize, SG_MAX_SEGMENTS); if (scsi_host_prot_dma(shost)) { shost->sg_prot_tablesize = min_not_zero(shost->sg_prot_tablesize, (unsigned short)SCSI_MAX_PROT_SG_SEGMENTS); BUG_ON(shost->sg_prot_tablesize < shost->sg_tablesize); - blk_queue_max_integrity_segments(q, shost->sg_prot_tablesize); + lim->max_integrity_segments = shost->sg_prot_tablesize; } - blk_queue_max_hw_sectors(q, shost->max_sectors); - blk_queue_bounce_limit(q, scsi_calculate_bounce_limit(shost)); - blk_queue_segment_boundary(q, shost->dma_boundary); - dma_set_seg_boundary(dev, shost->dma_boundary); - - blk_queue_max_segment_size(q, dma_get_max_seg_size(dev)); - - if (!shost->use_clustering) - q->limits.cluster = 0; + lim->max_hw_sectors = shost->max_sectors; + lim->seg_boundary_mask = shost->dma_boundary; + lim->max_segment_size = shost->max_segment_size; + lim->virt_boundary_mask = shost->virt_boundary_mask; + lim->dma_alignment = max_t(unsigned int, + shost->dma_alignment, dma_get_cache_alignment() - 1); /* - * set a reasonable default alignment on word boundaries: the - * host and device may alter it using - * blk_queue_update_dma_alignment() later. + * Propagate the DMA formation properties to the dma-mapping layer as + * a courtesy service to the LLDDs. This needs to check that the buses + * actually support the DMA API first, though. */ - blk_queue_dma_alignment(q, 0x03); -} -EXPORT_SYMBOL_GPL(__scsi_init_queue); - -static int scsi_init_rq(struct request_queue *q, struct request *rq, gfp_t gfp) -{ - struct Scsi_Host *shost = q->rq_alloc_data; - const bool unchecked_isa_dma = shost->unchecked_isa_dma; - struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); - - memset(cmd, 0, sizeof(*cmd)); - - if (unchecked_isa_dma) - cmd->flags |= SCMD_UNCHECKED_ISA_DMA; - cmd->sense_buffer = scsi_alloc_sense_buffer(unchecked_isa_dma, gfp, - NUMA_NO_NODE); - if (!cmd->sense_buffer) - goto fail; - cmd->req.sense = cmd->sense_buffer; - - if (scsi_host_get_prot(shost) >= SHOST_DIX_TYPE0_PROTECTION) { - cmd->prot_sdb = kmem_cache_zalloc(scsi_sdb_cache, gfp); - if (!cmd->prot_sdb) - goto fail_free_sense; + if (dev->dma_parms) { + dma_set_seg_boundary(dev, shost->dma_boundary); + dma_set_max_seg_size(dev, shost->max_segment_size); } - - return 0; - -fail_free_sense: - scsi_free_sense_buffer(unchecked_isa_dma, cmd->sense_buffer); -fail: - return -ENOMEM; } +EXPORT_SYMBOL_GPL(scsi_init_limits); -static void scsi_exit_rq(struct request_queue *q, struct request *rq) -{ - struct scsi_cmnd *cmd = blk_mq_rq_to_pdu(rq); +static const struct blk_mq_ops scsi_mq_ops_no_commit = { + .get_budget = scsi_mq_get_budget, + .put_budget = scsi_mq_put_budget, + .queue_rq = scsi_queue_rq, + .complete = scsi_complete, + .timeout = scsi_timeout, +#ifdef CONFIG_BLK_DEBUG_FS + .show_rq = scsi_show_rq, +#endif + .init_request = scsi_mq_init_request, + .exit_request = scsi_mq_exit_request, + .cleanup_rq = scsi_cleanup_rq, + .busy = scsi_mq_lld_busy, + .map_queues = scsi_map_queues, + .init_hctx = scsi_init_hctx, + .poll = scsi_mq_poll, + .set_rq_budget_token = scsi_mq_set_rq_budget_token, + .get_rq_budget_token = scsi_mq_get_rq_budget_token, +}; - if (cmd->prot_sdb) - kmem_cache_free(scsi_sdb_cache, cmd->prot_sdb); - scsi_free_sense_buffer(cmd->flags & SCMD_UNCHECKED_ISA_DMA, - cmd->sense_buffer); -} -struct request_queue *scsi_alloc_queue(struct scsi_device *sdev) +static void scsi_commit_rqs(struct blk_mq_hw_ctx *hctx) { - struct Scsi_Host *shost = sdev->host; - struct request_queue *q; - - q = blk_alloc_queue_node(GFP_KERNEL, NUMA_NO_NODE); - if (!q) - return NULL; - q->cmd_size = sizeof(struct scsi_cmnd) + shost->hostt->cmd_size; - q->rq_alloc_data = shost; - q->request_fn = scsi_request_fn; - q->init_rq_fn = scsi_init_rq; - q->exit_rq_fn = scsi_exit_rq; - q->initialize_rq_fn = scsi_initialize_rq; - - if (blk_init_allocated_queue(q) < 0) { - blk_cleanup_queue(q); - return NULL; - } + struct Scsi_Host *shost = hctx->driver_data; - __scsi_init_queue(shost, q); - blk_queue_prep_rq(q, scsi_prep_fn); - blk_queue_unprep_rq(q, scsi_unprep_fn); - blk_queue_softirq_done(q, scsi_softirq_done); - blk_queue_rq_timed_out(q, scsi_times_out); - blk_queue_lld_busy(q, scsi_lld_busy); - return q; + shost->hostt->commit_rqs(shost, hctx->queue_num); } static const struct blk_mq_ops scsi_mq_ops = { + .get_budget = scsi_mq_get_budget, + .put_budget = scsi_mq_put_budget, .queue_rq = scsi_queue_rq, - .complete = scsi_softirq_done, + .commit_rqs = scsi_commit_rqs, + .complete = scsi_complete, .timeout = scsi_timeout, #ifdef CONFIG_BLK_DEBUG_FS .show_rq = scsi_show_rq, #endif - .init_request = scsi_init_request, - .exit_request = scsi_exit_request, - .initialize_rq_fn = scsi_initialize_rq, + .init_request = scsi_mq_init_request, + .exit_request = scsi_mq_exit_request, + .cleanup_rq = scsi_cleanup_rq, + .busy = scsi_mq_lld_busy, .map_queues = scsi_map_queues, + .init_hctx = scsi_init_hctx, + .poll = scsi_mq_poll, + .set_rq_budget_token = scsi_mq_set_rq_budget_token, + .get_rq_budget_token = scsi_mq_get_rq_budget_token, }; -struct request_queue *scsi_mq_alloc_queue(struct scsi_device *sdev) -{ - sdev->request_queue = blk_mq_init_queue(&sdev->host->tag_set); - if (IS_ERR(sdev->request_queue)) - return NULL; - - sdev->request_queue->queuedata = sdev; - __scsi_init_queue(sdev->host, sdev->request_queue); - return sdev->request_queue; -} - int scsi_mq_setup_tags(struct Scsi_Host *shost) { unsigned int cmd_size, sgl_size; + struct blk_mq_tag_set *tag_set = &shost->tag_set; - sgl_size = scsi_mq_sgl_size(shost); + sgl_size = max_t(unsigned int, sizeof(struct scatterlist), + scsi_mq_inline_sgl_size(shost)); cmd_size = sizeof(struct scsi_cmnd) + shost->hostt->cmd_size + sgl_size; if (scsi_host_get_prot(shost)) - cmd_size += sizeof(struct scsi_data_buffer) + sgl_size; + cmd_size += sizeof(struct scsi_data_buffer) + + sizeof(struct scatterlist) * SCSI_INLINE_PROT_SG_CNT; - memset(&shost->tag_set, 0, sizeof(shost->tag_set)); - shost->tag_set.ops = &scsi_mq_ops; - shost->tag_set.nr_hw_queues = shost->nr_hw_queues ? : 1; - shost->tag_set.queue_depth = shost->can_queue; - shost->tag_set.cmd_size = cmd_size; - shost->tag_set.numa_node = NUMA_NO_NODE; - shost->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_SG_MERGE; - shost->tag_set.flags |= - BLK_ALLOC_POLICY_TO_MQ_FLAG(shost->hostt->tag_alloc_policy); - shost->tag_set.driver_data = shost; + memset(tag_set, 0, sizeof(*tag_set)); + if (shost->hostt->commit_rqs) + tag_set->ops = &scsi_mq_ops; + else + tag_set->ops = &scsi_mq_ops_no_commit; + tag_set->nr_hw_queues = shost->nr_hw_queues ? : 1; + tag_set->nr_maps = shost->nr_maps ? : 1; + tag_set->queue_depth = shost->can_queue + shost->nr_reserved_cmds; + tag_set->reserved_tags = shost->nr_reserved_cmds; + tag_set->cmd_size = cmd_size; + tag_set->numa_node = dev_to_node(shost->dma_dev); + if (shost->hostt->tag_alloc_policy_rr) + tag_set->flags |= BLK_MQ_F_TAG_RR; + if (shost->queuecommand_may_block) + tag_set->flags |= BLK_MQ_F_BLOCKING; + tag_set->driver_data = shost; + if (shost->host_tagset) + tag_set->flags |= BLK_MQ_F_TAG_HCTX_SHARED; - return blk_mq_alloc_tag_set(&shost->tag_set); + return blk_mq_alloc_tag_set(tag_set); } -void scsi_mq_destroy_tags(struct Scsi_Host *shost) +void scsi_mq_free_tags(struct kref *kref) { + struct Scsi_Host *shost = container_of(kref, typeof(*shost), + tagset_refcnt); + blk_mq_free_tag_set(&shost->tag_set); + complete(&shost->tagset_freed); } /** + * scsi_get_internal_cmd() - Allocate an internal SCSI command. + * @sdev: SCSI device from which to allocate the command + * @data_direction: Data direction for the allocated command + * @flags: request allocation flags, e.g. BLK_MQ_REQ_RESERVED or + * BLK_MQ_REQ_NOWAIT. + * + * Allocates a SCSI command for internal LLDD use. + */ +struct scsi_cmnd *scsi_get_internal_cmd(struct scsi_device *sdev, + enum dma_data_direction data_direction, + blk_mq_req_flags_t flags) +{ + enum req_op op = data_direction == DMA_TO_DEVICE ? REQ_OP_DRV_OUT : + REQ_OP_DRV_IN; + struct scsi_cmnd *scmd; + struct request *rq; + + rq = scsi_alloc_request(sdev->request_queue, op, flags); + if (IS_ERR(rq)) + return NULL; + scmd = blk_mq_rq_to_pdu(rq); + scmd->device = sdev; + + return scmd; +} +EXPORT_SYMBOL_GPL(scsi_get_internal_cmd); + +/** + * scsi_put_internal_cmd() - Free an internal SCSI command. + * @scmd: SCSI command to be freed + */ +void scsi_put_internal_cmd(struct scsi_cmnd *scmd) +{ + blk_mq_free_request(blk_mq_rq_from_pdu(scmd)); +} +EXPORT_SYMBOL_GPL(scsi_put_internal_cmd); + +/** * scsi_device_from_queue - return sdev associated with a request_queue * @q: The request queue to return the sdev from * @@ -2236,33 +2183,30 @@ struct scsi_device *scsi_device_from_queue(struct request_queue *q) { struct scsi_device *sdev = NULL; - if (q->mq_ops) { - if (q->mq_ops == &scsi_mq_ops) - sdev = q->queuedata; - } else if (q->request_fn == scsi_request_fn) + if (q->mq_ops == &scsi_mq_ops_no_commit || + q->mq_ops == &scsi_mq_ops) sdev = q->queuedata; if (!sdev || !get_device(&sdev->sdev_gendev)) sdev = NULL; return sdev; } +/* + * pktcdvd should have been integrated into the SCSI layers, but for historical + * reasons like the old IDE driver it isn't. This export allows it to safely + * probe if a given device is a SCSI one and only attach to that. + */ +#ifdef CONFIG_CDROM_PKTCDVD_MODULE EXPORT_SYMBOL_GPL(scsi_device_from_queue); +#endif -/* - * Function: scsi_block_requests() - * - * Purpose: Utility function used by low-level drivers to prevent further - * commands from being queued to the device. - * - * Arguments: shost - Host in question - * - * Returns: Nothing - * - * Lock status: No locks are assumed held. +/** + * scsi_block_requests - Utility function used by low-level drivers to prevent + * further commands from being queued to the device. + * @shost: host in question * - * Notes: There is no timer nor any other means by which the requests - * get unblocked other than the low-level driver calling - * scsi_unblock_requests(). + * There is no timer nor any other means by which the requests get unblocked + * other than the low-level driver calling scsi_unblock_requests(). */ void scsi_block_requests(struct Scsi_Host *shost) { @@ -2270,25 +2214,15 @@ void scsi_block_requests(struct Scsi_Host *shost) } EXPORT_SYMBOL(scsi_block_requests); -/* - * Function: scsi_unblock_requests() - * - * Purpose: Utility function used by low-level drivers to allow further - * commands from being queued to the device. - * - * Arguments: shost - Host in question - * - * Returns: Nothing - * - * Lock status: No locks are assumed held. - * - * Notes: There is no timer nor any other means by which the requests - * get unblocked other than the low-level driver calling - * scsi_unblock_requests(). - * - * This is done as an API function so that changes to the - * internals of the scsi mid-layer won't require wholesale - * changes to drivers that use this feature. +/** + * scsi_unblock_requests - Utility function used by low-level drivers to allow + * further commands to be queued to the device. + * @shost: host in question + * + * There is no timer nor any other means by which the requests get unblocked + * other than the low-level driver calling scsi_unblock_requests(). This is done + * as an API function so that changes to the internals of the scsi mid-layer + * won't require wholesale changes to drivers that use this feature. */ void scsi_unblock_requests(struct Scsi_Host *shost) { @@ -2297,24 +2231,9 @@ void scsi_unblock_requests(struct Scsi_Host *shost) } EXPORT_SYMBOL(scsi_unblock_requests); -int __init scsi_init_queue(void) -{ - scsi_sdb_cache = kmem_cache_create("scsi_data_buffer", - sizeof(struct scsi_data_buffer), - 0, 0, NULL); - if (!scsi_sdb_cache) { - printk(KERN_ERR "SCSI: can't init scsi sdb cache\n"); - return -ENOMEM; - } - - return 0; -} - void scsi_exit_queue(void) { kmem_cache_destroy(scsi_sense_cache); - kmem_cache_destroy(scsi_sense_isadma_cache); - kmem_cache_destroy(scsi_sdb_cache); } /** @@ -2322,7 +2241,6 @@ void scsi_exit_queue(void) * @sdev: SCSI device to be queried * @pf: Page format bit (1 == standard, 0 == vendor specific) * @sp: Save page bit (0 == don't save, 1 == save) - * @modepage: mode page being requested * @buffer: request buffer (may not be smaller than eight bytes) * @len: length of request buffer. * @timeout: command timeout @@ -2335,20 +2253,29 @@ void scsi_exit_queue(void) * status on error * */ -int -scsi_mode_select(struct scsi_device *sdev, int pf, int sp, int modepage, - unsigned char *buffer, int len, int timeout, int retries, - struct scsi_mode_data *data, struct scsi_sense_hdr *sshdr) +int scsi_mode_select(struct scsi_device *sdev, int pf, int sp, + unsigned char *buffer, int len, int timeout, int retries, + struct scsi_mode_data *data, struct scsi_sense_hdr *sshdr) { unsigned char cmd[10]; unsigned char *real_buffer; + const struct scsi_exec_args exec_args = { + .sshdr = sshdr, + }; int ret; memset(cmd, 0, sizeof(cmd)); cmd[1] = (pf ? 0x10 : 0) | (sp ? 0x01 : 0); - if (sdev->use_10_for_ms) { - if (len > 65535) + /* + * Use MODE SELECT(10) if the device asked for it or if the mode page + * and the mode select header cannot fit within the maximumm 255 bytes + * of the MODE SELECT(6) command. + */ + if (sdev->use_10_for_ms || + len + 4 > 255 || + data->block_descriptor_length > 255) { + if (len > 65535 - 8) return -EINVAL; real_buffer = kmalloc(8 + len, GFP_KERNEL); if (!real_buffer) @@ -2361,15 +2288,13 @@ scsi_mode_select(struct scsi_device *sdev, int pf, int sp, int modepage, real_buffer[3] = data->device_specific; real_buffer[4] = data->longlba ? 0x01 : 0; real_buffer[5] = 0; - real_buffer[6] = data->block_descriptor_length >> 8; - real_buffer[7] = data->block_descriptor_length; + put_unaligned_be16(data->block_descriptor_length, + &real_buffer[6]); cmd[0] = MODE_SELECT_10; - cmd[7] = len >> 8; - cmd[8] = len; + put_unaligned_be16(len, &cmd[7]); } else { - if (len > 255 || data->block_descriptor_length > 255 || - data->longlba) + if (data->longlba) return -EINVAL; real_buffer = kmalloc(4 + len, GFP_KERNEL); @@ -2381,14 +2306,13 @@ scsi_mode_select(struct scsi_device *sdev, int pf, int sp, int modepage, real_buffer[1] = data->medium_type; real_buffer[2] = data->device_specific; real_buffer[3] = data->block_descriptor_length; - cmd[0] = MODE_SELECT; cmd[4] = len; } - ret = scsi_execute_req(sdev, cmd, DMA_TO_DEVICE, real_buffer, len, - sshdr, timeout, retries, NULL); + ret = scsi_execute_cmd(sdev, cmd, REQ_OP_DRV_OUT, real_buffer, len, + timeout, retries, &exec_args); kfree(real_buffer); return ret; } @@ -2397,8 +2321,9 @@ EXPORT_SYMBOL_GPL(scsi_mode_select); /** * scsi_mode_sense - issue a mode sense, falling back from 10 to six bytes if necessary. * @sdev: SCSI device to be queried - * @dbd: set if mode sense will allow block descriptors to be returned + * @dbd: set to prevent mode sense from returning block descriptors * @modepage: mode page being requested + * @subpage: sub-page of the mode page being requested * @buffer: request buffer (may not be smaller than eight bytes) * @len: length of request buffer. * @timeout: command timeout @@ -2407,43 +2332,60 @@ EXPORT_SYMBOL_GPL(scsi_mode_select); * @sshdr: place to put sense data (or NULL if no sense to be collected). * must be SCSI_SENSE_BUFFERSIZE big. * - * Returns zero if unsuccessful, or the header offset (either 4 - * or 8 depending on whether a six or ten byte command was - * issued) if successful. + * Returns zero if successful, or a negative error number on failure */ int -scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, +scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, int subpage, unsigned char *buffer, int len, int timeout, int retries, struct scsi_mode_data *data, struct scsi_sense_hdr *sshdr) { unsigned char cmd[12]; int use_10_for_ms; int header_length; - int result, retry_count = retries; + int result; struct scsi_sense_hdr my_sshdr; + struct scsi_failure failure_defs[] = { + { + .sense = UNIT_ATTENTION, + .asc = SCMD_FAILURE_ASC_ANY, + .ascq = SCMD_FAILURE_ASCQ_ANY, + .allowed = retries, + .result = SAM_STAT_CHECK_CONDITION, + }, + {} + }; + struct scsi_failures failures = { + .failure_definitions = failure_defs, + }; + const struct scsi_exec_args exec_args = { + /* caller might not be interested in sense, but we need it */ + .sshdr = sshdr ? : &my_sshdr, + .failures = &failures, + }; memset(data, 0, sizeof(*data)); memset(&cmd[0], 0, 12); + + dbd = sdev->set_dbd_for_ms ? 8 : dbd; cmd[1] = dbd & 0x18; /* allows DBD and LLBA bits */ cmd[2] = modepage; + cmd[3] = subpage; - /* caller might not be interested in sense, but we need it */ - if (!sshdr) - sshdr = &my_sshdr; + sshdr = exec_args.sshdr; retry: - use_10_for_ms = sdev->use_10_for_ms; + use_10_for_ms = sdev->use_10_for_ms || len > 255; if (use_10_for_ms) { - if (len < 8) - len = 8; + if (len < 8 || len > 65535) + return -EINVAL; cmd[0] = MODE_SENSE_10; - cmd[8] = len; + put_unaligned_be16(len, &cmd[7]); header_length = 8; } else { if (len < 4) - len = 4; + return -EINVAL; cmd[0] = MODE_SENSE; cmd[4] = len; @@ -2452,60 +2394,61 @@ scsi_mode_sense(struct scsi_device *sdev, int dbd, int modepage, memset(buffer, 0, len); - result = scsi_execute_req(sdev, cmd, DMA_FROM_DEVICE, buffer, len, - sshdr, timeout, retries, NULL); + result = scsi_execute_cmd(sdev, cmd, REQ_OP_DRV_IN, buffer, len, + timeout, retries, &exec_args); + if (result < 0) + return result; /* This code looks awful: what it's doing is making sure an * ILLEGAL REQUEST sense return identifies the actual command * byte as the problem. MODE_SENSE commands can return * ILLEGAL REQUEST if the code page isn't supported */ - if (use_10_for_ms && !scsi_status_is_good(result) && - (driver_byte(result) & DRIVER_SENSE)) { + if (!scsi_status_is_good(result)) { if (scsi_sense_valid(sshdr)) { if ((sshdr->sense_key == ILLEGAL_REQUEST) && (sshdr->asc == 0x20) && (sshdr->ascq == 0)) { - /* - * Invalid command operation code + /* + * Invalid command operation code: retry using + * MODE SENSE(6) if this was a MODE SENSE(10) + * request, except if the request mode page is + * too large for MODE SENSE single byte + * allocation length field. */ - sdev->use_10_for_ms = 0; - goto retry; + if (use_10_for_ms) { + if (len > 255) + return -EIO; + sdev->use_10_for_ms = 0; + goto retry; + } } } + return -EIO; + } + if (unlikely(buffer[0] == 0x86 && buffer[1] == 0x0b && + (modepage == 6 || modepage == 8))) { + /* Initio breakage? */ + header_length = 0; + data->length = 13; + data->medium_type = 0; + data->device_specific = 0; + data->longlba = 0; + data->block_descriptor_length = 0; + } else if (use_10_for_ms) { + data->length = get_unaligned_be16(&buffer[0]) + 2; + data->medium_type = buffer[2]; + data->device_specific = buffer[3]; + data->longlba = buffer[4] & 0x01; + data->block_descriptor_length = get_unaligned_be16(&buffer[6]); + } else { + data->length = buffer[0] + 1; + data->medium_type = buffer[1]; + data->device_specific = buffer[2]; + data->block_descriptor_length = buffer[3]; } + data->header_length = header_length; - if(scsi_status_is_good(result)) { - if (unlikely(buffer[0] == 0x86 && buffer[1] == 0x0b && - (modepage == 6 || modepage == 8))) { - /* Initio breakage? */ - header_length = 0; - data->length = 13; - data->medium_type = 0; - data->device_specific = 0; - data->longlba = 0; - data->block_descriptor_length = 0; - } else if(use_10_for_ms) { - data->length = buffer[0]*256 + buffer[1] + 2; - data->medium_type = buffer[2]; - data->device_specific = buffer[3]; - data->longlba = buffer[4] & 0x01; - data->block_descriptor_length = buffer[6]*256 - + buffer[7]; - } else { - data->length = buffer[0] + 1; - data->medium_type = buffer[1]; - data->device_specific = buffer[2]; - data->block_descriptor_length = buffer[3]; - } - data->header_length = header_length; - } else if ((status_byte(result) == CHECK_CONDITION) && - scsi_sense_valid(sshdr) && - sshdr->sense_key == UNIT_ATTENTION && retry_count) { - retry_count--; - goto retry; - } - - return result; + return 0; } EXPORT_SYMBOL(scsi_mode_sense); @@ -2526,16 +2469,19 @@ scsi_test_unit_ready(struct scsi_device *sdev, int timeout, int retries, char cmd[] = { TEST_UNIT_READY, 0, 0, 0, 0, 0, }; + const struct scsi_exec_args exec_args = { + .sshdr = sshdr, + }; int result; /* try to eat the UNIT_ATTENTION if there are enough retries */ do { - result = scsi_execute_req(sdev, cmd, DMA_NONE, NULL, 0, sshdr, - timeout, retries, NULL); - if (sdev->removable && scsi_sense_valid(sshdr) && + result = scsi_execute_cmd(sdev, cmd, REQ_OP_DRV_IN, NULL, 0, + timeout, 1, &exec_args); + if (sdev->removable && result > 0 && scsi_sense_valid(sshdr) && sshdr->sense_key == UNIT_ATTENTION) sdev->changed = 1; - } while (scsi_sense_valid(sshdr) && + } while (result > 0 && scsi_sense_valid(sshdr) && sshdr->sense_key == UNIT_ATTENTION && --retries); return result; @@ -2547,7 +2493,7 @@ EXPORT_SYMBOL(scsi_test_unit_ready); * @sdev: scsi device to change the state of. * @state: state to change to. * - * Returns zero if unsuccessful or an error if the requested + * Returns zero if successful or an error if the requested * transition is illegal. */ int @@ -2567,7 +2513,7 @@ scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state) goto illegal; } break; - + case SDEV_RUNNING: switch (oldstate) { case SDEV_CREATED: @@ -2609,6 +2555,8 @@ scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state) switch (oldstate) { case SDEV_RUNNING: case SDEV_CREATED_BLOCK: + case SDEV_QUIESCE: + case SDEV_OFFLINE: break; default: goto illegal; @@ -2653,6 +2601,7 @@ scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state) break; } + sdev->offline_already = false; sdev->sdev_state = state; return 0; @@ -2668,7 +2617,7 @@ scsi_device_set_state(struct scsi_device *sdev, enum scsi_device_state state) EXPORT_SYMBOL(scsi_device_set_state); /** - * sdev_evt_emit - emit a single SCSI device uevent + * scsi_evt_emit - emit a single SCSI device uevent * @sdev: associated SCSI device * @evt: event to emit * @@ -2684,7 +2633,7 @@ static void scsi_evt_emit(struct scsi_device *sdev, struct scsi_event *evt) envp[idx++] = "SDEV_MEDIA_CHANGE=1"; break; case SDEV_EVT_INQUIRY_CHANGE_REPORTED: - scsi_rescan_device(&sdev->sdev_gendev); + scsi_rescan_device(sdev); envp[idx++] = "SDEV_UA=INQUIRY_DATA_HAS_CHANGED"; break; case SDEV_EVT_CAPACITY_CHANGE_REPORTED: @@ -2702,6 +2651,9 @@ static void scsi_evt_emit(struct scsi_device *sdev, struct scsi_event *evt) case SDEV_EVT_ALUA_STATE_CHANGE_REPORTED: envp[idx++] = "SDEV_UA=ASYMMETRIC_ACCESS_STATE_CHANGED"; break; + case SDEV_EVT_POWER_ON_RESET_OCCURRED: + envp[idx++] = "SDEV_UA=POWER_ON_RESET_OCCURRED"; + break; default: /* do nothing */ break; @@ -2713,7 +2665,7 @@ static void scsi_evt_emit(struct scsi_device *sdev, struct scsi_event *evt) } /** - * sdev_evt_thread - send a uevent for each scsi event + * scsi_evt_thread - send a uevent for each scsi event * @work: work struct for scsi_device * * Dispatch queued events to their associated scsi_device kobjects @@ -2806,6 +2758,7 @@ struct scsi_event *sdev_evt_alloc(enum scsi_device_event evt_type, case SDEV_EVT_MODE_PARAMETER_CHANGE_REPORTED: case SDEV_EVT_LUN_CHANGE_REPORTED: case SDEV_EVT_ALUA_STATE_CHANGE_REPORTED: + case SDEV_EVT_POWER_ON_RESET_OCCURRED: default: /* do nothing */ break; @@ -2838,48 +2791,13 @@ void sdev_evt_send_simple(struct scsi_device *sdev, EXPORT_SYMBOL_GPL(sdev_evt_send_simple); /** - * scsi_request_fn_active() - number of kernel threads inside scsi_request_fn() - * @sdev: SCSI device to count the number of scsi_request_fn() callers for. - */ -static int scsi_request_fn_active(struct scsi_device *sdev) -{ - struct request_queue *q = sdev->request_queue; - int request_fn_active; - - WARN_ON_ONCE(sdev->host->use_blk_mq); - - spin_lock_irq(q->queue_lock); - request_fn_active = q->request_fn_active; - spin_unlock_irq(q->queue_lock); - - return request_fn_active; -} - -/** - * scsi_wait_for_queuecommand() - wait for ongoing queuecommand() calls - * @sdev: SCSI device pointer. - * - * Wait until the ongoing shost->hostt->queuecommand() calls that are - * invoked from scsi_request_fn() have finished. - */ -static void scsi_wait_for_queuecommand(struct scsi_device *sdev) -{ - WARN_ON_ONCE(sdev->host->use_blk_mq); - - while (scsi_request_fn_active(sdev)) - msleep(20); -} - -/** - * scsi_device_quiesce - Block user issued commands. + * scsi_device_quiesce - Block all commands except power management. * @sdev: scsi device to quiesce. * * This works by trying to transition to the SDEV_QUIESCE state * (which must be a legal transition). When the device is in this - * state, only special requests will be accepted, all others will - * be deferred. Since special requests may also be requeued requests, - * a successful return doesn't guarantee the device will be - * totally quiescent. + * state, only power management requests will be accepted, all others will + * be deferred. * * Must be called with user context, may sleep. * @@ -2888,21 +2806,41 @@ static void scsi_wait_for_queuecommand(struct scsi_device *sdev) int scsi_device_quiesce(struct scsi_device *sdev) { + struct request_queue *q = sdev->request_queue; + unsigned int memflags; int err; + /* + * It is allowed to call scsi_device_quiesce() multiple times from + * the same context but concurrent scsi_device_quiesce() calls are + * not allowed. + */ + WARN_ON_ONCE(sdev->quiesced_by && sdev->quiesced_by != current); + + if (sdev->quiesced_by == current) + return 0; + + blk_set_pm_only(q); + + memflags = blk_mq_freeze_queue(q); + /* + * Ensure that the effect of blk_set_pm_only() will be visible + * for percpu_ref_tryget() callers that occur after the queue + * unfreeze even if the queue was already frozen before this function + * was called. See also https://lwn.net/Articles/573497/. + */ + synchronize_rcu(); + blk_mq_unfreeze_queue(q, memflags); + mutex_lock(&sdev->state_mutex); err = scsi_device_set_state(sdev, SDEV_QUIESCE); + if (err == 0) + sdev->quiesced_by = current; + else + blk_clear_pm_only(q); mutex_unlock(&sdev->state_mutex); - if (err) - return err; - - scsi_run_queue(sdev->request_queue); - while (atomic_read(&sdev->device_busy)) { - msleep_interruptible(200); - scsi_run_queue(sdev->request_queue); - } - return 0; + return err; } EXPORT_SYMBOL(scsi_device_quiesce); @@ -2922,9 +2860,12 @@ void scsi_device_resume(struct scsi_device *sdev) * device deleted during suspend) */ mutex_lock(&sdev->state_mutex); - if (sdev->sdev_state == SDEV_QUIESCE && - scsi_device_set_state(sdev, SDEV_RUNNING) == 0) - scsi_run_queue(sdev->request_queue); + if (sdev->sdev_state == SDEV_QUIESCE) + scsi_device_set_state(sdev, SDEV_RUNNING); + if (sdev->quiesced_by) { + sdev->quiesced_by = NULL; + blk_clear_pm_only(sdev->request_queue); + } mutex_unlock(&sdev->state_mutex); } EXPORT_SYMBOL(scsi_device_resume); @@ -2955,6 +2896,32 @@ scsi_target_resume(struct scsi_target *starget) } EXPORT_SYMBOL(scsi_target_resume); +static int __scsi_internal_device_block_nowait(struct scsi_device *sdev) +{ + if (scsi_device_set_state(sdev, SDEV_BLOCK)) + return scsi_device_set_state(sdev, SDEV_CREATED_BLOCK); + + return 0; +} + +void scsi_start_queue(struct scsi_device *sdev) +{ + if (cmpxchg(&sdev->queue_stopped, 1, 0)) + blk_mq_unquiesce_queue(sdev->request_queue); +} + +static void scsi_stop_queue(struct scsi_device *sdev) +{ + /* + * The atomic variable of ->queue_stopped covers that + * blk_mq_quiesce_queue* is balanced with blk_mq_unquiesce_queue. + * + * The caller needs to wait until quiesce is done. + */ + if (!cmpxchg(&sdev->queue_stopped, 0, 1)) + blk_mq_quiesce_queue_nowait(sdev->request_queue); +} + /** * scsi_internal_device_block_nowait - try to transition to the SDEV_BLOCK state * @sdev: device to block @@ -2971,84 +2938,54 @@ EXPORT_SYMBOL(scsi_target_resume); */ int scsi_internal_device_block_nowait(struct scsi_device *sdev) { - struct request_queue *q = sdev->request_queue; - unsigned long flags; - int err = 0; - - err = scsi_device_set_state(sdev, SDEV_BLOCK); - if (err) { - err = scsi_device_set_state(sdev, SDEV_CREATED_BLOCK); - - if (err) - return err; - } + int ret = __scsi_internal_device_block_nowait(sdev); - /* + /* * The device has transitioned to SDEV_BLOCK. Stop the * block layer from calling the midlayer with this device's - * request queue. + * request queue. */ - if (q->mq_ops) { - blk_mq_quiesce_queue_nowait(q); - } else { - spin_lock_irqsave(q->queue_lock, flags); - blk_stop_queue(q); - spin_unlock_irqrestore(q->queue_lock, flags); - } - - return 0; + if (!ret) + scsi_stop_queue(sdev); + return ret; } EXPORT_SYMBOL_GPL(scsi_internal_device_block_nowait); /** - * scsi_internal_device_block - try to transition to the SDEV_BLOCK state + * scsi_device_block - try to transition to the SDEV_BLOCK state * @sdev: device to block + * @data: dummy argument, ignored * - * Pause SCSI command processing on the specified device and wait until all - * ongoing scsi_request_fn() / scsi_queue_rq() calls have finished. May sleep. - * - * Returns zero if successful or a negative error code upon failure. + * Pause SCSI command processing on the specified device. Callers must wait + * until all ongoing scsi_queue_rq() calls have finished after this function + * returns. * * Note: * This routine transitions the device to the SDEV_BLOCK state (which must be * a legal transition). When the device is in this state, command processing * is paused until the device leaves the SDEV_BLOCK state. See also * scsi_internal_device_unblock(). - * - * To do: avoid that scsi_send_eh_cmnd() calls queuecommand() after - * scsi_internal_device_block() has blocked a SCSI device and also - * remove the rport mutex lock and unlock calls from srp_queuecommand(). */ -static int scsi_internal_device_block(struct scsi_device *sdev) +static void scsi_device_block(struct scsi_device *sdev, void *data) { - struct request_queue *q = sdev->request_queue; int err; + enum scsi_device_state state; mutex_lock(&sdev->state_mutex); - err = scsi_internal_device_block_nowait(sdev); - if (err == 0) { - if (q->mq_ops) - blk_mq_quiesce_queue(q); - else - scsi_wait_for_queuecommand(sdev); - } - mutex_unlock(&sdev->state_mutex); + err = __scsi_internal_device_block_nowait(sdev); + state = sdev->sdev_state; + if (err == 0) + /* + * scsi_stop_queue() must be called with the state_mutex + * held. Otherwise a simultaneous scsi_start_queue() call + * might unquiesce the queue before we quiesce it. + */ + scsi_stop_queue(sdev); - return err; -} - -void scsi_start_queue(struct scsi_device *sdev) -{ - struct request_queue *q = sdev->request_queue; - unsigned long flags; + mutex_unlock(&sdev->state_mutex); - if (q->mq_ops) { - blk_mq_unquiesce_queue(q); - } else { - spin_lock_irqsave(q->queue_lock, flags); - blk_start_queue(q); - spin_unlock_irqrestore(q->queue_lock, flags); - } + WARN_ONCE(err, "%s: failed to block %s in state %d\n", + __func__, dev_name(&sdev->sdev_gendev), state); } /** @@ -3069,23 +3006,36 @@ void scsi_start_queue(struct scsi_device *sdev) int scsi_internal_device_unblock_nowait(struct scsi_device *sdev, enum scsi_device_state new_state) { + switch (new_state) { + case SDEV_RUNNING: + case SDEV_TRANSPORT_OFFLINE: + break; + default: + return -EINVAL; + } + /* * Try to transition the scsi device to SDEV_RUNNING or one of the * offlined states and goose the device queue if successful. */ - if ((sdev->sdev_state == SDEV_BLOCK) || - (sdev->sdev_state == SDEV_TRANSPORT_OFFLINE)) + switch (sdev->sdev_state) { + case SDEV_BLOCK: + case SDEV_TRANSPORT_OFFLINE: sdev->sdev_state = new_state; - else if (sdev->sdev_state == SDEV_CREATED_BLOCK) { + break; + case SDEV_CREATED_BLOCK: if (new_state == SDEV_TRANSPORT_OFFLINE || new_state == SDEV_OFFLINE) sdev->sdev_state = new_state; else sdev->sdev_state = SDEV_CREATED; - } else if (sdev->sdev_state != SDEV_CANCEL && - sdev->sdev_state != SDEV_OFFLINE) + break; + case SDEV_CANCEL: + case SDEV_OFFLINE: + break; + default: return -EINVAL; - + } scsi_start_queue(sdev); return 0; @@ -3118,31 +3068,35 @@ static int scsi_internal_device_unblock(struct scsi_device *sdev, return ret; } -static void -device_block(struct scsi_device *sdev, void *data) -{ - scsi_internal_device_block(sdev); -} - static int target_block(struct device *dev, void *data) { if (scsi_is_target_device(dev)) starget_for_each_device(to_scsi_target(dev), NULL, - device_block); + scsi_device_block); return 0; } +/** + * scsi_block_targets - transition all SCSI child devices to SDEV_BLOCK state + * @dev: a parent device of one or more scsi_target devices + * @shost: the Scsi_Host to which this device belongs + * + * Iterate over all children of @dev, which should be scsi_target devices, + * and switch all subordinate scsi devices to SDEV_BLOCK state. Wait for + * ongoing scsi_queue_rq() calls to finish. May sleep. + * + * Note: + * @dev must not itself be a scsi_target device. + */ void -scsi_target_block(struct device *dev) +scsi_block_targets(struct Scsi_Host *shost, struct device *dev) { - if (scsi_is_target_device(dev)) - starget_for_each_device(to_scsi_target(dev), NULL, - device_block); - else - device_for_each_child(dev, NULL, target_block); + WARN_ON_ONCE(scsi_is_target_device(dev)); + device_for_each_child(dev, NULL, target_block); + blk_mq_wait_quiesce_done(&shost->tag_set); } -EXPORT_SYMBOL_GPL(scsi_target_block); +EXPORT_SYMBOL_GPL(scsi_block_targets); static void device_unblock(struct scsi_device *sdev, void *data) @@ -3171,6 +3125,59 @@ scsi_target_unblock(struct device *dev, enum scsi_device_state new_state) EXPORT_SYMBOL_GPL(scsi_target_unblock); /** + * scsi_host_block - Try to transition all logical units to the SDEV_BLOCK state + * @shost: device to block + * + * Pause SCSI command processing for all logical units associated with the SCSI + * host and wait until pending scsi_queue_rq() calls have finished. + * + * Returns zero if successful or a negative error code upon failure. + */ +int +scsi_host_block(struct Scsi_Host *shost) +{ + struct scsi_device *sdev; + int ret; + + /* + * Call scsi_internal_device_block_nowait so we can avoid + * calling synchronize_rcu() for each LUN. + */ + shost_for_each_device(sdev, shost) { + mutex_lock(&sdev->state_mutex); + ret = scsi_internal_device_block_nowait(sdev); + mutex_unlock(&sdev->state_mutex); + if (ret) { + scsi_device_put(sdev); + return ret; + } + } + + /* Wait for ongoing scsi_queue_rq() calls to finish. */ + blk_mq_wait_quiesce_done(&shost->tag_set); + + return 0; +} +EXPORT_SYMBOL_GPL(scsi_host_block); + +int +scsi_host_unblock(struct Scsi_Host *shost, int new_state) +{ + struct scsi_device *sdev; + int ret = 0; + + shost_for_each_device(sdev, shost) { + ret = scsi_internal_device_unblock(sdev, new_state); + if (ret) { + scsi_device_put(sdev); + break; + } + } + return ret; +} +EXPORT_SYMBOL_GPL(scsi_host_unblock); + +/** * scsi_kmap_atomic_sg - find and atomically map an sg-elemnt * @sgl: scatter-gather list * @sg_count: number of segments in sg @@ -3207,8 +3214,7 @@ void *scsi_kmap_atomic_sg(struct scatterlist *sgl, int sg_count, /* Offset starting from the beginning of first page in this sg-entry */ *offset = *offset - len_complete + sg->offset; - /* Assumption: contiguous pages can be accessed as "page + i" */ - page = nth_page(sg_page(sg), (*offset >> PAGE_SHIFT)); + page = sg_page(sg) + (*offset >> PAGE_SHIFT); *offset &= ~PAGE_MASK; /* Bytes in this sg-entry from *offset to the end of the page */ @@ -3244,6 +3250,78 @@ void sdev_enable_disk_events(struct scsi_device *sdev) } EXPORT_SYMBOL(sdev_enable_disk_events); +static unsigned char designator_prio(const unsigned char *d) +{ + if (d[1] & 0x30) + /* not associated with LUN */ + return 0; + + if (d[3] == 0) + /* invalid length */ + return 0; + + /* + * Order of preference for lun descriptor: + * - SCSI name string + * - NAA IEEE Registered Extended + * - EUI-64 based 16-byte + * - EUI-64 based 12-byte + * - NAA IEEE Registered + * - NAA IEEE Extended + * - EUI-64 based 8-byte + * - SCSI name string (truncated) + * - T10 Vendor ID + * as longer descriptors reduce the likelyhood + * of identification clashes. + */ + + switch (d[1] & 0xf) { + case 8: + /* SCSI name string, variable-length UTF-8 */ + return 9; + case 3: + switch (d[4] >> 4) { + case 6: + /* NAA registered extended */ + return 8; + case 5: + /* NAA registered */ + return 5; + case 4: + /* NAA extended */ + return 4; + case 3: + /* NAA locally assigned */ + return 1; + default: + break; + } + break; + case 2: + switch (d[3]) { + case 16: + /* EUI64-based, 16 byte */ + return 7; + case 12: + /* EUI64-based, 12 byte */ + return 6; + case 8: + /* EUI64-based, 8 byte */ + return 3; + default: + break; + } + break; + case 1: + /* T10 vendor ID */ + return 1; + default: + break; + } + + return 0; +} + /** * scsi_vpd_lun_id - return a unique device identification * @sdev: SCSI device @@ -3260,10 +3338,10 @@ EXPORT_SYMBOL(sdev_enable_disk_events); */ int scsi_vpd_lun_id(struct scsi_device *sdev, char *id, size_t id_len) { - u8 cur_id_type = 0xff; + u8 cur_id_prio = 0; u8 cur_id_size = 0; - unsigned char *d, *cur_id_str; - unsigned char __rcu *vpd_pg83; + const unsigned char *d, *cur_id_str; + const struct scsi_vpd *vpd_pg83; int id_size = -EINVAL; rcu_read_lock(); @@ -3273,20 +3351,6 @@ int scsi_vpd_lun_id(struct scsi_device *sdev, char *id, size_t id_len) return -ENXIO; } - /* - * Look for the correct descriptor. - * Order of preference for lun descriptor: - * - SCSI name string - * - NAA IEEE Registered Extended - * - EUI-64 based 16-byte - * - EUI-64 based 12-byte - * - NAA IEEE Registered - * - NAA IEEE Extended - * - T10 Vendor ID - * as longer descriptors reduce the likelyhood - * of identification clashes. - */ - /* The id string must be at least 20 bytes + terminating NULL byte */ if (id_len < 21) { rcu_read_unlock(); @@ -3294,39 +3358,32 @@ int scsi_vpd_lun_id(struct scsi_device *sdev, char *id, size_t id_len) } memset(id, 0, id_len); - d = vpd_pg83 + 4; - while (d < vpd_pg83 + sdev->vpd_pg83_len) { - /* Skip designators not referring to the LUN */ - if ((d[1] & 0x30) != 0x00) - goto next_desig; + for (d = vpd_pg83->data + 4; + d < vpd_pg83->data + vpd_pg83->len; + d += d[3] + 4) { + u8 prio = designator_prio(d); + + if (prio == 0 || cur_id_prio > prio) + continue; switch (d[1] & 0xf) { case 0x1: /* T10 Vendor ID */ if (cur_id_size > d[3]) break; - /* Prefer anything */ - if (cur_id_type > 0x01 && cur_id_type != 0xff) - break; + cur_id_prio = prio; cur_id_size = d[3]; if (cur_id_size + 4 > id_len) cur_id_size = id_len - 4; cur_id_str = d + 4; - cur_id_type = d[1] & 0xf; id_size = snprintf(id, id_len, "t10.%*pE", cur_id_size, cur_id_str); break; case 0x2: /* EUI-64 */ - if (cur_id_size > d[3]) - break; - /* Prefer NAA IEEE Registered Extended */ - if (cur_id_type == 0x3 && - cur_id_size == d[3]) - break; + cur_id_prio = prio; cur_id_size = d[3]; cur_id_str = d + 4; - cur_id_type = d[1] & 0xf; switch (cur_id_size) { case 8: id_size = snprintf(id, id_len, @@ -3344,17 +3401,14 @@ int scsi_vpd_lun_id(struct scsi_device *sdev, char *id, size_t id_len) cur_id_str); break; default: - cur_id_size = 0; break; } break; case 0x3: /* NAA */ - if (cur_id_size > d[3]) - break; + cur_id_prio = prio; cur_id_size = d[3]; cur_id_str = d + 4; - cur_id_type = d[1] & 0xf; switch (cur_id_size) { case 8: id_size = snprintf(id, id_len, @@ -3367,32 +3421,29 @@ int scsi_vpd_lun_id(struct scsi_device *sdev, char *id, size_t id_len) cur_id_str); break; default: - cur_id_size = 0; break; } break; case 0x8: /* SCSI name string */ - if (cur_id_size + 4 > d[3]) + if (cur_id_size > d[3]) break; /* Prefer others for truncated descriptor */ - if (cur_id_size && d[3] > id_len) - break; + if (d[3] > id_len) { + prio = 2; + if (cur_id_prio > prio) + break; + } + cur_id_prio = prio; cur_id_size = id_size = d[3]; cur_id_str = d + 4; - cur_id_type = d[1] & 0xf; if (cur_id_size >= id_len) cur_id_size = id_len - 1; memcpy(id, cur_id_str, cur_id_size); - /* Decrease priority for truncated descriptor */ - if (cur_id_size != id_size) - cur_id_size = 6; break; default: break; } -next_desig: - d += d[3] + 4; } rcu_read_unlock(); @@ -3400,19 +3451,21 @@ next_desig: } EXPORT_SYMBOL(scsi_vpd_lun_id); -/* +/** * scsi_vpd_tpg_id - return a target port group identifier * @sdev: SCSI device + * @rel_id: pointer to return relative target port in if not %NULL * * Returns the Target Port Group identifier from the information - * froom VPD page 0x83 of the device. + * from VPD page 0x83 of the device. + * Optionally sets @rel_id to the relative target port on success. * - * Returns the identifier or error on failure. + * Return: the identifier or error on failure. */ int scsi_vpd_tpg_id(struct scsi_device *sdev, int *rel_id) { - unsigned char *d; - unsigned char __rcu *vpd_pg83; + const unsigned char *d; + const struct scsi_vpd *vpd_pg83; int group_id = -EAGAIN, rel_port = -1; rcu_read_lock(); @@ -3422,8 +3475,8 @@ int scsi_vpd_tpg_id(struct scsi_device *sdev, int *rel_id) return -ENXIO; } - d = sdev->vpd_pg83 + 4; - while (d < sdev->vpd_pg83 + sdev->vpd_pg83_len) { + d = vpd_pg83->data + 4; + while (d < vpd_pg83->data + vpd_pg83->len) { switch (d[1] & 0xf) { case 0x4: /* Relative target port */ @@ -3446,3 +3499,24 @@ int scsi_vpd_tpg_id(struct scsi_device *sdev, int *rel_id) return group_id; } EXPORT_SYMBOL(scsi_vpd_tpg_id); + +/** + * scsi_build_sense - build sense data for a command + * @scmd: scsi command for which the sense should be formatted + * @desc: Sense format (non-zero == descriptor format, + * 0 == fixed format) + * @key: Sense key + * @asc: Additional sense code + * @ascq: Additional sense code qualifier + * + **/ +void scsi_build_sense(struct scsi_cmnd *scmd, int desc, u8 key, u8 asc, u8 ascq) +{ + scsi_build_sense_buffer(desc, scmd->sense_buffer, key, asc, ascq); + scmd->result = SAM_STAT_CHECK_CONDITION; +} +EXPORT_SYMBOL_GPL(scsi_build_sense); + +#ifdef CONFIG_SCSI_LIB_KUNIT_TEST +#include "scsi_lib_test.c" +#endif |