diff options
Diffstat (limited to 'drivers/block/nvme-core.c')
| -rw-r--r-- | drivers/block/nvme-core.c | 1271 |
1 files changed, 969 insertions, 302 deletions
diff --git a/drivers/block/nvme-core.c b/drivers/block/nvme-core.c index 1f14ac403945..7c64fa756cce 100644 --- a/drivers/block/nvme-core.c +++ b/drivers/block/nvme-core.c @@ -1,6 +1,6 @@ /* * NVM Express device driver - * Copyright (c) 2011, Intel Corporation. + * Copyright (c) 2011-2014, Intel Corporation. * * This program is free software; you can redistribute it and/or modify it * under the terms and conditions of the GNU General Public License, @@ -20,10 +20,12 @@ #include <linux/bio.h> #include <linux/bitops.h> #include <linux/blkdev.h> +#include <linux/cpu.h> #include <linux/delay.h> #include <linux/errno.h> #include <linux/fs.h> #include <linux/genhd.h> +#include <linux/hdreg.h> #include <linux/idr.h> #include <linux/init.h> #include <linux/interrupt.h> @@ -35,6 +37,7 @@ #include <linux/module.h> #include <linux/moduleparam.h> #include <linux/pci.h> +#include <linux/percpu.h> #include <linux/poison.h> #include <linux/ptrace.h> #include <linux/sched.h> @@ -46,8 +49,12 @@ #define NVME_Q_DEPTH 1024 #define SQ_SIZE(depth) (depth * sizeof(struct nvme_command)) #define CQ_SIZE(depth) (depth * sizeof(struct nvme_completion)) -#define NVME_MINORS 64 #define ADMIN_TIMEOUT (60 * HZ) +#define IOD_TIMEOUT (4 * NVME_IO_TIMEOUT) + +unsigned char io_timeout = 30; +module_param(io_timeout, byte, 0644); +MODULE_PARM_DESC(io_timeout, "timeout in seconds for I/O"); static int nvme_major; module_param(nvme_major, int, 0); @@ -58,14 +65,28 @@ module_param(use_threaded_interrupts, int, 0); static DEFINE_SPINLOCK(dev_list_lock); static LIST_HEAD(dev_list); static struct task_struct *nvme_thread; +static struct workqueue_struct *nvme_workq; +static wait_queue_head_t nvme_kthread_wait; + +static void nvme_reset_failed_dev(struct work_struct *ws); + +struct async_cmd_info { + struct kthread_work work; + struct kthread_worker *worker; + u32 result; + int status; + void *ctx; +}; /* * An NVM Express queue. Each device has at least two (one for admin * commands and one for I/O commands). */ struct nvme_queue { + struct rcu_head r_head; struct device *q_dmadev; struct nvme_dev *dev; + char irqname[24]; /* nvme4294967295-65535\0 */ spinlock_t q_lock; struct nvme_command *sq_cmds; volatile struct nvme_completion *cqes; @@ -74,15 +95,19 @@ struct nvme_queue { wait_queue_head_t sq_full; wait_queue_t sq_cong_wait; struct bio_list sq_cong; + struct list_head iod_bio; u32 __iomem *q_db; u16 q_depth; u16 cq_vector; u16 sq_head; u16 sq_tail; u16 cq_head; + u16 qid; u8 cq_phase; u8 cqe_seen; u8 q_suspended; + cpumask_var_t cpu_mask; + struct async_cmd_info cmdinfo; unsigned long cmdid_data[]; }; @@ -97,6 +122,7 @@ static inline void _nvme_check_size(void) BUILD_BUG_ON(sizeof(struct nvme_delete_queue) != 64); BUILD_BUG_ON(sizeof(struct nvme_features) != 64); BUILD_BUG_ON(sizeof(struct nvme_format_cmd) != 64); + BUILD_BUG_ON(sizeof(struct nvme_abort_cmd) != 64); BUILD_BUG_ON(sizeof(struct nvme_command) != 64); BUILD_BUG_ON(sizeof(struct nvme_id_ctrl) != 4096); BUILD_BUG_ON(sizeof(struct nvme_id_ns) != 4096); @@ -104,13 +130,14 @@ static inline void _nvme_check_size(void) BUILD_BUG_ON(sizeof(struct nvme_smart_log) != 512); } -typedef void (*nvme_completion_fn)(struct nvme_dev *, void *, +typedef void (*nvme_completion_fn)(struct nvme_queue *, void *, struct nvme_completion *); struct nvme_cmd_info { nvme_completion_fn fn; void *ctx; unsigned long timeout; + int aborted; }; static struct nvme_cmd_info *nvme_cmd_info(struct nvme_queue *nvmeq) @@ -154,6 +181,7 @@ static int alloc_cmdid(struct nvme_queue *nvmeq, void *ctx, info[cmdid].fn = handler; info[cmdid].ctx = ctx; info[cmdid].timeout = jiffies + timeout; + info[cmdid].aborted = 0; return cmdid; } @@ -172,28 +200,42 @@ static int alloc_cmdid_killable(struct nvme_queue *nvmeq, void *ctx, #define CMD_CTX_COMPLETED (0x310 + CMD_CTX_BASE) #define CMD_CTX_INVALID (0x314 + CMD_CTX_BASE) #define CMD_CTX_FLUSH (0x318 + CMD_CTX_BASE) +#define CMD_CTX_ABORT (0x31C + CMD_CTX_BASE) -static void special_completion(struct nvme_dev *dev, void *ctx, +static void special_completion(struct nvme_queue *nvmeq, void *ctx, struct nvme_completion *cqe) { if (ctx == CMD_CTX_CANCELLED) return; if (ctx == CMD_CTX_FLUSH) return; + if (ctx == CMD_CTX_ABORT) { + ++nvmeq->dev->abort_limit; + return; + } if (ctx == CMD_CTX_COMPLETED) { - dev_warn(&dev->pci_dev->dev, + dev_warn(nvmeq->q_dmadev, "completed id %d twice on queue %d\n", cqe->command_id, le16_to_cpup(&cqe->sq_id)); return; } if (ctx == CMD_CTX_INVALID) { - dev_warn(&dev->pci_dev->dev, + dev_warn(nvmeq->q_dmadev, "invalid id %d completed on queue %d\n", cqe->command_id, le16_to_cpup(&cqe->sq_id)); return; } - dev_warn(&dev->pci_dev->dev, "Unknown special completion %p\n", ctx); + dev_warn(nvmeq->q_dmadev, "Unknown special completion %p\n", ctx); +} + +static void async_completion(struct nvme_queue *nvmeq, void *ctx, + struct nvme_completion *cqe) +{ + struct async_cmd_info *cmdinfo = ctx; + cmdinfo->result = le32_to_cpup(&cqe->result); + cmdinfo->status = le16_to_cpup(&cqe->status) >> 1; + queue_kthread_work(cmdinfo->worker, &cmdinfo->work); } /* @@ -232,14 +274,34 @@ static void *cancel_cmdid(struct nvme_queue *nvmeq, int cmdid, return ctx; } -struct nvme_queue *get_nvmeq(struct nvme_dev *dev) +static struct nvme_queue *raw_nvmeq(struct nvme_dev *dev, int qid) +{ + return rcu_dereference_raw(dev->queues[qid]); +} + +static struct nvme_queue *get_nvmeq(struct nvme_dev *dev) __acquires(RCU) +{ + unsigned queue_id = get_cpu_var(*dev->io_queue); + rcu_read_lock(); + return rcu_dereference(dev->queues[queue_id]); +} + +static void put_nvmeq(struct nvme_queue *nvmeq) __releases(RCU) { - return dev->queues[get_cpu() + 1]; + rcu_read_unlock(); + put_cpu_var(nvmeq->dev->io_queue); } -void put_nvmeq(struct nvme_queue *nvmeq) +static struct nvme_queue *lock_nvmeq(struct nvme_dev *dev, int q_idx) + __acquires(RCU) { - put_cpu(); + rcu_read_lock(); + return rcu_dereference(dev->queues[q_idx]); +} + +static void unlock_nvmeq(struct nvme_queue *nvmeq) __releases(RCU) +{ + rcu_read_unlock(); } /** @@ -254,6 +316,10 @@ static int nvme_submit_cmd(struct nvme_queue *nvmeq, struct nvme_command *cmd) unsigned long flags; u16 tail; spin_lock_irqsave(&nvmeq->q_lock, flags); + if (nvmeq->q_suspended) { + spin_unlock_irqrestore(&nvmeq->q_lock, flags); + return -EBUSY; + } tail = nvmeq->sq_tail; memcpy(&nvmeq->sq_cmds[tail], cmd, sizeof(*cmd)); if (++tail == nvmeq->q_depth) @@ -293,6 +359,7 @@ nvme_alloc_iod(unsigned nseg, unsigned nbytes, gfp_t gfp) iod->npages = -1; iod->length = nbytes; iod->nents = 0; + iod->first_dma = 0ULL; iod->start_time = jiffies; } @@ -341,19 +408,31 @@ static void nvme_end_io_acct(struct bio *bio, unsigned long start_time) part_stat_unlock(); } -static void bio_completion(struct nvme_dev *dev, void *ctx, +static void bio_completion(struct nvme_queue *nvmeq, void *ctx, struct nvme_completion *cqe) { struct nvme_iod *iod = ctx; struct bio *bio = iod->private; u16 status = le16_to_cpup(&cqe->status) >> 1; + if (unlikely(status)) { + if (!(status & NVME_SC_DNR || + bio->bi_rw & REQ_FAILFAST_MASK) && + (jiffies - iod->start_time) < IOD_TIMEOUT) { + if (!waitqueue_active(&nvmeq->sq_full)) + add_wait_queue(&nvmeq->sq_full, + &nvmeq->sq_cong_wait); + list_add_tail(&iod->node, &nvmeq->iod_bio); + wake_up(&nvmeq->sq_full); + return; + } + } if (iod->nents) { - dma_unmap_sg(&dev->pci_dev->dev, iod->sg, iod->nents, + dma_unmap_sg(nvmeq->q_dmadev, iod->sg, iod->nents, bio_data_dir(bio) ? DMA_TO_DEVICE : DMA_FROM_DEVICE); nvme_end_io_acct(bio, iod->start_time); } - nvme_free_iod(dev, iod); + nvme_free_iod(nvmeq->dev, iod); if (status) bio_endio(bio, -EIO); else @@ -361,8 +440,8 @@ static void bio_completion(struct nvme_dev *dev, void *ctx, } /* length is in bytes. gfp flags indicates whether we may sleep. */ -int nvme_setup_prps(struct nvme_dev *dev, struct nvme_common_command *cmd, - struct nvme_iod *iod, int total_len, gfp_t gfp) +int nvme_setup_prps(struct nvme_dev *dev, struct nvme_iod *iod, int total_len, + gfp_t gfp) { struct dma_pool *pool; int length = total_len; @@ -375,7 +454,6 @@ int nvme_setup_prps(struct nvme_dev *dev, struct nvme_common_command *cmd, dma_addr_t prp_dma; int nprps, i; - cmd->prp1 = cpu_to_le64(dma_addr); length -= (PAGE_SIZE - offset); if (length <= 0) return total_len; @@ -390,7 +468,7 @@ int nvme_setup_prps(struct nvme_dev *dev, struct nvme_common_command *cmd, } if (length <= PAGE_SIZE) { - cmd->prp2 = cpu_to_le64(dma_addr); + iod->first_dma = dma_addr; return total_len; } @@ -405,13 +483,12 @@ int nvme_setup_prps(struct nvme_dev *dev, struct nvme_common_command *cmd, prp_list = dma_pool_alloc(pool, gfp, &prp_dma); if (!prp_list) { - cmd->prp2 = cpu_to_le64(dma_addr); + iod->first_dma = dma_addr; iod->npages = -1; return (total_len - length) + PAGE_SIZE; } list[0] = prp_list; iod->first_dma = prp_dma; - cmd->prp2 = cpu_to_le64(prp_dma); i = 0; for (;;) { if (i == PAGE_SIZE / 8) { @@ -450,10 +527,11 @@ static int nvme_split_and_submit(struct bio *bio, struct nvme_queue *nvmeq, bio_chain(split, bio); - if (bio_list_empty(&nvmeq->sq_cong)) + if (!waitqueue_active(&nvmeq->sq_full)) add_wait_queue(&nvmeq->sq_full, &nvmeq->sq_cong_wait); bio_list_add(&nvmeq->sq_cong, split); bio_list_add(&nvmeq->sq_cong, bio); + wake_up(&nvmeq->sq_full); return 0; } @@ -506,25 +584,13 @@ static int nvme_map_bio(struct nvme_queue *nvmeq, struct nvme_iod *iod, return length; } -/* - * We reuse the small pool to allocate the 16-byte range here as it is not - * worth having a special pool for these or additional cases to handle freeing - * the iod. - */ static int nvme_submit_discard(struct nvme_queue *nvmeq, struct nvme_ns *ns, struct bio *bio, struct nvme_iod *iod, int cmdid) { - struct nvme_dsm_range *range; + struct nvme_dsm_range *range = + (struct nvme_dsm_range *)iod_list(iod)[0]; struct nvme_command *cmnd = &nvmeq->sq_cmds[nvmeq->sq_tail]; - range = dma_pool_alloc(nvmeq->dev->prp_small_pool, GFP_ATOMIC, - &iod->first_dma); - if (!range) - return -ENOMEM; - - iod_list(iod)[0] = (__le64 *)range; - iod->npages = 0; - range->cattr = cpu_to_le32(0); range->nlb = cpu_to_le32(bio->bi_iter.bi_size >> ns->lba_shift); range->slba = cpu_to_le64(nvme_block_nr(ns, bio->bi_iter.bi_sector)); @@ -571,44 +637,22 @@ int nvme_submit_flush_data(struct nvme_queue *nvmeq, struct nvme_ns *ns) return nvme_submit_flush(nvmeq, ns, cmdid); } -/* - * Called with local interrupts disabled and the q_lock held. May not sleep. - */ -static int nvme_submit_bio_queue(struct nvme_queue *nvmeq, struct nvme_ns *ns, - struct bio *bio) +static int nvme_submit_iod(struct nvme_queue *nvmeq, struct nvme_iod *iod) { + struct bio *bio = iod->private; + struct nvme_ns *ns = bio->bi_bdev->bd_disk->private_data; struct nvme_command *cmnd; - struct nvme_iod *iod; - enum dma_data_direction dma_dir; - int cmdid, length, result; + int cmdid; u16 control; u32 dsmgmt; - int psegs = bio_phys_segments(ns->queue, bio); - - if ((bio->bi_rw & REQ_FLUSH) && psegs) { - result = nvme_submit_flush_data(nvmeq, ns); - if (result) - return result; - } - - result = -ENOMEM; - iod = nvme_alloc_iod(psegs, bio->bi_iter.bi_size, GFP_ATOMIC); - if (!iod) - goto nomem; - iod->private = bio; - result = -EBUSY; cmdid = alloc_cmdid(nvmeq, iod, bio_completion, NVME_IO_TIMEOUT); if (unlikely(cmdid < 0)) - goto free_iod; + return cmdid; - if (bio->bi_rw & REQ_DISCARD) { - result = nvme_submit_discard(nvmeq, ns, bio, iod, cmdid); - if (result) - goto free_cmdid; - return result; - } - if ((bio->bi_rw & REQ_FLUSH) && !psegs) + if (bio->bi_rw & REQ_DISCARD) + return nvme_submit_discard(nvmeq, ns, bio, iod, cmdid); + if ((bio->bi_rw & REQ_FLUSH) && !iod->nents) return nvme_submit_flush(nvmeq, ns, cmdid); control = 0; @@ -622,42 +666,85 @@ static int nvme_submit_bio_queue(struct nvme_queue *nvmeq, struct nvme_ns *ns, dsmgmt |= NVME_RW_DSM_FREQ_PREFETCH; cmnd = &nvmeq->sq_cmds[nvmeq->sq_tail]; - memset(cmnd, 0, sizeof(*cmnd)); - if (bio_data_dir(bio)) { - cmnd->rw.opcode = nvme_cmd_write; - dma_dir = DMA_TO_DEVICE; - } else { - cmnd->rw.opcode = nvme_cmd_read; - dma_dir = DMA_FROM_DEVICE; - } - - result = nvme_map_bio(nvmeq, iod, bio, dma_dir, psegs); - if (result <= 0) - goto free_cmdid; - length = result; + cmnd->rw.opcode = bio_data_dir(bio) ? nvme_cmd_write : nvme_cmd_read; cmnd->rw.command_id = cmdid; cmnd->rw.nsid = cpu_to_le32(ns->ns_id); - length = nvme_setup_prps(nvmeq->dev, &cmnd->common, iod, length, - GFP_ATOMIC); + cmnd->rw.prp1 = cpu_to_le64(sg_dma_address(iod->sg)); + cmnd->rw.prp2 = cpu_to_le64(iod->first_dma); cmnd->rw.slba = cpu_to_le64(nvme_block_nr(ns, bio->bi_iter.bi_sector)); - cmnd->rw.length = cpu_to_le16((length >> ns->lba_shift) - 1); + cmnd->rw.length = + cpu_to_le16((bio->bi_iter.bi_size >> ns->lba_shift) - 1); cmnd->rw.control = cpu_to_le16(control); cmnd->rw.dsmgmt = cpu_to_le32(dsmgmt); - nvme_start_io_acct(bio); if (++nvmeq->sq_tail == nvmeq->q_depth) nvmeq->sq_tail = 0; writel(nvmeq->sq_tail, nvmeq->q_db); return 0; +} + +/* + * Called with local interrupts disabled and the q_lock held. May not sleep. + */ +static int nvme_submit_bio_queue(struct nvme_queue *nvmeq, struct nvme_ns *ns, + struct bio *bio) +{ + struct nvme_iod *iod; + int psegs = bio_phys_segments(ns->queue, bio); + int result; + + if ((bio->bi_rw & REQ_FLUSH) && psegs) { + result = nvme_submit_flush_data(nvmeq, ns); + if (result) + return result; + } + + iod = nvme_alloc_iod(psegs, bio->bi_iter.bi_size, GFP_ATOMIC); + if (!iod) + return -ENOMEM; + + iod->private = bio; + if (bio->bi_rw & REQ_DISCARD) { + void *range; + /* + * We reuse the small pool to allocate the 16-byte range here + * as it is not worth having a special pool for these or + * additional cases to handle freeing the iod. + */ + range = dma_pool_alloc(nvmeq->dev->prp_small_pool, + GFP_ATOMIC, + &iod->first_dma); + if (!range) { + result = -ENOMEM; + goto free_iod; + } + iod_list(iod)[0] = (__le64 *)range; + iod->npages = 0; + } else if (psegs) { + result = nvme_map_bio(nvmeq, iod, bio, + bio_data_dir(bio) ? DMA_TO_DEVICE : DMA_FROM_DEVICE, + psegs); + if (result <= 0) + goto free_iod; + if (nvme_setup_prps(nvmeq->dev, iod, result, GFP_ATOMIC) != + result) { + result = -ENOMEM; + goto free_iod; + } + nvme_start_io_acct(bio); + } + if (unlikely(nvme_submit_iod(nvmeq, iod))) { + if (!waitqueue_active(&nvmeq->sq_full)) + add_wait_queue(&nvmeq->sq_full, &nvmeq->sq_cong_wait); + list_add_tail(&iod->node, &nvmeq->iod_bio); + } + return 0; - free_cmdid: - free_cmdid(nvmeq, cmdid, NULL); free_iod: nvme_free_iod(nvmeq->dev, iod); - nomem: return result; } @@ -681,7 +768,7 @@ static int nvme_process_cq(struct nvme_queue *nvmeq) } ctx = free_cmdid(nvmeq, cqe.command_id, &fn); - fn(nvmeq->dev, ctx, &cqe); + fn(nvmeq, ctx, &cqe); } /* If the controller ignores the cq head doorbell and continuously @@ -693,7 +780,7 @@ static int nvme_process_cq(struct nvme_queue *nvmeq) if (head == nvmeq->cq_head && phase == nvmeq->cq_phase) return 0; - writel(head, nvmeq->q_db + (1 << nvmeq->dev->db_stride)); + writel(head, nvmeq->q_db + nvmeq->dev->db_stride); nvmeq->cq_head = head; nvmeq->cq_phase = phase; @@ -717,7 +804,7 @@ static void nvme_make_request(struct request_queue *q, struct bio *bio) if (!nvmeq->q_suspended && bio_list_empty(&nvmeq->sq_cong)) result = nvme_submit_bio_queue(nvmeq, ns, bio); if (unlikely(result)) { - if (bio_list_empty(&nvmeq->sq_cong)) + if (!waitqueue_active(&nvmeq->sq_full)) add_wait_queue(&nvmeq->sq_full, &nvmeq->sq_cong_wait); bio_list_add(&nvmeq->sq_cong, bio); } @@ -761,7 +848,7 @@ struct sync_cmd_info { int status; }; -static void sync_completion(struct nvme_dev *dev, void *ctx, +static void sync_completion(struct nvme_queue *nvmeq, void *ctx, struct nvme_completion *cqe) { struct sync_cmd_info *cmdinfo = ctx; @@ -774,27 +861,46 @@ static void sync_completion(struct nvme_dev *dev, void *ctx, * Returns 0 on success. If the result is negative, it's a Linux error code; * if the result is positive, it's an NVM Express status code */ -int nvme_submit_sync_cmd(struct nvme_queue *nvmeq, struct nvme_command *cmd, +static int nvme_submit_sync_cmd(struct nvme_dev *dev, int q_idx, + struct nvme_command *cmd, u32 *result, unsigned timeout) { - int cmdid; + int cmdid, ret; struct sync_cmd_info cmdinfo; + struct nvme_queue *nvmeq; + + nvmeq = lock_nvmeq(dev, q_idx); + if (!nvmeq) { + unlock_nvmeq(nvmeq); + return -ENODEV; + } cmdinfo.task = current; cmdinfo.status = -EINTR; - cmdid = alloc_cmdid_killable(nvmeq, &cmdinfo, sync_completion, - timeout); - if (cmdid < 0) + cmdid = alloc_cmdid(nvmeq, &cmdinfo, sync_completion, timeout); + if (cmdid < 0) { + unlock_nvmeq(nvmeq); return cmdid; + } cmd->common.command_id = cmdid; set_current_state(TASK_KILLABLE); - nvme_submit_cmd(nvmeq, cmd); + ret = nvme_submit_cmd(nvmeq, cmd); + if (ret) { + free_cmdid(nvmeq, cmdid, NULL); + unlock_nvmeq(nvmeq); + set_current_state(TASK_RUNNING); + return ret; + } + unlock_nvmeq(nvmeq); schedule_timeout(timeout); if (cmdinfo.status == -EINTR) { - nvme_abort_command(nvmeq, cmdid); + nvmeq = lock_nvmeq(dev, q_idx); + if (nvmeq) + nvme_abort_command(nvmeq, cmdid); + unlock_nvmeq(nvmeq); return -EINTR; } @@ -804,10 +910,38 @@ int nvme_submit_sync_cmd(struct nvme_queue *nvmeq, struct nvme_command *cmd, return cmdinfo.status; } +static int nvme_submit_async_cmd(struct nvme_queue *nvmeq, + struct nvme_command *cmd, + struct async_cmd_info *cmdinfo, unsigned timeout) +{ + int cmdid; + + cmdid = alloc_cmdid_killable(nvmeq, cmdinfo, async_completion, timeout); + if (cmdid < 0) + return cmdid; + cmdinfo->status = -EINTR; + cmd->common.command_id = cmdid; + return nvme_submit_cmd(nvmeq, cmd); +} + int nvme_submit_admin_cmd(struct nvme_dev *dev, struct nvme_command *cmd, u32 *result) { - return nvme_submit_sync_cmd(dev->queues[0], cmd, result, ADMIN_TIMEOUT); + return nvme_submit_sync_cmd(dev, 0, cmd, result, ADMIN_TIMEOUT); +} + +int nvme_submit_io_cmd(struct nvme_dev *dev, struct nvme_command *cmd, + u32 *result) +{ + return nvme_submit_sync_cmd(dev, smp_processor_id() + 1, cmd, result, + NVME_IO_TIMEOUT); +} + +static int nvme_submit_admin_cmd_async(struct nvme_dev *dev, + struct nvme_command *cmd, struct async_cmd_info *cmdinfo) +{ + return nvme_submit_async_cmd(raw_nvmeq(dev, 0), cmd, cmdinfo, + ADMIN_TIMEOUT); } static int adapter_delete_queue(struct nvme_dev *dev, u8 opcode, u16 id) @@ -920,6 +1054,58 @@ int nvme_set_features(struct nvme_dev *dev, unsigned fid, unsigned dword11, } /** + * nvme_abort_cmd - Attempt aborting a command + * @cmdid: Command id of a timed out IO + * @queue: The queue with timed out IO + * + * Schedule controller reset if the command was already aborted once before and + * still hasn't been returned to the driver, or if this is the admin queue. + */ +static void nvme_abort_cmd(int cmdid, struct nvme_queue *nvmeq) +{ + int a_cmdid; + struct nvme_command cmd; + struct nvme_dev *dev = nvmeq->dev; + struct nvme_cmd_info *info = nvme_cmd_info(nvmeq); + struct nvme_queue *adminq; + + if (!nvmeq->qid || info[cmdid].aborted) { + if (work_busy(&dev->reset_work)) + return; + list_del_init(&dev->node); + dev_warn(&dev->pci_dev->dev, + "I/O %d QID %d timeout, reset controller\n", cmdid, + nvmeq->qid); + dev->reset_workfn = nvme_reset_failed_dev; + queue_work(nvme_workq, &dev->reset_work); + return; + } + + if (!dev->abort_limit) + return; + + adminq = rcu_dereference(dev->queues[0]); + a_cmdid = alloc_cmdid(adminq, CMD_CTX_ABORT, special_completion, + ADMIN_TIMEOUT); + if (a_cmdid < 0) + return; + + memset(&cmd, 0, sizeof(cmd)); + cmd.abort.opcode = nvme_admin_abort_cmd; + cmd.abort.cid = cmdid; + cmd.abort.sqid = cpu_to_le16(nvmeq->qid); + cmd.abort.command_id = a_cmdid; + + --dev->abort_limit; + info[cmdid].aborted = 1; + info[cmdid].timeout = jiffies + ADMIN_TIMEOUT; + + dev_warn(nvmeq->q_dmadev, "Aborting I/O %d QID %d\n", cmdid, + nvmeq->qid); + nvme_submit_cmd(adminq, &cmd); +} + +/** * nvme_cancel_ios - Cancel outstanding I/Os * @queue: The queue to cancel I/Os on * @timeout: True to only cancel I/Os which have timed out @@ -942,67 +1128,111 @@ static void nvme_cancel_ios(struct nvme_queue *nvmeq, bool timeout) continue; if (info[cmdid].ctx == CMD_CTX_CANCELLED) continue; - dev_warn(nvmeq->q_dmadev, "Cancelling I/O %d\n", cmdid); + if (timeout && nvmeq->dev->initialized) { + nvme_abort_cmd(cmdid, nvmeq); + continue; + } + dev_warn(nvmeq->q_dmadev, "Cancelling I/O %d QID %d\n", cmdid, + nvmeq->qid); ctx = cancel_cmdid(nvmeq, cmdid, &fn); - fn(nvmeq->dev, ctx, &cqe); + fn(nvmeq, ctx, &cqe); } } -static void nvme_free_queue(struct nvme_queue *nvmeq) +static void nvme_free_queue(struct rcu_head *r) { + struct nvme_queue *nvmeq = container_of(r, struct nvme_queue, r_head); + spin_lock_irq(&nvmeq->q_lock); while (bio_list_peek(&nvmeq->sq_cong)) { struct bio *bio = bio_list_pop(&nvmeq->sq_cong); bio_endio(bio, -EIO); } + while (!list_empty(&nvmeq->iod_bio)) { + static struct nvme_completion cqe = { + .status = cpu_to_le16( + (NVME_SC_ABORT_REQ | NVME_SC_DNR) << 1), + }; + struct nvme_iod *iod = list_first_entry(&nvmeq->iod_bio, + struct nvme_iod, + node); + list_del(&iod->node); + bio_completion(nvmeq, iod, &cqe); + } spin_unlock_irq(&nvmeq->q_lock); dma_free_coherent(nvmeq->q_dmadev, CQ_SIZE(nvmeq->q_depth), (void *)nvmeq->cqes, nvmeq->cq_dma_addr); dma_free_coherent(nvmeq->q_dmadev, SQ_SIZE(nvmeq->q_depth), nvmeq->sq_cmds, nvmeq->sq_dma_addr); + if (nvmeq->qid) + free_cpumask_var(nvmeq->cpu_mask); kfree(nvmeq); } -static void nvme_free_queues(struct nvme_dev *dev) +static void nvme_free_queues(struct nvme_dev *dev, int lowest) { int i; - for (i = dev->queue_count - 1; i >= 0; i--) { - nvme_free_queue(dev->queues[i]); + for (i = dev->queue_count - 1; i >= lowest; i--) { + struct nvme_queue *nvmeq = raw_nvmeq(dev, i); + rcu_assign_pointer(dev->queues[i], NULL); + call_rcu(&nvmeq->r_head, nvme_free_queue); dev->queue_count--; - dev->queues[i] = NULL; } } -static void nvme_disable_queue(struct nvme_dev *dev, int qid) +/** + * nvme_suspend_queue - put queue into suspended state + * @nvmeq - queue to suspend + * + * Returns 1 if already suspended, 0 otherwise. + */ +static int nvme_suspend_queue(struct nvme_queue *nvmeq) { - struct nvme_queue *nvmeq = dev->queues[qid]; - int vector = dev->entry[nvmeq->cq_vector].vector; + int vector = nvmeq->dev->entry[nvmeq->cq_vector].vector; spin_lock_irq(&nvmeq->q_lock); if (nvmeq->q_suspended) { spin_unlock_irq(&nvmeq->q_lock); - return; + return 1; } nvmeq->q_suspended = 1; + nvmeq->dev->online_queues--; spin_unlock_irq(&nvmeq->q_lock); irq_set_affinity_hint(vector, NULL); free_irq(vector, nvmeq); - /* Don't tell the adapter to delete the admin queue */ - if (qid) { - adapter_delete_sq(dev, qid); - adapter_delete_cq(dev, qid); - } + return 0; +} +static void nvme_clear_queue(struct nvme_queue *nvmeq) +{ spin_lock_irq(&nvmeq->q_lock); nvme_process_cq(nvmeq); nvme_cancel_ios(nvmeq, false); spin_unlock_irq(&nvmeq->q_lock); } +static void nvme_disable_queue(struct nvme_dev *dev, int qid) +{ + struct nvme_queue *nvmeq = raw_nvmeq(dev, qid); + + if (!nvmeq) + return; + if (nvme_suspend_queue(nvmeq)) + return; + + /* Don't tell the adapter to delete the admin queue. + * Don't tell a removed adapter to delete IO queues. */ + if (qid && readl(&dev->bar->csts) != -1) { + adapter_delete_sq(dev, qid); + adapter_delete_cq(dev, qid); + } + nvme_clear_queue(nvmeq); +} + static struct nvme_queue *nvme_alloc_queue(struct nvme_dev *dev, int qid, int depth, int vector) { @@ -1023,22 +1253,33 @@ static struct nvme_queue *nvme_alloc_queue(struct nvme_dev *dev, int qid, if (!nvmeq->sq_cmds) goto free_cqdma; + if (qid && !zalloc_cpumask_var(&nvmeq->cpu_mask, GFP_KERNEL)) + goto free_sqdma; + nvmeq->q_dmadev = dmadev; nvmeq->dev = dev; + snprintf(nvmeq->irqname, sizeof(nvmeq->irqname), "nvme%dq%d", + dev->instance, qid); spin_lock_init(&nvmeq->q_lock); nvmeq->cq_head = 0; nvmeq->cq_phase = 1; init_waitqueue_head(&nvmeq->sq_full); init_waitqueue_entry(&nvmeq->sq_cong_wait, nvme_thread); bio_list_init(&nvmeq->sq_cong); - nvmeq->q_db = &dev->dbs[qid << (dev->db_stride + 1)]; + INIT_LIST_HEAD(&nvmeq->iod_bio); + nvmeq->q_db = &dev->dbs[qid * 2 * dev->db_stride]; nvmeq->q_depth = depth; nvmeq->cq_vector = vector; + nvmeq->qid = qid; nvmeq->q_suspended = 1; dev->queue_count++; + rcu_assign_pointer(dev->queues[qid], nvmeq); return nvmeq; + free_sqdma: + dma_free_coherent(dmadev, SQ_SIZE(depth), (void *)nvmeq->sq_cmds, + nvmeq->sq_dma_addr); free_cqdma: dma_free_coherent(dmadev, CQ_SIZE(depth), (void *)nvmeq->cqes, nvmeq->cq_dma_addr); @@ -1052,11 +1293,10 @@ static int queue_request_irq(struct nvme_dev *dev, struct nvme_queue *nvmeq, { if (use_threaded_interrupts) return request_threaded_irq(dev->entry[nvmeq->cq_vector].vector, - nvme_irq_check, nvme_irq, - IRQF_DISABLED | IRQF_SHARED, + nvme_irq_check, nvme_irq, IRQF_SHARED, name, nvmeq); return request_irq(dev->entry[nvmeq->cq_vector].vector, nvme_irq, - IRQF_DISABLED | IRQF_SHARED, name, nvmeq); + IRQF_SHARED, name, nvmeq); } static void nvme_init_queue(struct nvme_queue *nvmeq, u16 qid) @@ -1067,11 +1307,12 @@ static void nvme_init_queue(struct nvme_queue *nvmeq, u16 qid) nvmeq->sq_tail = 0; nvmeq->cq_head = 0; nvmeq->cq_phase = 1; - nvmeq->q_db = &dev->dbs[qid << (dev->db_stride + 1)]; + nvmeq->q_db = &dev->dbs[qid * 2 * dev->db_stride]; memset(nvmeq->cmdid_data, 0, extra); memset((void *)nvmeq->cqes, 0, CQ_SIZE(nvmeq->q_depth)); nvme_cancel_ios(nvmeq, false); nvmeq->q_suspended = 0; + dev->online_queues++; } static int nvme_create_queue(struct nvme_queue *nvmeq, int qid) @@ -1087,13 +1328,13 @@ static int nvme_create_queue(struct nvme_queue *nvmeq, int qid) if (result < 0) goto release_cq; - result = queue_request_irq(dev, nvmeq, "nvme"); + result = queue_request_irq(dev, nvmeq, nvmeq->irqname); if (result < 0) goto release_sq; - spin_lock(&nvmeq->q_lock); + spin_lock_irq(&nvmeq->q_lock); nvme_init_queue(nvmeq, qid); - spin_unlock(&nvmeq->q_lock); + spin_unlock_irq(&nvmeq->q_lock); return result; @@ -1180,12 +1421,11 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev) if (result < 0) return result; - nvmeq = dev->queues[0]; + nvmeq = raw_nvmeq(dev, 0); if (!nvmeq) { nvmeq = nvme_alloc_queue(dev, 0, 64, 0); if (!nvmeq) return -ENOMEM; - dev->queues[0] = nvmeq; } aqa = nvmeq->q_depth - 1; @@ -1205,13 +1445,13 @@ static int nvme_configure_admin_queue(struct nvme_dev *dev) if (result) return result; - result = queue_request_irq(dev, nvmeq, "nvme admin"); + result = queue_request_irq(dev, nvmeq, nvmeq->irqname); if (result) return result; - spin_lock(&nvmeq->q_lock); + spin_lock_irq(&nvmeq->q_lock); nvme_init_queue(nvmeq, 0); - spin_unlock(&nvmeq->q_lock); + spin_unlock_irq(&nvmeq->q_lock); return result; } @@ -1287,7 +1527,6 @@ void nvme_unmap_user_pages(struct nvme_dev *dev, int write, static int nvme_submit_io(struct nvme_ns *ns, struct nvme_user_io __user *uio) { struct nvme_dev *dev = ns->dev; - struct nvme_queue *nvmeq; struct nvme_user_io io; struct nvme_command c; unsigned length, meta_len; @@ -1361,22 +1600,14 @@ static int nvme_submit_io(struct nvme_ns *ns, struct nvme_user_io __user *uio) c.rw.metadata = cpu_to_le64(meta_dma_addr); } - length = nvme_setup_prps(dev, &c.common, iod, length, GFP_KERNEL); + length = nvme_setup_prps(dev, iod, length, GFP_KERNEL); + c.rw.prp1 = cpu_to_le64(sg_dma_address(iod->sg)); + c.rw.prp2 = cpu_to_le64(iod->first_dma); - nvmeq = get_nvmeq(dev); - /* - * Since nvme_submit_sync_cmd sleeps, we can't keep preemption - * disabled. We may be preempted at any point, and be rescheduled - * to a different CPU. That will cause cacheline bouncing, but no - * additional races since q_lock already protects against other CPUs. - */ - put_nvmeq(nvmeq); if (length != (io.nblocks + 1) << ns->lba_shift) status = -ENOMEM; - else if (!nvmeq || nvmeq->q_suspended) - status = -EBUSY; else - status = nvme_submit_sync_cmd(nvmeq, &c, NULL, NVME_IO_TIMEOUT); + status = nvme_submit_io_cmd(dev, &c, NULL); if (meta_len) { if (status == NVME_SC_SUCCESS && !(io.opcode & 1)) { @@ -1441,8 +1672,9 @@ static int nvme_user_admin_cmd(struct nvme_dev *dev, length); if (IS_ERR(iod)) return PTR_ERR(iod); - length = nvme_setup_prps(dev, &c.common, iod, length, - GFP_KERNEL); + length = nvme_setup_prps(dev, iod, length, GFP_KERNEL); + c.common.prp1 = cpu_to_le64(sg_dma_address(iod->sg)); + c.common.prp2 = cpu_to_le64(iod->first_dma); } timeout = cmd.timeout_ms ? msecs_to_jiffies(cmd.timeout_ms) : @@ -1450,8 +1682,7 @@ static int nvme_user_admin_cmd(struct nvme_dev *dev, if (length != cmd.data_len) status = -ENOMEM; else - status = nvme_submit_sync_cmd(dev->queues[0], &c, &cmd.result, - timeout); + status = nvme_submit_sync_cmd(dev, 0, &c, &cmd.result, timeout); if (cmd.data_len) { nvme_unmap_user_pages(dev, cmd.opcode & 1, iod); @@ -1487,23 +1718,86 @@ static int nvme_ioctl(struct block_device *bdev, fmode_t mode, unsigned int cmd, } } +#ifdef CONFIG_COMPAT +static int nvme_compat_ioctl(struct block_device *bdev, fmode_t mode, + unsigned int cmd, unsigned long arg) +{ + struct nvme_ns *ns = bdev->bd_disk->private_data; + + switch (cmd) { + case SG_IO: + return nvme_sg_io32(ns, arg); + } + return nvme_ioctl(bdev, mode, cmd, arg); +} +#else +#define nvme_compat_ioctl NULL +#endif + +static int nvme_open(struct block_device *bdev, fmode_t mode) +{ + struct nvme_ns *ns = bdev->bd_disk->private_data; + struct nvme_dev *dev = ns->dev; + + kref_get(&dev->kref); + return 0; +} + +static void nvme_free_dev(struct kref *kref); + +static void nvme_release(struct gendisk *disk, fmode_t mode) +{ + struct nvme_ns *ns = disk->private_data; + struct nvme_dev *dev = ns->dev; + + kref_put(&dev->kref, nvme_free_dev); +} + +static int nvme_getgeo(struct block_device *bd, struct hd_geometry *geo) +{ + /* some standard values */ + geo->heads = 1 << 6; + geo->sectors = 1 << 5; + geo->cylinders = get_capacity(bd->bd_disk) >> 11; + return 0; +} + static const struct block_device_operations nvme_fops = { .owner = THIS_MODULE, .ioctl = nvme_ioctl, - .compat_ioctl = nvme_ioctl, + .compat_ioctl = nvme_compat_ioctl, + .open = nvme_open, + .release = nvme_release, + .getgeo = nvme_getgeo, }; +static void nvme_resubmit_iods(struct nvme_queue *nvmeq) +{ + struct nvme_iod *iod, *next; + + list_for_each_entry_safe(iod, next, &nvmeq->iod_bio, node) { + if (unlikely(nvme_submit_iod(nvmeq, iod))) + break; + list_del(&iod->node); + if (bio_list_empty(&nvmeq->sq_cong) && + list_empty(&nvmeq->iod_bio)) + remove_wait_queue(&nvmeq->sq_full, + &nvmeq->sq_cong_wait); + } +} + static void nvme_resubmit_bios(struct nvme_queue *nvmeq) { while (bio_list_peek(&nvmeq->sq_cong)) { struct bio *bio = bio_list_pop(&nvmeq->sq_cong); struct nvme_ns *ns = bio->bi_bdev->bd_disk->private_data; - if (bio_list_empty(&nvmeq->sq_cong)) + if (bio_list_empty(&nvmeq->sq_cong) && + list_empty(&nvmeq->iod_bio)) remove_wait_queue(&nvmeq->sq_full, &nvmeq->sq_cong_wait); if (nvme_submit_bio_queue(nvmeq, ns, bio)) { - if (bio_list_empty(&nvmeq->sq_cong)) + if (!waitqueue_active(&nvmeq->sq_full)) add_wait_queue(&nvmeq->sq_full, &nvmeq->sq_cong_wait); bio_list_add_head(&nvmeq->sq_cong, bio); @@ -1514,15 +1808,28 @@ static void nvme_resubmit_bios(struct nvme_queue *nvmeq) static int nvme_kthread(void *data) { - struct nvme_dev *dev; + struct nvme_dev *dev, *next; while (!kthread_should_stop()) { set_current_state(TASK_INTERRUPTIBLE); spin_lock(&dev_list_lock); - list_for_each_entry(dev, &dev_list, node) { + list_for_each_entry_safe(dev, next, &dev_list, node) { int i; + if (readl(&dev->bar->csts) & NVME_CSTS_CFS && + dev->initialized) { + if (work_busy(&dev->reset_work)) + continue; + list_del_init(&dev->node); + dev_warn(&dev->pci_dev->dev, + "Failed status, reset controller\n"); + dev->reset_workfn = nvme_reset_failed_dev; + queue_work(nvme_workq, &dev->reset_work); + continue; + } + rcu_read_lock(); for (i = 0; i < dev->queue_count; i++) { - struct nvme_queue *nvmeq = dev->queues[i]; + struct nvme_queue *nvmeq = + rcu_dereference(dev->queues[i]); if (!nvmeq) continue; spin_lock_irq(&nvmeq->q_lock); @@ -1531,9 +1838,11 @@ static int nvme_kthread(void *data) nvme_process_cq(nvmeq); nvme_cancel_ios(nvmeq, true); nvme_resubmit_bios(nvmeq); + nvme_resubmit_iods(nvmeq); unlock: spin_unlock_irq(&nvmeq->q_lock); } + rcu_read_unlock(); } spin_unlock(&dev_list_lock); schedule_timeout(round_jiffies_relative(HZ)); @@ -1541,33 +1850,6 @@ static int nvme_kthread(void *data) return 0; } -static DEFINE_IDA(nvme_index_ida); - -static int nvme_get_ns_idx(void) -{ - int index, error; - - do { - if (!ida_pre_get(&nvme_index_ida, GFP_KERNEL)) - return -1; - - spin_lock(&dev_list_lock); - error = ida_get_new(&nvme_index_ida, &index); - spin_unlock(&dev_list_lock); - } while (error == -EAGAIN); - - if (error) - index = -1; - return index; -} - -static void nvme_put_ns_idx(int index) -{ - spin_lock(&dev_list_lock); - ida_remove(&nvme_index_ida, index); - spin_unlock(&dev_list_lock); -} - static void nvme_config_discard(struct nvme_ns *ns) { u32 logical_block_size = queue_logical_block_size(ns->queue); @@ -1601,7 +1883,7 @@ static struct nvme_ns *nvme_alloc_ns(struct nvme_dev *dev, unsigned nsid, ns->dev = dev; ns->queue->queuedata = ns; - disk = alloc_disk(NVME_MINORS); + disk = alloc_disk(0); if (!disk) goto out_free_queue; ns->ns_id = nsid; @@ -1614,12 +1896,12 @@ static struct nvme_ns *nvme_alloc_ns(struct nvme_dev *dev, unsigned nsid, blk_queue_max_hw_sectors(ns->queue, dev->max_hw_sectors); disk->major = nvme_major; - disk->minors = NVME_MINORS; - disk->first_minor = NVME_MINORS * nvme_get_ns_idx(); + disk->first_minor = 0; disk->fops = &nvme_fops; disk->private_data = ns; disk->queue = ns->queue; disk->driverfs_dev = &dev->pci_dev->dev; + disk->flags = GENHD_FL_EXT_DEVT; sprintf(disk->disk_name, "nvme%dn%d", dev->instance, nsid); set_capacity(disk, le64_to_cpup(&id->nsze) << (ns->lba_shift - 9)); @@ -1635,13 +1917,141 @@ static struct nvme_ns *nvme_alloc_ns(struct nvme_dev *dev, unsigned nsid, return NULL; } -static void nvme_ns_free(struct nvme_ns *ns) +static int nvme_find_closest_node(int node) { - int index = ns->disk->first_minor / NVME_MINORS; - put_disk(ns->disk); - nvme_put_ns_idx(index); - blk_cleanup_queue(ns->queue); - kfree(ns); + int n, val, min_val = INT_MAX, best_node = node; + + for_each_online_node(n) { + if (n == node) + continue; + val = node_distance(node, n); + if (val < min_val) { + min_val = val; + best_node = n; + } + } + return best_node; +} + +static void nvme_set_queue_cpus(cpumask_t *qmask, struct nvme_queue *nvmeq, + int count) +{ + int cpu; + for_each_cpu(cpu, qmask) { + if (cpumask_weight(nvmeq->cpu_mask) >= count) + break; + if (!cpumask_test_and_set_cpu(cpu, nvmeq->cpu_mask)) + *per_cpu_ptr(nvmeq->dev->io_queue, cpu) = nvmeq->qid; + } +} + +static void nvme_add_cpus(cpumask_t *mask, const cpumask_t *unassigned_cpus, + const cpumask_t *new_mask, struct nvme_queue *nvmeq, int cpus_per_queue) +{ + int next_cpu; + for_each_cpu(next_cpu, new_mask) { + cpumask_or(mask, mask, get_cpu_mask(next_cpu)); + cpumask_or(mask, mask, topology_thread_cpumask(next_cpu)); + cpumask_and(mask, mask, unassigned_cpus); + nvme_set_queue_cpus(mask, nvmeq, cpus_per_queue); + } +} + +static void nvme_create_io_queues(struct nvme_dev *dev) +{ + unsigned i, max; + + max = min(dev->max_qid, num_online_cpus()); + for (i = dev->queue_count; i <= max; i++) + if (!nvme_alloc_queue(dev, i, dev->q_depth, i - 1)) + break; + + max = min(dev->queue_count - 1, num_online_cpus()); + for (i = dev->online_queues; i <= max; i++) + if (nvme_create_queue(raw_nvmeq(dev, i), i)) + break; +} + +/* + * If there are fewer queues than online cpus, this will try to optimally + * assign a queue to multiple cpus by grouping cpus that are "close" together: + * thread siblings, core, socket, closest node, then whatever else is + * available. + */ +static void nvme_assign_io_queues(struct nvme_dev *dev) +{ + unsigned cpu, cpus_per_queue, queues, remainder, i; + cpumask_var_t unassigned_cpus; + + nvme_create_io_queues(dev); + + queues = min(dev->online_queues - 1, num_online_cpus()); + if (!queues) + return; + + cpus_per_queue = num_online_cpus() / queues; + remainder = queues - (num_online_cpus() - queues * cpus_per_queue); + + if (!alloc_cpumask_var(&unassigned_cpus, GFP_KERNEL)) + return; + + cpumask_copy(unassigned_cpus, cpu_online_mask); + cpu = cpumask_first(unassigned_cpus); + for (i = 1; i <= queues; i++) { + struct nvme_queue *nvmeq = lock_nvmeq(dev, i); + cpumask_t mask; + + cpumask_clear(nvmeq->cpu_mask); + if (!cpumask_weight(unassigned_cpus)) { + unlock_nvmeq(nvmeq); + break; + } + + mask = *get_cpu_mask(cpu); + nvme_set_queue_cpus(&mask, nvmeq, cpus_per_queue); + if (cpus_weight(mask) < cpus_per_queue) + nvme_add_cpus(&mask, unassigned_cpus, + topology_thread_cpumask(cpu), + nvmeq, cpus_per_queue); + if (cpus_weight(mask) < cpus_per_queue) + nvme_add_cpus(&mask, unassigned_cpus, + topology_core_cpumask(cpu), + nvmeq, cpus_per_queue); + if (cpus_weight(mask) < cpus_per_queue) + nvme_add_cpus(&mask, unassigned_cpus, + cpumask_of_node(cpu_to_node(cpu)), + nvmeq, cpus_per_queue); + if (cpus_weight(mask) < cpus_per_queue) + nvme_add_cpus(&mask, unassigned_cpus, + cpumask_of_node( + nvme_find_closest_node( + cpu_to_node(cpu))), + nvmeq, cpus_per_queue); + if (cpus_weight(mask) < cpus_per_queue) + nvme_add_cpus(&mask, unassigned_cpus, + unassigned_cpus, + nvmeq, cpus_per_queue); + + WARN(cpumask_weight(nvmeq->cpu_mask) != cpus_per_queue, + "nvme%d qid:%d mis-matched queue-to-cpu assignment\n", + dev->instance, i); + + irq_set_affinity_hint(dev->entry[nvmeq->cq_vector].vector, + nvmeq->cpu_mask); + cpumask_andnot(unassigned_cpus, unassigned_cpus, + nvmeq->cpu_mask); + cpu = cpumask_next(cpu, unassigned_cpus); + if (remainder && !--remainder) + cpus_per_queue++; + unlock_nvmeq(nvmeq); + } + WARN(cpumask_weight(unassigned_cpus), "nvme%d unassigned online cpus\n", + dev->instance); + i = 0; + cpumask_andnot(unassigned_cpus, cpu_possible_mask, cpu_online_mask); + for_each_cpu(cpu, unassigned_cpus) + *per_cpu_ptr(dev->io_queue, cpu) = (i++ % queues) + 1; + free_cpumask_var(unassigned_cpus); } static int set_queue_count(struct nvme_dev *dev, int count) @@ -1659,15 +2069,29 @@ static int set_queue_count(struct nvme_dev *dev, int count) static size_t db_bar_size(struct nvme_dev *dev, unsigned nr_io_queues) { - return 4096 + ((nr_io_queues + 1) << (dev->db_stride + 3)); + return 4096 + ((nr_io_queues + 1) * 8 * dev->db_stride); +} + +static int nvme_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + struct nvme_dev *dev = container_of(self, struct nvme_dev, nb); + switch (action) { + case CPU_ONLINE: + case CPU_DEAD: + nvme_assign_io_queues(dev); + break; + } + return NOTIFY_OK; } static int nvme_setup_io_queues(struct nvme_dev *dev) { + struct nvme_queue *adminq = raw_nvmeq(dev, 0); struct pci_dev *pdev = dev->pci_dev; - int result, cpu, i, vecs, nr_io_queues, size, q_depth; + int result, i, vecs, nr_io_queues, size; - nr_io_queues = num_online_cpus(); + nr_io_queues = num_possible_cpus(); result = set_queue_count(dev, nr_io_queues); if (result < 0) return result; @@ -1686,37 +2110,22 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) size = db_bar_size(dev, nr_io_queues); } while (1); dev->dbs = ((void __iomem *)dev->bar) + 4096; - dev->queues[0]->q_db = dev->dbs; + adminq->q_db = dev->dbs; } /* Deregister the admin queue's interrupt */ - free_irq(dev->entry[0].vector, dev->queues[0]); + free_irq(dev->entry[0].vector, adminq); - vecs = nr_io_queues; - for (i = 0; i < vecs; i++) + for (i = 0; i < nr_io_queues; i++) dev->entry[i].entry = i; - for (;;) { - result = pci_enable_msix(pdev, dev->entry, vecs); - if (result <= 0) - break; - vecs = result; - } - - if (result < 0) { - vecs = nr_io_queues; - if (vecs > 32) - vecs = 32; - for (;;) { - result = pci_enable_msi_block(pdev, vecs); - if (result == 0) { - for (i = 0; i < vecs; i++) - dev->entry[i].vector = i + pdev->irq; - break; - } else if (result < 0) { - vecs = 1; - break; - } - vecs = result; + vecs = pci_enable_msix_range(pdev, dev->entry, 1, nr_io_queues); + if (vecs < 0) { + vecs = pci_enable_msi_range(pdev, 1, min(nr_io_queues, 32)); + if (vecs < 0) { + vecs = 1; + } else { + for (i = 0; i < vecs; i++) + dev->entry[i].vector = i + pdev->irq; } } @@ -1727,62 +2136,27 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) * number of interrupts. */ nr_io_queues = vecs; + dev->max_qid = nr_io_queues; - result = queue_request_irq(dev, dev->queues[0], "nvme admin"); + result = queue_request_irq(dev, adminq, adminq->irqname); if (result) { - dev->queues[0]->q_suspended = 1; + adminq->q_suspended = 1; goto free_queues; } /* Free previously allocated queues that are no longer usable */ - spin_lock(&dev_list_lock); - for (i = dev->queue_count - 1; i > nr_io_queues; i--) { - struct nvme_queue *nvmeq = dev->queues[i]; - - spin_lock(&nvmeq->q_lock); - nvme_cancel_ios(nvmeq, false); - spin_unlock(&nvmeq->q_lock); - - nvme_free_queue(nvmeq); - dev->queue_count--; - dev->queues[i] = NULL; - } - spin_unlock(&dev_list_lock); - - cpu = cpumask_first(cpu_online_mask); - for (i = 0; i < nr_io_queues; i++) { - irq_set_affinity_hint(dev->entry[i].vector, get_cpu_mask(cpu)); - cpu = cpumask_next(cpu, cpu_online_mask); - } + nvme_free_queues(dev, nr_io_queues + 1); + nvme_assign_io_queues(dev); - q_depth = min_t(int, NVME_CAP_MQES(readq(&dev->bar->cap)) + 1, - NVME_Q_DEPTH); - for (i = dev->queue_count - 1; i < nr_io_queues; i++) { - dev->queues[i + 1] = nvme_alloc_queue(dev, i + 1, q_depth, i); - if (!dev->queues[i + 1]) { - result = -ENOMEM; - goto free_queues; - } - } - - for (; i < num_possible_cpus(); i++) { - int target = i % rounddown_pow_of_two(dev->queue_count - 1); - dev->queues[i + 1] = dev->queues[target + 1]; - } - - for (i = 1; i < dev->queue_count; i++) { - result = nvme_create_queue(dev->queues[i], i); - if (result) { - for (--i; i > 0; i--) - nvme_disable_queue(dev, i); - goto free_queues; - } - } + dev->nb.notifier_call = &nvme_cpu_notify; + result = register_hotcpu_notifier(&dev->nb); + if (result) + goto free_queues; return 0; free_queues: - nvme_free_queues(dev); + nvme_free_queues(dev, 1); return result; } @@ -1794,6 +2168,7 @@ static int nvme_setup_io_queues(struct nvme_dev *dev) */ static int nvme_dev_add(struct nvme_dev *dev) { + struct pci_dev *pdev = dev->pci_dev; int res; unsigned nn, i; struct nvme_ns *ns; @@ -1803,8 +2178,7 @@ static int nvme_dev_add(struct nvme_dev *dev) dma_addr_t dma_addr; int shift = NVME_CAP_MPSMIN(readq(&dev->bar->cap)) + 12; - mem = dma_alloc_coherent(&dev->pci_dev->dev, 8192, &dma_addr, - GFP_KERNEL); + mem = dma_alloc_coherent(&pdev->dev, 8192, &dma_addr, GFP_KERNEL); if (!mem) return -ENOMEM; @@ -1817,13 +2191,14 @@ static int nvme_dev_add(struct nvme_dev *dev) ctrl = mem; nn = le32_to_cpup(&ctrl->nn); dev->oncs = le16_to_cpup(&ctrl->oncs); + dev->abort_limit = ctrl->acl + 1; memcpy(dev->serial, ctrl->sn, sizeof(ctrl->sn)); memcpy(dev->model, ctrl->mn, sizeof(ctrl->mn)); memcpy(dev->firmware_rev, ctrl->fr, sizeof(ctrl->fr)); if (ctrl->mdts) dev->max_hw_sectors = 1 << (ctrl->mdts + shift - 9); - if ((dev->pci_dev->vendor == PCI_VENDOR_ID_INTEL) && - (dev->pci_dev->device == 0x0953) && ctrl->vs[3]) + if ((pdev->vendor == PCI_VENDOR_ID_INTEL) && + (pdev->device == 0x0953) && ctrl->vs[3]) dev->stripe_size = 1 << (ctrl->vs[3] + shift); id_ns = mem; @@ -1855,6 +2230,7 @@ static int nvme_dev_add(struct nvme_dev *dev) static int nvme_dev_map(struct nvme_dev *dev) { + u64 cap; int bars, result = -ENOMEM; struct pci_dev *pdev = dev->pci_dev; @@ -1871,16 +2247,23 @@ static int nvme_dev_map(struct nvme_dev *dev) dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32))) goto disable; - pci_set_drvdata(pdev, dev); dev->bar = ioremap(pci_resource_start(pdev, 0), 8192); if (!dev->bar) goto disable; - - dev->db_stride = NVME_CAP_STRIDE(readq(&dev->bar->cap)); + if (readl(&dev->bar->csts) == -1) { + result = -ENODEV; + goto unmap; + } + cap = readq(&dev->bar->cap); + dev->q_depth = min_t(int, NVME_CAP_MQES(cap) + 1, NVME_Q_DEPTH); + dev->db_stride = 1 << NVME_CAP_STRIDE(cap); dev->dbs = ((void __iomem *)dev->bar) + 4096; return 0; + unmap: + iounmap(dev->bar); + dev->bar = NULL; disable: pci_release_regions(pdev); disable_pci: @@ -1898,37 +2281,202 @@ static void nvme_dev_unmap(struct nvme_dev *dev) if (dev->bar) { iounmap(dev->bar); dev->bar = NULL; + pci_release_regions(dev->pci_dev); } - pci_release_regions(dev->pci_dev); if (pci_is_enabled(dev->pci_dev)) pci_disable_device(dev->pci_dev); } -static void nvme_dev_shutdown(struct nvme_dev *dev) +struct nvme_delq_ctx { + struct task_struct *waiter; + struct kthread_worker *worker; + atomic_t refcount; +}; + +static void nvme_wait_dq(struct nvme_delq_ctx *dq, struct nvme_dev *dev) +{ + dq->waiter = current; + mb(); + + for (;;) { + set_current_state(TASK_KILLABLE); + if (!atomic_read(&dq->refcount)) + break; + if (!schedule_timeout(ADMIN_TIMEOUT) || + fatal_signal_pending(current)) { + set_current_state(TASK_RUNNING); + + nvme_disable_ctrl(dev, readq(&dev->bar->cap)); + nvme_disable_queue(dev, 0); + + send_sig(SIGKILL, dq->worker->task, 1); + flush_kthread_worker(dq->worker); + return; + } + } + set_current_state(TASK_RUNNING); +} + +static void nvme_put_dq(struct nvme_delq_ctx *dq) +{ + atomic_dec(&dq->refcount); + if (dq->waiter) + wake_up_process(dq->waiter); +} + +static struct nvme_delq_ctx *nvme_get_dq(struct nvme_delq_ctx *dq) +{ + atomic_inc(&dq->refcount); + return dq; +} + +static void nvme_del_queue_end(struct nvme_queue *nvmeq) +{ + struct nvme_delq_ctx *dq = nvmeq->cmdinfo.ctx; + + nvme_clear_queue(nvmeq); + nvme_put_dq(dq); +} + +static int adapter_async_del_queue(struct nvme_queue *nvmeq, u8 opcode, + kthread_work_func_t fn) +{ + struct nvme_command c; + + memset(&c, 0, sizeof(c)); + c.delete_queue.opcode = opcode; + c.delete_queue.qid = cpu_to_le16(nvmeq->qid); + + init_kthread_work(&nvmeq->cmdinfo.work, fn); + return nvme_submit_admin_cmd_async(nvmeq->dev, &c, &nvmeq->cmdinfo); +} + +static void nvme_del_cq_work_handler(struct kthread_work *work) +{ + struct nvme_queue *nvmeq = container_of(work, struct nvme_queue, + cmdinfo.work); + nvme_del_queue_end(nvmeq); +} + +static int nvme_delete_cq(struct nvme_queue *nvmeq) +{ + return adapter_async_del_queue(nvmeq, nvme_admin_delete_cq, + nvme_del_cq_work_handler); +} + +static void nvme_del_sq_work_handler(struct kthread_work *work) +{ + struct nvme_queue *nvmeq = container_of(work, struct nvme_queue, + cmdinfo.work); + int status = nvmeq->cmdinfo.status; + + if (!status) + status = nvme_delete_cq(nvmeq); + if (status) + nvme_del_queue_end(nvmeq); +} + +static int nvme_delete_sq(struct nvme_queue *nvmeq) +{ + return adapter_async_del_queue(nvmeq, nvme_admin_delete_sq, + nvme_del_sq_work_handler); +} + +static void nvme_del_queue_start(struct kthread_work *work) +{ + struct nvme_queue *nvmeq = container_of(work, struct nvme_queue, + cmdinfo.work); + allow_signal(SIGKILL); + if (nvme_delete_sq(nvmeq)) + nvme_del_queue_end(nvmeq); +} + +static void nvme_disable_io_queues(struct nvme_dev *dev) { int i; + DEFINE_KTHREAD_WORKER_ONSTACK(worker); + struct nvme_delq_ctx dq; + struct task_struct *kworker_task = kthread_run(kthread_worker_fn, + &worker, "nvme%d", dev->instance); + + if (IS_ERR(kworker_task)) { + dev_err(&dev->pci_dev->dev, + "Failed to create queue del task\n"); + for (i = dev->queue_count - 1; i > 0; i--) + nvme_disable_queue(dev, i); + return; + } + + dq.waiter = NULL; + atomic_set(&dq.refcount, 0); + dq.worker = &worker; + for (i = dev->queue_count - 1; i > 0; i--) { + struct nvme_queue *nvmeq = raw_nvmeq(dev, i); + + if (nvme_suspend_queue(nvmeq)) + continue; + nvmeq->cmdinfo.ctx = nvme_get_dq(&dq); + nvmeq->cmdinfo.worker = dq.worker; + init_kthread_work(&nvmeq->cmdinfo.work, nvme_del_queue_start); + queue_kthread_work(dq.worker, &nvmeq->cmdinfo.work); + } + nvme_wait_dq(&dq, dev); + kthread_stop(kworker_task); +} - for (i = dev->queue_count - 1; i >= 0; i--) - nvme_disable_queue(dev, i); +/* +* Remove the node from the device list and check +* for whether or not we need to stop the nvme_thread. +*/ +static void nvme_dev_list_remove(struct nvme_dev *dev) +{ + struct task_struct *tmp = NULL; spin_lock(&dev_list_lock); list_del_init(&dev->node); + if (list_empty(&dev_list) && !IS_ERR_OR_NULL(nvme_thread)) { + tmp = nvme_thread; + nvme_thread = NULL; + } spin_unlock(&dev_list_lock); - if (dev->bar) + if (tmp) + kthread_stop(tmp); +} + +static void nvme_dev_shutdown(struct nvme_dev *dev) +{ + int i; + + dev->initialized = 0; + unregister_hotcpu_notifier(&dev->nb); + + nvme_dev_list_remove(dev); + + if (!dev->bar || (dev->bar && readl(&dev->bar->csts) == -1)) { + for (i = dev->queue_count - 1; i >= 0; i--) { + struct nvme_queue *nvmeq = raw_nvmeq(dev, i); + nvme_suspend_queue(nvmeq); + nvme_clear_queue(nvmeq); + } + } else { + nvme_disable_io_queues(dev); nvme_shutdown_ctrl(dev); + nvme_disable_queue(dev, 0); + } nvme_dev_unmap(dev); } static void nvme_dev_remove(struct nvme_dev *dev) { - struct nvme_ns *ns, *next; + struct nvme_ns *ns; - list_for_each_entry_safe(ns, next, &dev->namespaces, list) { - list_del(&ns->list); - del_gendisk(ns->disk); - nvme_ns_free(ns); + list_for_each_entry(ns, &dev->namespaces, list) { + if (ns->disk->flags & GENHD_FL_UP) + del_gendisk(ns->disk); + if (!blk_queue_dying(ns->queue)) + blk_cleanup_queue(ns->queue); } } @@ -1985,14 +2533,23 @@ static void nvme_release_instance(struct nvme_dev *dev) spin_unlock(&dev_list_lock); } +static void nvme_free_namespaces(struct nvme_dev *dev) +{ + struct nvme_ns *ns, *next; + + list_for_each_entry_safe(ns, next, &dev->namespaces, list) { + list_del(&ns->list); + put_disk(ns->disk); + kfree(ns); + } +} + static void nvme_free_dev(struct kref *kref) { struct nvme_dev *dev = container_of(kref, struct nvme_dev, kref); - nvme_dev_remove(dev); - nvme_dev_shutdown(dev); - nvme_free_queues(dev); - nvme_release_instance(dev); - nvme_release_prp_pools(dev); + + nvme_free_namespaces(dev); + free_percpu(dev->io_queue); kfree(dev->queues); kfree(dev->entry); kfree(dev); @@ -2036,6 +2593,7 @@ static const struct file_operations nvme_dev_fops = { static int nvme_dev_start(struct nvme_dev *dev) { int result; + bool start_thread = false; result = nvme_dev_map(dev); if (result) @@ -2046,9 +2604,24 @@ static int nvme_dev_start(struct nvme_dev *dev) goto unmap; spin_lock(&dev_list_lock); + if (list_empty(&dev_list) && IS_ERR_OR_NULL(nvme_thread)) { + start_thread = true; + nvme_thread = NULL; + } list_add(&dev->node, &dev_list); spin_unlock(&dev_list_lock); + if (start_thread) { + nvme_thread = kthread_run(nvme_kthread, NULL, "nvme"); + wake_up(&nvme_kthread_wait); + } else + wait_event_killable(nvme_kthread_wait, nvme_thread); + + if (IS_ERR_OR_NULL(nvme_thread)) { + result = nvme_thread ? PTR_ERR(nvme_thread) : -EINTR; + goto disable; + } + result = nvme_setup_io_queues(dev); if (result && result != -EBUSY) goto disable; @@ -2056,14 +2629,76 @@ static int nvme_dev_start(struct nvme_dev *dev) return result; disable: - spin_lock(&dev_list_lock); - list_del_init(&dev->node); - spin_unlock(&dev_list_lock); + nvme_disable_queue(dev, 0); + nvme_dev_list_remove(dev); unmap: nvme_dev_unmap(dev); return result; } +static int nvme_remove_dead_ctrl(void *arg) +{ + struct nvme_dev *dev = (struct nvme_dev *)arg; + struct pci_dev *pdev = dev->pci_dev; + + if (pci_get_drvdata(pdev)) + pci_stop_and_remove_bus_device(pdev); + kref_put(&dev->kref, nvme_free_dev); + return 0; +} + +static void nvme_remove_disks(struct work_struct *ws) +{ + struct nvme_dev *dev = container_of(ws, struct nvme_dev, reset_work); + + nvme_dev_remove(dev); + nvme_free_queues(dev, 1); +} + +static int nvme_dev_resume(struct nvme_dev *dev) +{ + int ret; + + ret = nvme_dev_start(dev); + if (ret && ret != -EBUSY) + return ret; + if (ret == -EBUSY) { + spin_lock(&dev_list_lock); + dev->reset_workfn = nvme_remove_disks; + queue_work(nvme_workq, &dev->reset_work); + spin_unlock(&dev_list_lock); + } + dev->initialized = 1; + return 0; +} + +static void nvme_dev_reset(struct nvme_dev *dev) +{ + nvme_dev_shutdown(dev); + if (nvme_dev_resume(dev)) { + dev_err(&dev->pci_dev->dev, "Device failed to resume\n"); + kref_get(&dev->kref); + if (IS_ERR(kthread_run(nvme_remove_dead_ctrl, dev, "nvme%d", + dev->instance))) { + dev_err(&dev->pci_dev->dev, + "Failed to start controller remove task\n"); + kref_put(&dev->kref, nvme_free_dev); + } + } +} + +static void nvme_reset_failed_dev(struct work_struct *ws) +{ + struct nvme_dev *dev = container_of(ws, struct nvme_dev, reset_work); + nvme_dev_reset(dev); +} + +static void nvme_reset_workfn(struct work_struct *work) +{ + struct nvme_dev *dev = container_of(work, struct nvme_dev, reset_work); + dev->reset_workfn(work); +} + static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id) { int result = -ENOMEM; @@ -2080,10 +2715,15 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id) GFP_KERNEL); if (!dev->queues) goto free; + dev->io_queue = alloc_percpu(unsigned short); + if (!dev->io_queue) + goto free; INIT_LIST_HEAD(&dev->namespaces); + dev->reset_workfn = nvme_reset_failed_dev; + INIT_WORK(&dev->reset_work, nvme_reset_workfn); dev->pci_dev = pdev; - + pci_set_drvdata(pdev, dev); result = nvme_set_instance(dev); if (result) goto free; @@ -2092,6 +2732,7 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id) if (result) goto release; + kref_init(&dev->kref); result = nvme_dev_start(dev); if (result) { if (result == -EBUSY) @@ -2113,29 +2754,50 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id) if (result) goto remove; - kref_init(&dev->kref); + dev->initialized = 1; return 0; remove: nvme_dev_remove(dev); + nvme_free_namespaces(dev); shutdown: nvme_dev_shutdown(dev); release_pools: - nvme_free_queues(dev); + nvme_free_queues(dev, 0); nvme_release_prp_pools(dev); release: nvme_release_instance(dev); free: + free_percpu(dev->io_queue); kfree(dev->queues); kfree(dev->entry); kfree(dev); return result; } +static void nvme_shutdown(struct pci_dev *pdev) +{ + struct nvme_dev *dev = pci_get_drvdata(pdev); + nvme_dev_shutdown(dev); +} + static void nvme_remove(struct pci_dev *pdev) { struct nvme_dev *dev = pci_get_drvdata(pdev); + + spin_lock(&dev_list_lock); + list_del_init(&dev->node); + spin_unlock(&dev_list_lock); + + pci_set_drvdata(pdev, NULL); + flush_work(&dev->reset_work); misc_deregister(&dev->miscdev); + nvme_dev_remove(dev); + nvme_dev_shutdown(dev); + nvme_free_queues(dev, 0); + rcu_barrier(); + nvme_release_instance(dev); + nvme_release_prp_pools(dev); kref_put(&dev->kref, nvme_free_dev); } @@ -2146,6 +2808,7 @@ static void nvme_remove(struct pci_dev *pdev) #define nvme_slot_reset NULL #define nvme_error_resume NULL +#ifdef CONFIG_PM_SLEEP static int nvme_suspend(struct device *dev) { struct pci_dev *pdev = to_pci_dev(dev); @@ -2159,14 +2822,14 @@ static int nvme_resume(struct device *dev) { struct pci_dev *pdev = to_pci_dev(dev); struct nvme_dev *ndev = pci_get_drvdata(pdev); - int ret; - ret = nvme_dev_start(ndev); - /* XXX: should remove gendisks if resume fails */ - if (ret) - nvme_free_queues(ndev); - return ret; + if (nvme_dev_resume(ndev) && !work_busy(&ndev->reset_work)) { + ndev->reset_workfn = nvme_reset_failed_dev; + queue_work(nvme_workq, &ndev->reset_work); + } + return 0; } +#endif static SIMPLE_DEV_PM_OPS(nvme_dev_pm_ops, nvme_suspend, nvme_resume); @@ -2181,7 +2844,7 @@ static const struct pci_error_handlers nvme_err_handler = { /* Move to pci_ids.h later */ #define PCI_CLASS_STORAGE_EXPRESS 0x010802 -static DEFINE_PCI_DEVICE_TABLE(nvme_id_table) = { +static const struct pci_device_id nvme_id_table[] = { { PCI_DEVICE_CLASS(PCI_CLASS_STORAGE_EXPRESS, 0xffffff) }, { 0, } }; @@ -2192,6 +2855,7 @@ static struct pci_driver nvme_driver = { .id_table = nvme_id_table, .probe = nvme_probe, .remove = nvme_remove, + .shutdown = nvme_shutdown, .driver = { .pm = &nvme_dev_pm_ops, }, @@ -2202,13 +2866,15 @@ static int __init nvme_init(void) { int result; - nvme_thread = kthread_run(nvme_kthread, NULL, "nvme"); - if (IS_ERR(nvme_thread)) - return PTR_ERR(nvme_thread); + init_waitqueue_head(&nvme_kthread_wait); + + nvme_workq = create_singlethread_workqueue("nvme"); + if (!nvme_workq) + return -ENOMEM; result = register_blkdev(nvme_major, "nvme"); if (result < 0) - goto kill_kthread; + goto kill_workq; else if (result > 0) nvme_major = result; @@ -2219,8 +2885,8 @@ static int __init nvme_init(void) unregister_blkdev: unregister_blkdev(nvme_major, "nvme"); - kill_kthread: - kthread_stop(nvme_thread); + kill_workq: + destroy_workqueue(nvme_workq); return result; } @@ -2228,11 +2894,12 @@ static void __exit nvme_exit(void) { pci_unregister_driver(&nvme_driver); unregister_blkdev(nvme_major, "nvme"); - kthread_stop(nvme_thread); + destroy_workqueue(nvme_workq); + BUG_ON(nvme_thread && !IS_ERR(nvme_thread)); } MODULE_AUTHOR("Matthew Wilcox <willy@linux.intel.com>"); MODULE_LICENSE("GPL"); -MODULE_VERSION("0.8"); +MODULE_VERSION("0.9"); module_init(nvme_init); module_exit(nvme_exit); |
