Merge tag 'for-linus-20190202' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "A few fixes that should go into this release. This contains:

   - MD pull request from Song, fixing a recovery OOM issue (Alexei)

   - Fix for a sync related stall (Jianchao)

   - Dummy callback for timeouts (Tetsuo)

   - IDE atapi sense ordering fix (me)"

* tag 'for-linus-20190202' of git://git.kernel.dk/linux-block:
  ide: ensure atapi sense request aren't preempted
  blk-mq: fix a hung issue when fsync
  block: pass no-op callback to INIT_WORK().
  md/raid5: fix 'out of memory' during raid cache recovery

9 files changed, 93 insertions, 53 deletions

diff --git a/block/blk-core.c b/block/blk-core.c
index 3c5f61ceeb67..6b78ec56a4f2 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -462,6 +462,10 @@ static void blk_rq_timed_out_timer(struct timer_list *t)
         kblockd_schedule_work(&q->timeout_work);
 }
 
+static void blk_timeout_work(struct work_struct *work)
+{
+}
+
 /**
  * blk_alloc_queue_node - allocate a request queue
  * @gfp_mask: memory allocation flags
@@ -505,7 +509,7 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
         timer_setup(&q->backing_dev_info->laptop_mode_wb_timer,
                     laptop_mode_timer_fn, 0);
         timer_setup(&q->timeout, blk_rq_timed_out_timer, 0);
-        INIT_WORK(&q->timeout_work, NULL);
+        INIT_WORK(&q->timeout_work, blk_timeout_work);
         INIT_LIST_HEAD(&q->icq_list);
 #ifdef CONFIG_BLK_CGROUP
         INIT_LIST_HEAD(&q->blkg_list);
diff --git a/block/blk-flush.c b/block/blk-flush.c
index a3fc7191c694..6e0f2d97fc6d 100644
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -335,7 +335,7 @@ static void mq_flush_data_end_io(struct request *rq, blk_status_t error)
         blk_flush_complete_seq(rq, fq, REQ_FSEQ_DATA, error);
         spin_unlock_irqrestore(&fq->mq_flush_lock, flags);
 
-        blk_mq_run_hw_queue(hctx, true);
+        blk_mq_sched_restart(hctx);
 }
 
 /**
diff --git a/drivers/ide/ide-atapi.c b/drivers/ide/ide-atapi.c
index da58020a144e..33a28cde126c 100644
--- a/drivers/ide/ide-atapi.c
+++ b/drivers/ide/ide-atapi.c
@@ -235,21 +235,28 @@ EXPORT_SYMBOL_GPL(ide_prep_sense);
 
 int ide_queue_sense_rq(ide_drive_t *drive, void *special)
 {
-        struct request *sense_rq = drive->sense_rq;
+        ide_hwif_t *hwif = drive->hwif;
+        struct request *sense_rq;
+        unsigned long flags;
+
+        spin_lock_irqsave(&hwif->lock, flags);
 
         /* deferred failure from ide_prep_sense() */
         if (!drive->sense_rq_armed) {
                 printk(KERN_WARNING PFX "%s: error queuing a sense request\n",
                        drive->name);
+                spin_unlock_irqrestore(&hwif->lock, flags);
                 return -ENOMEM;
         }
 
+        sense_rq = drive->sense_rq;
         ide_req(sense_rq)->special = special;
         drive->sense_rq_armed = false;
 
         drive->hwif->rq = NULL;
 
         ide_insert_request_head(drive, sense_rq);
+        spin_unlock_irqrestore(&hwif->lock, flags);
         return 0;
 }
 EXPORT_SYMBOL_GPL(ide_queue_sense_rq);
diff --git a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c
index 8445b484ae69..b137f27a34d5 100644
--- a/drivers/ide/ide-io.c
+++ b/drivers/ide/ide-io.c
@@ -68,8 +68,10 @@ int ide_end_rq(ide_drive_t *drive, struct request *rq, blk_status_t error,
         }
 
         if (!blk_update_request(rq, error, nr_bytes)) {
-                if (rq == drive->sense_rq)
+                if (rq == drive->sense_rq) {
                         drive->sense_rq = NULL;
+                        drive->sense_rq_active = false;
+                }
 
                 __blk_mq_end_request(rq, error);
                 return 0;
@@ -451,16 +453,11 @@ void ide_requeue_and_plug(ide_drive_t *drive, struct request *rq)
                 blk_mq_delay_run_hw_queue(q->queue_hw_ctx[0], 3);
 }
 
-/*
- * Issue a new request to a device.
- */
-blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx,
-                          const struct blk_mq_queue_data *bd)
+blk_status_t ide_issue_rq(ide_drive_t *drive, struct request *rq,
+                          bool local_requeue)
 {
-        ide_drive_t     *drive = hctx->queue->queuedata;
-        ide_hwif_t      *hwif = drive->hwif;
+        ide_hwif_t *hwif = drive->hwif;
         struct ide_host *host = hwif->host;
-        struct request  *rq = bd->rq;
         ide_startstop_t startstop;
 
         if (!blk_rq_is_passthrough(rq) && !(rq->rq_flags & RQF_DONTPREP)) {
@@ -474,8 +471,6 @@ blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx,
         if (ide_lock_host(host, hwif))
                 return BLK_STS_DEV_RESOURCE;
 
-        blk_mq_start_request(rq);
-
         spin_lock_irq(&hwif->lock);
 
         if (!ide_lock_port(hwif)) {
@@ -511,18 +506,6 @@ repeat:
                 drive->dev_flags &= ~(IDE_DFLAG_SLEEPING | IDE_DFLAG_PARKED);
 
                 /*
-                 * we know that the queue isn't empty, but this can happen
-                 * if ->prep_rq() decides to kill a request
-                 */
-                if (!rq) {
-                        rq = bd->rq;
-                        if (!rq) {
-                                ide_unlock_port(hwif);
-                                goto out;
-                        }
-                }
-
-                /*
                  * Sanity: don't accept a request that isn't a PM request
                  * if we are currently power managed. This is very important as
                  * blk_stop_queue() doesn't prevent the blk_fetch_request()
@@ -560,9 +543,12 @@ repeat:
                 }
         } else {
 plug_device:
+                if (local_requeue)
+                        list_add(&rq->queuelist, &drive->rq_list);
                 spin_unlock_irq(&hwif->lock);
                 ide_unlock_host(host);
-                ide_requeue_and_plug(drive, rq);
+                if (!local_requeue)
+                        ide_requeue_and_plug(drive, rq);
                 return BLK_STS_OK;
         }
 
@@ -573,6 +559,26 @@ out:
         return BLK_STS_OK;
 }
 
+/*
+ * Issue a new request to a device.
+ */
+blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *hctx,
+                          const struct blk_mq_queue_data *bd)
+{
+        ide_drive_t *drive = hctx->queue->queuedata;
+        ide_hwif_t *hwif = drive->hwif;
+
+        spin_lock_irq(&hwif->lock);
+        if (drive->sense_rq_active) {
+                spin_unlock_irq(&hwif->lock);
+                return BLK_STS_DEV_RESOURCE;
+        }
+        spin_unlock_irq(&hwif->lock);
+
+        blk_mq_start_request(bd->rq);
+        return ide_issue_rq(drive, bd->rq, false);
+}
+
 static int drive_is_ready(ide_drive_t *drive)
 {
         ide_hwif_t *hwif = drive->hwif;
@@ -893,13 +899,8 @@ EXPORT_SYMBOL_GPL(ide_pad_transfer);
 
 void ide_insert_request_head(ide_drive_t *drive, struct request *rq)
 {
-        ide_hwif_t *hwif = drive->hwif;
-        unsigned long flags;
-
-        spin_lock_irqsave(&hwif->lock, flags);
+        drive->sense_rq_active = true;
         list_add_tail(&rq->queuelist, &drive->rq_list);
-        spin_unlock_irqrestore(&hwif->lock, flags);
-
         kblockd_schedule_work(&drive->rq_work);
 }
 EXPORT_SYMBOL_GPL(ide_insert_request_head);
diff --git a/drivers/ide/ide-park.c b/drivers/ide/ide-park.c
index 102aa3bc3e7f..8af7af6001eb 100644
--- a/drivers/ide/ide-park.c
+++ b/drivers/ide/ide-park.c
@@ -54,7 +54,9 @@ static void issue_park_cmd(ide_drive_t *drive, unsigned long timeout)
         scsi_req(rq)->cmd[0] = REQ_UNPARK_HEADS;
         scsi_req(rq)->cmd_len = 1;
         ide_req(rq)->type = ATA_PRIV_MISC;
+        spin_lock_irq(&hwif->lock);
         ide_insert_request_head(drive, rq);
+        spin_unlock_irq(&hwif->lock);
 
 out:
         return;
diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c
index 63627be0811a..5aeaca24a28f 100644
--- a/drivers/ide/ide-probe.c
+++ b/drivers/ide/ide-probe.c
@@ -1159,18 +1159,27 @@ static void drive_rq_insert_work(struct work_struct *work)
         ide_drive_t *drive = container_of(work, ide_drive_t, rq_work);
         ide_hwif_t *hwif = drive->hwif;
         struct request *rq;
+        blk_status_t ret;
         LIST_HEAD(list);
 
-        spin_lock_irq(&hwif->lock);
-        if (!list_empty(&drive->rq_list))
-                list_splice_init(&drive->rq_list, &list);
-        spin_unlock_irq(&hwif->lock);
+        blk_mq_quiesce_queue(drive->queue);
 
-        while (!list_empty(&list)) {
-                rq = list_first_entry(&list, struct request, queuelist);
+        ret = BLK_STS_OK;
+        spin_lock_irq(&hwif->lock);
+        while (!list_empty(&drive->rq_list)) {
+                rq = list_first_entry(&drive->rq_list, struct request, queuelist);
                 list_del_init(&rq->queuelist);
-                blk_execute_rq_nowait(drive->queue, rq->rq_disk, rq, true, NULL);
+
+                spin_unlock_irq(&hwif->lock);
+                ret = ide_issue_rq(drive, rq, true);
+                spin_lock_irq(&hwif->lock);
         }
+        spin_unlock_irq(&hwif->lock);
+
+        blk_mq_unquiesce_queue(drive->queue);
+
+        if (ret != BLK_STS_OK)
+                kblockd_schedule_work(&drive->rq_work);
 }
 
 static const u8 ide_hwif_to_major[] =
diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c
index ec3a5ef7fee0..cbbe6b6535be 100644
--- a/drivers/md/raid5-cache.c
+++ b/drivers/md/raid5-cache.c
@@ -1935,12 +1935,14 @@ out:
 }
 
 static struct stripe_head *
-r5c_recovery_alloc_stripe(struct r5conf *conf,
-                          sector_t stripe_sect)
+r5c_recovery_alloc_stripe(
+                struct r5conf *conf,
+                sector_t stripe_sect,
+                int noblock)
 {
         struct stripe_head *sh;
 
-        sh = raid5_get_active_stripe(conf, stripe_sect, 0, 1, 0);
+        sh = raid5_get_active_stripe(conf, stripe_sect, 0, noblock, 0);
         if (!sh)
                 return NULL;  /* no more stripe available */
 
@@ -2150,7 +2152,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log,
                                                 stripe_sect);
 
                 if (!sh) {
-                        sh = r5c_recovery_alloc_stripe(conf, stripe_sect);
+                        sh = r5c_recovery_alloc_stripe(conf, stripe_sect, 1);
                         /*
                          * cannot get stripe from raid5_get_active_stripe
                          * try replay some stripes
@@ -2159,20 +2161,29 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log,
                                 r5c_recovery_replay_stripes(
                                         cached_stripe_list, ctx);
                                 sh = r5c_recovery_alloc_stripe(
-                                        conf, stripe_sect);
+                                        conf, stripe_sect, 1);
                         }
                         if (!sh) {
+                                int new_size = conf->min_nr_stripes * 2;
                                 pr_debug("md/raid:%s: Increasing stripe cache size to %d to recovery data on journal.\n",
                                         mdname(mddev),
-                                        conf->min_nr_stripes * 2);
-                                raid5_set_cache_size(mddev,
-                                                     conf->min_nr_stripes * 2);
-                                sh = r5c_recovery_alloc_stripe(conf,
-                                                               stripe_sect);
+                                        new_size);
+                                ret = raid5_set_cache_size(mddev, new_size);
+                                if (conf->min_nr_stripes <= new_size / 2) {
+                                        pr_err("md/raid:%s: Cannot increase cache size, ret=%d, new_size=%d, min_nr_stripes=%d, max_nr_stripes=%d\n",
+                                                mdname(mddev),
+                                                ret,
+                                                new_size,
+                                                conf->min_nr_stripes,
+                                                conf->max_nr_stripes);
+                                        return -ENOMEM;
+                                }
+                                sh = r5c_recovery_alloc_stripe(
+                                        conf, stripe_sect, 0);
                         }
                         if (!sh) {
                                 pr_err("md/raid:%s: Cannot get enough stripes due to memory pressure. Recovery failed.\n",
-                                       mdname(mddev));
+                                        mdname(mddev));
                                 return -ENOMEM;
                         }
                         list_add_tail(&sh->lru, cached_stripe_list);
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 4990f0319f6c..cecea901ab8c 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -6369,6 +6369,7 @@ raid5_show_stripe_cache_size(struct mddev *mddev, char *page)
 int
 raid5_set_cache_size(struct mddev *mddev, int size)
 {
+        int result = 0;
         struct r5conf *conf = mddev->private;
 
         if (size <= 16 || size > 32768)
@@ -6385,11 +6386,14 @@ raid5_set_cache_size(struct mddev *mddev, int size)
 
         mutex_lock(&conf->cache_size_mutex);
         while (size > conf->max_nr_stripes)
-                if (!grow_one_stripe(conf, GFP_KERNEL))
+                if (!grow_one_stripe(conf, GFP_KERNEL)) {
+                        conf->min_nr_stripes = conf->max_nr_stripes;
+                        result = -ENOMEM;
                         break;
+                }
         mutex_unlock(&conf->cache_size_mutex);
 
-        return 0;
+        return result;
 }
 EXPORT_SYMBOL(raid5_set_cache_size);
 
diff --git a/include/linux/ide.h b/include/linux/ide.h
index e7d29ae633cd..971cf76a78a0 100644
--- a/include/linux/ide.h
+++ b/include/linux/ide.h
@@ -615,6 +615,7 @@ struct ide_drive_s {
 
         /* current sense rq and buffer */
         bool sense_rq_armed;
+        bool sense_rq_active;
         struct request *sense_rq;
         struct request_sense sense_data;
 
@@ -1219,6 +1220,7 @@ extern void ide_stall_queue(ide_drive_t *drive, unsigned long timeout);
 extern void ide_timer_expiry(struct timer_list *t);
 extern irqreturn_t ide_intr(int irq, void *dev_id);
 extern blk_status_t ide_queue_rq(struct blk_mq_hw_ctx *, const struct blk_mq_queue_data *);
+extern blk_status_t ide_issue_rq(ide_drive_t *, struct request *, bool);
 extern void ide_requeue_and_plug(ide_drive_t *drive, struct request *rq);
 
 void ide_init_disk(struct gendisk *, ide_drive_t *);