From 9fd097b14918875bd6f125ed699d7bbbba5893ee Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 21 Apr 2011 21:32:55 +0200 Subject: block: unexport DISK_EVENT_MEDIA_CHANGE for legacy/fringe drivers In-kernel disk event polling doesn't matter for legacy/fringe drivers and may lead to infinite event loop if ->check_events() implementation generates events on level condition instead of edge. Now that block layer supports suppressing exporting unlisted events, simply leaving disk->events cleared allows these drivers to keep the internal revalidation behavior intact while avoiding weird interactions with userland event handler. Signed-off-by: Tejun Heo Cc: Kay Sievers Signed-off-by: Jens Axboe --- drivers/block/DAC960.c | 1 - drivers/block/amiflop.c | 1 - drivers/block/ataflop.c | 1 - drivers/block/floppy.c | 1 - drivers/block/paride/pcd.c | 1 - drivers/block/paride/pd.c | 1 - drivers/block/paride/pf.c | 1 - drivers/block/swim.c | 1 - drivers/block/swim3.c | 1 - drivers/block/ub.c | 1 - drivers/block/xsysace.c | 1 - drivers/cdrom/gdrom.c | 1 - drivers/cdrom/viocd.c | 1 - drivers/message/i2o/i2o_block.c | 1 - drivers/s390/char/tape_block.c | 1 - 15 files changed, 15 deletions(-) diff --git a/drivers/block/DAC960.c b/drivers/block/DAC960.c index 8066d086578..e086fbbbe85 100644 --- a/drivers/block/DAC960.c +++ b/drivers/block/DAC960.c @@ -2547,7 +2547,6 @@ static bool DAC960_RegisterBlockDevice(DAC960_Controller_T *Controller) disk->major = MajorNumber; disk->first_minor = n << DAC960_MaxPartitionsBits; disk->fops = &DAC960_BlockDeviceOperations; - disk->events = DISK_EVENT_MEDIA_CHANGE; } /* Indicate the Block Device Registration completed successfully, diff --git a/drivers/block/amiflop.c b/drivers/block/amiflop.c index 456c0cc90dc..8eba86bba59 100644 --- a/drivers/block/amiflop.c +++ b/drivers/block/amiflop.c @@ -1736,7 +1736,6 @@ static int __init fd_probe_drives(void) disk->major = FLOPPY_MAJOR; disk->first_minor = drive; disk->fops = &floppy_fops; - disk->events = DISK_EVENT_MEDIA_CHANGE; sprintf(disk->disk_name, "fd%d", drive); disk->private_data = &unit[drive]; set_capacity(disk, 880*2); diff --git a/drivers/block/ataflop.c b/drivers/block/ataflop.c index c871eae1412..ede16c64ff0 100644 --- a/drivers/block/ataflop.c +++ b/drivers/block/ataflop.c @@ -1964,7 +1964,6 @@ static int __init atari_floppy_init (void) unit[i].disk->first_minor = i; sprintf(unit[i].disk->disk_name, "fd%d", i); unit[i].disk->fops = &floppy_fops; - unit[i].disk->events = DISK_EVENT_MEDIA_CHANGE; unit[i].disk->private_data = &unit[i]; unit[i].disk->queue = blk_init_queue(do_fd_request, &ataflop_lock); diff --git a/drivers/block/floppy.c b/drivers/block/floppy.c index 301d7a9a41a..db8f88586c8 100644 --- a/drivers/block/floppy.c +++ b/drivers/block/floppy.c @@ -4205,7 +4205,6 @@ static int __init floppy_init(void) disks[dr]->major = FLOPPY_MAJOR; disks[dr]->first_minor = TOMINOR(dr); disks[dr]->fops = &floppy_fops; - disks[dr]->events = DISK_EVENT_MEDIA_CHANGE; sprintf(disks[dr]->disk_name, "fd%d", dr); init_timer(&motor_off_timer[dr]); diff --git a/drivers/block/paride/pcd.c b/drivers/block/paride/pcd.c index 2f2ccf68625..8690e31d993 100644 --- a/drivers/block/paride/pcd.c +++ b/drivers/block/paride/pcd.c @@ -320,7 +320,6 @@ static void pcd_init_units(void) disk->first_minor = unit; strcpy(disk->disk_name, cd->name); /* umm... */ disk->fops = &pcd_bdops; - disk->events = DISK_EVENT_MEDIA_CHANGE; } } diff --git a/drivers/block/paride/pd.c b/drivers/block/paride/pd.c index 21dfdb77686..869e7676d46 100644 --- a/drivers/block/paride/pd.c +++ b/drivers/block/paride/pd.c @@ -837,7 +837,6 @@ static void pd_probe_drive(struct pd_unit *disk) p->fops = &pd_fops; p->major = major; p->first_minor = (disk - pd) << PD_BITS; - p->events = DISK_EVENT_MEDIA_CHANGE; disk->gd = p; p->private_data = disk; p->queue = pd_queue; diff --git a/drivers/block/paride/pf.c b/drivers/block/paride/pf.c index 7adeb1edbf4..f21b520ef41 100644 --- a/drivers/block/paride/pf.c +++ b/drivers/block/paride/pf.c @@ -294,7 +294,6 @@ static void __init pf_init_units(void) disk->first_minor = unit; strcpy(disk->disk_name, pf->name); disk->fops = &pf_fops; - disk->events = DISK_EVENT_MEDIA_CHANGE; if (!(*drives[unit])[D_PRT]) pf_drive_count++; } diff --git a/drivers/block/swim.c b/drivers/block/swim.c index 24a482f2fbd..fd5adcd5594 100644 --- a/drivers/block/swim.c +++ b/drivers/block/swim.c @@ -858,7 +858,6 @@ static int __devinit swim_floppy_init(struct swim_priv *swd) swd->unit[drive].disk->first_minor = drive; sprintf(swd->unit[drive].disk->disk_name, "fd%d", drive); swd->unit[drive].disk->fops = &floppy_fops; - swd->unit[drive].disk->events = DISK_EVENT_MEDIA_CHANGE; swd->unit[drive].disk->private_data = &swd->unit[drive]; swd->unit[drive].disk->queue = swd->queue; set_capacity(swd->unit[drive].disk, 2880); diff --git a/drivers/block/swim3.c b/drivers/block/swim3.c index 4c10f56facb..773bfa79277 100644 --- a/drivers/block/swim3.c +++ b/drivers/block/swim3.c @@ -1163,7 +1163,6 @@ static int __devinit swim3_attach(struct macio_dev *mdev, const struct of_device disk->major = FLOPPY_MAJOR; disk->first_minor = i; disk->fops = &floppy_fops; - disk->events = DISK_EVENT_MEDIA_CHANGE; disk->private_data = &floppy_states[i]; disk->queue = swim3_queue; disk->flags |= GENHD_FL_REMOVABLE; diff --git a/drivers/block/ub.c b/drivers/block/ub.c index 68b9430c7cf..0e376d46bdd 100644 --- a/drivers/block/ub.c +++ b/drivers/block/ub.c @@ -2334,7 +2334,6 @@ static int ub_probe_lun(struct ub_dev *sc, int lnum) disk->major = UB_MAJOR; disk->first_minor = lun->id * UB_PARTS_PER_LUN; disk->fops = &ub_bd_fops; - disk->events = DISK_EVENT_MEDIA_CHANGE; disk->private_data = lun; disk->driverfs_dev = &sc->intf->dev; diff --git a/drivers/block/xsysace.c b/drivers/block/xsysace.c index 645ff765cd1..6c7fd7db6df 100644 --- a/drivers/block/xsysace.c +++ b/drivers/block/xsysace.c @@ -1005,7 +1005,6 @@ static int __devinit ace_setup(struct ace_device *ace) ace->gd->major = ace_major; ace->gd->first_minor = ace->id * ACE_NUM_MINORS; ace->gd->fops = &ace_fops; - ace->gd->events = DISK_EVENT_MEDIA_CHANGE; ace->gd->queue = ace->queue; ace->gd->private_data = ace; snprintf(ace->gd->disk_name, 32, "xs%c", ace->id + 'a'); diff --git a/drivers/cdrom/gdrom.c b/drivers/cdrom/gdrom.c index b2b034fea34..3ceaf006e7f 100644 --- a/drivers/cdrom/gdrom.c +++ b/drivers/cdrom/gdrom.c @@ -803,7 +803,6 @@ static int __devinit probe_gdrom(struct platform_device *devptr) goto probe_fail_cdrom_register; } gd.disk->fops = &gdrom_bdops; - gd.disk->events = DISK_EVENT_MEDIA_CHANGE; /* latch on to the interrupt */ err = gdrom_set_interrupt_handlers(); if (err) diff --git a/drivers/cdrom/viocd.c b/drivers/cdrom/viocd.c index 4e874c5fa60..e427fbe4599 100644 --- a/drivers/cdrom/viocd.c +++ b/drivers/cdrom/viocd.c @@ -626,7 +626,6 @@ static int viocd_probe(struct vio_dev *vdev, const struct vio_device_id *id) gendisk->queue = q; gendisk->fops = &viocd_fops; gendisk->flags = GENHD_FL_CD|GENHD_FL_REMOVABLE; - gendisk->events = DISK_EVENT_MEDIA_CHANGE; set_capacity(gendisk, 0); gendisk->private_data = d; d->viocd_disk = gendisk; diff --git a/drivers/message/i2o/i2o_block.c b/drivers/message/i2o/i2o_block.c index 643ad52e3ca..4796bbf0ae4 100644 --- a/drivers/message/i2o/i2o_block.c +++ b/drivers/message/i2o/i2o_block.c @@ -1000,7 +1000,6 @@ static struct i2o_block_device *i2o_block_device_alloc(void) gd->major = I2O_MAJOR; gd->queue = queue; gd->fops = &i2o_block_fops; - gd->events = DISK_EVENT_MEDIA_CHANGE; gd->private_data = dev; dev->gd = gd; diff --git a/drivers/s390/char/tape_block.c b/drivers/s390/char/tape_block.c index 83cea9a55e2..1b3924c2fff 100644 --- a/drivers/s390/char/tape_block.c +++ b/drivers/s390/char/tape_block.c @@ -236,7 +236,6 @@ tapeblock_setup_device(struct tape_device * device) disk->major = tapeblock_major; disk->first_minor = device->first_minor; disk->fops = &tapeblock_fops; - disk->events = DISK_EVENT_MEDIA_CHANGE; disk->private_data = tape_get_device(device); disk->queue = blkdat->request_queue; set_capacity(disk, 0); -- cgit v1.2.3-70-g09d2 From bf2253a6f00e8fea5b026e471e9f0d0a1b3621f2 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 29 Apr 2011 10:15:14 +0200 Subject: cdrom: always check_disk_change() on open cdrom_open() called check_disk_change() after the rest of open path succeeded which leads to the following bizarre behavior. * After media change, if the device opened without O_NONBLOCK, open_for_data() naturally fails with -ENOMEDIA and check_disk_change() is never called. The media is known to be gone and the open failure makes it obvious to the userland but device invalidation never happens. * But if the device is opened with O_NONBLOCK, all the checks are bypassed and cdrom_open() doesn't notice that the media is not there and check_disk_change() is called and invalidation happens. There's nothing to be gained by avoiding calling check_disk_change() on open failure. Common cases end up calling check_disk_change() anyway. All we get is inconsistent behavior. Fix it by moving check_disk_change() invocation to the top of cdrom_open() so that it always gets called regardless of how the rest of open proceeds. Stable: 2.6.38 Signed-off-by: Tejun Heo Reported-by: Amit Shah Tested-by: Amit Shah Cc: stable@kernel.org Signed-off-by: Jens Axboe --- drivers/cdrom/cdrom.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/cdrom/cdrom.c b/drivers/cdrom/cdrom.c index 514dd8efaf7..75fb965b8f7 100644 --- a/drivers/cdrom/cdrom.c +++ b/drivers/cdrom/cdrom.c @@ -986,6 +986,9 @@ int cdrom_open(struct cdrom_device_info *cdi, struct block_device *bdev, fmode_t cdinfo(CD_OPEN, "entering cdrom_open\n"); + /* open is event synchronization point, check events first */ + check_disk_change(bdev); + /* if this was a O_NONBLOCK open and we should honor the flags, * do a quick open without drive/disc integrity checks. */ cdi->use_count++; @@ -1012,9 +1015,6 @@ int cdrom_open(struct cdrom_device_info *cdi, struct block_device *bdev, fmode_t cdinfo(CD_OPEN, "Use count for \"/dev/%s\" now %d\n", cdi->name, cdi->use_count); - /* Do this on open. Don't wait for mount, because they might - not be mounting, but opening with O_NONBLOCK */ - check_disk_change(bdev); return 0; err_release: if (CDROM_CAN(CDC_LOCK) && cdi->options & CDO_LOCK) { -- cgit v1.2.3-70-g09d2 From 02e352287a40bd456eb78df705bf888bc3161d3f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Fri, 29 Apr 2011 10:15:20 +0200 Subject: block: rescan partitions on invalidated devices on -ENOMEDIA too __blkdev_get() doesn't rescan partitions if disk->fops->open() fails, which leads to ghost partition devices lingering after medimum removal is known to both the kernel and userland. The behavior also creates a subtle inconsistency where O_NONBLOCK open, which doesn't fail even if there's no medium, clears the ghots partitions, which is exploited to work around the problem from userland. Fix it by updating __blkdev_get() to issue partition rescan after -ENOMEDIA too. This was reported in the following bz. https://bugzilla.kernel.org/show_bug.cgi?id=13029 Stable: 2.6.38 Signed-off-by: Tejun Heo Reported-by: David Zeuthen Reported-by: Martin Pitt Reported-by: Kay Sievers Tested-by: Kay Sievers Cc: Alan Cox Cc: stable@kernel.org Signed-off-by: Jens Axboe --- fs/block_dev.c | 27 ++++++++++++++++++--------- 1 file changed, 18 insertions(+), 9 deletions(-) diff --git a/fs/block_dev.c b/fs/block_dev.c index 5147bdd3b8e..257b00e9842 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1102,6 +1102,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) if (!bdev->bd_part) goto out_clear; + ret = 0; if (disk->fops->open) { ret = disk->fops->open(bdev, mode); if (ret == -ERESTARTSYS) { @@ -1118,9 +1119,18 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) put_disk(disk); goto restart; } - if (ret) - goto out_clear; } + /* + * If the device is invalidated, rescan partition + * if open succeeded or failed with -ENOMEDIUM. + * The latter is necessary to prevent ghost + * partitions on a removed medium. + */ + if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM)) + rescan_partitions(disk, bdev); + if (ret) + goto out_clear; + if (!bdev->bd_openers) { bd_set_size(bdev,(loff_t)get_capacity(disk)<<9); bdi = blk_get_backing_dev_info(bdev); @@ -1128,8 +1138,6 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) bdi = &default_backing_dev_info; bdev_inode_switch_bdi(bdev->bd_inode, bdi); } - if (bdev->bd_invalidated) - rescan_partitions(disk, bdev); } else { struct block_device *whole; whole = bdget_disk(disk, 0); @@ -1153,13 +1161,14 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) } } else { if (bdev->bd_contains == bdev) { - if (bdev->bd_disk->fops->open) { + ret = 0; + if (bdev->bd_disk->fops->open) ret = bdev->bd_disk->fops->open(bdev, mode); - if (ret) - goto out_unlock_bdev; - } - if (bdev->bd_invalidated) + /* the same as first opener case, read comment there */ + if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM)) rescan_partitions(bdev->bd_disk, bdev); + if (ret) + goto out_unlock_bdev; } /* only one opener holds refs to the module and disk */ module_put(disk->fops->owner); -- cgit v1.2.3-70-g09d2 From 70087dc38cc77ca8f46059564c00338777734762 Mon Sep 17 00:00:00 2001 From: Vivek Goyal Date: Mon, 16 May 2011 15:24:08 +0200 Subject: blk-throttle: Use task_subsys_state() to determine a task's blkio_cgroup Currentlly we first map the task to cgroup and then cgroup to blkio_cgroup. There is a more direct way to get to blkio_cgroup from task using task_subsys_state(). Use that. The real reason for the fix is that it also avoids a race in generic cgroup code. During remount/umount rebind_subsystems() is called and it can do following with and rcu protection. cgrp->subsys[i] = NULL; That means if somebody got hold of cgroup under rcu and then it tried to do cgroup->subsys[] to get to blkio_cgroup, it would get NULL which is wrong. I was running into this race condition with ltp running on a upstream derived kernel and that lead to crash. So ideally we should also fix cgroup generic code to wait for rcu grace period before setting pointer to NULL. Li Zefan is not very keen on introducing synchronize_wait() as he thinks it will slow down moun/remount/umount operations. So for the time being atleast fix the kernel crash by taking a more direct route to blkio_cgroup. One tester had reported a crash while running LTP on a derived kernel and with this fix crash is no more seen while the test has been running for over 6 days. Signed-off-by: Vivek Goyal Reviewed-by: Li Zefan Signed-off-by: Jens Axboe --- block/blk-cgroup.c | 7 +++++++ block/blk-cgroup.h | 3 +++ block/blk-throttle.c | 9 ++++----- block/cfq-iosched.c | 11 +++++------ 4 files changed, 19 insertions(+), 11 deletions(-) diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c index f0605ab2a76..471fdcc5df8 100644 --- a/block/blk-cgroup.c +++ b/block/blk-cgroup.c @@ -114,6 +114,13 @@ struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup) } EXPORT_SYMBOL_GPL(cgroup_to_blkio_cgroup); +struct blkio_cgroup *task_blkio_cgroup(struct task_struct *tsk) +{ + return container_of(task_subsys_state(tsk, blkio_subsys_id), + struct blkio_cgroup, css); +} +EXPORT_SYMBOL_GPL(task_blkio_cgroup); + static inline void blkio_update_group_weight(struct blkio_group *blkg, unsigned int weight) { diff --git a/block/blk-cgroup.h b/block/blk-cgroup.h index 10919fae2d3..c774930cc20 100644 --- a/block/blk-cgroup.h +++ b/block/blk-cgroup.h @@ -291,6 +291,7 @@ static inline void blkiocg_set_start_empty_time(struct blkio_group *blkg) {} #if defined(CONFIG_BLK_CGROUP) || defined(CONFIG_BLK_CGROUP_MODULE) extern struct blkio_cgroup blkio_root_cgroup; extern struct blkio_cgroup *cgroup_to_blkio_cgroup(struct cgroup *cgroup); +extern struct blkio_cgroup *task_blkio_cgroup(struct task_struct *tsk); extern void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg, struct blkio_group *blkg, void *key, dev_t dev, enum blkio_policy_id plid); @@ -314,6 +315,8 @@ void blkiocg_update_io_remove_stats(struct blkio_group *blkg, struct cgroup; static inline struct blkio_cgroup * cgroup_to_blkio_cgroup(struct cgroup *cgroup) { return NULL; } +static inline struct blkio_cgroup * +task_blkio_cgroup(struct task_struct *tsk) { return NULL; } static inline void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg, struct blkio_group *blkg, void *key, dev_t dev, diff --git a/block/blk-throttle.c b/block/blk-throttle.c index 0475a22a420..252a81a306f 100644 --- a/block/blk-throttle.c +++ b/block/blk-throttle.c @@ -160,9 +160,8 @@ static void throtl_put_tg(struct throtl_grp *tg) } static struct throtl_grp * throtl_find_alloc_tg(struct throtl_data *td, - struct cgroup *cgroup) + struct blkio_cgroup *blkcg) { - struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup); struct throtl_grp *tg = NULL; void *key = td; struct backing_dev_info *bdi = &td->queue->backing_dev_info; @@ -229,12 +228,12 @@ done: static struct throtl_grp * throtl_get_tg(struct throtl_data *td) { - struct cgroup *cgroup; struct throtl_grp *tg = NULL; + struct blkio_cgroup *blkcg; rcu_read_lock(); - cgroup = task_cgroup(current, blkio_subsys_id); - tg = throtl_find_alloc_tg(td, cgroup); + blkcg = task_blkio_cgroup(current); + tg = throtl_find_alloc_tg(td, blkcg); if (!tg) tg = &td->root_tg; rcu_read_unlock(); diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 5b52011e3a4..ab7a9e6a9b1 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -1014,10 +1014,9 @@ void cfq_update_blkio_group_weight(void *key, struct blkio_group *blkg, cfqg->needs_update = true; } -static struct cfq_group * -cfq_find_alloc_cfqg(struct cfq_data *cfqd, struct cgroup *cgroup, int create) +static struct cfq_group * cfq_find_alloc_cfqg(struct cfq_data *cfqd, + struct blkio_cgroup *blkcg, int create) { - struct blkio_cgroup *blkcg = cgroup_to_blkio_cgroup(cgroup); struct cfq_group *cfqg = NULL; void *key = cfqd; int i, j; @@ -1079,12 +1078,12 @@ done: */ static struct cfq_group *cfq_get_cfqg(struct cfq_data *cfqd, int create) { - struct cgroup *cgroup; + struct blkio_cgroup *blkcg; struct cfq_group *cfqg = NULL; rcu_read_lock(); - cgroup = task_cgroup(current, blkio_subsys_id); - cfqg = cfq_find_alloc_cfqg(cfqd, cgroup, create); + blkcg = task_blkio_cgroup(current); + cfqg = cfq_find_alloc_cfqg(cfqd, blkcg, create); if (!cfqg && create) cfqg = &cfqd->root_group; rcu_read_unlock(); -- cgit v1.2.3-70-g09d2 From 9937a5e2f32892db0dbeefc2b3bc74b3ae3ea9c7 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Tue, 17 May 2011 11:04:44 +0200 Subject: scsi: remove performance regression due to async queue run Commit c21e6beb removed our queue request_fn re-enter protection, and defaulted to always running the queues from kblockd to be safe. This was a known potential slow down, but should be safe. Unfortunately this is causing big performance regressions for some, so we need to improve this logic. Looking into the details of the re-enter, the real issue is on requeue of requests. Requeue of requests upon seeing a BUSY condition from the device ends up re-running the queue, causing traces like this: scsi_request_fn() scsi_dispatch_cmd() scsi_queue_insert() __scsi_queue_insert() scsi_run_queue() scsi_request_fn() ... potentially causing the issue we want to avoid. So special case the requeue re-run of the queue, but improve it to offload the entire run of local queue and starved queue from a single workqueue callback. This is a lot better than potentially kicking off a workqueue run for each device seen. This also fixes the issue of the local device going into recursion, since the above mentioned commit never moved that queue run out of line. Signed-off-by: Jens Axboe --- drivers/scsi/scsi_lib.c | 20 ++++++++++++++++---- drivers/scsi/scsi_scan.c | 2 ++ include/scsi/scsi_device.h | 1 + 3 files changed, 19 insertions(+), 4 deletions(-) diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index e9901b8f844..01e4e51c4b6 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -74,8 +74,6 @@ struct kmem_cache *scsi_sdb_cache; */ #define SCSI_QUEUE_DELAY 3 -static void scsi_run_queue(struct request_queue *q); - /* * Function: scsi_unprep_request() * @@ -161,7 +159,7 @@ static int __scsi_queue_insert(struct scsi_cmnd *cmd, int reason, int unbusy) blk_requeue_request(q, cmd->request); spin_unlock_irqrestore(q->queue_lock, flags); - scsi_run_queue(q); + kblockd_schedule_work(q, &device->requeue_work); return 0; } @@ -433,7 +431,11 @@ static void scsi_run_queue(struct request_queue *q) continue; } - blk_run_queue_async(sdev->request_queue); + spin_unlock(shost->host_lock); + spin_lock(sdev->request_queue->queue_lock); + __blk_run_queue(sdev->request_queue); + spin_unlock(sdev->request_queue->queue_lock); + spin_lock(shost->host_lock); } /* put any unprocessed entries back */ list_splice(&starved_list, &shost->starved_list); @@ -442,6 +444,16 @@ static void scsi_run_queue(struct request_queue *q) blk_run_queue(q); } +void scsi_requeue_run_queue(struct work_struct *work) +{ + struct scsi_device *sdev; + struct request_queue *q; + + sdev = container_of(work, struct scsi_device, requeue_work); + q = sdev->request_queue; + scsi_run_queue(q); +} + /* * Function: scsi_requeue_command() * diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c index 087821fac8f..58584dc0724 100644 --- a/drivers/scsi/scsi_scan.c +++ b/drivers/scsi/scsi_scan.c @@ -242,6 +242,7 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget, int display_failure_msg = 1, ret; struct Scsi_Host *shost = dev_to_shost(starget->dev.parent); extern void scsi_evt_thread(struct work_struct *work); + extern void scsi_requeue_run_queue(struct work_struct *work); sdev = kzalloc(sizeof(*sdev) + shost->transportt->device_size, GFP_ATOMIC); @@ -264,6 +265,7 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget, INIT_LIST_HEAD(&sdev->event_list); spin_lock_init(&sdev->list_lock); INIT_WORK(&sdev->event_work, scsi_evt_thread); + INIT_WORK(&sdev->requeue_work, scsi_requeue_run_queue); sdev->sdev_gendev.parent = get_device(&starget->dev); sdev->sdev_target = starget; diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h index 2d3ec509468..dd82e02ddde 100644 --- a/include/scsi/scsi_device.h +++ b/include/scsi/scsi_device.h @@ -169,6 +169,7 @@ struct scsi_device { sdev_dev; struct execute_work ew; /* used to get process context on put */ + struct work_struct requeue_work; struct scsi_dh_data *scsi_dh_data; enum scsi_device_state sdev_state; -- cgit v1.2.3-70-g09d2 From 3ec717b7ca4ee1d75d77e4f6286430d8f01d1dbd Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Wed, 18 May 2011 11:22:43 +0200 Subject: block: don't delay blk_run_queue_async Let's check a scenario: 1. blk_delay_queue(q, SCSI_QUEUE_DELAY); 2. blk_run_queue_async(); the second one will became a noop, because q->delay_work already has WORK_STRUCT_PENDING_BIT set, so the delayed work will still run after SCSI_QUEUE_DELAY. But blk_run_queue_async actually hopes the delayed work runs immediately. Fix this by doing a cancel on potentially pending delayed work before queuing an immediate run of the workqueue. Signed-off-by: Shaohua Li Signed-off-by: Jens Axboe --- block/blk-core.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/block/blk-core.c b/block/blk-core.c index a2e58eeb354..3fe00a14822 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -316,8 +316,10 @@ EXPORT_SYMBOL(__blk_run_queue); */ void blk_run_queue_async(struct request_queue *q) { - if (likely(!blk_queue_stopped(q))) + if (likely(!blk_queue_stopped(q))) { + __cancel_delayed_work(&q->delay_work); queue_delayed_work(kblockd_workqueue, &q->delay_work, 0); + } } EXPORT_SYMBOL(blk_run_queue_async); -- cgit v1.2.3-70-g09d2