archive/linux
1
0
Fork 0
mirror of https://github.com/torvalds/linux.git synced 2025-08-16 06:31:34 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

block: protect wbt_lat_usec using q->elevator_lock

Browse source 
The wbt latency and state could be updated while initializing the
elevator or exiting the elevator. It could be also updated while
configuring IO latency QoS parameters using cgroup. The elevator
code path is now protected with q->elevator_lock. So we should
protect the access to sysfs attribute wbt_lat_usec using q->elevator
_lock instead of q->sysfs_lock. White we're at it, also protect
ioc_qos_write(), which configures wbt parameters via cgroup, using
q->elevator_lock.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Nilay Shroff <nilay@linux.ibm.com>
Link: https://lore.kernel.org/r/20250304102551.2533767-7-nilay@linux.ibm.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Nilay Shroff 2025-03-04 15:52:35 +05:30 • committed by Jens Axboe
parent 3efe7571c3
commit 245618f8e4
3 changed files with 12 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

2
block/blk-iocost.c
View file

@ -3248,6 +3248,7 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
} }
memflags = blk_mq_freeze_queue(disk->queue); memflags = blk_mq_freeze_queue(disk->queue);
mutex_lock(&disk->queue->elevator_lock);
blk_mq_quiesce_queue(disk->queue); blk_mq_quiesce_queue(disk->queue);
spin_lock_irq(&ioc->lock); spin_lock_irq(&ioc->lock);
@ -3355,6 +3356,7 @@ einval:
spin_unlock_irq(&ioc->lock); spin_unlock_irq(&ioc->lock);
blk_mq_unquiesce_queue(disk->queue); blk_mq_unquiesce_queue(disk->queue);
mutex_unlock(&disk->queue->elevator_lock);
blk_mq_unfreeze_queue(disk->queue, memflags); blk_mq_unfreeze_queue(disk->queue, memflags);
ret = -EINVAL; ret = -EINVAL;

20
block/blk-sysfs.c
View file

@ -557,7 +557,7 @@ static ssize_t queue_wb_lat_show(struct gendisk *disk, char *page)
ssize_t ret; ssize_t ret;
struct request_queue *q = disk->queue; struct request_queue *q = disk->queue;
mutex_lock(&q->sysfs_lock); mutex_lock(&q->elevator_lock);
if (!wbt_rq_qos(q)) { if (!wbt_rq_qos(q)) {
ret = -EINVAL; ret = -EINVAL;
goto out; goto out;
@ -570,7 +570,7 @@ static ssize_t queue_wb_lat_show(struct gendisk *disk, char *page)
ret = sysfs_emit(page, "%llu\n", div_u64(wbt_get_min_lat(q), 1000)); ret = sysfs_emit(page, "%llu\n", div_u64(wbt_get_min_lat(q), 1000));
out: out:
mutex_unlock(&q->sysfs_lock); mutex_unlock(&q->elevator_lock);
return ret; return ret;
} }
@ -589,8 +589,8 @@ static ssize_t queue_wb_lat_store(struct gendisk *disk, const char *page,
if (val < -1) if (val < -1)
return -EINVAL; return -EINVAL;
mutex_lock(&q->sysfs_lock);
memflags = blk_mq_freeze_queue(q); memflags = blk_mq_freeze_queue(q);
mutex_lock(&q->elevator_lock);
rqos = wbt_rq_qos(q); rqos = wbt_rq_qos(q);
if (!rqos) { if (!rqos) {
@ -619,8 +619,8 @@ static ssize_t queue_wb_lat_store(struct gendisk *disk, const char *page,
blk_mq_unquiesce_queue(q); blk_mq_unquiesce_queue(q);
out: out:
mutex_unlock(&q->elevator_lock);
blk_mq_unfreeze_queue(q, memflags); blk_mq_unfreeze_queue(q, memflags);
mutex_unlock(&q->sysfs_lock);
return ret; return ret;
} }
@ -689,19 +689,15 @@ static struct attribute *queue_attrs[] = {
/* Request-based queue attributes that are not relevant for bio-based queues. */ /* Request-based queue attributes that are not relevant for bio-based queues. */
static struct attribute *blk_mq_queue_attrs[] = { static struct attribute *blk_mq_queue_attrs[] = {
/*
* Attributes which are protected with q->sysfs_lock.
*/
#ifdef CONFIG_BLK_WBT
&queue_wb_lat_entry.attr,
#endif
/* /*
* Attributes which require some form of locking other than * Attributes which require some form of locking other than
* q->sysfs_lock. * q->sysfs_lock.
*/ */
&elv_iosched_entry.attr, &elv_iosched_entry.attr,
&queue_requests_entry.attr, &queue_requests_entry.attr,
#ifdef CONFIG_BLK_WBT
&queue_wb_lat_entry.attr,
#endif
/* /*
* Attributes which don't require locking. * Attributes which don't require locking.
*/ */
@ -882,10 +878,10 @@ int blk_register_queue(struct gendisk *disk)
goto out_crypto_sysfs_unregister; goto out_crypto_sysfs_unregister;
} }
} }
wbt_enable_default(disk);
mutex_unlock(&q->elevator_lock); mutex_unlock(&q->elevator_lock);
blk_queue_flag_set(QUEUE_FLAG_REGISTERED, q); blk_queue_flag_set(QUEUE_FLAG_REGISTERED, q);
wbt_enable_default(disk);
/* Now everything is ready and send out KOBJ_ADD uevent */ /* Now everything is ready and send out KOBJ_ADD uevent */
kobject_uevent(&disk->queue_kobj, KOBJ_ADD); kobject_uevent(&disk->queue_kobj, KOBJ_ADD);

4
include/linux/blkdev.h
View file

@ -563,8 +563,8 @@ struct request_queue {
/* /*
* Protects against I/O scheduler switching, particularly when * Protects against I/O scheduler switching, particularly when
* updating q->elevator. Since the elevator update code path may * updating q->elevator. Since the elevator update code path may
* also modify q->nr_requests, this lock also protects the sysfs * also modify q->nr_requests and wbt latency, this lock also
* attribute nr_requests. * protects the sysfs attributes nr_requests and wbt_lat_usec.
* To ensure proper locking order during an elevator update, first * To ensure proper locking order during an elevator update, first
* freeze the queue, then acquire ->elevator_lock. * freeze the queue, then acquire ->elevator_lock.
*/ */