* [PATCH] blk-mq-fs: invoke queue_for_each_hw_ctx under sysfs_lock
@ 2018-06-25 6:21 Jianchao Wang
2018-06-26 15:29 ` Bart Van Assche
0 siblings, 1 reply; 2+ messages in thread
From: Jianchao Wang @ 2018-06-25 6:21 UTC (permalink / raw)
To: axboe; +Cc: linux-block, linux-kernel
blk_mq_run_hw_queues and blk_mq_start_stopped_hw_queues in
queue_state_write will invoke queue_for_each_hw_ctx. It will race
with blk_mq_realloc_hw_ctxs and incur NULL pointer reference. Put
them under sysfs_lock to serialize the accessing to queue_hw_ctx
and nr_hw_queues.
Signed-off-by: Jianchao Wang <jianchao.w.wang@oracle.com>
---
block/blk-mq-debugfs.c | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/block/blk-mq-debugfs.c b/block/blk-mq-debugfs.c
index ffa6223..2e0c444 100644
--- a/block/blk-mq-debugfs.c
+++ b/block/blk-mq-debugfs.c
@@ -151,6 +151,7 @@ static ssize_t queue_state_write(void *data, const char __user *buf,
{
struct request_queue *q = data;
char opbuf[16] = { }, *op;
+ int res;
/*
* The "state" attribute is removed after blk_cleanup_queue() has called
@@ -169,9 +170,17 @@ static ssize_t queue_state_write(void *data, const char __user *buf,
return -EFAULT;
op = strstrip(opbuf);
if (strcmp(op, "run") == 0) {
+ res = mutex_lock_interruptible(&q->sysfs_lock);
+ if (res)
+ goto out;
blk_mq_run_hw_queues(q, true);
+ mutex_unlock(&q->sysfs_lock);
} else if (strcmp(op, "start") == 0) {
+ res = mutex_lock_interruptible(&q->sysfs_lock);
+ if (res)
+ goto out;
blk_mq_start_stopped_hw_queues(q, true);
+ mutex_unlock(&q->sysfs_lock);
} else if (strcmp(op, "kick") == 0) {
blk_mq_kick_requeue_list(q);
} else {
@@ -180,6 +189,7 @@ static ssize_t queue_state_write(void *data, const char __user *buf,
pr_err("%s: use 'run', 'start' or 'kick'\n", __func__);
return -EINVAL;
}
+out:
return count;
}
--
2.7.4
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] blk-mq-fs: invoke queue_for_each_hw_ctx under sysfs_lock
2018-06-25 6:21 [PATCH] blk-mq-fs: invoke queue_for_each_hw_ctx under sysfs_lock Jianchao Wang
@ 2018-06-26 15:29 ` Bart Van Assche
0 siblings, 0 replies; 2+ messages in thread
From: Bart Van Assche @ 2018-06-26 15:29 UTC (permalink / raw)
To: Jianchao Wang, axboe; +Cc: linux-block, linux-kernel
On 06/24/18 23:20, Jianchao Wang wrote:
> blk_mq_run_hw_queues and blk_mq_start_stopped_hw_queues in
> queue_state_write will invoke queue_for_each_hw_ctx. It will race
> with blk_mq_realloc_hw_ctxs and incur NULL pointer reference. Put
> them under sysfs_lock to serialize the accessing to queue_hw_ctx
> and nr_hw_queues.
The above looks wrong to me. blk_mq_realloc_hw_ctxs() starts with
calling blk_mq_unregister_hctx(). That last function uses kobject_del()
and kobject_del() waits until all pending sysfs callback functions for
the deleted objects have finished.
Bart.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2018-06-26 15:29 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-06-25 6:21 [PATCH] blk-mq-fs: invoke queue_for_each_hw_ctx under sysfs_lock Jianchao Wang
2018-06-26 15:29 ` Bart Van Assche
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).