Re: [PATCH v2 1/7] md/raid10: fix r10bio leak in raid10_write_request() error paths

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] Execution paths in raid10_write_request and raid10_read_request unconditionally block on REQ_NOWAIT bios.
- [High] Unconditional use of GFP_NOIO for r10_bio allocation can block, violating REQ_NOWAIT semantics.
--

commit d4f532acb9d5fd7506dfbbb4a12a97248c3fb429
Author: Abd-Alrhman Masalkhi <[email protected]>

md/raid10: fix r10bio leak in raid10_write_request() error paths
    
When raid10_write_request() fails because REQ_NOWAIT is set, the
allocated r10_bio is not freed before returning, resulting in a memory
leak. Free r10_bio before returning from the REQ_NOWAIT error paths.
    
Link: https://lore.kernel.org/linux-raid/[email protected]/

> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index 0a3cfdd3f5df8..bd322eccdc3ff 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -1365,6 +1365,7 @@ static bool raid10_write_request(struct mddev *mddev, struct bio *bio,
>  		/* Bail out if REQ_NOWAIT is set for the bio */
>  		if (bio->bi_opf & REQ_NOWAIT) {
>  			bio_wouldblock_error(bio);
> +			free_r10bio(r10_bio);
>  			return false;
>  		}

[ ... ]

> @@ -1398,6 +1399,7 @@ static bool raid10_write_request(struct mddev *mddev, struct bio *bio,
>  		if (bio->bi_opf & REQ_NOWAIT) {
>  			allow_barrier(conf);
>  			bio_wouldblock_error(bio);
> +			free_r10bio(r10_bio);
>  			return false;
>  		}

[ ... ]

[Severity: High]
This is a pre-existing issue, but does wait_blocked_dev() still block
unconditionally later in raid10_write_request() even when REQ_NOWAIT is set?

raid10_write_request() {
    ...
    r10_bio->read_slot = -1;
    raid10_find_phys(conf, r10_bio);

    wait_blocked_dev(mddev, r10_bio);
    ...
}

Inside wait_blocked_dev(), the code calls md_wait_for_blocked_rdev() and
wait_barrier(), both of which can block indefinitely. Can this cause
unintended latency spikes or thread starvation for userspace applications
issuing async I/O?

Similarly, in raid10_read_request(), when a bio is split across a chunk
boundary, it calls wait_barrier(conf, false) without checking the
parent bio's REQ_NOWAIT flag.

[Severity: High]
This is another pre-existing issue, but do the mempool_alloc() calls in
__make_request() and raid10_handle_discard() bypass REQ_NOWAIT semantics?

__make_request() {
    ...
    r10_bio = mempool_alloc(&conf->r10bio_pool, GFP_NOIO);
    ...
}

If the mempool is depleted and memory is low, this GFP_NOIO allocation can
block. Should this use GFP_NOWAIT for REQ_NOWAIT bios and return
BLK_STS_AGAIN when the pool is exhausted to preserve the non-blocking
contract?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1