| Message ID | 20191217155407.928386-1-rpenyaev@suse.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/1] block: end bio with BLK_STS_AGAIN in case of non-mq devs and REQ_NOWAIT | expand |
On 12/17/19 8:54 AM, Roman Penyaev wrote: > Non-mq devs do not honor REQ_NOWAIT so give a chance to the caller to repeat > request gracefully on -EAGAIN error. > > The problem is well reproduced using io_uring: > > mkfs.ext4 /dev/ram0 > mount /dev/ram0 /mnt > > # Preallocate a file > dd if=/dev/zero of=/mnt/file bs=1M count=1 > > # Start fio with io_uring and get -EIO > fio --rw=write --ioengine=io_uring --size=1M --direct=1 --name=job --filename=/mnt/file Thanks, this makes a lot of sense, doing -EIO for that case is wrong. Applied for 5.5.
diff --git a/block/blk-core.c b/block/blk-core.c index d5e668ec751b..1075aaff606d 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -886,11 +886,14 @@ generic_make_request_checks(struct bio *bio) } /* - * For a REQ_NOWAIT based request, return -EOPNOTSUPP - * if queue is not a request based queue. + * Non-mq queues do not honor REQ_NOWAIT, so complete a bio + * with BLK_STS_AGAIN status in order to catch -EAGAIN and + * to give a chance to the caller to repeat request gracefully. */ - if ((bio->bi_opf & REQ_NOWAIT) && !queue_is_mq(q)) - goto not_supported; + if ((bio->bi_opf & REQ_NOWAIT) && !queue_is_mq(q)) { + status = BLK_STS_AGAIN; + goto end_io; + } if (should_fail_bio(bio)) goto end_io;
Non-mq devs do not honor REQ_NOWAIT so give a chance to the caller to repeat request gracefully on -EAGAIN error. The problem is well reproduced using io_uring: mkfs.ext4 /dev/ram0 mount /dev/ram0 /mnt # Preallocate a file dd if=/dev/zero of=/mnt/file bs=1M count=1 # Start fio with io_uring and get -EIO fio --rw=write --ioengine=io_uring --size=1M --direct=1 --name=job --filename=/mnt/file Signed-off-by: Roman Penyaev <rpenyaev@suse.de> --- block/blk-core.c | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-)