Commit e879a0d9 authored by Yu Kuai's avatar Yu Kuai Committed by Yu Kuai
Browse files

md/raid1,raid10: don't ignore IO flags 

If blk-wbt is enabled by default, it's found that raid write performance
is quite bad because all IO are throttled by wbt of underlying disks,
due to flag REQ_IDLE is ignored. And turns out this behaviour exist since
blk-wbt is introduced.

Other than REQ_IDLE, other flags should not be ignored as well, for
example REQ_META can be set for filesystems, clearing it can cause priority
reverse problems; And REQ_NOWAIT should not be cleared as well, because
io will wait instead of failing directly in underlying disks.

Fix those problems by keep IO flags from master bio.

Fises: f51d46d0 ("md: add support for REQ_NOWAIT")
Fixes: e34cbd30 ("blk-wbt: add general throttling mechanism")
Fixes: 5404bc7a ("[PATCH] Allow file systems to differentiate between data and meta reads")
Link: https://lore.kernel.org/linux-raid/20250227121657.832356-1-yukuai1@huaweicloud.com


Signed-off-by: default avatarYu Kuai <yukuai3@huawei.com>
parent 1320fe87

drivers/md/raid1.c

+0 −5
Original line number Diff line number Diff line
@@ -1317,8 +1317,6 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio, 
	struct r1conf *conf = mddev->private;

	struct raid1_info *mirror;

	struct bio *read_bio;

	const enum req_op op = bio_op(bio);

	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;

	int max_sectors;

	int rdisk, error;

	bool r1bio_existed = !!r1_bio;
@@ -1406,7 +1404,6 @@ static void raid1_read_request(struct mddev *mddev, struct bio *bio, 
	read_bio->bi_iter.bi_sector = r1_bio->sector +

		mirror->rdev->data_offset;

	read_bio->bi_end_io = raid1_end_read_request;

	read_bio->bi_opf = op | do_sync;

	if (test_bit(FailFast, &mirror->rdev->flags) &&

	    test_bit(R1BIO_FailFast, &r1_bio->state))

	        read_bio->bi_opf |= MD_FAILFAST;
@@ -1655,8 +1652,6 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio, 


		mbio->bi_iter.bi_sector	= (r1_bio->sector + rdev->data_offset);

		mbio->bi_end_io	= raid1_end_write_request;

		mbio->bi_opf = bio_op(bio) |

			(bio->bi_opf & (REQ_SYNC | REQ_FUA | REQ_ATOMIC));

		if (test_bit(FailFast, &rdev->flags) &&

		    !test_bit(WriteMostly, &rdev->flags) &&

		    conf->raid_disks - mddev->degraded > 1)

drivers/md/raid10.c

+0 −8
Original line number Diff line number Diff line
@@ -1147,8 +1147,6 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio, 
{

	struct r10conf *conf = mddev->private;

	struct bio *read_bio;

	const enum req_op op = bio_op(bio);

	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;

	int max_sectors;

	struct md_rdev *rdev;

	char b[BDEVNAME_SIZE];
@@ -1229,7 +1227,6 @@ static void raid10_read_request(struct mddev *mddev, struct bio *bio, 
	read_bio->bi_iter.bi_sector = r10_bio->devs[slot].addr +

		choose_data_offset(r10_bio, rdev);

	read_bio->bi_end_io = raid10_end_read_request;

	read_bio->bi_opf = op | do_sync;

	if (test_bit(FailFast, &rdev->flags) &&

	    test_bit(R10BIO_FailFast, &r10_bio->state))

	        read_bio->bi_opf |= MD_FAILFAST;
@@ -1248,10 +1245,6 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio, 
				  struct bio *bio, bool replacement,

				  int n_copy)

{

	const enum req_op op = bio_op(bio);

	const blk_opf_t do_sync = bio->bi_opf & REQ_SYNC;

	const blk_opf_t do_fua = bio->bi_opf & REQ_FUA;

	const blk_opf_t do_atomic = bio->bi_opf & REQ_ATOMIC;

	unsigned long flags;

	struct r10conf *conf = mddev->private;

	struct md_rdev *rdev;
@@ -1270,7 +1263,6 @@ static void raid10_write_one_disk(struct mddev *mddev, struct r10bio *r10_bio, 
	mbio->bi_iter.bi_sector	= (r10_bio->devs[n_copy].addr +

				   choose_data_offset(r10_bio, rdev));

	mbio->bi_end_io	= raid10_end_write_request;

	mbio->bi_opf = op | do_sync | do_fua | do_atomic;

	if (!replacement && test_bit(FailFast,

				     &conf->mirrors[devnum].rdev->flags)

			 && enough(conf, devnum))