[PATCH --resend 1/2] md: disable WRITE SAME if it fails for linear/raid0
From: Shaohua Li <hidden>
Date: 2017-02-13 19:55:47
Subsystem:
software raid (multiple disks) support, the rest · Maintainers:
Song Liu, Yu Kuai, Linus Torvalds
This makes md do the same thing as dm for write same IO failure. Please see 7eee4ae(dm: disable WRITE SAME if it fails) for details why we need this. Also reported here: https://bugzilla.kernel.org/show_bug.cgi?id=118581 Signed-off-by: Shaohua Li <redacted> --- drivers/md/linear.c | 6 +++++- drivers/md/md.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ drivers/md/md.h | 2 ++ drivers/md/raid0.c | 6 +++++- 4 files changed, 57 insertions(+), 2 deletions(-)
diff --git a/drivers/md/linear.c b/drivers/md/linear.c
index 26a73b2..bebc834 100644
--- a/drivers/md/linear.c
+++ b/drivers/md/linear.c@@ -291,7 +291,11 @@ static void linear_make_request(struct mddev *mddev, struct bio *bio) trace_block_bio_remap(bdev_get_queue(split->bi_bdev), split, disk_devt(mddev->gendisk), bio_sector); - generic_make_request(split); + if (bio_op(split) == REQ_OP_WRITE_SAME) + generic_make_request(md_writesame_setup(mddev, + split)); + else + generic_make_request(split); } } while (split != bio); return;
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 13020e5..7354f0b 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c@@ -312,6 +312,51 @@ static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio) return BLK_QC_T_NONE; } +struct md_writesame_data { + struct bio *orig_bio; + struct mddev *mddev; + struct bio cloned_bio; +}; + +static void md_writesame_endio(struct bio *bio) +{ + struct md_writesame_data *data = bio->bi_private; + + if (bio->bi_error == -EREMOTEIO && + !bdev_get_queue(bio->bi_bdev)->limits.max_write_same_sectors) + data->mddev->queue->limits.max_write_same_sectors = 0; + + data->orig_bio->bi_error = bio->bi_error; + bio_endio(data->orig_bio); + + kfree(data); +} + +struct bio *md_writesame_setup(struct mddev *mddev, struct bio *bio) +{ + struct md_writesame_data *data; + struct bio *cloned_bio; + + /* + * this failure means we ignore a chance to handle writesame failure, + * which isn't critcal, we can handle the failure if new writesame IO + * comes + */ + data = kmalloc(sizeof(*data), GFP_NOIO | __GFP_NORETRY); + if (!data) + return bio; + cloned_bio = &data->cloned_bio; + data->mddev = mddev; + data->orig_bio = bio; + bio_init(cloned_bio, NULL, 0); + __bio_clone_fast(cloned_bio, bio); + + cloned_bio->bi_private = data; + cloned_bio->bi_end_io = md_writesame_endio; + return cloned_bio; +} +EXPORT_SYMBOL_GPL(md_writesame_setup); + /* mddev_suspend makes sure no new requests are submitted * to the device, and that any requests that have been submitted * are completely handled.
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 2a51403..5c983e9 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h@@ -710,4 +710,6 @@ static inline void mddev_clear_unsupported_flags(struct mddev *mddev, { mddev->flags &= ~unsupported_flags; } + +extern struct bio *md_writesame_setup(struct mddev *mddev, struct bio *bio); #endif /* _MD_MD_H */
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 848365d..e38636e 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c@@ -503,7 +503,11 @@ static void raid0_make_request(struct mddev *mddev, struct bio *bio) trace_block_bio_remap(bdev_get_queue(split->bi_bdev), split, disk_devt(mddev->gendisk), bio_sector); - generic_make_request(split); + if (bio_op(split) == REQ_OP_WRITE_SAME) + generic_make_request(md_writesame_setup(mddev, + split)); + else + generic_make_request(split); } } while (split != bio); }
--
2.9.3