summaryrefslogtreecommitdiff
path: root/drivers/md
diff options
context:
space:
mode:
authorNeilBrown <neilb@suse.com>2016-11-18 16:16:12 +1100
committerShaohua Li <shli@fb.com>2016-11-22 09:14:42 -0800
commit1919cbb23bf1b3e0fdb7b6edfb7369f920744087 (patch)
tree72ee0e0ec5d72c323a451f64ee9c61970a64bcaf /drivers/md
parent8d3ca83dcf9ca3d58822eddd279918d46f41e9ff (diff)
md/raid10: add failfast handling for writes.
When writing to a fastfail device, we use MD_FASTFAIL unless it is the only device being written to. For resync/recovery, assume there was a working device to read from so always use MD_FASTFAIL. If a write for resync/recovery fails, we just fail the device - there is not much else to do. If a normal write fails, but the device cannot be marked Faulty (must be only one left), we queue for write error handling which calls narrow_write_error() to write the block synchronously without any failfast flags. Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Shaohua Li <shli@fb.com>
Diffstat (limited to 'drivers/md')
-rw-r--r--drivers/md/raid10.c29
1 files changed, 28 insertions, 1 deletions
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 7cdc9bcd21b0..525ca9923707 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -100,6 +100,7 @@ static int max_queued_requests = 1024;
static void allow_barrier(struct r10conf *conf);
static void lower_barrier(struct r10conf *conf);
static int _enough(struct r10conf *conf, int previous, int ignore);
+static int enough(struct r10conf *conf, int ignore);
static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr,
int *skipped);
static void reshape_request_write(struct mddev *mddev, struct r10bio *r10_bio);
@@ -450,6 +451,7 @@ static void raid10_end_write_request(struct bio *bio)
struct r10conf *conf = r10_bio->mddev->private;
int slot, repl;
struct md_rdev *rdev = NULL;
+ struct bio *to_put = NULL;
bool discard_error;
discard_error = bio->bi_error && bio_op(bio) == REQ_OP_DISCARD;
@@ -477,8 +479,24 @@ static void raid10_end_write_request(struct bio *bio)
if (!test_and_set_bit(WantReplacement, &rdev->flags))
set_bit(MD_RECOVERY_NEEDED,
&rdev->mddev->recovery);
- set_bit(R10BIO_WriteError, &r10_bio->state);
+
dec_rdev = 0;
+ if (test_bit(FailFast, &rdev->flags) &&
+ (bio->bi_opf & MD_FAILFAST)) {
+ md_error(rdev->mddev, rdev);
+ if (!test_bit(Faulty, &rdev->flags))
+ /* This is the only remaining device,
+ * We need to retry the write without
+ * FailFast
+ */
+ set_bit(R10BIO_WriteError, &r10_bio->state);
+ else {
+ r10_bio->devs[slot].bio = NULL;
+ to_put = bio;
+ dec_rdev = 1;
+ }
+ } else
+ set_bit(R10BIO_WriteError, &r10_bio->state);
}
} else {
/*
@@ -528,6 +546,8 @@ static void raid10_end_write_request(struct bio *bio)
one_write_done(r10_bio);
if (dec_rdev)
rdev_dec_pending(rdev, conf->mddev);
+ if (to_put)
+ bio_put(to_put);
}
/*
@@ -1390,6 +1410,9 @@ retry_write:
mbio->bi_bdev = rdev->bdev;
mbio->bi_end_io = raid10_end_write_request;
bio_set_op_attrs(mbio, op, do_sync | do_fua);
+ if (test_bit(FailFast, &conf->mirrors[d].rdev->flags) &&
+ enough(conf, d))
+ mbio->bi_opf |= MD_FAILFAST;
mbio->bi_private = r10_bio;
if (conf->mddev->gendisk)
@@ -2052,6 +2075,8 @@ static void sync_request_write(struct mddev *mddev, struct r10bio *r10_bio)
atomic_inc(&r10_bio->remaining);
md_sync_acct(conf->mirrors[d].rdev->bdev, bio_sectors(tbio));
+ if (test_bit(FailFast, &conf->mirrors[d].rdev->flags))
+ tbio->bi_opf |= MD_FAILFAST;
tbio->bi_iter.bi_sector += conf->mirrors[d].rdev->data_offset;
tbio->bi_bdev = conf->mirrors[d].rdev->bdev;
generic_make_request(tbio);
@@ -3341,6 +3366,8 @@ static sector_t raid10_sync_request(struct mddev *mddev, sector_t sector_nr,
bio->bi_private = r10_bio;
bio->bi_end_io = end_sync_write;
bio_set_op_attrs(bio, REQ_OP_WRITE, 0);
+ if (test_bit(FailFast, &conf->mirrors[d].rdev->flags))
+ bio->bi_opf |= MD_FAILFAST;
bio->bi_iter.bi_sector = sector + rdev->data_offset;
bio->bi_bdev = rdev->bdev;
count++;