md/raid5: writes should get directed to replacement as well as original.
When writing, we need to submit two writes, one to the original, and
one to the replacement - if there is a replacement.
If the write to the replacement results in a write error, we just fail
the device. We only try to record write errors to the original.
When writing for recovery, we shouldn't write to the original. This
will be addressed in a subsequent patch that generally addresses
recovery.
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index e5795d3..14878a9 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -503,8 +503,8 @@
for (i = disks; i--; ) {
int rw;
- struct bio *bi;
- struct md_rdev *rdev;
+ struct bio *bi, *rbi;
+ struct md_rdev *rdev, *rrdev = NULL;
if (test_and_clear_bit(R5_Wantwrite, &sh->dev[i].flags)) {
if (test_and_clear_bit(R5_WantFUA, &sh->dev[i].flags))
rw = WRITE_FUA;
@@ -516,27 +516,36 @@
continue;
bi = &sh->dev[i].req;
+ rbi = &sh->dev[i].rreq; /* For writing to replacement */
bi->bi_rw = rw;
- if (rw & WRITE)
+ rbi->bi_rw = rw;
+ if (rw & WRITE) {
bi->bi_end_io = raid5_end_write_request;
- else
+ rbi->bi_end_io = raid5_end_write_request;
+ } else
bi->bi_end_io = raid5_end_read_request;
rcu_read_lock();
- if (rw == READ &&
- test_bit(R5_ReadRepl, &sh->dev[i].flags))
+ rdev = rcu_dereference(conf->disks[i].rdev);
+ if (rw & WRITE)
+ rrdev = rcu_dereference(conf->disks[i].replacement);
+ else if (test_bit(R5_ReadRepl, &sh->dev[i].flags))
rdev = rcu_dereference(conf->disks[i].replacement);
- else
- rdev = rcu_dereference(conf->disks[i].rdev);
+
if (rdev && test_bit(Faulty, &rdev->flags))
rdev = NULL;
if (rdev)
atomic_inc(&rdev->nr_pending);
+ if (rrdev && test_bit(Faulty, &rrdev->flags))
+ rrdev = NULL;
+ if (rrdev)
+ atomic_inc(&rrdev->nr_pending);
rcu_read_unlock();
/* We have already checked bad blocks for reads. Now
- * need to check for writes.
+ * need to check for writes. We never accept write errors
+ * on the replacement, so we don't to check rrdev.
*/
while ((rw & WRITE) && rdev &&
test_bit(WriteErrorSeen, &rdev->flags)) {
@@ -583,8 +592,32 @@
bi->bi_io_vec[0].bv_offset = 0;
bi->bi_size = STRIPE_SIZE;
bi->bi_next = NULL;
+ if (rrdev)
+ set_bit(R5_DOUBLE_LOCKED, &sh->dev[i].flags);
generic_make_request(bi);
- } else {
+ }
+ if (rrdev) {
+ if (s->syncing || s->expanding || s->expanded)
+ md_sync_acct(rrdev->bdev, STRIPE_SECTORS);
+
+ set_bit(STRIPE_IO_STARTED, &sh->state);
+
+ rbi->bi_bdev = rrdev->bdev;
+ pr_debug("%s: for %llu schedule op %ld on "
+ "replacement disc %d\n",
+ __func__, (unsigned long long)sh->sector,
+ rbi->bi_rw, i);
+ atomic_inc(&sh->count);
+ rbi->bi_sector = sh->sector + rrdev->data_offset;
+ rbi->bi_flags = 1 << BIO_UPTODATE;
+ rbi->bi_idx = 0;
+ rbi->bi_io_vec[0].bv_len = STRIPE_SIZE;
+ rbi->bi_io_vec[0].bv_offset = 0;
+ rbi->bi_size = STRIPE_SIZE;
+ rbi->bi_next = NULL;
+ generic_make_request(rbi);
+ }
+ if (!rdev && !rrdev) {
if (rw & WRITE)
set_bit(STRIPE_DEGRADED, &sh->state);
pr_debug("skip op %ld on disc %d for sector %llu\n",
@@ -1695,14 +1728,23 @@
struct stripe_head *sh = bi->bi_private;
struct r5conf *conf = sh->raid_conf;
int disks = sh->disks, i;
+ struct md_rdev *uninitialized_var(rdev);
int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags);
sector_t first_bad;
int bad_sectors;
+ int replacement = 0;
- for (i=0 ; i<disks; i++)
- if (bi == &sh->dev[i].req)
+ for (i = 0 ; i < disks; i++) {
+ if (bi == &sh->dev[i].req) {
+ rdev = conf->disks[i].rdev;
break;
-
+ }
+ if (bi == &sh->dev[i].rreq) {
+ rdev = conf->disks[i].replacement;
+ replacement = 1;
+ break;
+ }
+ }
pr_debug("end_write_request %llu/%d, count %d, uptodate: %d.\n",
(unsigned long long)sh->sector, i, atomic_read(&sh->count),
uptodate);
@@ -1711,21 +1753,30 @@
return;
}
- if (!uptodate) {
- set_bit(WriteErrorSeen, &conf->disks[i].rdev->flags);
- set_bit(R5_WriteError, &sh->dev[i].flags);
- } else if (is_badblock(conf->disks[i].rdev, sh->sector, STRIPE_SECTORS,
- &first_bad, &bad_sectors))
- set_bit(R5_MadeGood, &sh->dev[i].flags);
+ if (replacement) {
+ if (!uptodate)
+ md_error(conf->mddev, rdev);
+ else if (is_badblock(rdev, sh->sector,
+ STRIPE_SECTORS,
+ &first_bad, &bad_sectors))
+ set_bit(R5_MadeGoodRepl, &sh->dev[i].flags);
+ } else {
+ if (!uptodate) {
+ set_bit(WriteErrorSeen, &rdev->flags);
+ set_bit(R5_WriteError, &sh->dev[i].flags);
+ } else if (is_badblock(rdev, sh->sector,
+ STRIPE_SECTORS,
+ &first_bad, &bad_sectors))
+ set_bit(R5_MadeGood, &sh->dev[i].flags);
+ }
+ rdev_dec_pending(rdev, conf->mddev);
- rdev_dec_pending(conf->disks[i].rdev, conf->mddev);
-
- clear_bit(R5_LOCKED, &sh->dev[i].flags);
+ if (!test_and_clear_bit(R5_DOUBLE_LOCKED, &sh->dev[i].flags))
+ clear_bit(R5_LOCKED, &sh->dev[i].flags);
set_bit(STRIPE_HANDLE, &sh->state);
release_stripe(sh);
}
-
static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous);
static void raid5_build_block(struct stripe_head *sh, int i, int previous)
@@ -1739,6 +1790,13 @@
dev->req.bi_private = sh;
dev->vec.bv_page = dev->page;
+ bio_init(&dev->rreq);
+ dev->rreq.bi_io_vec = &dev->rvec;
+ dev->rreq.bi_vcnt++;
+ dev->rreq.bi_max_vecs++;
+ dev->rreq.bi_private = sh;
+ dev->rvec.bv_page = dev->page;
+
dev->flags = 0;
dev->sector = compute_blocknr(sh, i, previous);
}
@@ -3132,6 +3190,15 @@
} else
clear_bit(R5_MadeGood, &dev->flags);
}
+ if (test_bit(R5_MadeGoodRepl, &dev->flags)) {
+ struct md_rdev *rdev2 = rcu_dereference(
+ conf->disks[i].replacement);
+ if (rdev2 && !test_bit(Faulty, &rdev2->flags)) {
+ s->handle_bad_blocks = 1;
+ atomic_inc(&rdev2->nr_pending);
+ } else
+ clear_bit(R5_MadeGoodRepl, &dev->flags);
+ }
if (!test_bit(R5_Insync, &dev->flags)) {
/* The ReadError flag will just be confusing now */
clear_bit(R5_ReadError, &dev->flags);
@@ -3404,6 +3471,12 @@
STRIPE_SECTORS);
rdev_dec_pending(rdev, conf->mddev);
}
+ if (test_and_clear_bit(R5_MadeGoodRepl, &dev->flags)) {
+ rdev = conf->disks[i].replacement;
+ rdev_clear_badblocks(rdev, sh->sector,
+ STRIPE_SECTORS);
+ rdev_dec_pending(rdev, conf->mddev);
+ }
}
if (s.ops_request)
diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h
index 4cfd801..f6faaa1 100644
--- a/drivers/md/raid5.h
+++ b/drivers/md/raid5.h
@@ -259,6 +259,7 @@
enum r5dev_flags {
R5_UPTODATE, /* page contains current data */
R5_LOCKED, /* IO has been submitted on "req" */
+ R5_DOUBLE_LOCKED,/* Cannot clear R5_LOCKED until 2 writes complete */
R5_OVERWRITE, /* towrite covers whole page */
/* and some that are internal to handle_stripe */
R5_Insync, /* rdev && rdev->in_sync at start */