diff options
author | Mike Snitzer <snitzer@redhat.com> | 2015-07-15 20:52:04 (GMT) |
---|---|---|
committer | Mike Snitzer <snitzer@redhat.com> | 2015-07-16 14:23:49 (GMT) |
commit | bcc696fac11fe13e59dda5aaec6322a25b7c9a3a (patch) | |
tree | 956a9236835e48e108c0547eca1da8b0e8a1e2e9 /drivers/md/dm-thin.c | |
parent | b06075a98d595b761881fb2d7b8a557ea2f8b7ac (diff) | |
download | linux-bcc696fac11fe13e59dda5aaec6322a25b7c9a3a.tar.xz |
dm thin: stay in out-of-data-space mode once no_space_timeout expires
This fixes an issue where running out of data space would cause the
thin-pool's metadata to become read-only. There was no reason to make
metadata read-only -- calling set_pool_mode() with PM_READ_ONLY was a
misguided way to error all queued and future write IOs. We can
accomplish the same by degrading from PM_OUT_OF_DATA_SPACE to
PM_OUT_OF_DATA_SPACE with error_if_no_space enabled.
Otherwise, the use of PM_READ_ONLY could cause a race where commit() was
started before the PM_READ_ONLY transition but dm_pool_commit_metadata()
would go on to fail because the block manager had transitioned to
read-only. The return of -EPERM from dm_pool_commit_metadata(), due to
attempting to commit while in read-only mode, caused the thin-pool to
set 'needs_check' because a metadata_operation_failed(). This needless
cascade of failures makes life for users more difficult than needed.
Reported-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Diffstat (limited to 'drivers/md/dm-thin.c')
-rw-r--r-- | drivers/md/dm-thin.c | 21 |
1 files changed, 17 insertions, 4 deletions
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c index 8f015d9..34e7953 100644 --- a/drivers/md/dm-thin.c +++ b/drivers/md/dm-thin.c @@ -2282,18 +2282,23 @@ static void do_waker(struct work_struct *ws) queue_delayed_work(pool->wq, &pool->waker, COMMIT_PERIOD); } +static void notify_of_pool_mode_change_to_oods(struct pool *pool); + /* * We're holding onto IO to allow userland time to react. After the * timeout either the pool will have been resized (and thus back in - * PM_WRITE mode), or we degrade to PM_READ_ONLY and start erroring IO. + * PM_WRITE mode), or we degrade to PM_OUT_OF_DATA_SPACE w/ error_if_no_space. */ static void do_no_space_timeout(struct work_struct *ws) { struct pool *pool = container_of(to_delayed_work(ws), struct pool, no_space_timeout); - if (get_pool_mode(pool) == PM_OUT_OF_DATA_SPACE && !pool->pf.error_if_no_space) - set_pool_mode(pool, PM_READ_ONLY); + if (get_pool_mode(pool) == PM_OUT_OF_DATA_SPACE && !pool->pf.error_if_no_space) { + pool->pf.error_if_no_space = true; + notify_of_pool_mode_change_to_oods(pool); + error_retry_list(pool); + } } /*----------------------------------------------------------------*/ @@ -2371,6 +2376,14 @@ static void notify_of_pool_mode_change(struct pool *pool, const char *new_mode) dm_device_name(pool->pool_md), new_mode); } +static void notify_of_pool_mode_change_to_oods(struct pool *pool) +{ + if (!pool->pf.error_if_no_space) + notify_of_pool_mode_change(pool, "out-of-data-space (queue IO)"); + else + notify_of_pool_mode_change(pool, "out-of-data-space (error IO)"); +} + static bool passdown_enabled(struct pool_c *pt) { return pt->adjusted_pf.discard_passdown; @@ -2455,7 +2468,7 @@ static void set_pool_mode(struct pool *pool, enum pool_mode new_mode) * frequently seeing this mode. */ if (old_mode != new_mode) - notify_of_pool_mode_change(pool, "out-of-data-space"); + notify_of_pool_mode_change_to_oods(pool); pool->process_bio = process_bio_read_only; pool->process_discard = process_discard_bio; pool->process_cell = process_cell_read_only; |