diff options
author | Sage Weil <sage@newdream.net> | 2011-05-24 18:46:31 (GMT) |
---|---|---|
committer | Sage Weil <sage@newdream.net> | 2011-05-24 18:52:12 (GMT) |
commit | db3540522e955c1ebb391f4f5324dff4f20ecd09 (patch) | |
tree | 8c25b07caa8614345c71f09e8872e60b68af4c31 /fs/ceph/mds_client.h | |
parent | cd634fb6eec72ef8e6dd677546b8d0ffdd2501eb (diff) | |
download | linux-db3540522e955c1ebb391f4f5324dff4f20ecd09.tar.xz |
ceph: fix cap flush race reentrancy
In e9964c10 we change cap flushing to do a delicate dance because some
inodes on the cap_dirty list could be in a migrating state (got EXPORT but
not IMPORT) in which we couldn't actually flush and move from
dirty->flushing, breaking the while (!empty) { process first } loop
structure. It worked for a single sync thread, but was not reentrant and
triggered infinite loops when multiple syncers came along.
Instead, move inodes with dirty to a separate cap_dirty_migrating list
when in the limbo export-but-no-import state, allowing us to go back to
the simple loop structure (which was reentrant). This is cleaner and more
robust.
Audited the cap_dirty users and this looks fine:
list_empty(&ci->i_dirty_item) is still a reliable indicator of whether we
have dirty caps (which list we're on is irrelevant) and list_del_init()
calls still do the right thing.
Signed-off-by: Sage Weil <sage@newdream.net>
Diffstat (limited to 'fs/ceph/mds_client.h')
-rw-r--r-- | fs/ceph/mds_client.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 4e3a9cc..7d8a0d6 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -278,6 +278,7 @@ struct ceph_mds_client { u64 cap_flush_seq; struct list_head cap_dirty; /* inodes with dirty caps */ + struct list_head cap_dirty_migrating; /* ...that are migration... */ int num_cap_flushing; /* # caps we are flushing */ spinlock_t cap_dirty_lock; /* protects above items */ wait_queue_head_t cap_flushing_wq; |