diff options
| author | Jeff Layton <jlayton@kernel.org> | 2020-04-01 23:07:52 +0200 |
|---|---|---|
| committer | Ilya Dryomov <idryomov@gmail.com> | 2020-06-01 13:22:52 +0200 |
| commit | 1cf03a68e791b1673bc4daaa88a0820f34f538f8 (patch) | |
| tree | f264e092d23e310ce86bf25052dd592e7cb5e141 /fs/ceph/super.h | |
| parent | ceph: reset i_requested_max_size if file write is not wanted (diff) | |
| download | linux-1cf03a68e791b1673bc4daaa88a0820f34f538f8.tar.xz linux-1cf03a68e791b1673bc4daaa88a0820f34f538f8.zip | |
ceph: convert mdsc->cap_dirty to a per-session list
This is a per-sb list now, but that makes it difficult to tell when
the cap is the last dirty one associated with the session. Switch
this to be a per-session list, but continue using the
mdsc->cap_dirty_lock to protect the lists.
This list is only ever walked in ceph_flush_dirty_caps, so change that
to walk the sessions array and then flush the caps for inodes on each
session's list.
If the auth cap ever changes while the inode has dirty caps, then
move the inode to the appropriate session for the new auth_cap. Also,
ensure that we never remove an auth cap while the inode is still on the
s_cap_dirty list.
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: "Yan, Zheng" <zyan@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Diffstat (limited to 'fs/ceph/super.h')
| -rw-r--r-- | fs/ceph/super.h | 19 |
1 files changed, 16 insertions, 3 deletions
diff --git a/fs/ceph/super.h b/fs/ceph/super.h index bb372859c0ad..3a95395a4217 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -351,9 +351,22 @@ struct ceph_inode_info { struct rb_root i_caps; /* cap list */ struct ceph_cap *i_auth_cap; /* authoritative cap, if any */ unsigned i_dirty_caps, i_flushing_caps; /* mask of dirtied fields */ - struct list_head i_dirty_item, i_flushing_item; /* protected by - * mdsc->cap_dirty_lock - */ + + /* + * Link to the the auth cap's session's s_cap_dirty list. s_cap_dirty + * is protected by the mdsc->cap_dirty_lock, but each individual item + * is also protected by the inode's i_ceph_lock. Walking s_cap_dirty + * requires the mdsc->cap_dirty_lock. List presence for an item can + * be tested under the i_ceph_lock. Changing anything requires both. + */ + struct list_head i_dirty_item; + + /* + * Link to session's s_cap_flushing list. Protected by + * mdsc->cap_dirty_lock. + */ + struct list_head i_flushing_item; + /* we need to track cap writeback on a per-cap-bit basis, to allow * overlapping, pipelined cap flushes to the mds. we can probably * reduce the tid to 8 bits if we're concerned about inode size. */ |