]>
git.apps.os.sepia.ceph.com Git - ceph.git/log
Yehuda Sadeh [Mon, 20 Oct 2008 21:49:35 +0000 (14:49 -0700)]
kclient: some more cleanup
Yehuda Sadeh [Mon, 20 Oct 2008 21:39:46 +0000 (14:39 -0700)]
kclient: cleanup
Yehuda Sadeh [Mon, 20 Oct 2008 21:20:35 +0000 (14:20 -0700)]
kclient: some leak tool cleanup
Yehuda Sadeh [Mon, 20 Oct 2008 20:53:33 +0000 (13:53 -0700)]
kclient: remove extra debug messages
Yehuda Sadeh [Mon, 20 Oct 2008 20:15:34 +0000 (13:15 -0700)]
kclient: analyze crush correctly, kzalloc
Yehuda Sadeh [Mon, 20 Oct 2008 18:34:08 +0000 (11:34 -0700)]
Merge branch 'unstable' into kleak
Conflicts:
src/kernel/super.h
Yehuda Sadeh [Mon, 20 Oct 2008 17:19:17 +0000 (10:19 -0700)]
kclient: leak output shows filename and line
Yehuda Sadeh [Sat, 18 Oct 2008 00:16:49 +0000 (17:16 -0700)]
kclient: add missing files
Yehuda Sadeh [Sat, 18 Oct 2008 00:14:25 +0000 (17:14 -0700)]
kclient: leak detections tool
Sage Weil [Mon, 20 Oct 2008 22:37:49 +0000 (15:37 -0700)]
mds todo
Sage Weil [Mon, 20 Oct 2008 22:36:42 +0000 (15:36 -0700)]
filestore: fix getattrs to interpret listxattr return value correctly
Sage Weil [Mon, 20 Oct 2008 22:31:40 +0000 (15:31 -0700)]
filestore: clean up xattr routines using nice helpers
Conflicts:
src/os/FileStore.cc
Sage Weil [Mon, 20 Oct 2008 22:07:08 +0000 (15:07 -0700)]
mds: kill off inodes that get recovered but no longer have any parents
Sage Weil [Mon, 20 Oct 2008 21:54:33 +0000 (14:54 -0700)]
mds: don't clobber ctime on snap removal
Sage Weil [Mon, 20 Oct 2008 21:26:14 +0000 (14:26 -0700)]
kclient: fix snap_realm leak
The snap tree carries no reference. Instead, when the refcount hits 0, we remove
the realm from the tree.
Alternatively, we could attach the realm to it's parent inode and avoid the snap
tree entirely... hmm.
Sage Weil [Mon, 20 Oct 2008 21:06:37 +0000 (14:06 -0700)]
osd: clean up write_info vs write_log vs write_state
Sage Weil [Mon, 20 Oct 2008 20:56:43 +0000 (13:56 -0700)]
kclient: fix possible null pointer deref
Sage Weil [Mon, 20 Oct 2008 20:55:19 +0000 (13:55 -0700)]
crush: include cleanup
Sage Weil [Mon, 20 Oct 2008 20:54:56 +0000 (13:54 -0700)]
kclient: include crush in patch series
Sage Weil [Mon, 20 Oct 2008 20:35:01 +0000 (13:35 -0700)]
mds: update log segment tid list on journaled tableclient ack
Yehuda Sadeh [Mon, 20 Oct 2008 20:20:31 +0000 (13:20 -0700)]
Merge branch 'unstable' of ssh://ceph.newdream.net/git/ceph into unstable
Yehuda Sadeh [Mon, 20 Oct 2008 20:18:06 +0000 (13:18 -0700)]
kclient: free sessions container on exit
Sage Weil [Mon, 20 Oct 2008 19:49:20 +0000 (12:49 -0700)]
kclient: whitespace
Sage Weil [Mon, 20 Oct 2008 17:36:26 +0000 (10:36 -0700)]
kclient: remove useless __ for __u32 et al
Sage Weil [Mon, 20 Oct 2008 17:16:59 +0000 (10:16 -0700)]
kclient: fix else as per CodingStyle
Yehuda Sadeh [Mon, 20 Oct 2008 18:25:08 +0000 (11:25 -0700)]
kclient: wrong parameter passed to mdsmap_destroy
Yehuda Sadeh [Mon, 20 Oct 2008 18:23:42 +0000 (11:23 -0700)]
kclient: fix leak in osd reply messages
Yehuda Sadeh [Mon, 20 Oct 2008 17:41:52 +0000 (10:41 -0700)]
kclient: mdsmap leak fix
Yehuda Sadeh [Mon, 20 Oct 2008 17:24:54 +0000 (10:24 -0700)]
kclient: free monmap on exit
Sage Weil [Mon, 20 Oct 2008 17:14:53 +0000 (10:14 -0700)]
kclient: move dout etc. defs to ceph_debug.h
Sage Weil [Mon, 20 Oct 2008 17:06:08 +0000 (10:06 -0700)]
kclient: some osd, mds client comments
Sage Weil [Mon, 20 Oct 2008 16:54:23 +0000 (09:54 -0700)]
kclient: last bit of mon_client cleanup
Sage Weil [Mon, 20 Oct 2008 16:24:52 +0000 (09:24 -0700)]
kclient: script to build patch set for upstream
Sage Weil [Mon, 20 Oct 2008 16:24:25 +0000 (09:24 -0700)]
kclient: some mon client cleanup
Sage Weil [Sun, 19 Oct 2008 20:43:56 +0000 (13:43 -0700)]
kclient: kbuild patch update
Sage Weil [Sat, 18 Oct 2008 04:12:52 +0000 (21:12 -0700)]
kclient: more messenger cleanups, comments
Sage Weil [Sat, 18 Oct 2008 03:47:07 +0000 (20:47 -0700)]
kclient: messenger.c cleanup, reorganization, comments
Sage Weil [Fri, 17 Oct 2008 21:12:35 +0000 (14:12 -0700)]
kclient: fix cap_snap flush to reference correct snap
The client was passing the snapid the data was to fall into
instead of the snap it _follows_ (which is now the mds tends
to look at things). Use snapid-1 instead. This fixes the
problem of files being written while the snapshot was taken
appearing as 0 bytes in the snapshot.
Sage Weil [Fri, 17 Oct 2008 20:36:15 +0000 (13:36 -0700)]
osd todos
Sage Weil [Fri, 17 Oct 2008 20:30:33 +0000 (13:30 -0700)]
mds: only check osdmap against snap table while active
Sage Weil [Fri, 17 Oct 2008 20:27:11 +0000 (13:27 -0700)]
osd: track snap collections per pg, and clean them up
Explicitly track which snap collections exist for each pg. Clean them
up when the PG is destroyed.
Sage Weil [Fri, 17 Oct 2008 19:55:23 +0000 (12:55 -0700)]
osd: fast path if snap collection doesn't exist in snap_trimmer
Sage Weil [Fri, 17 Oct 2008 19:53:46 +0000 (12:53 -0700)]
filestore: return error code (instead of asserting) if collection doesn't exist on collection_list
The OSD doesn't pre-create the snap collections.
Sage Weil [Fri, 17 Oct 2008 19:44:17 +0000 (12:44 -0700)]
vstartnew.sh: turn up debugging on osd
Sage Weil [Fri, 17 Oct 2008 19:43:15 +0000 (12:43 -0700)]
mds: cleanly update snapserver table after snap purge
Any modifications to the table need to be journaled. Create MDSTableServer infrastructure
for a simple update done by the server (no prepare/commit), and use it to
remove the snapid(s). Replay no works.
Sage Weil [Fri, 17 Oct 2008 19:17:28 +0000 (12:17 -0700)]
kclient: more snap.c comments, cleanup
Sage Weil [Fri, 17 Oct 2008 17:17:13 +0000 (10:17 -0700)]
kclient: a few more snap.c comments
Sage Weil [Fri, 17 Oct 2008 14:26:03 +0000 (07:26 -0700)]
kclient: adjust ceph_finish_cap behavior
Look at capsnap->dirty instead of used cap refs, as this directly indicates
whether _this_ capsnap can be flushed (used bit indicates whether _all_
capsnaps can be flushed.. no need to wait that long!). Avoid an extra
call to ceph_flush_snaps by waiting until after to call it.
Sage Weil [Fri, 17 Oct 2008 14:12:04 +0000 (07:12 -0700)]
kclient: some snap.c comments, cleanup
Yehuda Sadeh [Fri, 17 Oct 2008 19:31:00 +0000 (12:31 -0700)]
kclient: osd messages timeout retriggering
Yehuda Sadeh [Fri, 17 Oct 2008 16:58:25 +0000 (09:58 -0700)]
kclient: refine osd messages timeout handling
Yehuda Sadeh [Thu, 16 Oct 2008 22:00:29 +0000 (15:00 -0700)]
kclient: debug message
Sage Weil [Fri, 17 Oct 2008 01:27:57 +0000 (18:27 -0700)]
debian: include debian dir in usr/src/modules/ceph
Sage Weil [Thu, 16 Oct 2008 23:39:43 +0000 (16:39 -0700)]
kclient: some export.c comments
Sage Weil [Thu, 16 Oct 2008 23:34:44 +0000 (16:34 -0700)]
kclient: addr.c comments, cleanup
I'm leaving the ability to redirty a page within a given
snapc in ceph_set_page_dirty (even though it isn't used)
because I'm not sure about the failure paths.. I suspect
we will need it later.
Sage Weil [Thu, 16 Oct 2008 22:33:54 +0000 (15:33 -0700)]
debian: make ceph-source include module source
There's a bunch of weirdness in the rules file, though, that I don't understand.
What's rules.modules.in? Why was it being populated (along with a debian dir)
in /usr/src/modules/ceph?
Sage Weil [Thu, 16 Oct 2008 22:33:10 +0000 (15:33 -0700)]
kclient: mdsmap cleanups
Sage Weil [Thu, 16 Oct 2008 22:26:24 +0000 (15:26 -0700)]
kclient: osdmap.c cleanups
Sage Weil [Thu, 16 Oct 2008 22:01:35 +0000 (15:01 -0700)]
kclient: osd_client comments, cleanup
Sage Weil [Thu, 16 Oct 2008 20:53:56 +0000 (13:53 -0700)]
debian: fix kernel module in source package
Sage Weil [Thu, 16 Oct 2008 20:20:41 +0000 (13:20 -0700)]
ceph_fs.h: whitespace
Sage Weil [Thu, 16 Oct 2008 19:13:09 +0000 (12:13 -0700)]
kclient: more mds_client comments, cleanup
Yehuda Sadeh [Thu, 16 Oct 2008 18:54:25 +0000 (11:54 -0700)]
kclient: more endianity fixes
Yehuda Sadeh [Thu, 16 Oct 2008 16:29:59 +0000 (09:29 -0700)]
kclient: endianity handling fixes
Yehuda Sadeh [Wed, 15 Oct 2008 23:08:33 +0000 (16:08 -0700)]
kclient: endianity fixes
Sage Weil [Thu, 16 Oct 2008 16:14:57 +0000 (09:14 -0700)]
kclient: some mds_client.c comments, cleanup
Sage Weil [Thu, 16 Oct 2008 02:50:03 +0000 (19:50 -0700)]
kclient: send osd, ping messages in lossless mode
This is what they should have been doing all along. Oops.
Sage Weil [Thu, 16 Oct 2008 00:17:39 +0000 (17:17 -0700)]
kclient: super.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 23:46:18 +0000 (16:46 -0700)]
kclient: move inode alloc/destroy into inode.c
Sage Weil [Wed, 15 Oct 2008 23:33:35 +0000 (16:33 -0700)]
kclient: caps.c comments, cleanup
Yehuda Sadeh [Wed, 15 Oct 2008 22:15:09 +0000 (15:15 -0700)]
kclient: remove compilation warning
Yehuda Sadeh [Wed, 15 Oct 2008 22:10:56 +0000 (15:10 -0700)]
kclient: sparse code cleanup
Sage Weil [Wed, 15 Oct 2008 22:03:29 +0000 (15:03 -0700)]
kclient: file.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 21:46:26 +0000 (14:46 -0700)]
kclient: dir.c comments, cleanup
Sage Weil [Wed, 15 Oct 2008 21:01:22 +0000 (14:01 -0700)]
vstartnew.sh: -n (norestart) to avoid crun restart loop
Sage Weil [Wed, 15 Oct 2008 20:52:33 +0000 (13:52 -0700)]
kclient: fix inode refcount problems in ceph_fill_trace
If we fail to take i_mutex, we get the inode via ceph_get_inode, which bumps
i_count. Make sure we iput() it before jumping to update_inode label. Note
that d_instantiate consumes the ref, but the other paths (d_alloc_anon,
d_find_alias) do not.
Yehuda Sadeh [Wed, 15 Oct 2008 20:34:24 +0000 (13:34 -0700)]
kclient: some more code cleanup
Yehuda Sadeh [Wed, 15 Oct 2008 20:33:38 +0000 (13:33 -0700)]
fix crc issue (when len < 4)
Yehuda Sadeh [Wed, 15 Oct 2008 18:54:22 +0000 (11:54 -0700)]
kclient: some sparse cleaning
Sage Weil [Wed, 15 Oct 2008 19:12:40 +0000 (12:12 -0700)]
use flags for mds chown requests
This avoids colliding with uids of -1 (with the old, lame method).
Clean up client's utimes() code while we're at it to explicitly specify
a mask.
Sage Weil [Wed, 15 Oct 2008 19:02:20 +0000 (12:02 -0700)]
kclient: inode.c comments, cleanup
Yehuda Sadeh [Wed, 15 Oct 2008 18:52:20 +0000 (11:52 -0700)]
kclient: fix umount_begin for 2.6.26+
Yehuda Sadeh [Wed, 15 Oct 2008 16:59:31 +0000 (09:59 -0700)]
fix wrong byte ordering handling (thanks brent!)
Yehuda Sadeh [Wed, 15 Oct 2008 16:58:30 +0000 (09:58 -0700)]
remove debug helper code for crc32
Sage Weil [Tue, 14 Oct 2008 23:53:38 +0000 (16:53 -0700)]
kclient: preallocate cap if we expect it
Preallocate memory for a capability when doing an open request. If we don't
use it, free it with the mds_request. This avoids the possibility of an ENOMEM
in the reply handler, which is a little difficult to deal with intelligently.
Sage Weil [Tue, 14 Oct 2008 23:18:24 +0000 (16:18 -0700)]
kclient: remove static caps from inode
Since most inodes will never have a capability, there is no point in embedding them
in the ceph_inode_info and paying the memory overhead in the general case.
Sage Weil [Tue, 14 Oct 2008 23:15:54 +0000 (16:15 -0700)]
kclient: super.h comments, cleanup, some minor cruft removed
Sage Weil [Tue, 14 Oct 2008 22:17:14 +0000 (15:17 -0700)]
ceph_fs: some cleanup, including mon statfs messages
Yehuda Sadeh [Tue, 14 Oct 2008 21:53:56 +0000 (14:53 -0700)]
checking buffer before calculating the crc (for debug)
Yehuda Sadeh [Tue, 14 Oct 2008 21:49:58 +0000 (14:49 -0700)]
kclient: some more error handling
Sage Weil [Tue, 14 Oct 2008 19:37:32 +0000 (12:37 -0700)]
kclient: remove ping_ack entirely
Sage Weil [Tue, 14 Oct 2008 20:43:51 +0000 (13:43 -0700)]
dstart.sh: use unstable btrfs
Sage Weil [Tue, 14 Oct 2008 20:05:30 +0000 (13:05 -0700)]
filestore: fix btrfs version detection
Yehuda Sadeh [Tue, 14 Oct 2008 19:48:52 +0000 (12:48 -0700)]
kclient: lower debug level on debug message
Yehuda Sadeh [Tue, 14 Oct 2008 19:34:18 +0000 (12:34 -0700)]
kclient: fix race condition
Yehuda Sadeh [Tue, 14 Oct 2008 16:20:25 +0000 (09:20 -0700)]
kclient: remove ping ack
Yehuda Sadeh [Tue, 14 Oct 2008 01:37:39 +0000 (18:37 -0700)]
kclient: preliminary forced umount
Sage Weil [Tue, 14 Oct 2008 00:42:44 +0000 (17:42 -0700)]
kclient: ping osds whose requests are timing out
This ensures the client detects when the osd tcp session has closed and
a reply may have been lost. If that does happen, the handle_reset handler
will be called and the request(s) will be resent.
Sage Weil [Tue, 14 Oct 2008 00:41:07 +0000 (17:41 -0700)]
msgr: ping needs no payload. lose ping_ack.
They were unused/ignored.
Sage Weil [Mon, 13 Oct 2008 23:57:44 +0000 (16:57 -0700)]
kclient: convert spinlock to mutex in osdc