summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* messenger: zero pages infrastructure (not working yet)historic/msgr_zeropagesYehuda Sadeh2009-04-083-11/+91
|
* todoSage Weil2009-04-071-1/+0
|
* config: remove INSTALL_PREFIXSage Weil2009-04-071-3/+3
| | | | | On a 'norma' install, it's /usr, which isn't helpful. If anything, we need sysconfdir, etc. individually. Blech.
* uclient: fix up caps some. make dentry leases workSage Weil2009-04-072-98/+180
|
* uclient: set fsid in statfs requestSage Weil2009-04-072-2/+2
|
* kclient: fix ino returned in fstat for 32 bit archYehuda Sadeh2009-04-071-1/+1
|
* kclient: don't alter mds_wanted caps if dropping capsYehuda Sadeh2009-04-071-1/+4
|
* vstart: kill bashismsSage Weil2009-04-071-30/+29
|
* kclient: remove alignment restrictions on O_DIRECT reads and writesSage Weil2009-04-071-4/+0
| | | | | | These aren't needed, since we aren't restricted by DMA to a hardware disk device or any such thing. (And even if they were, it'd probably be sector alignment, not page alignment.)
* mds: fix compilation errorYehuda Sadeh2009-04-071-1/+1
|
* mds: bump mds protocolYehuda Sadeh2009-04-071-1/+1
|
* mds: add CEPH_CAP_OP_REVOKEYehuda Sadeh2009-04-074-1/+5
|
* kclient: clear di->lease_session when dropping sessionYehuda Sadeh2009-04-071-0/+1
|
* mds: flush log when expanding caps and filelock is unstableYehuda Sadeh2009-04-061-1/+8
|
* mds, kclient: add CEPH_MDS_OP_CREATE operationYehuda Sadeh2009-04-064-5/+15
|
* kclient: implement permission iopSage Weil2009-04-053-0/+17
| | | | | | | | | The permission iop verfies we have a valid uid/gid/mode by checking for the AUTH cap. Then calls the generic helper. This is 100% consistent because we don't carry a cap ref over the permission checkout. OTOH, if we don't hold the cap and the mds doesn't actually issue it, we never will anyway. Maybe fix this eventually.
* kclient: make do_getattr take inode, not dentrySage Weil2009-04-053-13/+10
|
* kclient: d_alloc_anon on older kernelsSage Weil2009-04-051-0/+4
|
* mds: make sure kick wasn't immediatley fruitful in rdlock_trySage Weil2009-04-041-0/+3
|
* mds: simplify EMetaBlob interface - don't pass explicit inode_t pointerSage Weil2009-04-045-26/+25
| | | | It is always in->get_projected_inode().
* mds: fix journal replaySage Weil2009-04-045-16/+53
| | | | | Load root inode before replaying journal. Kill root special case in journal replay. Replay root inode if it is journaled.
* mds: journal projected inode values by defaultSage Weil2009-04-041-10/+15
| | | | There was a bug in here that was journaling _old_ inode values. Doh!
* vstart: run valgrind properlySage Weil2009-04-041-37/+45
|
* todoSage Weil2009-04-041-6/+4
|
* mds: make mds restart workSage Weil2009-04-041-38/+80
| | | | Root mds needs to load root inode, .ceph directories.
* mon: fix mds stopping and restartSage Weil2009-04-042-10/+12
| | | | | Remove stopped mds from last_beacon (monitoring) set. Fix mdsmap dump output to include stopping set. Remove mds from stopped set on start.
* mds: fix permissions checks on root, system inodesSage Weil2009-04-041-3/+12
|
* mds: journal root inode changesSage Weil2009-04-044-17/+69
| | | | Journal root inode changes, and flush them when trimming the log segment.
* mds: inode fetch/storeSage Weil2009-04-042-0/+110
|
* objecter: only send one ping per laggy osdSage Weil2009-04-041-1/+6
|
* objecter: add support for compound read ops (ObjectRead)Sage Weil2009-04-043-75/+105
|
* osd: implement GETXATTR read opSage Weil2009-04-041-0/+15
|
* kclient: better readdir debug outputSage Weil2009-04-032-5/+8
|
* mds: make rdlock_try kick the lock if it blocksSage Weil2009-04-032-13/+23
| | | | | We were strangely getting by even though it wasn't changing the lock state. Weird.
* mds: issue AUTH_EXCL on mkdirSage Weil2009-04-031-0/+2
| | | | Client is clearly an exclusive user, and will often chmod.
* kclient: drop LINK_RDCACHE cap on rename source inodeSage Weil2009-04-031-0/+2
| | | | | | | This is only needed because the mds locks the srci linklock. It has to lock _some_ field on the source inode in order to make and propagate the ctime change; the linklock is generally the least contended. So humor the mds...
* kclient: update cap->mds_wanted on revocation replySage Weil2009-04-031-0/+1
| | | | | Make sure our local mds_wanted reflects the value sent off to the mds during a revocation.
* mds: do not want LINK caps for any open file; no non-FILE caps when readonlySage Weil2009-04-031-8/+3
| | | | | | | | When readonly, we don't want anything other than FILE caps. (The mds will of course give us more when it can.) For WR and RDWR, we still want AUTH and XATTR caps, for now, since there is some possibility that the client will want to change them.
* kclient: drop any unwanted caps when unlinking a fileSage Weil2009-04-031-1/+20
| | | | | | | | | Since we can't reopen an unlinked file, there's no reason to hold on to them. And the MDS will often want to lock to update max_size, which could mean spurious revocations. At the very least, make sure we drop the LINK caps. Do the same for the rename target, if it exists.
* kclient: send messages with default priority (to match mds)Sage Weil2009-04-031-1/+1
| | | | | | | Make our default priority match the mds. This is only important because we occasionally resend and incoming message out again without fixing the priority to match our default, which means out-of-order processing on the other end.
* msgr: show seq # of incoming messagesSage Weil2009-04-031-0/+1
|
* mds: alloc wrlock by EXCL clientSage Weil2009-04-031-1/+1
| | | | Not sure why this was off. It is allowed in EXCL_MIX...
* mds: don't issue new caps with a revocationSage Weil2009-04-031-1/+5
| | | | This is a really just a heuristic...
* kclient: queue inode for cap check if mds_wanted mismatchSage Weil2009-04-032-1/+18
| | | | | | | | If mds_wanted in add_cap is more than we actually want (for example, on a getattr that races with a cap wanted release), requeue a cap check. We don't want to release immediately if we can help it because something like readdir would prematurely (?) release caps we're holding on to for good measure...
* mds: fix wanted updatesSage Weil2009-04-031-5/+11
| | | | | | | | Check issue_seq, not seq, for wanted updates. We're primarily worried about racing against a lookup+open, which means we only need to worry about an _issue_ (not a cap grant or revocation). Allow wanted _expansion_ even if seq mismatch, though!
* mds: clean up issue_new_caps interfaceSage Weil2009-04-033-21/+10
|
* kclient: fix session leakYehuda Sadeh2009-04-031-5/+6
|
* kclient: grab inode before queue_cap_snap; cleanupYehuda Sadeh2009-04-032-3/+6
|
* kclient: fix some log messagesYehuda Sadeh2009-04-031-2/+6
|
* uclient: mksnap/rmsnap via mkdir/rmdirSage Weil2009-04-032-40/+4
|
View Lorry Controller Status