Whamcloud - gitweb
tappro [Sat, 26 Feb 2005 08:56:34 +0000 (08:56 +0000)]
HLD document for SMFS(Storage Management Filesystem)
adilger [Sat, 26 Feb 2005 08:14:35 +0000 (08:14 +0000)]
Add b_cmd* to buildcvs to get correct portals version.
yury [Fri, 25 Feb 2005 19:45:18 +0000 (19:45 +0000)]
- locks and dentries should be released in the case request is being handled by wrong MDS
and should be restared with correct one in mds_getattr_lock().
eeb [Thu, 24 Feb 2005 17:32:39 +0000 (17:32 +0000)]
* vibnal HIGHMEM hacking
eeb [Thu, 24 Feb 2005 15:35:33 +0000 (15:35 +0000)]
* simpified vibnal rdma setup
eeb [Thu, 24 Feb 2005 14:57:37 +0000 (14:57 +0000)]
* Fixed the vibnal resilience-in-the-face-of-peer-crashes issue
alex [Thu, 24 Feb 2005 11:57:06 +0000 (11:57 +0000)]
- tuning ported from b1_4
eeb [Thu, 24 Feb 2005 11:01:55 +0000 (11:01 +0000)]
* lustre updates for recent portals HEAD changes
- added vib
- zeroconf vib, ra
- lconf/lmc openib,vib changes
eeb [Wed, 23 Feb 2005 21:43:25 +0000 (21:43 +0000)]
* First cut working vibnal
eeb [Wed, 23 Feb 2005 21:34:42 +0000 (21:34 +0000)]
* removed unused #defines in RANAL
* fixed a misleading comment in RANAL
yury [Wed, 23 Feb 2005 14:59:02 +0000 (14:59 +0000)]
- fix from b1_4 in MDS_CHECK_RESENT()
- more fixes about fid extracting on MDS. By now correct fid is always should
be returned to client.
adilger [Wed, 23 Feb 2005 12:05:27 +0000 (12:05 +0000)]
Branch: b1_4
Sort list alphabetically, like it says in the comment.
liuduo [Wed, 23 Feb 2005 06:59:26 +0000 (06:59 +0000)]
add some code related to setxid
liuduo [Wed, 23 Feb 2005 03:35:36 +0000 (03:35 +0000)]
to make b_hd_remote_uid portals based on b_hd_sec portals
alex [Tue, 22 Feb 2005 19:28:49 +0000 (19:28 +0000)]
- commit from zzeng
yury [Tue, 22 Feb 2005 16:37:48 +0000 (16:37 +0000)]
- changes with names on exports and another fields in llite and mds. lov_exp is
renamed to dt_exp and lmv_exp is renamed to md_exp. Related changes.
- removed OBD_MD_FID. And fid is now fetched on MDS on all constructive requests. It
is not fetched on unlink(), rename(), etc. This is because OBD_MD_FID flag lost
the sense as lmv layer needs it for all operations (to check if object is splitted).
adilger [Fri, 18 Feb 2005 23:38:51 +0000 (23:38 +0000)]
Branch: b1_4
Ignore generated files.
nic [Thu, 17 Feb 2005 20:43:11 +0000 (20:43 +0000)]
ia64 needs make vmlinux on 2.6-suse, and it works on rhel-2.4 as well
jacob [Tue, 15 Feb 2005 21:31:18 +0000 (21:31 +0000)]
b=5684
r=adilger
First stab at improving some user visible error messages.
nic [Tue, 15 Feb 2005 06:07:26 +0000 (06:07 +0000)]
need these changes on HEAD for build too
eeb [Mon, 14 Feb 2005 15:30:19 +0000 (15:30 +0000)]
* Added loopback optimisation to lib-move.c
lctl --net ??? loopback # show current state
lctl --net ??? loopback on # enable loopback optimisation
lctl --net ??? loopback off # disable loopback optimisation
By default loopback optimisation is turned OFF
eeb [Sat, 12 Feb 2005 11:11:53 +0000 (11:11 +0000)]
* fixed bug in handling timed out RapkSetRiParams()
phil [Sat, 12 Feb 2005 06:23:38 +0000 (06:23 +0000)]
I keep finding these trees that I have, with little uncommitted copyright
changes, and wondering how that happened.
eeb [Sat, 12 Feb 2005 03:23:36 +0000 (03:23 +0000)]
* Coded for new connection handshake.
eeb [Fri, 11 Feb 2005 20:22:44 +0000 (20:22 +0000)]
* ghastly hack to serialise RapkSetRiParams with comms scheduler thread
* Ranal "working" @ 1.2GBytes/sec single channel, 2.4GBytes/sec double
nic [Fri, 11 Feb 2005 02:52:13 +0000 (02:52 +0000)]
only build ksocknal with modules enabled
jacob [Fri, 11 Feb 2005 01:24:30 +0000 (01:24 +0000)]
originally by phil
do not dist non-linux (darwin) files by default.
phil [Thu, 10 Feb 2005 23:09:22 +0000 (23:09 +0000)]
removed some empty files (authors, news, readme)
added a license file
fixed up some stray copyright notices
adilger [Thu, 10 Feb 2005 22:23:29 +0000 (22:23 +0000)]
Fix LC_COLLATE so that [A-Z] only matches upper-case letters.
eeb [Wed, 9 Feb 2005 12:33:10 +0000 (12:33 +0000)]
* Bug 5676 fix: keep extra ref on peer until all done in
ksocknal_del_peer_locked CVS:
eeb [Tue, 8 Feb 2005 01:15:18 +0000 (01:15 +0000)]
* removed spurious 'inline' from extern portal_ioctl_getdata() decl
liangzhen [Sat, 5 Feb 2005 08:47:01 +0000 (08:47 +0000)]
b_port_step is a branch only for portals
jacob [Sat, 5 Feb 2005 01:18:13 +0000 (01:18 +0000)]
i did not mean for this to get committed
jacob [Sat, 5 Feb 2005 00:53:52 +0000 (00:53 +0000)]
sort alphabetically
jacob [Sat, 5 Feb 2005 00:53:16 +0000 (00:53 +0000)]
remove arch symlink
jacob [Sat, 5 Feb 2005 00:52:19 +0000 (00:52 +0000)]
Add empty install-data-hook rule on non-darwin
jacob [Sat, 5 Feb 2005 00:50:12 +0000 (00:50 +0000)]
portals no longer installs list.h
nic [Fri, 4 Feb 2005 23:06:04 +0000 (23:06 +0000)]
add missing headers to dist - fixes problems with rpm builds
eeb [Fri, 4 Feb 2005 19:25:06 +0000 (19:25 +0000)]
* updates to HEAD lustre since landing b_port_step on portals
eeb [Fri, 4 Feb 2005 18:50:00 +0000 (18:50 +0000)]
* Landed portals:b_port_step as follows...
- removed CFS_DECL_SPIN*
just use 'spinlock_t' and initialise with spin_lock_init()
- removed CFS_DECL_MUTEX*
just use 'struct semaphore' and initialise with init_mutex()
- removed CFS_DECL_RWSEM*
just use 'struct rw_semaphore' and initialise with init_rwsem()
- renamed cfs_sleep_chan -> cfs_waitq
cfs_sleep_link -> cfs_waitlink
- fixed race in linux version of arch-independent socknal
(the ENOMEM/EAGAIN decision).
- Didn't fix problems in Darwin version of arch-independent socknal
(resetting socket callbacks, eager ack hack, ENOMEM/EAGAIN decision)
- removed libcfs types from non-socknal header files (only some types
in the header files had been changed; the .c files hadn't been
updated at all).
- Updated lustre b1_4 to match
eeb [Fri, 4 Feb 2005 12:46:52 +0000 (12:46 +0000)]
* ranal passes netregression
jacob [Fri, 4 Feb 2005 00:28:05 +0000 (00:28 +0000)]
b=5576
add LB_CHECK_FILE(S), which work while cross-compiling.
jacob [Wed, 2 Feb 2005 20:07:58 +0000 (20:07 +0000)]
apparently os x's cvs includes the full cvs repo path in
CVS/Repository, but we only want the basename.
eeb [Wed, 2 Feb 2005 17:17:53 +0000 (17:17 +0000)]
* 5639 fix by having qswnal detect when portals is discarding
jacob [Wed, 2 Feb 2005 04:29:51 +0000 (04:29 +0000)]
complain if more than one argument was passed, so that people who do ./lustrecvs co -r foo get an error as soon as practical
eeb [Tue, 1 Feb 2005 11:58:01 +0000 (11:58 +0000)]
* ranal debugging and lconf support
eeb [Mon, 31 Jan 2005 14:34:05 +0000 (14:34 +0000)]
* 5630 fix takes ibnal global lock at raised IRQ priority
eeb [Mon, 31 Jan 2005 12:30:58 +0000 (12:30 +0000)]
* 5602 fix improves checks that NID is set correctly and causes incorrect
configurations to fail rather than trip a server assertion.
* 5606 fix ensures both peers agree on maximum "immediate" get.
* ranal updates
jacob [Fri, 28 Jan 2005 23:18:13 +0000 (23:18 +0000)]
fix copy/paste error
jacob [Fri, 28 Jan 2005 19:48:29 +0000 (19:48 +0000)]
b=5584; only install module if it is built
yury [Fri, 28 Jan 2005 17:20:46 +0000 (17:20 +0000)]
- fixed problem with sid of virtual directory "FIDS". By now it is saved to store just like last_rcvd and last_fid and read from there on startup.
- a lot of changes about last_fid management. last_fid_changed flag is removed and last_fid_sem is
replaced by spin lock.
nkj [Fri, 28 Jan 2005 15:11:42 +0000 (15:11 +0000)]
re-applying my initial changes. forgot that the HEAD is for 2.6 kernels only
so there is no need to accomodate for 2.4 kernel config differences.
nkj [Fri, 28 Jan 2005 12:03:17 +0000 (12:03 +0000)]
2.4 kernels do not have the ext3 xattr config flag that 2.6 kernels have.
So it must only warn, not abort.
yury [Thu, 27 Jan 2005 09:21:55 +0000 (09:21 +0000)]
- removed optimization related to fid's updating on clients.
huzf [Thu, 27 Jan 2005 08:32:04 +0000 (08:32 +0000)]
add b_port_test branch
wangdi [Wed, 26 Jan 2005 16:12:09 +0000 (16:12 +0000)]
Branch:HEAD
remove unnecessary error msg, cleanup
nkj [Wed, 26 Jan 2005 14:29:44 +0000 (14:29 +0000)]
changed from warn to error when ext3 attributes are not enabled in the kernel
yury [Wed, 26 Jan 2005 10:28:58 +0000 (10:28 +0000)]
- initializing parent_lockh in mds_open() the same way as in other places.
- calling mds_finish_transno() in mds_reint_unlink() in all cases, not only in
cleanup_release == 1. This may cure problem we have observed yesterday (about
duplicating fids for parent and child).
- removed redundant lockh initializing in few places.
- release locks the same way as in other places (using lockh + 1, not &lockh[1]).
yury [Wed, 26 Jan 2005 09:57:45 +0000 (09:57 +0000)]
- removed redundant check and lock release. If name is not passed to mds_id2locked_dentry()
parent lock is not taken, so we do not need to release it.
yury [Wed, 26 Jan 2005 08:37:37 +0000 (08:37 +0000)]
- fixed releasing lock with correct mode in mds_id2locked_dentry()
jacob [Tue, 25 Jan 2005 18:34:23 +0000 (18:34 +0000)]
b=5588
add check for sys/ioctl.h that was missed when moving to portals-split.
(also fix longstanding typo when checking for running kernel header)
jacob [Tue, 25 Jan 2005 17:15:51 +0000 (17:15 +0000)]
b=5592; set CC in lbuild to work around this problem
wangdi [Tue, 25 Jan 2005 06:46:01 +0000 (06:46 +0000)]
Branch: HEAD
fix debug_daemon size problem
wangdi [Mon, 24 Jan 2005 09:01:29 +0000 (09:01 +0000)]
Branch: HEAD
a typo fix(forget goto) for debug_daemon utils
wangdi [Mon, 24 Jan 2005 03:50:48 +0000 (03:50 +0000)]
Branch: HEAD
temp fix debug_daemon bug on HEAD
niu [Fri, 21 Jan 2005 04:19:18 +0000 (04:19 +0000)]
let child branch of bcray use same libsysio.
jacob [Fri, 21 Jan 2005 01:36:30 +0000 (01:36 +0000)]
b=5578
add pseries64 support to rpms etc.
jacob [Thu, 20 Jan 2005 21:12:46 +0000 (21:12 +0000)]
b=5573
fix savedheaders when we have multiple target-archs in a given target.
this probably breaks srource rpms if we were to build i586 and i686,
but we don't do that (yet).
wangdi [Thu, 20 Jan 2005 14:04:11 +0000 (14:04 +0000)]
Branch: HEAD
set o_fid in mdc_obj_create for LBUG in mdt_obj_create
b:5234
jacob [Wed, 19 Jan 2005 22:43:54 +0000 (22:43 +0000)]
b=5568
don't install lonal when building with cray portals
adilger [Wed, 19 Jan 2005 19:39:44 +0000 (19:39 +0000)]
Branch: b_cray
Fix typo from CVS reorg.
b=5568
alex [Wed, 19 Jan 2005 16:40:55 +0000 (16:40 +0000)]
b=5209
the patch contains:
1) journal_release_buffer() doesn't cause journal overflow
2) journal_commit_thread() uses persistent buffer instead of local wbuf
(could cause journal overflow also)
3) optimization in journal_forget() allows to avoid needless commits
alex [Wed, 19 Jan 2005 16:24:42 +0000 (16:24 +0000)]
- minor correction to the patch for gcc-3.4.2
adilger [Mon, 17 Jan 2005 23:45:04 +0000 (23:45 +0000)]
Branch: b_cray
Add CRAY_PORTALS-specific debugging hooks.
brian [Mon, 17 Jan 2005 21:32:24 +0000 (21:32 +0000)]
Fixes from b_port_step needed to get HEAD to build.
r=jacob
r=alex
brian [Mon, 17 Jan 2005 20:49:37 +0000 (20:49 +0000)]
Remove trailing /lustre from lustre/ldlm/lustre pathnames.
r=jacob.
nic [Wed, 12 Jan 2005 20:54:10 +0000 (20:54 +0000)]
r=boc
use the kernelsmp.files for smp kernels
adilger [Wed, 12 Jan 2005 19:46:31 +0000 (19:46 +0000)]
Branch: b1_4_smallfix
Two compile fixes for recent obdfilter changes.
Compile warning fixups for 2.6/x86_64.
b=5545
jacob [Wed, 12 Jan 2005 04:41:15 +0000 (04:41 +0000)]
- fix .config file not found false error message
- work around red hat bug #14463: nested %if doesn't work right with
rpm circa rhl 6.2-9 (2000-2003)
jacob [Tue, 11 Jan 2005 21:57:25 +0000 (21:57 +0000)]
update all files in the lustre root dir; not just lustrecvs
eeb [Tue, 11 Jan 2005 20:53:51 +0000 (20:53 +0000)]
* Added b_ptl_smallfix into build/buildcvs
jacob [Tue, 11 Jan 2005 19:09:59 +0000 (19:09 +0000)]
prune directories on co
jacob [Tue, 11 Jan 2005 18:17:59 +0000 (18:17 +0000)]
Remove always false expression that may be confusing rpm on rhl 9
jacob [Tue, 11 Jan 2005 17:12:44 +0000 (17:12 +0000)]
- lustrecvs now updates and reexecs itself
- set LC_COLLATE=C otherwise [a-z] and [A-Z] don't work how we expect
them to
adilger [Tue, 11 Jan 2005 16:52:31 +0000 (16:52 +0000)]
Remove [A-Z]* regexp since it matches b* for some reason.
eeb [Tue, 11 Jan 2005 15:41:58 +0000 (15:41 +0000)]
* added more openib automagic compile checks
eeb [Tue, 11 Jan 2005 14:09:40 +0000 (14:09 +0000)]
* Minor openibnal cleanups
phil [Tue, 11 Jan 2005 13:10:59 +0000 (13:10 +0000)]
We only really care about this at one place (NERSC, which uses x86_64), so
do the reasonable thing and enable multi-frag only on non-HIGHMEM platforms,
where there's no kmap-deadlock risk.
eeb [Tue, 11 Jan 2005 03:37:43 +0000 (03:37 +0000)]
* Made openib not use the subnet manager to discover connection parameters
+ Openib listens on a TCP/IP port for service queries and responds with the
service id, port GID and pkey.
+ Openib peer table entries have become (NID, IP, port) tuples, where IP
and port connect to the peer's service query server.
+ /proc interfaces for port, listener timeout and backlog (restarts kernel
listener on update).
+ lmc/lconf support for new openib peers, including only running acceptor
if net == 'tcp'
+ Changed connection daemon to conduct both sides of the service query as
well as IB connection establishment. Spawning several since tcp/ip reads
can block for a timeout.
+ Added a reaper thread to do connection cleanup and timeout checks (the
single connection daemon used to do that too).
* Removed some unused lconf default constants
* Better openib automagic compilation check.
* Consistent usage (flipping + optional checksum) of all openib messages
(service query, connection requests and "normal" message flow).
* Exhaustive openib "are-you-still-the-same-person" checks
* Fixed bug which stopped idle persistent peers from getting removed from the
peer table.
* Fixed some ranal bugs which became obvious when similar problems were
debugged in openibnal
adilger [Mon, 10 Jan 2005 23:44:02 +0000 (23:44 +0000)]
Add support for building with a specific CVS Tag. For versions before v1_4_1
we assume no tag in libsysio, for future versions they will have a specific
libsysio tag.
jacob [Mon, 10 Jan 2005 18:03:16 +0000 (18:03 +0000)]
Create -obj's version.h with the correct version in the kernel-source rpm
jacob [Mon, 10 Jan 2005 18:02:21 +0000 (18:02 +0000)]
Don't run make dep for 2.6
eeb [Sun, 9 Jan 2005 19:59:19 +0000 (19:59 +0000)]
* propagated openibnal fix (deleting an idle persistent peer) to voltaire
and infinicon NALs
eeb [Sun, 9 Jan 2005 19:52:51 +0000 (19:52 +0000)]
* Fixed bug in destroying idle persistent peer
nic [Sat, 8 Jan 2005 00:30:25 +0000 (00:30 +0000)]
r=boc
update lbuild to play nicely with the new CVS layout
jacob [Fri, 7 Jan 2005 19:42:55 +0000 (19:42 +0000)]
b=5514
r=adilger
Remove full paths of files in debug logs.
eeb [Fri, 7 Jan 2005 16:41:51 +0000 (16:41 +0000)]
* fixed openib comment typo
* better checking/error messages on OpenIB automagic
phil [Fri, 7 Jan 2005 13:55:31 +0000 (13:55 +0000)]
- pull the new libsysio tag for b_cray
- we've only had this a week and it's already out of alphabetical order?
can't you people read?
phil [Fri, 7 Jan 2005 13:46:45 +0000 (13:46 +0000)]
This commit was generated by cvs2svn to compensate for changes in r46154,
which included commits to RCS files with non-trunk default branches.
phil [Fri, 7 Jan 2005 13:46:45 +0000 (13:46 +0000)]
new upstream libsysio snapshot (
20041101)