Whamcloud - gitweb
jitendra [Wed, 13 Jun 2007 07:28:20 +0000 (07:28 +0000)]
This update includes Hex error ID's and checksum calculation for console error messages.
It also includes code fixes for first five improved error messages.
b=10595
i=adilger
i=braam
scjody [Wed, 13 Jun 2007 01:09:07 +0000 (01:09 +0000)]
Branch HEAD
Move to new ldiskfs series, formerly known as ldiskfs2.
b=12068
scjody [Wed, 13 Jun 2007 00:37:07 +0000 (00:37 +0000)]
Update .cvsignore files
scjody [Tue, 12 Jun 2007 23:39:31 +0000 (23:39 +0000)]
Branch b1_6
Add --disable-modules to fix LTS build failure.
b=12068
scjody [Tue, 12 Jun 2007 22:34:31 +0000 (22:34 +0000)]
Branch b1_6
Remove CVS test file
scjody [Tue, 12 Jun 2007 22:29:18 +0000 (22:29 +0000)]
Branch b1_6
Add files missed in last commit. AFAIK I did things right but CVS shows
them as "Status: Entry Invalid" and won't let me add them except in a
fresh checkout.
b=12068
i=brian
i=mjmac
scjody [Tue, 12 Jun 2007 22:25:53 +0000 (22:25 +0000)]
cvs test
scjody [Tue, 12 Jun 2007 22:00:37 +0000 (22:00 +0000)]
Branch b1_6
Move ldiskfs to its own package and CVS module.
b=12068
i=mjmac
i=brian
isaac [Tue, 12 Jun 2007 19:17:12 +0000 (19:17 +0000)]
- should init kptllnd_data.kptl_idle_txs before calling kptllnd_shutdown.
- should init kptllnd_data.kptl_ptlid2str_lock before calling kptllnd_ptlid2str.
scjody [Mon, 11 Jun 2007 21:51:43 +0000 (21:51 +0000)]
Branch HEAD
Fix weird build issue
isaac [Mon, 11 Jun 2007 09:42:20 +0000 (09:42 +0000)]
i=eeb:
- gmlnd ignored some transmit errors when finalizing lnet messages.
mjmac [Fri, 8 Jun 2007 18:53:20 +0000 (18:53 +0000)]
b=12348
i=scjody
i=adilger
* Create a lustre-tests RPM that includes stuff from lustre/tests
scjody [Fri, 8 Jun 2007 17:44:49 +0000 (17:44 +0000)]
Branch HEAD
Initial commit of new ldiskfs module. This is based on ldiskfs from
b1_6; patches and series files are identical to those found there.
b=12068
i=mjmac
i=brian
mjmac [Fri, 8 Jun 2007 15:49:31 +0000 (15:49 +0000)]
b=12348
i=scjody
i=adilger
* Create a lustre-tests RPM that includes stuff from lustre/tests
liangzhen [Thu, 7 Jun 2007 19:15:43 +0000 (19:15 +0000)]
Remove semicolon after MACRO do {...} while (0)
isaac [Thu, 31 May 2007 17:02:20 +0000 (17:02 +0000)]
i=eeb
- ptllnd logs a piece of incorrect debug info in kptllnd_peer_handle_hello.
isaac [Wed, 30 May 2007 08:35:45 +0000 (08:35 +0000)]
i=eeb
- should set the_lnet.ln_finalizing before completing messages.
eeb [Fri, 25 May 2007 13:42:27 +0000 (13:42 +0000)]
* changed /proc/sys/lnet/peers and debug peer output so that only routers
show state as "up" or "down" - non-routers show "~rtr".
wangdi [Thu, 24 May 2007 14:01:50 +0000 (14:01 +0000)]
Branch:b_hd_ctxt
update from b1_6
jxiong [Thu, 24 May 2007 03:01:02 +0000 (03:01 +0000)]
file ext3-raid5-zerocopy.patch was initially added on branch b_HEAD_bug11890.
jxiong [Thu, 24 May 2007 03:01:01 +0000 (03:01 +0000)]
file raid5-zerocopy.patch was initially added on branch b_HEAD_bug11890.
eeb [Wed, 23 May 2007 11:56:33 +0000 (11:56 +0000)]
* ptllnd
- revert default ptltrace_on_timeout to 0 - timeouts can be routine in
redundant router configurations.
- merged minor fixes from Cray XT 2.0
eeb [Wed, 23 May 2007 11:25:46 +0000 (11:25 +0000)]
* Print portals error string in ptllnd warnings/errors
eeb [Mon, 21 May 2007 11:48:14 +0000 (11:48 +0000)]
* userspace (catamount) ptllnd changes
- Error handling
Ensure all communications complete in finite time. Ensure errors cause
clean peer state teardown so that communications can be re-established
after a peer crash.
Note that this does NOT handle reconnection to a failed LNET router, which
is required for routed configurations.
- Environment tunables
PTLLND_DEBUG (boolean, dflt 0) is a global switch to enable/disable debug
features.
PTLLND_TX_HISTORY (int, dflt debug?1024:0) sets the size of the history
buffer.
PTLLND_ABORT_ON_PROTOCOL_MISMATCH (boolean, dflt 1) calls abort on
connecting to a peer running a different version of the ptllnd protocol.
PTLLND_ABORT_ON_NAK (boolean, dflt 0) abort when a peer sends a NAK
(e.g. because it has timed out this node).
PTLLND_DUMP_ON_NAK (boolean, dflt debug?1:0) dumps peer debug and the
history on receiving a NAK
PTLLND_WATCHDOG_INTERVAL (int, dflt 1) sets how often to check some peers
for timed-out communications while the application blocks for
communications to complete.
PTLLND_TIMEOUT (int, dflt 50) is the communications timeout in seconds.
PTLLND_LONG_WAIT (int, dflt debug?5:PTLLND_TIMEOUT) is a time in seconds
after which the ptllnd prints a warning if it blocks for longer during
connection establishment, cleanup after an error or cleanup during shutdown.
eeb [Mon, 21 May 2007 10:25:30 +0000 (10:25 +0000)]
* cleanup after removing lnet/tests
shadow [Thu, 17 May 2007 06:47:03 +0000 (06:47 +0000)]
update patchless client for support 2.6.20, 2.6.21 and RHEL5 kernels.
b=11647
i=green
i=adilger
adilger [Wed, 16 May 2007 23:19:08 +0000 (23:19 +0000)]
Branch HEAD
Quiet compiler warnings due to printing __u8 instead of char.
adilger [Wed, 16 May 2007 22:58:27 +0000 (22:58 +0000)]
Branch HEAD
Look for dbench client.txt file in /usr/share/dbench also.
adilger [Wed, 16 May 2007 22:57:44 +0000 (22:57 +0000)]
Branch HEAD
Remove extra blank lines.
adilger [Wed, 16 May 2007 22:57:11 +0000 (22:57 +0000)]
Branch HEAD
Run tests in $DIR and not in $MOUNT.
adilger [Wed, 16 May 2007 22:56:23 +0000 (22:56 +0000)]
Branch HEAD
Quiet compile warning when CONFIG_LSF is on (large single file).
adilger [Wed, 16 May 2007 22:33:25 +0000 (22:33 +0000)]
Branch HEAD
Quiet compile warning when CONFIG_LSF is on (large single file).
adilger [Wed, 16 May 2007 22:31:28 +0000 (22:31 +0000)]
Branch HEAD
Quiet warning about use of uninitialized variable.
adilger [Wed, 16 May 2007 22:29:44 +0000 (22:29 +0000)]
Branch HEAD
Fix backticks to be apostrophes in comments.
adilger [Wed, 16 May 2007 22:25:38 +0000 (22:25 +0000)]
Branch HEAD
Change comment so that nobody gets the idea that "#define CDEBUG_ENABLED 0"
is the way to turn this off.
adilger [Wed, 16 May 2007 22:23:07 +0000 (22:23 +0000)]
Branch HEAD
Clarify comment.
zam [Wed, 16 May 2007 05:13:06 +0000 (05:13 +0000)]
own build tag for uOSS branch
zam [Wed, 16 May 2007 03:08:45 +0000 (03:08 +0000)]
undo uOSS-related changes from HEAD branch of the build module.
eeb [Tue, 15 May 2007 17:04:48 +0000 (17:04 +0000)]
Severity : normal
Frequency : rare
Bugzilla : 11472
Description: Changed the default kqswlnd ntxmsg=512
eeb [Tue, 15 May 2007 16:37:46 +0000 (16:37 +0000)]
* Removed lnet/tests - these tests were trivial and not even in use.
eeb [Tue, 15 May 2007 13:58:09 +0000 (13:58 +0000)]
Severity : major
Frequency : rare
Bugzilla : 12458
Description: Assertion failure in kernel ptllnd caused by posting passive
bulk buffers before connection establishment complete.
zam [Mon, 14 May 2007 08:30:04 +0000 (08:30 +0000)]
add LLCPP flags to enable extra functions to use in user-level libcfs code.
adilger [Fri, 11 May 2007 06:36:54 +0000 (06:36 +0000)]
Branch HEAD
Don't run lfscktest.sh if the OSTs are remote.
adilger [Fri, 11 May 2007 00:39:11 +0000 (00:39 +0000)]
Branch HEAD
Fix whitespace difference that caused patch reject.
b=10651
cliffw [Thu, 10 May 2007 23:26:56 +0000 (23:26 +0000)]
Fix missing '$' in scp
cliffw [Thu, 10 May 2007 23:21:04 +0000 (23:21 +0000)]
Replaced portals with LNET
adilger [Thu, 10 May 2007 23:11:58 +0000 (23:11 +0000)]
Branch HEAD
Update release version to 1.1.
cliffw [Thu, 10 May 2007 23:07:27 +0000 (23:07 +0000)]
Updated
adilger [Thu, 10 May 2007 22:48:19 +0000 (22:48 +0000)]
Branch HEAD
Minor doc updates?
adilger [Thu, 10 May 2007 19:14:40 +0000 (19:14 +0000)]
Branch HEAD
Fix the nanosecond timestamp patches to declare all of the missing fields
in the kernel patches.
b=10651
i=green
i=kalpak
adilger [Thu, 10 May 2007 18:38:51 +0000 (18:38 +0000)]
Branch HEAD
Apply the rest of the __iget() changes for patchless clients to
compile out the now-unnecessary parts of the code.
b=12391
i=scjody
i=green
scjody [Thu, 10 May 2007 16:31:22 +0000 (16:31 +0000)]
Branch HEAD
Fix compiler warning on Cray XT3.
Original patch by walter@clusterfs.com.
b=12359
i=scjody
i=adilger
scjody [Thu, 10 May 2007 16:31:00 +0000 (16:31 +0000)]
Branch HEAD
Fix compiler warning.
Original patch by walter@clusterfs.com.
b=12346
i=adilger
r=scjody
scjody [Thu, 10 May 2007 16:21:22 +0000 (16:21 +0000)]
Branch b1_4
Fix compiler warning on Cray XT3.
Original patch by walter@clusterfs.com.
b=12360
i=adilger
i=scjody
johann [Thu, 10 May 2007 15:44:31 +0000 (15:44 +0000)]
Branch HEAD
b=12229
i=adilger
i=nathan
eeb [Thu, 10 May 2007 15:44:28 +0000 (15:44 +0000)]
Severity : major
Frequency : rare
Bugzilla : 12455
Description: A race in kernel ptllnd between deleting a peer and posting
new communications for it could hang communications -
manifesting as "Unexpectedly long timeout" messages.
Severity : major
Frequency : rare
Bugzilla : 12432
Description: Kernel ptllnd lock ordering issue could hang a node.
eeb [Thu, 10 May 2007 11:57:25 +0000 (11:57 +0000)]
* removed lnet router upcalls
adilger [Thu, 10 May 2007 03:41:20 +0000 (03:41 +0000)]
Branch HEAD
Quiet compiler warning.
b=11757
scjody [Thu, 10 May 2007 03:27:34 +0000 (03:27 +0000)]
Branch HEAD
Add 'creds.h', a new include file, so it will appear in the 'make dist'
tarball. Fixes a build problem.
adilger [Wed, 9 May 2007 23:14:30 +0000 (23:14 +0000)]
Branch HEAD
Move #ifdef above now-redundant checks if __iget() is not exported, per
Oleg's inspection comments.
b=12391
i=green
adilger [Wed, 9 May 2007 23:06:10 +0000 (23:06 +0000)]
Branch HEAD
Reverse the default so that health_check writes are off normally.
i=adilger, i=nathan
b=11512
Add missing ChangeLog entries.
scjody [Wed, 9 May 2007 22:44:41 +0000 (22:44 +0000)]
Branch HEAD
Explicitly set inode type on liblustre mdc_create() call
Original patch from bogl@cray.com
b=12253
i=scjody
i=nathan
scjody [Wed, 9 May 2007 22:28:54 +0000 (22:28 +0000)]
Branch b1_4
Clean up libsysio compiler warnings on XT3.
b=12332
i=adilger
i=walter
scjody [Wed, 9 May 2007 22:27:48 +0000 (22:27 +0000)]
Branch b1_4
Update libsysio to latest version.
b=12331
cliffw [Wed, 9 May 2007 20:40:22 +0000 (20:40 +0000)]
b=11512
Updated Change Log
cliffw [Wed, 9 May 2007 20:37:34 +0000 (20:37 +0000)]
b=11512
i=wangdi
i=shadow
Removes write from health check, adds option to configure.
green [Tue, 8 May 2007 19:08:43 +0000 (19:08 +0000)]
b=11546
r=adilger,shadow
Check that request is still eligible for replay in mdc_set_replay_data()
before doing anything.
eeb [Tue, 8 May 2007 14:52:08 +0000 (14:52 +0000)]
Severity : major
Frequency : rare
Bugzilla : 11706
Description: Added LNetSetAsync() to ensure single-threaded userspace
clients can be eager LNET receivers even when the application
is not executing in the filesystem.
nathan [Fri, 4 May 2007 23:32:31 +0000 (23:32 +0000)]
remove b1_6 aliases in anticipation of 1.8 landing
eeb [Fri, 4 May 2007 17:34:17 +0000 (17:34 +0000)]
* Changed the ptllnd module parameter ptltrace_on_timeout default to 1
scjody [Fri, 4 May 2007 16:44:50 +0000 (16:44 +0000)]
Branch HEAD
- Remove RHEL 3 from supported kernels list
- Remove duplicated 12123 entry
- Re-add missing 3244 entry to 1.4.10 section.
nathan [Fri, 4 May 2007 00:05:34 +0000 (00:05 +0000)]
i=adilger
i=johann
b=12404
fix uninitialized flags in cfg_llog_instance
shadow [Thu, 3 May 2007 12:25:57 +0000 (12:25 +0000)]
fix size for generate uuid.
b=10802
i=green
i=adilger
kalpak [Thu, 3 May 2007 12:25:10 +0000 (12:25 +0000)]
b=11757
i=adilger
i=girish
Change llapi_lov_get_uuids() to read the UUIDs from /proc instead
of using an ioctl. This allows lfsck for > 160 OSTs to succeed.
scjody [Wed, 2 May 2007 23:12:52 +0000 (23:12 +0000)]
Branch b1_4
Fix build error on XT3 related to type of length variables.
Original patch by walter.
b=12358
i=adilger
i=scjody
scjody [Wed, 2 May 2007 22:58:06 +0000 (22:58 +0000)]
Branch b1_4
Fix build errors on catamount related to HAVE_LIBPTHREAD not being defined.
Original patch by walter.
b=12147
i=adilger
i=scjody
scjody [Wed, 2 May 2007 22:54:30 +0000 (22:54 +0000)]
Branch b1_4
Fix compiler warnings related to CDEBUG_ENABLED and CDEBUG_ENTRY_EXIT
on Cray XT3.
Original patch by walter.
b=12146
i=scjody
i=adilger
scjody [Wed, 2 May 2007 21:39:14 +0000 (21:39 +0000)]
Branch HEAD
Add LNET router traceability for debug purposes
If a checksum failure occurs with a router as part of the IO path, the NID of
the last router that forwarded the bulk data is printed so it can be
identified.
Original patch by eeb.
b=11548
i=eeb
i=adilger
scjody [Wed, 2 May 2007 19:18:43 +0000 (19:18 +0000)]
Update HEAD ChangeLog based on other branches
cliffw [Wed, 2 May 2007 18:37:00 +0000 (18:37 +0000)]
b=11424
r=scjody
rc=alex
Merge from LLNL chaos tree.
nathan [Wed, 2 May 2007 18:31:35 +0000 (18:31 +0000)]
b=12054
i=scjody
i=green
bounds check all strcat/strcpy for limited buffer length
eeb [Tue, 1 May 2007 15:46:50 +0000 (15:46 +0000)]
* 12016 - fixed race in patchless zero-copy socket teardown
adilger [Tue, 1 May 2007 09:22:38 +0000 (09:22 +0000)]
Branch HEAD
Quiet -Wundef compile warning.
zam [Tue, 1 May 2007 07:29:50 +0000 (07:29 +0000)]
undo autoMakefile.am.toplevel changes that were done for debugging
zam [Tue, 1 May 2007 07:24:53 +0000 (07:24 +0000)]
multi-threaded user-level libcfs: cfs_kernel_tread implemented
using pthread_create.
wangdi [Mon, 30 Apr 2007 23:55:41 +0000 (23:55 +0000)]
Branch: HEAD
Fix bug number and ALWAYS_EXCEPT according to nathan's advices, which should
be line up vertically. Disable test_42 again, it seems bug9789 still there.
lsy [Mon, 30 Apr 2007 08:58:43 +0000 (08:58 +0000)]
b=11315
i=adilger
unmatched spin_lock/unlock of imp_lock might cause deadlock.
scjody [Sat, 28 Apr 2007 02:19:37 +0000 (02:19 +0000)]
Branch HEAD
Patch from nic@cray.com: add spin locks around import/export bit flag changes.
b=11315
i=adilger
i=alex
scjody [Sat, 28 Apr 2007 01:49:11 +0000 (01:49 +0000)]
Branch HEAD
Type-check arguments against format string in CDEBUG/CERROR, fix warnings
introduced.
Original patch by adilger.
b=12327
i=green
i=isaac
shadow [Fri, 27 Apr 2007 17:58:41 +0000 (17:58 +0000)]
LC_CONFIG_CDEBUG don't run while build liblustre on XT3.
b=12146
i=green
i=adilger
yury [Fri, 27 Apr 2007 16:08:36 +0000 (16:08 +0000)]
- removed duplicated LC_CONFIG_CDEBUG from lustre, renamed LC_CONFIG_CDEBUG in lnet to LN_CONFIG_CDEBUG - from Umka and Shadow;
- fixed deadlock in mdt_close() - patch from Mike;
- fixed unknown mode in mdt_open() - patch from Fanyong;
- fixed freeing of not allocated context in ptlrpc_main() - patch from Umka;
- fixed hanging in OST and further issues with module unloading - patch from Shadow;
- fixed hanging on umount with -f - patch from Shadow;
zam [Thu, 26 Apr 2007 20:23:07 +0000 (20:23 +0000)]
wait queue primitives implementation using pthread API
nathan [Thu, 26 Apr 2007 18:07:07 +0000 (18:07 +0000)]
Report which tests were run and the total time taken
nathan [Thu, 26 Apr 2007 16:46:00 +0000 (16:46 +0000)]
b=12007
b=11013
Back out bz 11013 due to chronic test failures
scjody [Wed, 25 Apr 2007 19:47:39 +0000 (19:47 +0000)]
Branch b1_4
Type-check arguments against format string in CDEBUG/CERROR, fix warnings
introduced.
b=12327
i=green
i=isaac
cliffw [Wed, 25 Apr 2007 18:32:08 +0000 (18:32 +0000)]
b=11653
r=adliger,shadow
lctl df output sorted by timestamp. From LLNL.
scjody [Tue, 24 Apr 2007 01:12:07 +0000 (01:12 +0000)]
Branch HEAD
Run these with "bash" not "sh" since they use bash-only features disabled
when newer versions of bash are invoked as "sh".
zam [Mon, 23 Apr 2007 20:21:46 +0000 (20:21 +0000)]
b=11875
modify top-level lnet/include/libcfs/libcfs.h,kp30.h,portals_utils.h
files to conditionally include header files for solaris port.
create Solaris port files and shared posix-*.[ch] files for both
Linux and Solaris ports.
The posix-*.[ch] files are currently noop
(posix threads API calls should be added there).
adilger [Mon, 23 Apr 2007 19:02:51 +0000 (19:02 +0000)]
Branch HEAD
Add inode nr_unused refcount patch to kernel series.
b=12203, b=12181
i=green (original patch), i=shadow
adilger [Mon, 23 Apr 2007 18:26:39 +0000 (18:26 +0000)]
Branch HEAD
Add separator.