Whamcloud - gitweb
fs/lustre-release.git
21 years agoMerge from posix_stable.
adilger [Thu, 20 Jun 2002 18:46:01 +0000 (18:46 +0000)]
Merge from posix_stable.

21 years agoMerge from posix_stable.
adilger [Thu, 20 Jun 2002 18:27:12 +0000 (18:27 +0000)]
Merge from posix_stable.

21 years agoMerge new failure test points from posix_stable.
adilger [Thu, 20 Jun 2002 18:10:37 +0000 (18:10 +0000)]
Merge new failure test points from posix_stable.

21 years agoMerge from posix_stable.
adilger [Thu, 20 Jun 2002 18:05:00 +0000 (18:05 +0000)]
Merge from posix_stable.

21 years agoMerge posix_stable to the head.
adilger [Thu, 20 Jun 2002 17:58:46 +0000 (17:58 +0000)]
Merge posix_stable to the head.

21 years agoAdd ll_pack to the head.
adilger [Thu, 20 Jun 2002 17:55:43 +0000 (17:55 +0000)]
Add ll_pack to the head.

21 years agoMerge from posix_stable branch.
adilger [Thu, 20 Jun 2002 17:48:49 +0000 (17:48 +0000)]
Merge from posix_stable branch.

21 years ago- bring the chaos patch forward to where user mode linux one is.
braam [Thu, 20 Jun 2002 17:46:07 +0000 (17:46 +0000)]
- bring the chaos patch forward to where user mode linux one is.

21 years agohtree patch fix for the head.
adilger [Thu, 20 Jun 2002 17:16:52 +0000 (17:16 +0000)]
htree patch fix for the head.

21 years agoBring stock kernel 2.4.18 intent patch up to date.
braam [Thu, 20 Jun 2002 16:46:11 +0000 (16:46 +0000)]
Bring stock kernel 2.4.18 intent patch up to date.
 - with this patch the intent-test passes on user mode linux

21 years agoRemove debugging from htree patch, sync with branch.
adilger [Wed, 19 Jun 2002 18:41:35 +0000 (18:41 +0000)]
Remove debugging from htree patch, sync with branch.

21 years agofile ll_pack.c was initially added on branch posix_stable.
adilger [Tue, 18 Jun 2002 16:48:36 +0000 (16:48 +0000)]
file ll_pack.c was initially added on branch posix_stable.

21 years agoAnother attempted DLM deadlock fixup
pschwan [Tue, 18 Jun 2002 03:38:03 +0000 (03:38 +0000)]
Another attempted DLM deadlock fixup

21 years ago- Fixed a DLM deadlock bug
pschwan [Tue, 18 Jun 2002 03:12:12 +0000 (03:12 +0000)]
- Fixed a DLM deadlock bug
- Added the leak finding script

21 years ago- Add more DLM threads
pschwan [Mon, 17 Jun 2002 23:51:51 +0000 (23:51 +0000)]
- Add more DLM threads
- Fix unlink() of nonexistant file

21 years ago- Fixed some screwy ldlm resource refcount issues
pschwan [Mon, 17 Jun 2002 22:29:04 +0000 (22:29 +0000)]
- Fixed some screwy ldlm resource refcount issues

21 years agoMany dlm and intent lock fixes.
pschwan [Mon, 17 Jun 2002 19:53:26 +0000 (19:53 +0000)]
Many dlm and intent lock fixes.

 - initial unlink support
 - a DLM trace mode for debugging
 - serious DLM concurrency bugfixes
 - removed vestigial ``local callback'' code from ldlm
 - the beginnings of multi-RPC infrastructure

21 years agoupdated patch to current CVS head of Lustre code.
braam [Mon, 17 Jun 2002 19:38:05 +0000 (19:38 +0000)]
updated patch to current CVS head of Lustre code.

21 years agoMakefile fix for libxml2 headers.
braam [Mon, 17 Jun 2002 15:40:44 +0000 (15:40 +0000)]
Makefile fix for libxml2 headers.

21 years agoSimple example XML file for new obdctl --xml option.
behlendo [Mon, 17 Jun 2002 04:13:53 +0000 (04:13 +0000)]
Simple example XML file for new obdctl --xml option.

21 years agoBarebones XML support added to obdctl via the --xml option. Validation of
brian [Mon, 17 Jun 2002 04:13:19 +0000 (04:13 +0000)]
Barebones XML support added to obdctl via the --xml option.  Validation of
the XML should be performed against the lustre.dtd (but isn't yet).  Also
of note we now need the libxml2 library to be installed.

21 years agoBarebones XML support added to obdctl via the --xml option. Validation of
behlendo [Mon, 17 Jun 2002 04:13:18 +0000 (04:13 +0000)]
Barebones XML support added to obdctl via the --xml option.  Validation of
the XML should be performed against the lustre.dtd (but isn't yet).  Also
of note we now need the libxml2 library to be installed.

21 years agoWMM - Restored difftime use of doubles.
wmarcusm [Sun, 16 Jun 2002 07:19:29 +0000 (07:19 +0000)]
WMM - Restored difftime use of doubles.

21 years agoMove some variable setup to a common function that can be called at cleanup
pschwan [Sun, 16 Jun 2002 06:54:00 +0000 (06:54 +0000)]
Move some variable setup to a common function that can be called at cleanup
in addition to setup--this (I think) fixes the problem with del_uuid being
called with different arguments from the original add_uuid

21 years ago- Fixed serious LDLM bugs, including mistaking the lock _type_ for the lock
pschwan [Sat, 15 Jun 2002 23:38:26 +0000 (23:38 +0000)]
- Fixed serious LDLM bugs, including mistaking the lock _type_ for the lock
_mode_ in llight's file data lock conversion
- Fixed delayed-create-on-open bug
- prepare_write was zeroing out way too much of the page.  fixed.
- when mdc locks are revoked, don't invalidate a non-directory inode's pages
- increase timeouts to 100

21 years agoChaos patch is now good.
braam [Sat, 15 Jun 2002 20:17:14 +0000 (20:17 +0000)]
Chaos patch is now good.

21 years agoNow the patch applies too...
braam [Sat, 15 Jun 2002 14:56:04 +0000 (14:56 +0000)]
Now the patch applies too...

21 years agoAdd slab patches to chaos patch.
braam [Sat, 15 Jun 2002 14:51:24 +0000 (14:51 +0000)]
Add slab patches to chaos patch.

21 years agoThis has a kernel patch for handle validation. All that is missing
braam [Sat, 15 Jun 2002 06:55:32 +0000 (06:55 +0000)]
This has a kernel patch for handle validation.  All that is missing
now for secure pointers is:
 - the random numbers in the handles
 - careful management of when handles are freed

21 years agoInstead of making two 128k request buffers, make ten 64k buffers. Each thread
pschwan [Sat, 15 Jun 2002 06:50:44 +0000 (06:50 +0000)]
Instead of making two 128k request buffers, make ten 64k buffers.  Each thread
can currently pin at most one request buffer, and with six OST threads it was
trivial to pin both request buffers and exhaust our RPC handling ability.

21 years agoStart six ost threads at startup instead of two
pschwan [Sat, 15 Jun 2002 05:35:03 +0000 (05:35 +0000)]
Start six ost threads at startup instead of two

21 years agothe ost was doing commitrw in a callback, which runs in interrupt context
pschwan [Sat, 15 Jun 2002 04:46:38 +0000 (04:46 +0000)]
the ost was doing commitrw in a callback, which runs in interrupt context
in the new elan NAL.  Now we delay the commitrw until all pages have been
received, and hold up the ost handler thread to perform the commitrw there.

21 years agoIncrease RPC timeouts
pschwan [Sat, 15 Jun 2002 03:57:31 +0000 (03:57 +0000)]
Increase RPC timeouts

21 years agoBack out bad kunmap "fix"
pschwan [Sat, 15 Jun 2002 03:21:47 +0000 (03:21 +0000)]
Back out bad kunmap "fix"

21 years ago- Fix two leaks
pschwan [Fri, 14 Jun 2002 23:36:50 +0000 (23:36 +0000)]
- Fix two leaks
- Fix mknod regression
- Fixed one connection leak

21 years agofix broken doc makefile
braam [Fri, 14 Jun 2002 22:57:10 +0000 (22:57 +0000)]
fix broken doc makefile

21 years ago- Fixed remote dlm lock conversion
pschwan [Fri, 14 Jun 2002 21:32:12 +0000 (21:32 +0000)]
- Fixed remote dlm lock conversion
- More intent lock progress

21 years ago- clear dentry->d_it in intent_release
pschwan [Fri, 14 Jun 2002 18:08:27 +0000 (18:08 +0000)]
- clear dentry->d_it in intent_release
- new patches to fix open() crash

21 years ago- We allow the intent policy function to abort lock enqueue
pschwan [Fri, 14 Jun 2002 17:57:38 +0000 (17:57 +0000)]
- We allow the intent policy function to abort lock enqueue
- Return -ENOENT from getattr, not -EEXIST
- add page.c to obdfilter build
- new kernel patches

21 years agoLanding the mds_lock_devel branch on the trunk. Notables:
pschwan [Fri, 14 Jun 2002 17:13:23 +0000 (17:13 +0000)]
Landing the mds_lock_devel branch on the trunk.  Notables:

- Considerable intent locking infrastructure, but it's not working perfectly
  yet.
- Updated documentation
- Many bugs.  Hold on.
- Test script changes to allow multiple mounts

I'll leave the branch around for a day or two in case we somehow missed
anything, then it will die.

21 years agoFix a crash in the POSIX test.
adilger [Fri, 14 Jun 2002 07:49:18 +0000 (07:49 +0000)]
Fix a crash in the POSIX test.

21 years agoAnother part of the statfs story, disabled.
adilger [Fri, 14 Jun 2002 07:30:16 +0000 (07:30 +0000)]
Another part of the statfs story, disabled.

21 years agoProbable fix for storing 64-bit objids on the MDS for ext3. Probably needs
adilger [Fri, 14 Jun 2002 07:26:20 +0000 (07:26 +0000)]
Probable fix for storing 64-bit objids on the MDS for ext3.  Probably needs
to be revised to fit the new LOV format (if there is enough space for it).

21 years agoCreate the filesystem ROOT directory with 0755 permissions (the same as a
adilger [Fri, 14 Jun 2002 03:24:50 +0000 (03:24 +0000)]
Create the filesystem ROOT directory with 0755 permissions (the same as a
normal filesystem).

21 years agoadd intent related update to NFS server.
braam [Thu, 13 Jun 2002 23:08:15 +0000 (23:08 +0000)]
add intent related update to NFS server.

21 years ago- Update md intent locking for chaos and lum patches to latest
braam [Thu, 13 Jun 2002 21:32:32 +0000 (21:32 +0000)]
- Update md intent locking for chaos and lum patches to latest
  version.  Nothing terribly new apart from a first, failed, shot at
  handling renames.
- Slightly improved locking documentation

21 years ago- Added the obdo inline code to the OSC, for the non-lov-using case
pschwan [Thu, 13 Jun 2002 21:05:11 +0000 (21:05 +0000)]
- Added the obdo inline code to the OSC, for the non-lov-using case
- use extN MDS
- fixups to the EA code that I broke earlier

21 years ago- Fixes obdecho
pschwan [Thu, 13 Jun 2002 19:56:41 +0000 (19:56 +0000)]
- Fixes obdecho
- Fixes server-side write leaks
- removes ext2 and ext3 MDS module insertion from common.sh

21 years agoWARNING: This commit breaks everything. It will be back in shape within 12
pschwan [Thu, 13 Jun 2002 19:25:10 +0000 (19:25 +0000)]
WARNING: This commit breaks everything.  It will be back in shape within 12
hours, we hope, but update at your own (extreme) risk.

- Adds more unfinished LOV functionality
- Changes MDS EAs to contain LOV data
- Adds new brw callbacks
- Fixes to work with the Portals tip

21 years agoDo kunmap() on pages kmapp()ed in echo_preprw().
adilger [Wed, 12 Jun 2002 20:12:05 +0000 (20:12 +0000)]
Do kunmap() on pages kmapp()ed in echo_preprw().

21 years agoDon't override MDS return code with transaction close code on error.
adilger [Wed, 12 Jun 2002 20:06:50 +0000 (20:06 +0000)]
Don't override MDS return code with transaction close code on error.

21 years agoDon't open files (i.e. directories) on the MDS with O_DIRECT.
adilger [Wed, 12 Jun 2002 20:05:47 +0000 (20:05 +0000)]
Don't open files (i.e. directories) on the MDS with O_DIRECT.

21 years agoMake sure we clean up all the generated extN files on "make distclean"
adilger [Wed, 12 Jun 2002 20:04:21 +0000 (20:04 +0000)]
Make sure we clean up all the generated extN files on "make distclean"

21 years agoCommented-out mdc_statfs method. Not sure of how to set up all the portals
adilger [Wed, 12 Jun 2002 19:21:58 +0000 (19:21 +0000)]
Commented-out mdc_statfs method.  Not sure of how to set up all the portals
stuff, but it should be close.

21 years agoMinor cleanups.
adilger [Wed, 12 Jun 2002 19:20:53 +0000 (19:20 +0000)]
Minor cleanups.
Add commented-out call to mdc_statfs() for ll_statfs() - it isn't complete.

21 years agoPart of the O_DIRECT fix. The page indices iobuf->maplist[i]->index
adilger [Wed, 12 Jun 2002 19:20:00 +0000 (19:20 +0000)]
Part of the O_DIRECT fix.  The page indices  iobuf->maplist[i]->index
do not appear to contain valid offsets into the file.  Use the "blocknr"
value per generic_direct_IO instead.

21 years agoWhitespace cleanup
adilger [Wed, 12 Jun 2002 19:11:30 +0000 (19:11 +0000)]
Whitespace cleanup

21 years agoDon't dereference niobuf_local page if it is not set in debug code.
adilger [Wed, 12 Jun 2002 19:10:20 +0000 (19:10 +0000)]
Don't dereference niobuf_local page if it is not set in debug code.

21 years agoAdd fsx binary to list of ignored files
adilger [Wed, 12 Jun 2002 19:07:45 +0000 (19:07 +0000)]
Add fsx binary to list of ignored files

21 years agoRemove typo file.
adilger [Wed, 12 Jun 2002 08:27:38 +0000 (08:27 +0000)]
Remove typo file.

21 years agocommon.sh: tell us when we are mounting
adilger [Wed, 12 Jun 2002 07:25:46 +0000 (07:25 +0000)]
common.sh: tell us when we are mounting
runiozone: allow the size parameter to be given on the command-line

21 years agoCheck the f_mode field instead of the f_flags field when using FMODE_ masks.
adilger [Wed, 12 Jun 2002 00:17:40 +0000 (00:17 +0000)]
Check the f_mode field instead of the f_flags field when using FMODE_ masks.

21 years agoMerge changes made in portals/linux/utils/parser.c into lustre/utils/parser.c
adilger [Tue, 11 Jun 2002 19:40:12 +0000 (19:40 +0000)]
Merge changes made in portals/linux/utils/parser.c into lustre/utils/parser.c

21 years agoClean up the right portals network type if it is set (default tcp).
adilger [Tue, 11 Jun 2002 17:24:10 +0000 (17:24 +0000)]
Clean up the right portals network type if it is set (default tcp).

21 years agoRemove compiler warning for fsx on linux.
adilger [Tue, 11 Jun 2002 17:23:29 +0000 (17:23 +0000)]
Remove compiler warning for fsx on linux.

21 years agoAdd connid and xid to timeout messages.
adilger [Mon, 10 Jun 2002 23:38:04 +0000 (23:38 +0000)]
Add connid and xid to timeout messages.

21 years agoSame fix for lustre_get_page_write().
adilger [Mon, 10 Jun 2002 22:39:21 +0000 (22:39 +0000)]
Same fix for lustre_get_page_write().

21 years agoDon't kmap() a this page twice. We kunmap() in lustre_put_page() so we
adilger [Mon, 10 Jun 2002 22:37:52 +0000 (22:37 +0000)]
Don't kmap() a this page twice.  We kunmap() in lustre_put_page() so we
do the kmap() in all cases.

21 years agoAdd prototypes for the debug code in lib/debug.c.
adilger [Mon, 10 Jun 2002 22:35:12 +0000 (22:35 +0000)]
Add prototypes for the debug code in lib/debug.c.

21 years agoDump ->addr as a pointer instead of a long long.
adilger [Mon, 10 Jun 2002 22:34:38 +0000 (22:34 +0000)]
Dump ->addr as a pointer instead of a long long.

21 years agoUse OBD_ALLOC for BRW vector test code.
adilger [Mon, 10 Jun 2002 22:31:34 +0000 (22:31 +0000)]
Use OBD_ALLOC for BRW vector test code.

Turn device name error into a debug message to avoid confusion.

21 years agoAdd fsx test.
adilger [Sat, 8 Jun 2002 01:11:26 +0000 (01:11 +0000)]
Add fsx test.

21 years agoMove VM stuff into lib/page.c.
adilger [Sat, 8 Jun 2002 01:10:25 +0000 (01:10 +0000)]
Move VM stuff into lib/page.c.

21 years agoExit out if there are setup errors.
adilger [Sat, 8 Jun 2002 01:05:58 +0000 (01:05 +0000)]
Exit out if there are setup errors.

21 years agoSave filesystem context before writing to last_rcvd file.
adilger [Sat, 8 Jun 2002 01:03:09 +0000 (01:03 +0000)]
Save filesystem context before writing to last_rcvd file.

21 years agoMajor fixups for multi-page I/Os in filterobd.
adilger [Sat, 8 Jun 2002 00:58:00 +0000 (00:58 +0000)]
Major fixups for multi-page I/Os in filterobd.
A few stubs for statfs on the filesystem.

21 years agoFix some minor llite bugs. Should fix fsx problems.
adilger [Sat, 8 Jun 2002 00:44:02 +0000 (00:44 +0000)]
Fix some minor llite bugs.  Should fix fsx problems.

21 years agoMake obdctl return the error code from the most recently executed command
adilger [Sat, 8 Jun 2002 00:39:35 +0000 (00:39 +0000)]
Make obdctl return the error code from the most recently executed command
in interactive mode.  This allows scripts using HERE documents to abort
properly during setup.

21 years agoTranscription error.
adilger [Fri, 7 Jun 2002 17:22:33 +0000 (17:22 +0000)]
Transcription error.

21 years agoUpdated patch to remove extraneous changes.
braam [Fri, 7 Jun 2002 17:17:49 +0000 (17:17 +0000)]
Updated patch to remove extraneous changes.

21 years agoDon't stop checking for operations to commit if we hit an operation
adilger [Fri, 7 Jun 2002 00:15:44 +0000 (00:15 +0000)]
Don't stop checking for operations to commit if we hit an operation
with reqno == 0.

Peter - the question remains why we aren't already skipping these
open operations because of the PTL_RPC_FL_REPLAY flag?  Is it because
they are closes or something?

21 years agoSubtract the OBD_FREE'd memory from the total before printing it.
adilger [Thu, 6 Jun 2002 21:40:27 +0000 (21:40 +0000)]
Subtract the OBD_FREE'd memory from the total before printing it.

21 years ago- add more handling of intents (for chmod style functions)
braam [Thu, 6 Jun 2002 21:37:58 +0000 (21:37 +0000)]
- add more handling of intents (for chmod style functions)
- update to the head

21 years agoRename "bulk" to "desc" to avoid confusion with the local "bulk" variable.
adilger [Thu, 6 Jun 2002 21:31:39 +0000 (21:31 +0000)]
Rename "bulk" to "desc" to avoid confusion with the local "bulk" variable.

21 years agoChange paths to use Lustre pathname instead of obd
braam [Thu, 6 Jun 2002 19:26:17 +0000 (19:26 +0000)]
Change paths to use Lustre pathname instead of obd

21 years agoBring Makefile.am forward to move extN before mds.
braam [Thu, 6 Jun 2002 19:23:16 +0000 (19:23 +0000)]
Bring Makefile.am forward to move extN before mds.

21 years agoWe weren't freeing local_nb in ost_brw_read(). This becomes obvious if you
adilger [Thu, 6 Jun 2002 08:16:58 +0000 (08:16 +0000)]
We weren't freeing local_nb in ost_brw_read().  This becomes obvious if you
run test_brw with larger vector sizes and large numbers of iterations.

21 years agoRemove ___wait_on_page entirely. If we ever need it again, it will be
adilger [Thu, 6 Jun 2002 08:13:38 +0000 (08:13 +0000)]
Remove ___wait_on_page entirely.  If we ever need it again, it will be
different anyways.

21 years agoGet rid of local definition of ___wait_on_page. It existed at least as far
adilger [Wed, 5 Jun 2002 23:38:04 +0000 (23:38 +0000)]
Get rid of local definition of ___wait_on_page.  It existed at least as far
back as 2.4.0-prerelease and all 2.4.0-test kernels.  It also breaks with
current RH RMAP kernels as well.

21 years ago- fix Marcus' bug. Will move t_last_stable forward.
braam [Wed, 5 Jun 2002 22:48:02 +0000 (22:48 +0000)]
- fix Marcus' bug.  Will move t_last_stable forward.

21 years agoChange the cleanup scripts to use the debug_kernel instead of get_debug.
adilger [Wed, 5 Jun 2002 20:27:41 +0000 (20:27 +0000)]
Change the cleanup scripts to use the debug_kernel instead of get_debug.

21 years agoMove ioctl values out of the way of standard ext2/ext3 ioctl numbers.
adilger [Mon, 3 Jun 2002 22:36:53 +0000 (22:36 +0000)]
Move ioctl values out of the way of standard ext2/ext3 ioctl numbers.

21 years agoAdd llite file flag to allow skipping the locking for file read/write ops.
adilger [Mon, 3 Jun 2002 22:18:57 +0000 (22:18 +0000)]
Add llite file flag to allow skipping the locking for file read/write ops.

The ioctl LL_IOC_SETFLAGS and LL_IOC_CLRFLAGS can be used to set and clear
the LL_FILE_IGNORE_LOCK flag on a file, and LL_IOC_GETFLAGS can be used to
retrieve the current flags value.  These values are defined in lustre_lite.h.

Also add a small optimization - we do not update the file atime on reads if
it would not change the actual atime value (i.e. do not update this field
more often than once per second).  We also avoid updating the atime on R/O
mounted filesystems or if the "noatime" mount option is used.

21 years agoReplace usage of "page" with "bulk" where "page" was actually a "struct
adilger [Mon, 3 Jun 2002 19:57:11 +0000 (19:57 +0000)]
Replace usage of "page" with "bulk" where "page" was actually a "struct
ptlrpc_bulk_page" to avoid (mental) confusion with the kernel "struct page".

21 years agoChange object "refcounts" to be dgets instead of igets, where possible,
adilger [Mon, 3 Jun 2002 19:03:26 +0000 (19:03 +0000)]
Change object "refcounts" to be dgets instead of igets, where possible,
because igets are not only more expensive (need to lock the inode, etc),
but you need to do the dentry lookup to find the inode in the first place.

Also cleaned up access to atomic counters to use "atomic_read(count)"
instead of count.value.  Purely cosmetic, since we don't really need
the atomicness of the counters for debug output.

21 years agoFix block leak on truncate - we weren't setting the correct mode on the
adilger [Fri, 31 May 2002 21:14:49 +0000 (21:14 +0000)]
Fix block leak on truncate - we weren't setting the correct mode on the
file and that was causing the truncate to be skipped.

21 years agoFix the RH-chaos kernel oops. This was caused by trying to enable slab
adilger [Fri, 31 May 2002 11:05:03 +0000 (11:05 +0000)]
Fix the RH-chaos kernel oops.  This was caused by trying to enable slab
poisoning (in order to catch errors where we dereference freed objects).
Sadly, the slab.c code checks for "invalid" flags and BUGs if any are set
(only doing so at runtime) rather than just not defining them and catching
the error at compile time or just ignoring it entirely.

21 years agoFix minor typo in chaos1 patch.
adilger [Fri, 31 May 2002 09:57:03 +0000 (09:57 +0000)]
Fix minor typo in chaos1 patch.

21 years agoRemove debugging that slipped into checkin.
adilger [Thu, 30 May 2002 22:22:04 +0000 (22:22 +0000)]
Remove debugging that slipped into checkin.

21 years agoCheck RPC reply status for both MDC and OSC. This allows us to return
adilger [Thu, 30 May 2002 22:18:41 +0000 (22:18 +0000)]
Check RPC reply status for both MDC and OSC.  This allows us to return
errors "properly" from the MDS and OST to the filesystem/kernel/application.
This finally appears to handle out-of-space errors on the MDS and OST
reasonably well.

Still to do at some point: OS agnostic error value encoding/decoding.