Whamcloud - gitweb
fs/lustre-release.git
21 years agoMask off the type bits and not the permission bits.
adilger [Fri, 27 Sep 2002 21:11:53 +0000 (21:11 +0000)]
Mask off the type bits and not the permission bits.

21 years agoSilence bogus compiler warning by moving code closer together.
adilger [Fri, 27 Sep 2002 00:12:52 +0000 (00:12 +0000)]
Silence bogus compiler warning by moving code closer together.

21 years agoAdd update for v0_5_12
adilger [Fri, 27 Sep 2002 00:03:08 +0000 (00:03 +0000)]
Add update for v0_5_12

21 years agoRemove unused variable.
adilger [Thu, 26 Sep 2002 23:59:40 +0000 (23:59 +0000)]
Remove unused variable.

21 years agob=611336
pschwan [Thu, 26 Sep 2002 20:56:00 +0000 (20:56 +0000)]
b=611336
Until we really fix 611336 to never send larger IOV requests than the server
can handle, we should at least avoid _crashing_ the server.

21 years agoI don't much like crashing
pschwan [Thu, 26 Sep 2002 17:55:55 +0000 (17:55 +0000)]
I don't much like crashing

21 years agob=614867
pschwan [Thu, 26 Sep 2002 17:45:30 +0000 (17:45 +0000)]
b=614867
- Remove replies from lock callbacks; Monish verified that this would not harm
the BlueArcs.
- Fixup an overlooked instrumentation detail
- Add an LBUG() to a lock conversion case that I don't think we should ever hit
yet.

21 years agob=613941
pschwan [Thu, 26 Sep 2002 16:37:17 +0000 (16:37 +0000)]
b=613941
Previously, if a filesystem user called readlink() without enough buffer space,
Lustre would allocate a reply packet that's too small and we'd drop the packet
and go into recovery.

Now Lustre sends the maximum size to the server, which will truncate its own
readlink and reply appropriately.

21 years agoDon't forget to distribute the ext3 fixes patch.
gord-fig [Thu, 26 Sep 2002 03:29:58 +0000 (03:29 +0000)]
Don't forget to distribute the ext3 fixes patch.

21 years ago- #include good, implicit-decl warning bad.
shaver [Wed, 25 Sep 2002 23:32:33 +0000 (23:32 +0000)]
- #include good, implicit-decl warning bad.

21 years ago- Use l_wait_event in ldlm_completion_ast to both trigger recovery and make
shaver [Wed, 25 Sep 2002 22:58:09 +0000 (22:58 +0000)]
- Use l_wait_event in ldlm_completion_ast to both trigger recovery and make
  hangs there due to dead MDSs/OSTs interruptible.

21 years agoRemove "helpful" error message, because it screws things up.
adilger [Wed, 25 Sep 2002 22:20:17 +0000 (22:20 +0000)]
Remove "helpful" error message, because it screws things up.

21 years agoAdd more obvious error message after a patch failure, instead of before.
adilger [Wed, 25 Sep 2002 21:43:12 +0000 (21:43 +0000)]
Add more obvious error message after a patch failure, instead of before.

21 years agoReverse order of i_version and extN_mark_inode_dirty tests, for stock kernels.
adilger [Wed, 25 Sep 2002 21:27:19 +0000 (21:27 +0000)]
Reverse order of i_version and extN_mark_inode_dirty tests, for stock kernels.
List the chaos22 patch as an alternative.

21 years agoAutomatically detect when to apply the fixes patch.
gord-fig [Wed, 25 Sep 2002 20:49:08 +0000 (20:49 +0000)]
Automatically detect when to apply the fixes patch.

21 years agoUpdate UML fire-putting-out patch.
gord-fig [Wed, 25 Sep 2002 20:23:49 +0000 (20:23 +0000)]
Update UML fire-putting-out patch.

21 years agoDon't automatically add patches to CVS.
adilger [Wed, 25 Sep 2002 19:59:54 +0000 (19:59 +0000)]
Don't automatically add patches to CVS.
Don't run with set -vx, but echo some status.
Remove "EXTRA" files before trying to apply existing patches.
Formatting.

21 years ago- Remove cancelled locks from the waiting list. Revoking locks from a dead
shaver [Wed, 25 Sep 2002 19:43:33 +0000 (19:43 +0000)]
- Remove cancelled locks from the waiting list.  Revoking locks from a dead
  client to service another client now seems to work quite well.  (The dead
  client can even reconnect successfully.)

21 years agoMake patch generation more accurate with new files.
gord-fig [Wed, 25 Sep 2002 17:52:29 +0000 (17:52 +0000)]
Make patch generation more accurate with new files.

21 years agoFixed Bug #614157, not uncommentig the registration/deregistration part
thantry [Wed, 25 Sep 2002 16:23:00 +0000 (16:23 +0000)]
Fixed Bug #614157, not uncommentig the registration/deregistration part

21 years agoFirst cut at uml40 extN patch.
gord-fig [Wed, 25 Sep 2002 04:06:49 +0000 (04:06 +0000)]
First cut at uml40 extN patch.

21 years ago- add ucred structure. MDS now sets the fsuid and fsgid to that of
braam [Wed, 25 Sep 2002 03:20:04 +0000 (03:20 +0000)]
- add ucred structure.  MDS now sets the fsuid and fsgid to that of
  the caller to enforce permissions, with a new argument to push_ctxt.

21 years ago- Cancel any and all outstanding locks when an export is disconnected.
shaver [Wed, 25 Sep 2002 02:25:26 +0000 (02:25 +0000)]
- Cancel any and all outstanding locks when an export is disconnected.
- Remove icky and unused ldlm_destroy_export and mds_destroy_export hooks.
- Trigger recovery from timed-out lock callbacks.
- Support for replyless requests:
 - add 1 to initial request refcount, balanced in request_out_callback
 - don't set up reply portal buffer if replen is 0
- Ignore reply to DLM blocking/completed ASTs.  (Note: we still very much care
  about cancellation in response to blocking ASTs.)
- Server-side recovery now "simply" forces a disconnect of every export using
  the failed connection.
- Handle (better, not perfectly) the case where we signal failure on a connection
  that is already undergoing recovery.  We need to do more here, but this will
  keep us from going too deeply insane for now.

21 years agoCertified patch hacker I am not. Get the line counts right.
gord-fig [Tue, 24 Sep 2002 22:55:59 +0000 (22:55 +0000)]
Certified patch hacker I am not.  Get the line counts right.

21 years agoAdapt ../extN/ext3-2.4.18-fixes.diff to chaos22.
gord-fig [Tue, 24 Sep 2002 22:45:53 +0000 (22:45 +0000)]
Adapt ../extN/ext3-2.4.18-fixes.diff to chaos22.

21 years agoEnable make diff target for creating kernel-specific patches.
gord-fig [Tue, 24 Sep 2002 22:07:15 +0000 (22:07 +0000)]
Enable make diff target for creating kernel-specific patches.

21 years agoThe old patch was failing because sometimes the inodes were gone by the time
pschwan [Tue, 24 Sep 2002 20:04:00 +0000 (20:04 +0000)]
The old patch was failing because sometimes the inodes were gone by the time
the callback was called with CB_DYING.  I also realized that the client would
believe that it had valid data for too long.

Instead of an extra callback when the lock is freed, I moved that extra
callback to just before we cancel the lock.  Because there's two paths for that
now (cancelling with server notification and cancelling without), I had to
introduce _another_ l_flag, which indicates whether we've done this call.  Oh
well.  It works much better now.

21 years agoUse sizeof(var) instead of sizeof(type).
adilger [Tue, 24 Sep 2002 19:26:10 +0000 (19:26 +0000)]
Use sizeof(var) instead of sizeof(type).

21 years agoAdd debugging to show where we are getting refcounts on obdclass, to aid in
adilger [Tue, 24 Sep 2002 19:25:09 +0000 (19:25 +0000)]
Add debugging to show where we are getting refcounts on obdclass, to aid in
cleanup debugging (sometimes it has refcounts on an error cleanup).

21 years agoFix simple_mknod() to handle existing files properly.
adilger [Tue, 24 Sep 2002 19:23:31 +0000 (19:23 +0000)]
Fix simple_mknod() to handle existing files properly.
Specify the file type when creating a file in obdfilter() to allow checking.

21 years agoUse the OBD_MD_FLTYPE and OBD_MD_FLMODE flags where appropriate.
adilger [Tue, 24 Sep 2002 19:18:17 +0000 (19:18 +0000)]
Use the OBD_MD_FLTYPE and OBD_MD_FLMODE flags where appropriate.

21 years agoAnother minor change to the extN makefile, so it works if the fixes patch
adilger [Tue, 24 Sep 2002 19:01:33 +0000 (19:01 +0000)]
Another minor change to the extN makefile, so it works if the fixes patch
is uncommented.

21 years agoRemove a fix from the "fixes" patch for vanilla kernels which require changes
adilger [Tue, 24 Sep 2002 18:58:32 +0000 (18:58 +0000)]
Remove a fix from the "fixes" patch for vanilla kernels which require changes
to the core VFS code that cannot be copied into our code.

21 years ago- documentation updates for the collaborative cache
braam [Tue, 24 Sep 2002 18:54:54 +0000 (18:54 +0000)]
- documentation updates for the collaborative cache
- NAL document with beginning of I/B NAL design
- all MDS requests have an fsuid and fsgid now to start permission
  checking on the MDS.

21 years agoIgnore the --reformat option after the initial setup step.
adilger [Tue, 24 Sep 2002 18:44:21 +0000 (18:44 +0000)]
Ignore the --reformat option after the initial setup step.
Allow for a small amount of internal filesystem usage for status file.

21 years agoAdd this include for a clean compile.
braam [Tue, 24 Sep 2002 18:01:40 +0000 (18:01 +0000)]
Add this include for a clean compile.

21 years agoMakefile microfix.
shaver [Tue, 24 Sep 2002 17:57:05 +0000 (17:57 +0000)]
Makefile microfix.

21 years agoDamn those continuation characters in Makefiles.
adilger [Tue, 24 Sep 2002 17:51:53 +0000 (17:51 +0000)]
Damn those continuation characters in Makefiles.

21 years agoAdd comment to point people to ext3-2.4.18-fixes.diff in extN/Makefile.am.
adilger [Tue, 24 Sep 2002 17:20:57 +0000 (17:20 +0000)]
Add comment to point people to ext3-2.4.18-fixes.diff in extN/Makefile.am.

21 years agoBack out the piece of the locking fix that caused the problem, until I can
pschwan [Tue, 24 Sep 2002 17:06:33 +0000 (17:06 +0000)]
Back out the piece of the locking fix that caused the problem, until I can
actually fix it.

21 years ago- Adds an int 'local_only' to the cancel_unused function and methods.
pschwan [Tue, 24 Sep 2002 16:24:45 +0000 (16:24 +0000)]
- Adds an int 'local_only' to the cancel_unused function and methods.
  Pass in a non-zero value to cancel the locks locally without notifying the
  server.  Used in recovery.
- Documentation update for the same.

21 years agoAdd definitions of the valid flags.
adilger [Tue, 24 Sep 2002 08:36:01 +0000 (08:36 +0000)]
Add definitions of the valid flags.

21 years agoUpdated to apply to "fixed" ext3 tree from RH kernel.
adilger [Tue, 24 Sep 2002 06:38:36 +0000 (06:38 +0000)]
Updated to apply to "fixed" ext3 tree from RH kernel.
Include patch which can be applied in extN as a prelude to the other
patches, or which can be applied in the source ext3 to fix that also.

21 years agoUpdate patch to apply to RH kernel.
adilger [Mon, 23 Sep 2002 23:50:43 +0000 (23:50 +0000)]
Update patch to apply to RH kernel.

21 years agoUpdate htree patch to apply to the RH kernel. To keep it usable on
adilger [Mon, 23 Sep 2002 23:44:13 +0000 (23:44 +0000)]
Update htree patch to apply to the RH kernel.  To keep it usable on
other kernels (since this change is not yet even in 2.4.20-pre7) we should
have a separate pre-patch which does only the RH fixes themselves to a
stock kernel.

21 years ago* add ignore_error cmd to lctl
rread [Mon, 23 Sep 2002 21:27:35 +0000 (21:27 +0000)]
* add ignore_error cmd to lctl
* use it in lconf during cleanup so errors are non-fatal
* return lctl exit status in lconf properly

21 years agob=612449
pschwan [Sun, 22 Sep 2002 03:32:28 +0000 (03:32 +0000)]
b=612449
Fix two bugs related to file attributes:

1. We were accidentally clobbering the mode with the mode from the OST in
   ll_file_size
2. We were not updating the file attributes in ll_find_inode

Both are fixed, and inode attributes are once again coherent.

21 years agob=605627
pschwan [Sun, 22 Sep 2002 02:47:48 +0000 (02:47 +0000)]
b=605627
Fixes two bugs related to 605627 that have been present since we started
correctly reusing inodes.

- We were giving up locks at file close time, but not throwing away file data
pages.  Fixed, by adding a flag to the blocking callback.  It's now called
under two different circumstances: when a lock needs to be given up
(LDLM_CB_BLOCKING) and when a lock is about to be freed (LDLM_CB_DYING)

- We were not refreshing inode attributes (notably size) correctly.  I brute
force this by always calling ll_file_size() in ll_inode_revalidate, but this
needs some obvious immediate refinement.

As an aside, I noticed that the DLM API documentation gives almost no mention
to the arguments or calling conventions of the callback functions.

21 years ago Modifications for "circulating" request buffers (sized in lustre_net.h)
eeb [Fri, 20 Sep 2002 21:05:50 +0000 (21:05 +0000)]
  Modifications for "circulating" request buffers (sized in lustre_net.h)

21 years agoUse slightly more robust way of making configure depend on lustre.spec.in
gord-fig [Fri, 20 Sep 2002 19:43:07 +0000 (19:43 +0000)]
Use slightly more robust way of making configure depend on lustre.spec.in

21 years ago* attempt to return proper exit codes when errors occur
rread [Fri, 20 Sep 2002 18:43:49 +0000 (18:43 +0000)]
* attempt to return proper exit codes when errors occur
* remove unnecessary --get <url> option

21 years agoFix typo.
adilger [Fri, 20 Sep 2002 16:15:42 +0000 (16:15 +0000)]
Fix typo.

21 years agoAllow the user to specify lmc and config filenames.
gord-fig [Fri, 20 Sep 2002 15:00:05 +0000 (15:00 +0000)]
Allow the user to specify lmc and config filenames.

21 years agoAdd in IT_LINK2.
adilger [Fri, 20 Sep 2002 09:25:15 +0000 (09:25 +0000)]
Add in IT_LINK2.

21 years agoUpdates for v0_5_11
adilger [Fri, 20 Sep 2002 09:18:19 +0000 (09:18 +0000)]
Updates for v0_5_11

21 years agoDon't bug if we run out of space on the MDS when storing an EA.
adilger [Fri, 20 Sep 2002 09:13:28 +0000 (09:13 +0000)]
Don't bug if we run out of space on the MDS when storing an EA.

21 years agoDon't just barf if we can't allocate memory.
adilger [Fri, 20 Sep 2002 09:10:47 +0000 (09:10 +0000)]
Don't just barf if we can't allocate memory.

21 years ago64-bit printout cleanups.
adilger [Fri, 20 Sep 2002 09:09:34 +0000 (09:09 +0000)]
64-bit printout cleanups.

21 years agoDon't return an error code if we already have a reply buffer.
adilger [Fri, 20 Sep 2002 09:08:17 +0000 (09:08 +0000)]
Don't return an error code if we already have a reply buffer.

21 years agoReturn the rename tests to runtests.
adilger [Fri, 20 Sep 2002 09:06:00 +0000 (09:06 +0000)]
Return the rename tests to runtests.
Check for allocation leaks between the start and end of the run (this will
probably complain on BA OSTs, because presumably they are consuming space
for their subdirectories.
Add "-v" verbose option.

21 years agoRe-add memset() of request struct, since we re-use the same struct many times.
adilger [Fri, 20 Sep 2002 08:05:06 +0000 (08:05 +0000)]
Re-add memset() of request struct, since we re-use the same struct many times.

21 years ago- Parallel recovery implementation, to match documented design.
shaver [Fri, 20 Sep 2002 00:38:00 +0000 (00:38 +0000)]
- Parallel recovery implementation, to match documented design.
- Handle failed client recovery by simply restarting recovery.  Not perfect, but
  surprisingly useful for testing.

21 years agoAdd link support to vanilla kernel patch.
adilger [Thu, 19 Sep 2002 23:19:55 +0000 (23:19 +0000)]
Add link support to vanilla kernel patch.

21 years agoSet the EASIZE valid flag in the MDS body, so the filesystem knows there
adilger [Thu, 19 Sep 2002 22:35:34 +0000 (22:35 +0000)]
Set the EASIZE valid flag in the MDS body, so the filesystem knows there
is EA data available.

21 years agoAdd a comment on why we don't return a lock for some operations.
adilger [Thu, 19 Sep 2002 22:33:19 +0000 (22:33 +0000)]
Add a comment on why we don't return a lock for some operations.

21 years agoFix comments, add declaration for simple_mknod().
adilger [Thu, 19 Sep 2002 22:32:37 +0000 (22:32 +0000)]
Fix comments, add declaration for simple_mknod().

21 years ago* add support for gmnal to ptlrpc
rread [Thu, 19 Sep 2002 22:27:27 +0000 (22:27 +0000)]
* add support for gmnal to ptlrpc

21 years ago- add hardlink support
braam [Thu, 19 Sep 2002 20:53:53 +0000 (20:53 +0000)]
- add hardlink support
- update kernel patch
- simple create method for obdfilter instead of filp open
- add mcr-* config files
- add cobd document draft 1.

21 years ago- add collaborative caching document. Radhika: one more pass and add
braam [Thu, 19 Sep 2002 20:50:26 +0000 (20:50 +0000)]
- add collaborative caching document. Radhika: one more pass and add
  pictures before we distribute.

21 years agoAdd --dump option to usage output.
adilger [Thu, 19 Sep 2002 20:12:53 +0000 (20:12 +0000)]
Add --dump option to usage output.

21 years ago- Added preventative locking in two areas, where we're walking shared lists
pschwan [Thu, 19 Sep 2002 18:28:18 +0000 (18:28 +0000)]
- Added preventative locking in two areas, where we're walking shared lists
- Fixed a subtle race condition in ldlm_local_lock_enqueue: we checked to see if
  a completion AST had been received and then acted, with no lock.  This fixes
  the assertion that I was seeing yesterday.

21 years agoadd a note for 0.5.9
braam [Thu, 19 Sep 2002 17:50:26 +0000 (17:50 +0000)]
add a note for 0.5.9

21 years agoDon't allocate so much space on the stack!!!
adilger [Thu, 19 Sep 2002 08:15:23 +0000 (08:15 +0000)]
Don't allocate so much space on the stack!!!

21 years agoAvoid allocating 'event' and 'request' on the stack (saving some 350 bytes),
pschwan [Thu, 19 Sep 2002 08:09:35 +0000 (08:09 +0000)]
Avoid allocating 'event' and 'request' on the stack (saving some 350 bytes),
because there is strong evidence that we are running out of space in some call
paths

21 years agoUpdate uml.xml as generated by a more recent lmc.
adilger [Wed, 18 Sep 2002 23:27:17 +0000 (23:27 +0000)]
Update uml.xml as generated by a more recent lmc.

21 years agoUse $TMP variable, if set, for storing temp loopback files.
adilger [Wed, 18 Sep 2002 23:25:01 +0000 (23:25 +0000)]
Use $TMP variable, if set, for storing temp loopback files.

21 years agoIgnore PtlPut error, so that we will still timeout and trigger recovery.
shaver [Wed, 18 Sep 2002 21:36:41 +0000 (21:36 +0000)]
Ignore PtlPut error, so that we will still timeout and trigger recovery.
(This case occurs when the MDS or OST drops off the elan network, for
example.)

21 years agoChange port numbers to 988, now that BA OSTs also use that port.
adilger [Wed, 18 Sep 2002 20:07:30 +0000 (20:07 +0000)]
Change port numbers to 988, now that BA OSTs also use that port.

21 years agoUpdates for 0.5.8 release.
adilger [Wed, 18 Sep 2002 18:18:52 +0000 (18:18 +0000)]
Updates for 0.5.8 release.

21 years agoincreased request buffering for the OST service
eeb [Wed, 18 Sep 2002 18:17:26 +0000 (18:17 +0000)]
increased request buffering for the OST service

21 years agoProbably memory corruption bug: the file lock bits weren't aware that
pschwan [Wed, 18 Sep 2002 17:47:05 +0000 (17:47 +0000)]
Probably memory corruption bug: the file lock bits weren't aware that
stripe_count==0 is legal, so they were allocating a 0-byte lock array that
would hold one 16-byte lock handle.  Fixed.

21 years agoFixup readme a bit to reflect reality.
adilger [Wed, 18 Sep 2002 16:51:42 +0000 (16:51 +0000)]
Fixup readme a bit to reflect reality.
Add object type to test_getattr obdo, so it works with obdfilter.
Fix up threading again, for "asked-for-threads-but-not-running-threads" case.

21 years agoAllow caller to specify autoconf program to use, in order to support systems
shaver [Wed, 18 Sep 2002 14:39:12 +0000 (14:39 +0000)]
Allow caller to specify autoconf program to use, in order to support systems
with both 2.13 and 2.53.

21 years ago- add manual pages with configuration help.
braam [Wed, 18 Sep 2002 05:44:26 +0000 (05:44 +0000)]
- add manual pages with configuration help.
- of particular importance is lustre-conf.
- this is included as a new chapter in the book.

21 years ago- add templates for manpages.
braam [Wed, 18 Sep 2002 03:33:19 +0000 (03:33 +0000)]
- add templates for manpages.

21 years agoFix "thread" parameters for test_brw and test_getattr for non-threaded tests.
adilger [Tue, 17 Sep 2002 22:38:27 +0000 (22:38 +0000)]
Fix "thread" parameters for test_brw and test_getattr for non-threaded tests.

21 years agoFix problem with shell escaping of variables.
adilger [Tue, 17 Sep 2002 22:30:26 +0000 (22:30 +0000)]
Fix problem with shell escaping of variables.

21 years agoFix badness in test scripts.
adilger [Tue, 17 Sep 2002 22:06:20 +0000 (22:06 +0000)]
Fix badness in test scripts.

21 years agoClean up runtests for inclusion in ltest.
gord-fig [Tue, 17 Sep 2002 21:11:18 +0000 (21:11 +0000)]
Clean up runtests for inclusion in ltest.

21 years agoTodo comment.
adilger [Tue, 17 Sep 2002 20:19:12 +0000 (20:19 +0000)]
Todo comment.

21 years agoLook for compat flag in GS_OPTIONS.
gord-fig [Tue, 17 Sep 2002 20:11:35 +0000 (20:11 +0000)]
Look for compat flag in GS_OPTIONS.

21 years ago* now current with lmc and lconf
rread [Tue, 17 Sep 2002 19:19:39 +0000 (19:19 +0000)]
* now current with lmc and lconf

21 years agob=596503
pschwan [Tue, 17 Sep 2002 19:00:13 +0000 (19:00 +0000)]
b=596503
b=608680
Reenable bits of dentry caching by doing lock matching for LOOKUP intents in
ll_revalidate2.

21 years agoRemove the locking warning added yesterday.
adilger [Tue, 17 Sep 2002 16:43:01 +0000 (16:43 +0000)]
Remove the locking warning added yesterday.

I now see that this was fixed by the "re-get file attributes after we
get the file lock" change, which seemed sort of overkill at the time...

21 years agoUse make distdir to generate a pristine source tree because we need to keep around...
gord-fig [Tue, 17 Sep 2002 16:23:44 +0000 (16:23 +0000)]
Use make distdir to generate a pristine source tree because we need to keep around some files that are clobbered by distclean.

21 years agoAppease tex2pdf.
gord-fig [Tue, 17 Sep 2002 15:24:57 +0000 (15:24 +0000)]
Appease tex2pdf.

21 years agoShould only happen with lctl, but just to be safe...
adilger [Mon, 16 Sep 2002 23:17:29 +0000 (23:17 +0000)]
Should only happen with lctl, but just to be safe...

21 years agoReplace all of the "char[37]" uses with obd_uuid_t.
adilger [Mon, 16 Sep 2002 23:10:40 +0000 (23:10 +0000)]
Replace all of the "char[37]" uses with obd_uuid_t.

21 years agoPrevent false "leak" because of OBD memory accounting.
adilger [Mon, 16 Sep 2002 22:58:58 +0000 (22:58 +0000)]
Prevent false "leak" because of OBD memory accounting.

21 years agoRemove the extra magic from mds_extN.c for now (we will get it back soon).
adilger [Mon, 16 Sep 2002 22:52:30 +0000 (22:52 +0000)]
Remove the extra magic from mds_extN.c for now (we will get it back soon).