Whamcloud - gitweb
fs/lustre-release.git
22 years ago- remove unused ptlrpc_uuid2conn
shaver [Mon, 24 Mar 2003 20:49:14 +0000 (20:49 +0000)]
- remove unused ptlrpc_uuid2conn
- lose extra dlm-hook-put CERROR
- don't need to cleanup DLM ns or schedule when forcing disconnect: it's already
  cleaned up, or we don't get unmounted
- export and import extra-put assertion checking
- fix connection switching to work when we don't already have a connection created

22 years agopc file for Zach's truncate_complete_page() export patch
pschwan [Mon, 24 Mar 2003 17:41:36 +0000 (17:41 +0000)]
pc file for Zach's truncate_complete_page() export patch

22 years agoAdd Zach's patch to export truncate_complete_page(); still needs an
pschwan [Mon, 24 Mar 2003 17:41:08 +0000 (17:41 +0000)]
Add Zach's patch to export truncate_complete_page(); still needs an
hp-pnnl and vanilla equivalent

22 years agoFixed the patch for intent_release. After this fix lock refs are in order now.
amrutjoshi [Mon, 24 Mar 2003 15:52:19 +0000 (15:52 +0000)]
Fixed the patch for intent_release. After this fix lock refs are in order now.
Most of the metadata ops are working.

22 years agounlinkmany a-la createmany and statmany
pschwan [Sat, 22 Mar 2003 20:18:29 +0000 (20:18 +0000)]
unlinkmany a-la createmany and statmany

22 years ago- don't need to unregister the recovd, since it's gone
rread [Fri, 21 Mar 2003 22:57:24 +0000 (22:57 +0000)]
- don't need to unregister the recovd, since it's gone
- class_destroy_import() is doing an import_put, so
  I removed an extra put from class_obd_cleanup().

22 years agoRefreshed patche with intent_release calls
amrutjoshi [Fri, 21 Mar 2003 20:41:58 +0000 (20:41 +0000)]
Refreshed patche with intent_release calls

22 years agoFixing IT_GETATTR intents and many small changes.
amrutjoshi [Fri, 21 Mar 2003 13:28:52 +0000 (13:28 +0000)]
Fixing IT_GETATTR intents and many small changes.

22 years ago- MDS/client open code changes, to open directories on the MDS.
braam [Fri, 21 Mar 2003 03:59:05 +0000 (03:59 +0000)]
- MDS/client open code changes, to open directories on the MDS.
- no more slab validation on the MDS

22 years agoThese are the kernel patches for Peter's open directory fixes. Old
braam [Fri, 21 Mar 2003 03:56:20 +0000 (03:56 +0000)]
These are the kernel patches for Peter's open directory fixes.  Old
code can probably work with this (we'll allow it in through a small
change in obd_class.c)

22 years agoI believe that this will fix the remaining threaded unlink issues, although
adilger [Thu, 20 Mar 2003 11:02:16 +0000 (11:02 +0000)]
I believe that this will fix the remaining threaded unlink issues, although
I wasn't able to get a chance to test it...

Basically, uncomment extN-delete_thread.diff in extN/Makefile.am and give
it a whirl under dbench and/or runtests or whatever, and if it passes we
are golden.

22 years ago- move the xml/ldap handling class from lconf into an external module
rread [Wed, 19 Mar 2003 20:51:37 +0000 (20:51 +0000)]
- move the xml/ldap handling class from lconf into an external module
- load_ldap.sh - loads a lustre xml config into ldap
- lactive - updates failover targets with new active devices

22 years agoPatches for 2.5.63. This wont work until lustre is patched for 2.5.
amrutjoshi [Sat, 15 Mar 2003 16:20:08 +0000 (16:20 +0000)]
Patches for 2.5.63. This wont work until lustre is patched for 2.5.

22 years agoBugzilla 895
thantry [Sat, 15 Mar 2003 02:36:06 +0000 (02:36 +0000)]
Bugzilla 895

22 years agoDelete thread patch. First, tried to "fake out" the VFS by twiddling bits in
adilger [Sat, 15 Mar 2003 01:32:50 +0000 (01:32 +0000)]
Delete thread patch.  First, tried to "fake out" the VFS by twiddling bits in
the inode to keep it around after it should have been destroyed, but no dice.

Then, I tried to allocate a "mock inode" and copy over the existing inode to
that and use it only for the unlink code.  Sadly, copying list_head,
semaphore, etc does not work, so you have to end up re-initializing the whole
thing anyways, and it would just break on 2.5 anyways.

Finally, I did the "right" thing - read the same inode into a new struct
inode with iget(), and then flag that inode for "real" destruction and
have the delete thread just do an iput.  Very simple, very easy.[*]

I also split the orphan list handling out of the superblock lock into
its own lock, so that we don't get stuck behind the delete thread (which
holds it for long periods doing truncates) when we are trying to add new
inodes to the truncate list.

This code passes basic acceptance testing under UML, but I'm not checking
in the Makefile.am changes that activate it until I give it a shot with
dbench 20 or "rm -r directory_full_of_large_files" so on DEV.  Other
people testing it is of course welcome (just add extN-delete_thread.diff
and ext3-orphan_lock.diff to the end of EXTNP).

[*] It reminds me about a story I heard once, where an engineer who had
    retired, but was on retainer for his old company in case they needed
    him for consulting.  Sure enough, the company's complex oil refinery
    was not working properly, and after the company engineers couldn't
    figure out what was wrong they called the retiree for assistance.

    The retiree walked around the refinery, asking questions, looking at
    valves and guages, etc., until finally he asked for a hammer, gave a
    pipe a swift blow, and told them to fire up the plant again.  Sure
    enough, all was working properly again, and the company was happy.

    Until they got the invoice - $25,000.  In an outrage, they called the
    retiree up and asked how he could charge $25,000 for just hitting a
    pipe with a hammer.  In reply, the engineer said "Hitting the pipe
    with the hammer was only $10, the other $24,990 was for knowing where
    to hit it."

22 years agoBackport of bugfix from 2.5/2.4.21-pre5 which calculates the correct value
adilger [Fri, 14 Mar 2003 21:12:07 +0000 (21:12 +0000)]
Backport of bugfix from 2.5/2.4.21-pre5 which calculates the correct value
for the number of blocks to reserve for a truncate.  We were asking for 8x
as many blocks as we needed, although for large files this was capped at
EXT3_MAX_TRANS_BLOCKS anyways, so not much harm done.

Conditionally applied so when it appears in our upstream kernels we will
not die (when I commit corresponding changes to Makefile.am).

22 years ago- b=959: add connection-switching to import recovery. (Not yet tested: one of
shaver [Wed, 12 Mar 2003 12:51:15 +0000 (12:51 +0000)]
- b=959: add connection-switching to import recovery. (Not yet tested: one of
  my MDS nodes isn't coming up right now.)

22 years agoScript to grab asm output and look for heavy stack (ab)users.
adilger [Wed, 12 Mar 2003 00:58:12 +0000 (00:58 +0000)]
Script to grab asm output and look for heavy stack (ab)users.
Could be improved a bit by having it take a list of files as args and
running objdump and then preceding each line with the filename, so we can
run it on modules more easily than having a shell script call a perl script...

22 years agoadd a extra param to mmap unless it always return error.
nfshp [Tue, 11 Mar 2003 14:45:06 +0000 (14:45 +0000)]
add a extra param to mmap unless it always return error.

22 years agoBug 828 is fixed by Peng Zhao, arrpoved by Andreas.
pengzhao [Tue, 11 Mar 2003 07:17:08 +0000 (07:17 +0000)]
Bug 828 is fixed by Peng Zhao, arrpoved by Andreas.
runas.c allows the root to "runas" another user to do things.

22 years agoRemove liblustre-build-breaking l_wait_event define, and get rid of wait_event
shaver [Mon, 10 Mar 2003 07:22:15 +0000 (07:22 +0000)]
Remove liblustre-build-breaking l_wait_event define, and get rid of wait_event
while I'm at it.

22 years ago- b=977: C-z (and other non-fatal signals) send us into an infinite loop if
shaver [Sun, 9 Mar 2003 21:01:41 +0000 (21:01 +0000)]
- b=977: C-z (and other non-fatal signals) send us into an infinite loop if
         we wait for recovery.  We now use signal blocking to prevent signal
         delivery until we're interested in some or all signals.

- b=988: umount -f hangs when MDS is dead

- b=722: Lustre kernel threads cause load average to skyrocket. (Still get
         a little boost from the socknal threads, but it's much better.)

22 years agooops, fix my last checkin.
nfshp [Sun, 9 Mar 2003 07:56:20 +0000 (07:56 +0000)]
oops, fix my last checkin.

22 years agofix IS_ERR macro: return 0/NULL is sign of success.
nfshp [Sun, 9 Mar 2003 07:53:03 +0000 (07:53 +0000)]
fix IS_ERR macro: return 0/NULL is sign of success.

22 years agoCompletely untested (but compiled) port of noread-creates patch to
adilger [Sat, 8 Mar 2003 15:59:30 +0000 (15:59 +0000)]
Completely untested (but compiled) port of noread-creates patch to
2.5.current.  I suspect it is OK, since the ext3 code hasn't changed much,
but needs testing.

22 years ago- Remove the now-unnecessary connection chaining bits from imports and exports.
shaver [Sat, 8 Mar 2003 02:09:49 +0000 (02:09 +0000)]
- Remove the now-unnecessary connection chaining bits from imports and exports.

- In related news, always drop the exports' and imports' connection refs when
  they're destroyed.

- class_connect was leaking its "caller" (compare: "handle lifetime") ref to
  the new export.  We're going to try without the leak for a while, see how
  it goes.

- Bulk descs now have import and export pointers, so that they can play nicely
  with recovery.

- Clean, informative console diagnostics for recovery cases.  (Hi, Terry!)

- Do the ldlm-hook i_m_g/i_m_p based on emptiness of the conn_list, since it's
  far too late by the time the __exit routines run.

- Put an ugly-but-serviceable UUID in the server's connection's remote_uuid.

- Remember that we can already have an export in req->rq_export upon entry to
  target_handle_connect, if the client and server are the same node, and
  therefore share a handle table.

- Get rid of class_signal_connection_failure, which was vestigial at best.

- Welcome back to the world of connection-sharing.

- Call the upcall like:
  /path/to/upcall $FAILED_IMPORT_UUID $FAILED_OBD_UUID $CONN_UUID

22 years ago- b=958: recovery per export/import, not per connection
shaver [Fri, 7 Mar 2003 20:16:05 +0000 (20:16 +0000)]
- b=958: recovery per export/import, not per connection
  - no more recovd thread:
    - import failure and upcall-execution is done by the failing thread
    - export failure and client eviction is done by the thread that's going
      to be waiting for it to complete anyway
    - import reconnection runs synchronously in the context of the lctl
      invocation, which means that we can now return errors to the upcall
  - no more fancy recovery state machine
  - no more RPCDEV service

- b=954: zero-impact OBD_PING opcode for MDS and OST

- Grand Unified Theory of Client Recovery
  - mdc_recover/osc_recovery unified in ptlrpc_recover_import

22 years agomoved procbridge.h to include/portals
rread [Tue, 4 Mar 2003 21:42:30 +0000 (21:42 +0000)]
moved procbridge.h to include/portals

22 years agofix make rpms, at least for me.
rread [Tue, 4 Mar 2003 10:00:10 +0000 (10:00 +0000)]
fix make rpms, at least for me.
- new configure option:
--enable-efence: turns on -lefence support for liblustre
  This option is OFF by default, so rpms will build on a machine
  that doesn't have efence-devel.

- add liblustre to DIST_SUBDIRS

22 years agob=941: zab's fix for ia64 compiles. untested, but it does compile
rread [Tue, 4 Mar 2003 00:38:11 +0000 (00:38 +0000)]
b=941: zab's fix for ia64 compiles. untested, but it does compile

22 years agoone small modification for linking libtest
shorthair [Sun, 2 Mar 2003 04:36:47 +0000 (04:36 +0000)]
one small modification for linking libtest

22 years ago- fixes for builds of liblustre (leave out kernel include path)
braam [Sat, 1 Mar 2003 22:45:07 +0000 (22:45 +0000)]
- fixes for builds of liblustre (leave out kernel include path)
- get ptlrpc working in the library

22 years ago- push common ldlm_handle_foo_ setup paths into their caller so it can send
zab [Sat, 1 Mar 2003 00:32:51 +0000 (00:32 +0000)]
- push common ldlm_handle_foo_ setup paths into their caller so it can send
  the reply before descending into the callbacks
- avoid doing 0 len ll_brws in writepage by unlocking pages outside i_size

22 years ago- minor build fixes for liblustre
braam [Fri, 28 Feb 2003 18:43:05 +0000 (18:43 +0000)]
- minor build fixes for liblustre

22 years agoAdd patch for handling largefile growth bug.
adilger [Thu, 27 Feb 2003 23:44:52 +0000 (23:44 +0000)]
Add patch for handling largefile growth bug.
Remove bogus warning case from Makefile for patch application, we already
do a bunch of conditional patch application stuff so use that.

22 years ago- miniature buildfixes for liblustre.
braam [Wed, 26 Feb 2003 23:35:57 +0000 (23:35 +0000)]
- miniature buildfixes for liblustre.

22 years agoFixes to recovery-cleanup.sh. We clean up pretty well, but the interrupted-open
shaver [Tue, 25 Feb 2003 17:55:17 +0000 (17:55 +0000)]
Fixes to recovery-cleanup.sh.  We clean up pretty well, but the interrupted-open
case trips an assertion (bug 912).

22 years agoQuick test for client cleanup while in recovery.
shaver [Tue, 25 Feb 2003 16:40:50 +0000 (16:40 +0000)]
Quick test for client cleanup while in recovery.

22 years ago- bring b_devel changes into b_io in preparation for file size fixes
zab [Tue, 25 Feb 2003 01:08:41 +0000 (01:08 +0000)]
- bring b_devel changes into b_io in preparation for file size fixes

22 years ago- store the FID for an open file in the request, so that replay can open by
shaver [Mon, 24 Feb 2003 21:48:20 +0000 (21:48 +0000)]
- store the FID for an open file in the request, so that replay can open by
  FID, in case the file has been deleted, etc.

- reconstruct replies for close (untested), link (tested), unlink (tested, but
  will leak OST objects until open-unlink works).

- quickie test program for link(2)

22 years agosome fixes for compilation with linux2.5.59
shorthair [Sat, 22 Feb 2003 15:07:55 +0000 (15:07 +0000)]
some fixes for compilation with linux2.5.59

22 years agoAdd return value to init_timer() to quiet compiler warning.
adilger [Sat, 22 Feb 2003 08:53:08 +0000 (08:53 +0000)]
Add return value to init_timer() to quiet compiler warning.

22 years ago- to keep things tidy on the screen
braam [Sat, 22 Feb 2003 03:15:57 +0000 (03:15 +0000)]
- to keep things tidy on the screen

22 years agofill in some sample code in libtest.c to demonstrate how to use
rread [Fri, 21 Feb 2003 18:24:42 +0000 (18:24 +0000)]
fill in some sample code in libtest.c to demonstrate how to use
the ioctl dumps created by lctl.

22 years agoAdd a new test which compiles portals and lustre in a lustre mount.
adilger [Fri, 21 Feb 2003 11:44:56 +0000 (11:44 +0000)]
Add a new test which compiles portals and lustre in a lustre mount.

22 years ago- another day of good progress with liblustre. We now have all the
braam [Fri, 21 Feb 2003 11:05:45 +0000 (11:05 +0000)]
- another day of good progress with liblustre.  We now have all the
  modules initialized, ready to start playing with lctl dumpfiles.

22 years agoLanding b_malt onto b_devel
rread [Thu, 20 Feb 2003 20:42:34 +0000 (20:42 +0000)]
Landing b_malt onto b_devel
- move uuid and handles to lustre, with config changes
- b=204,667: fix router config so network interfaces can be
  created in any order.
- Make sure all the ioctl calls  in obd.c are packed.

22 years ago- fix the uuid ioctls to work on the lustre side
rread [Thu, 20 Feb 2003 09:42:19 +0000 (09:42 +0000)]
- fix the uuid ioctls to work on the lustre side
- add IOC_PACK() to all the ioctls in lctl
- fix a format string in super.c
- rename IOCINIT to IOC_INIT, to match other macros
- looks ready to land on b_devel

22 years agoupdate b_malt
rread [Thu, 20 Feb 2003 07:25:24 +0000 (07:25 +0000)]
update b_malt
- includes multinet diffs, which will are needed on malt

22 years ago- completing the move of UUID registration from portals to lustre.
rread [Thu, 20 Feb 2003 06:52:56 +0000 (06:52 +0000)]
- completing the move of UUID registration from portals to lustre.
  Need to merge with multinet to finish this.

22 years agoUntested implementation of reconstruct_getattr_name.
shaver [Wed, 19 Feb 2003 22:01:25 +0000 (22:01 +0000)]
Untested implementation of reconstruct_getattr_name.
Test program for single invocation of GETATTR_NAME.

22 years agoDeleting the dummy file.
thantry [Wed, 19 Feb 2003 19:58:03 +0000 (19:58 +0000)]
Deleting the dummy file.

22 years agoThis is a test file. We have been moved to behind a firewall, and we want
thantry [Wed, 19 Feb 2003 19:53:21 +0000 (19:53 +0000)]
This is a test file. We have been moved to behind a firewall, and we want
to ensure that we can perform regular CVS operations. Please excuse.

22 years agonew files with api's we'd initially put in portals
braam [Wed, 19 Feb 2003 04:30:18 +0000 (04:30 +0000)]
new files with api's we'd initially put in portals

22 years ago- add new files move over from portals
braam [Wed, 19 Feb 2003 04:29:19 +0000 (04:29 +0000)]
- add new files move over from portals

22 years agoa-m-double for 2 mounts
pschwan [Tue, 18 Feb 2003 05:45:11 +0000 (05:45 +0000)]
a-m-double for 2 mounts

22 years ago- the bulk of the build fixes for liblustre. Note
braam [Mon, 17 Feb 2003 10:23:25 +0000 (10:23 +0000)]
- the bulk of the build fixes for liblustre. Note
  that all user level programs now need to include
  <liblustre.h> (as is done in the utils and test
  directories.

22 years agobring b_io up to the lastest write caching code. fsx and rundbench 1 pass in a
zab [Sat, 15 Feb 2003 21:45:46 +0000 (21:45 +0000)]
bring b_io up to the lastest write caching code.  fsx and rundbench 1 pass in a
96M all-in-one UML.

- prepare_write is throttled by finding dirty pages on the super block's
  dirty inodes and writing them to the network
- commit_write marks the page dirty and updates i_size
- writepage blocks writing the page and other dirty pages to the network
- sort the pages within a obd_brw batch so that block allocation isn't hosed
  on the OST
- don't change s_dirty's position on the list during writeback, that seems to
  be the job of writepage's callers
- don't try and mess with page's list membership after obd_brw completes,
  filemap_fdata{sync,wait} take care of that
- put a hack in obdo_to_inode that tricks ll_file_size into preferring the
  local i_size when there are cached pages on the inode
- add license blurb and editor instructions
- get rid of the management of vm lru pages and liod thread, prepare_write
  throttling and kupdate serve the same task (hopefully)
- remove unused ll_flush_inode_pages
- throw in a OSC-side "count > 0" assert to match a similar OST assert that
  I couldn't reproduce
- writeback will try to batch PTL_MD_MAX_IOV pages on the wire

22 years ago- rebase b_io against HEAD in preparation for the latest write cache code
zab [Sat, 15 Feb 2003 20:21:59 +0000 (20:21 +0000)]
- rebase b_io against HEAD in preparation for the latest write cache code

22 years agoAdd patch (already in 2.4.21-pre and 2.5.current) to quiet extN ino_t warnings.
adilger [Tue, 11 Feb 2003 23:43:27 +0000 (23:43 +0000)]
Add patch (already in 2.4.21-pre and 2.5.current) to quiet extN ino_t warnings.

22 years ago- add a fix to execute binaries to the kernel code (sorry folks, new kernel!)
braam [Sun, 9 Feb 2003 09:11:00 +0000 (09:11 +0000)]
- add a fix to execute binaries to the kernel code (sorry folks, new kernel!)
  ONLY did rh-2.4.18-18
- add test 30 to sanity.sh to test the same
- check in pc and series file for 2.5

22 years agoFix sys_link in vfs_intent_hp.patch to match vfs_intent-2.4.18.patch
pschwan [Fri, 7 Feb 2003 16:15:15 +0000 (16:15 +0000)]
Fix sys_link in vfs_intent_hp.patch to match vfs_intent-2.4.18.patch

22 years ago* Added a barrier test program in utils 'obdbarrier'
eeb [Fri, 7 Feb 2003 15:59:15 +0000 (15:59 +0000)]
*  Added a barrier test program in utils 'obdbarrier'

*  Changed lock cleanup code in echo_client to use obd_cancel_unused(),
   since the LOV's lock callbacks happen for each stripe, passing the
   stripe locks rather than the single lock who's handle was returned by
   obd_enqueue()

*  moved obdio/obdbarrier common procedures into obdiolib.c

22 years agoRemove CONFIG_DEV_RDONLY so Mike doesn't lose his sanity.
adilger [Fri, 7 Feb 2003 09:48:29 +0000 (09:48 +0000)]
Remove CONFIG_DEV_RDONLY so Mike doesn't lose his sanity.

22 years agoUpdate for l10 kernel version.
adilger [Thu, 6 Feb 2003 23:25:55 +0000 (23:25 +0000)]
Update for l10 kernel version.

22 years agoFixups to sanity.sh and createtest.c so it passes when run as a non-root
adilger [Wed, 5 Feb 2003 22:37:18 +0000 (22:37 +0000)]
Fixups to sanity.sh and createtest.c so it passes when run as a non-root
user (to keep Terry happy ;-).

We're not doing extensive permission testing here - leave that to POSIX.

22 years agoFix for bug 695, and a regression test to go with it (get rid of hard-coded
adilger [Wed, 5 Feb 2003 21:55:30 +0000 (21:55 +0000)]
Fix for bug 695, and a regression test to go with it (get rid of hard-coded
pathnames in sanity.sh while I'm at it ;-).

These are only client-side checks (the MDS does it's own checking for
both mds_open() and mds_reint_create() so that we don't create a bogus
file type).

22 years agoRegression test for bug 695.
adilger [Wed, 5 Feb 2003 21:45:43 +0000 (21:45 +0000)]
Regression test for bug 695.

22 years ago* Added lock enqueue/cancel ioctl interface to echo_client, with proper
eeb [Wed, 5 Feb 2003 17:45:18 +0000 (17:45 +0000)]
*  Added lock enqueue/cancel ioctl interface to echo_client, with proper
   teardown on (possibly unexpected) client exit.

*  added '-l' flag to utils/obdio, so that it locks the extent it writes,
   then reads back.

*  took 'filter' prefix off /proc stats interface in filter and copied it
   into the echo OBD; obdstats takes obd type name parameter as well as
   optional repeat interval.

22 years agoFix typo for patch series.
adilger [Mon, 3 Feb 2003 19:21:17 +0000 (19:21 +0000)]
Fix typo for patch series.

22 years agoUpdate kernel patches for hp kernel.
adilger [Mon, 3 Feb 2003 18:31:10 +0000 (18:31 +0000)]
Update kernel patches for hp kernel.

22 years agoFix recovery underpinnings:
shaver [Mon, 3 Feb 2003 18:25:50 +0000 (18:25 +0000)]
Fix recovery underpinnings:
 - Store the client's UUID (from last_rcvd) in exp_client_uuid, so
   target_handle_connect can match it up.
 - Don't obd_connect if target_handle_reconnect matched us up with an existing
   export.
 - Don't invalidate everything so eagerly: only if we can't recover and it
   wasn't a partition-reconnect.
 - Clear the in-recovery flag before we send clients their delayed replies.
 - Add "force" argument to lctl cleanup, and remove it from detach.
 - Use new lmc API for recovery-small.sh (more test cleanup coming).

22 years ago- temporary debugging stuff - this will slow you down.
braam [Sat, 1 Feb 2003 21:58:03 +0000 (21:58 +0000)]
- temporary debugging stuff - this will slow you down.

22 years agoAdd invalidate-show patch to kernel.
adilger [Sat, 1 Feb 2003 07:42:42 +0000 (07:42 +0000)]
Add invalidate-show patch to kernel.
Move docs to a more noticable place.

22 years agoMerge b_intent into b_md:
pschwan [Fri, 31 Jan 2003 21:37:21 +0000 (21:37 +0000)]
Merge b_intent into b_md:

- New kernel patch (version 9)
- DLM hooks to revalidate locked data, once the lock is granted (604)
- Further MDS reorganization, particularly of the open and o_creat paths

22 years ago- merge b_md into b_intent
pschwan [Fri, 31 Jan 2003 21:15:10 +0000 (21:15 +0000)]
- merge b_md into b_intent
- rename invalidate-show.diff, at andreas's request

22 years ago- Clean up, but don't error out if some parts fail.
shaver [Fri, 31 Jan 2003 19:06:33 +0000 (19:06 +0000)]
- Clean up, but don't error out if some parts fail.
- ONLY for easier controlled-reproduction.
- Error codes count.

22 years ago- error handling fixes from Andreas
braam [Thu, 30 Jan 2003 22:27:04 +0000 (22:27 +0000)]
- error handling fixes from Andreas

22 years agoan open(O_LOV_CREATE_DELAY) test
pschwan [Thu, 30 Jan 2003 21:09:09 +0000 (21:09 +0000)]
an open(O_LOV_CREATE_DELAY) test

22 years ago- return 0, not ENOENT, from a failed lookup in mds_open
pschwan [Thu, 30 Jan 2003 20:27:07 +0000 (20:27 +0000)]
- return 0, not ENOENT, from a failed lookup in mds_open

22 years ago- fill inodes if files exist
braam [Thu, 30 Jan 2003 19:40:16 +0000 (19:40 +0000)]
- fill inodes if files exist

22 years agoNew recovery regression -- hard-coded node names, probably other badness,
shaver [Thu, 30 Jan 2003 19:24:58 +0000 (19:24 +0000)]
New recovery regression -- hard-coded node names, probably other badness,
but it's finding bugs now, so I'm checking it in.

22 years ago- more error handling
braam [Thu, 30 Jan 2003 07:18:38 +0000 (07:18 +0000)]
- more error handling

22 years ago- add iod rmap patches to b_intent
braam [Thu, 30 Jan 2003 05:17:01 +0000 (05:17 +0000)]
- add iod rmap patches to b_intent

22 years ago- fix typo
braam [Thu, 30 Jan 2003 05:07:54 +0000 (05:07 +0000)]
- fix typo

22 years agoAdd patch to be more verbose about which inodes are busy at unmount time
adilger [Wed, 29 Jan 2003 22:13:06 +0000 (22:13 +0000)]
Add patch to be more verbose about which inodes are busy at unmount time
(for "VFS: Busy inodes" message).  Not added to any patchsets yet (I leave
it up to Peter to decide if he wants it applied to our kernels by default).

22 years ago* added /proc stats to obdfilter; should be general purpose and in lprocfs
eeb [Wed, 29 Jan 2003 21:17:33 +0000 (21:17 +0000)]
*  added /proc stats to obdfilter; should be general purpose and in lprocfs
*  added utils/obdstat for monitoring stats
*  added utils/obdio.c for exercising servers via echo_client.
*  echo_client does auto close when user proc exits

22 years ago- fix for O_EXCL case in mds_open
braam [Wed, 29 Jan 2003 19:51:34 +0000 (19:51 +0000)]
- fix for O_EXCL case in mds_open
- updates to cray plan
- remove unnecessary error handling from
  mdc_completion_callback

22 years ago- don't call mds_pack_md for non-regular files in mds_validate_dentry (doh)
pschwan [Wed, 29 Jan 2003 15:47:52 +0000 (15:47 +0000)]
- don't call mds_pack_md for non-regular files in mds_validate_dentry (doh)
- add an mcreate/open test to sanity

b_intent gets back to sanity #27 again

22 years ago- fill in ea in mds open when file exists.
braam [Wed, 29 Jan 2003 06:39:51 +0000 (06:39 +0000)]
- fill in ea in mds open when file exists.

22 years agoAdd the sync-on-unmount fix, and another fix which avoids accessing freed
adilger [Wed, 29 Jan 2003 01:08:24 +0000 (01:08 +0000)]
Add the sync-on-unmount fix, and another fix which avoids accessing freed
inodes upon ENOSPC or other errors (from AKPM).

The Makefile should be smart enough to detect if they need to be applied.

22 years agoUpdates to vfs_intent_hp.patch for PNNL (bug 727).
adilger [Tue, 28 Jan 2003 06:35:12 +0000 (06:35 +0000)]
Updates to vfs_intent_hp.patch for PNNL (bug 727).
This may or may not work, as I haven't had a chance to test, but it is 95%.

22 years ago- don't clobber flags
braam [Tue, 28 Jan 2003 05:26:51 +0000 (05:26 +0000)]
- don't clobber flags

22 years ago- lib lustre planning document
braam [Tue, 28 Jan 2003 05:02:22 +0000 (05:02 +0000)]
- lib lustre planning document
- mds server document
- return detailed status in body->flags

22 years agodon't dereference NULL or freed mfd in mds_open
pschwan [Tue, 28 Jan 2003 04:17:24 +0000 (04:17 +0000)]
don't dereference NULL or freed mfd in mds_open

22 years ago- do an mntget() before dentry_open()
pschwan [Tue, 28 Jan 2003 04:11:16 +0000 (04:11 +0000)]
- do an mntget() before dentry_open()

22 years ago- remove unused mfd_clienthandle
pschwan [Mon, 27 Jan 2003 22:50:29 +0000 (22:50 +0000)]
- remove unused mfd_clienthandle
- fix double semaphore up() in mds_open

22 years ago- added ldlm_lock_decref_and_cancel, to eliminate _that_ particular race
pschwan [Mon, 27 Jan 2003 21:24:48 +0000 (21:24 +0000)]
- added ldlm_lock_decref_and_cancel, to eliminate _that_ particular race
- don't send an extra buffer to the MDS if the EA length is zero
- andreas's fix to initialize rc to 0 in mds_open
- removed some tabvs
- in mds_reint_setattr, don't set the MD if the client didn't send one
- in mds_reint_create, ericm's fix for the POSIX nametoolong bug
- fix mds_open prototype in mds_reint.c

22 years ago- probable fix for mds refcount problem. touch /mnt/lustre/f cleans up now.
braam [Fri, 24 Jan 2003 19:22:23 +0000 (19:22 +0000)]
- probable fix for mds refcount problem.  touch /mnt/lustre/f cleans up now.

22 years agoanother bug fix.
braam [Thu, 23 Jan 2003 22:04:25 +0000 (22:04 +0000)]
another bug fix.