Whamcloud - gitweb
fs/lustre-release.git
21 years agoRemove two printks from startup; every line that we don't print is
pschwan [Mon, 7 Jul 2003 00:06:46 +0000 (00:06 +0000)]
Remove two printks from startup; every line that we don't print is
1100 fewer lines in the MCR /var/log/messages

21 years agoSome kernels apparently need an explicit include for timeval here
pschwan [Sun, 6 Jul 2003 22:55:42 +0000 (22:55 +0000)]
Some kernels apparently need an explicit include for timeval here

21 years agob=1141
pschwan [Sun, 6 Jul 2003 20:07:45 +0000 (20:07 +0000)]
b=1141
Another mostly-reorganization commit, to reduce the size of my working
patch set.

 - remove gratuitous lustre_put_page() inline
 - split filter_preprw() into filter_preprw_read() and filter_preprw_write()
 - reverse the 2.5 waitfor_one_page macro into a 2.4 wait_on_page_locked macro
 - unimportant changes galore

21 years agob=1141
pschwan [Sun, 6 Jul 2003 18:49:43 +0000 (18:49 +0000)]
b=1141
A reorganization of obdfilter/ before I launch into the real work.
There should be no functional changes.
 - removes unused 'niobuf_remote' argument from fsfilt_brw_start
 - splits filter.c into filter.c, filter_io.c, filter_log.c, and filter_san.c
   More of the same is on the way, but this is a good start.

21 years agob=1502
pschwan [Sun, 6 Jul 2003 01:16:48 +0000 (01:16 +0000)]
b=1502
- Discard incredibly old requests (older than "obd_timeout" seconds) in
  ptlrpc_main without even trying to process them, because the client
  has already given up on us.  See the bug text for more details on why
  this is the right thing to do.
- Clean up deeply-nested ptlrpc_main()
- On Zach's good advice, I also fixed Portals to record when a packet
  arrives in useful microseconds instead of relatively useless cycles

21 years agoland b_merge on b_devel. Includes b_mount and b_orphan, which includes:
pschwan [Sat, 5 Jul 2003 22:47:31 +0000 (22:47 +0000)]
land b_merge on b_devel.  Includes b_mount and b_orphan, which includes:

 - orphans are moved into the PENDING directory for possible recovery
 - replayed opens now open by fid for orphan/rename safety (1042)
 - last close of an orphan inode generates a transno (683)
 - avoid CERROR in normal ll_setattr_raw() error case (1500)

Added code which is currently disabled:

 - dentry pinning from b_mount for cwd and mountpoints (1020)
 - logging code for orphans (and many other things)

21 years agosmall mgmt_svc tweaks
shaver [Sat, 5 Jul 2003 14:50:51 +0000 (14:50 +0000)]
small mgmt_svc tweaks

21 years agoAllow some "slop" in the timestamps, because UML isn't very good at keeping
adilger [Sat, 5 Jul 2003 07:03:38 +0000 (07:03 +0000)]
Allow some "slop" in the timestamps, because UML isn't very good at keeping
accurate times if it is busy running.

21 years agonew ioctl statfs has been added as part of the globalhealth. Now the lctl have
sravi [Fri, 4 Jul 2003 07:17:50 +0000 (07:17 +0000)]
new ioctl statfs has been added as part of the globalhealth. Now the lctl have
command statfs, which is used for stat  on MDS or OST.
This is a utility get the MDS/OST status and generate the snmp trap.

21 years agob=1469
rread [Thu, 3 Jul 2003 23:05:04 +0000 (23:05 +0000)]
b=1469

Create a simple wrapper around gethostbyname, so we can print a sane
error message when the host is not found, instead of "No such file or
directory."

21 years agook, no, really, commit the qla2xxx driver.
zab [Thu, 3 Jul 2003 22:36:21 +0000 (22:36 +0000)]
ok, no, really, commit the qla2xxx driver.

21 years agoinitial mgmt/monitoring service skeleton
shaver [Thu, 3 Jul 2003 19:18:33 +0000 (19:18 +0000)]
initial mgmt/monitoring service skeleton

21 years agoThis file is to check the health of the OST as part of the global-health.
sravi [Thu, 3 Jul 2003 17:16:38 +0000 (17:16 +0000)]
This file is to check the health of the OST as part of the global-health.
This does not modify any other files lustre/portals.
May need to optimize this a bit.

21 years ago* Added missing kportal_put_ni() when forcing portals failures
eeb [Thu, 3 Jul 2003 15:40:30 +0000 (15:40 +0000)]
*  Added missing kportal_put_ni() when forcing portals failures

21 years ago- the filter doesn't have a cli_obd, so pass an import instead of a
pschwan [Thu, 3 Jul 2003 05:16:48 +0000 (05:16 +0000)]
- the filter doesn't have a cli_obd, so pass an import instead of a
  conn in the llcd
- avoid double obd_transno_commit_cb()
- initialize the fo_sem for great justice

21 years agob=1501
pschwan [Wed, 2 Jul 2003 23:26:55 +0000 (23:26 +0000)]
b=1501
- return -ENOTDIR from mds_open() if we're trying to open a
  non-directory with O_DIRECTORY
- add tests/o_directory.c and make it in sanity test 38

21 years ago- add the b1 qla2xxx 2.5 driver that we're using. b3 panics.
zab [Wed, 2 Jul 2003 21:36:16 +0000 (21:36 +0000)]
- add the b1 qla2xxx 2.5 driver that we're using.  b3 panics.

21 years ago- add set_info calls, to inform the filter when the MDS connects
pschwan [Wed, 2 Jul 2003 18:40:59 +0000 (18:40 +0000)]
- add set_info calls, to inform the filter when the MDS connects
- make OST_LOG_CANCEL into a more generic OBD_LOG_CANCEL packet
- the filter can't use the normal obd methods unless we also give it an MDC
- lov_connect now initializes the export lists unconditionally
- the filter needs to register for the commit callback if there are
  cookies to be cancelled, regardless of whether this is a failover OST
- hacked around in the lprocfs sewer until it built again

21 years agoClient-side portions of bug 974: push cache and dirty-page management down into
shaver [Wed, 2 Jul 2003 13:32:34 +0000 (13:32 +0000)]
Client-side portions of bug 974: push cache and dirty-page management down into
    OSC, in preparation for pre-allocation grants.  Also fixes 975 by providing
    /proc/fs/lustre/llite/fs0/max_dirty_pages to limit client-side cache.

21 years agoadd tg3_netconsole.pc
wangdi [Tue, 1 Jul 2003 11:04:09 +0000 (11:04 +0000)]
add tg3_netconsole.pc

21 years agoadd netconsole_sysrq
wangdi [Tue, 1 Jul 2003 11:02:45 +0000 (11:02 +0000)]
add netconsole_sysrq

21 years agoadd tg3_netconsole.patch
wangdi [Tue, 1 Jul 2003 11:01:41 +0000 (11:01 +0000)]
add tg3_netconsole.patch

21 years agob=1424
rread [Tue, 1 Jul 2003 07:35:18 +0000 (07:35 +0000)]
b=1424
r=shaver,others

- cleanup import level and generation
- fixes several bugs:

b=1420 ost and client both timeout, client never reconnects
b=1440 very small change to call the lov set_active in recovery
b=1448 another small change to send DISCONNECT at a normal level

21 years ago- move e1000_Poll() into a less broken place
zab [Mon, 30 Jun 2003 23:24:37 +0000 (23:24 +0000)]
- move e1000_Poll() into a less broken place
- s/static// in front of stop_this_cpu so lkcd compiles under smp
+ mask these changes with tons of line numbering and diff -p noise

21 years ago- add netdump patches to 2.5.73 kgdb series.
braam [Mon, 30 Jun 2003 21:30:46 +0000 (21:30 +0000)]
- add netdump patches to 2.5.73 kgdb series.

21 years ago- add the kexec patch to 2.5.73
braam [Mon, 30 Jun 2003 21:26:49 +0000 (21:26 +0000)]
- add the kexec patch to 2.5.73

21 years ago- move things forward to 2.5.73, include lkcd patches
braam [Mon, 30 Jun 2003 20:52:50 +0000 (20:52 +0000)]
- move things forward to 2.5.73, include lkcd patches

21 years agob=957: add pinger for recoverable clients, in order to promptly detect server
shaver [Mon, 30 Jun 2003 20:11:27 +0000 (20:11 +0000)]
b=957: add pinger for recoverable clients, in order to promptly detect server
       failure and join recovery.  Currently DISABLED (#ifdef ENABLE_PINGER)
       pending some discussion about deployment.

Also: use the --config (or file.xml) instead of 'none' in the mount line, so
that mounted filesystems have a little more context.

21 years ago- major cleanup of the 2.5 kernel patch
braam [Mon, 30 Jun 2003 17:55:07 +0000 (17:55 +0000)]
- major cleanup of the 2.5 kernel patch
  - we indicate now with two flags when the last path name component
    is being looked up, ending almost all confusion about the code
    path.
  - many small fixes.
  - sanity finishes without failures, until at cleanup something hangs.

21 years agoFix utime.c so it calls utime() on the right file and print valid errors.
adilger [Fri, 27 Jun 2003 23:56:27 +0000 (23:56 +0000)]
Fix utime.c so it calls utime() on the right file and print valid errors.
Fix mkdirdeep.c so it is generally useful
 - don't dump/trace by default
 - exit on errors
 - allow deeper than 100 tree (up to PATH_MAX).
 - don't clobber stack generating stat pathnames.

21 years ago- fix up super.c
braam [Fri, 27 Jun 2003 20:42:20 +0000 (20:42 +0000)]
- fix up super.c
- minor change to llite_lib.c to avoid a warning.

21 years agoFix lock-match fallout from 1228: use the inode against which the intent lock
shaver [Fri, 27 Jun 2003 17:25:31 +0000 (17:25 +0000)]
Fix lock-match fallout from 1228: use the inode against which the intent lock
was granted, often not the one against which the lock was requested, when
matching l_data.

Also, match locks in ll_dir_readpage.

Remove some deadwood fops-functions.

21 years ago[liblustre]: fix an segfault when running iozone. don't more checking
meijia [Fri, 27 Jun 2003 04:23:47 +0000 (04:23 +0000)]
[liblustre]: fix an segfault when running iozone. don't more checking
before remove an inode.

21 years agoadding otree files (not yet used)
shaver [Thu, 26 Jun 2003 20:20:01 +0000 (20:20 +0000)]
adding otree files (not yet used)

22 years agoScript to test ls speed on a single host using two mount points.
mmex [Thu, 26 Jun 2003 19:37:36 +0000 (19:37 +0000)]
Script to test ls speed on a single host using two mount points.
Not a great test but can give basic information on ls speed issues.

22 years ago- Many unused variable warning fixes (several of which were not really a
pschwan [Thu, 26 Jun 2003 17:49:35 +0000 (17:49 +0000)]
- Many unused variable warning fixes (several of which were not really a
  problem, I think)
- list_splice_init() isn't defined in 2.4.18
- FSTYPE change for lov.sh

22 years ago- fixes the maximum OST packet size, a lot of list management, and a
pschwan [Wed, 25 Jun 2003 22:38:32 +0000 (22:38 +0000)]
- fixes the maximum OST packet size, a lot of list management, and a
  couple other bugs
- starts threads at OST and MDS startup
- able to complete a round of create/setattr/cancel if SENDNOW is set
- some corruption happens when I ask it to build up and cancel whole
  pages; I need to run on multiple systems to start narrowing down where.

22 years ago[liblustre]: add back the script
meijia [Wed, 25 Jun 2003 05:09:24 +0000 (05:09 +0000)]
[liblustre]: add back the script

22 years agoFix for bug 1466, delete thread locking up machine during orphan recovery.
adilger [Tue, 24 Jun 2003 23:34:21 +0000 (23:34 +0000)]
Fix for bug 1466, delete thread locking up machine during orphan recovery.

22 years agoSimple utime testing script, as I can't concentrate on anything more complex.
adilger [Tue, 24 Jun 2003 00:04:20 +0000 (00:04 +0000)]
Simple utime testing script, as I can't concentrate on anything more complex.

22 years agoUse a better variable name than "i" for the open file descriptor.
adilger [Mon, 23 Jun 2003 18:20:54 +0000 (18:20 +0000)]
Use a better variable name than "i" for the open file descriptor.

22 years ago[liblustre]:
meijia [Mon, 23 Jun 2003 11:56:54 +0000 (11:56 +0000)]
[liblustre]:
 - treat an closed inode as invalid.
 - reap inode right after unlink/rmdir and when found stale ones.

22 years agoadd patches for 2.5.72. Minor Portals miscompilation still needs a
braam [Sun, 22 Jun 2003 21:12:01 +0000 (21:12 +0000)]
add patches for 2.5.72. Minor Portals miscompilation still needs a
fix, but I cannot do that now I think.

22 years agomany changes to the intent patch. Most of sanity runs now.
braam [Sat, 21 Jun 2003 20:34:46 +0000 (20:34 +0000)]
many changes to the intent patch.  Most of sanity runs now.

22 years ago- Ensure stdout is flushed after each new line, this makes it possible to
behlendo [Fri, 20 Jun 2003 23:57:46 +0000 (23:57 +0000)]
- Ensure stdout is flushed after each new line, this makes it possible to
  redirect the resulting output to a file.
- %Lu doesn't format properly, changed to %lu

22 years agoFixes Connectathon bigfile(30 MB)test bug
girish_c [Fri, 20 Jun 2003 22:43:18 +0000 (22:43 +0000)]
Fixes Connectathon bigfile(30 MB)test bug

22 years agob_orphan builds again; now for a some testing
pschwan [Fri, 20 Jun 2003 16:54:24 +0000 (16:54 +0000)]
b_orphan builds again; now for a some testing

22 years ago* 1439 fix
eeb [Wed, 18 Jun 2003 22:08:38 +0000 (22:08 +0000)]
*   1439 fix

22 years agoFix patches that were eaten by patch scripts.
adilger [Wed, 18 Jun 2003 18:10:06 +0000 (18:10 +0000)]
Fix patches that were eaten by patch scripts.

22 years agoNFS export patch for lustre
girish_c [Wed, 18 Jun 2003 16:39:43 +0000 (16:39 +0000)]
NFS export patch for lustre
1) Patch is updated with review comments
2) All Connectathon tests passes except bigfile(30 MB)test

22 years agoFix delete thread so that it doesn't sleep uninterruptibly and raise load avg.
adilger [Wed, 18 Jun 2003 09:07:19 +0000 (09:07 +0000)]
Fix delete thread so that it doesn't sleep uninterruptibly and raise load avg.
Do not start delete thread if asyncdel mount option is not given.

22 years agohaven't even tried to build, but getting closer:
pschwan [Tue, 17 Jun 2003 09:58:09 +0000 (09:58 +0000)]
haven't even tried to build, but getting closer:
- move the recovd_thread to ptlrpc, because it uses ptlrpc functions
- have one global llog_commit_master, in ptlrpc

22 years ago[liblustre]: add script to generate libraries
meijia [Tue, 17 Jun 2003 03:12:53 +0000 (03:12 +0000)]
[liblustre]: add script to generate libraries

22 years agoThese patches bring mcore style dumping to 2.5 via lkcd. Suparana did this
zab [Sat, 14 Jun 2003 20:45:52 +0000 (20:45 +0000)]
These patches bring mcore style dumping to 2.5 via lkcd.  Suparana did this
work a while back and committed it to lkcd.  She describes the
functionality in http://lwn.net/Articles/22085/ before it was committed.

The patches aren't added to a series/pc quite yet as they're against the
stand-alone 2.5.69 tree.  But they give an idea of what is required and
needed to be recorded somewhere.

22 years agob=1383
coopcoopbware [Fri, 13 Jun 2003 23:25:01 +0000 (23:25 +0000)]
b=1383
r=robert

- a less invasive way of propagating OBD_MAX_IOCTL_BUFFER. Still set by passing the --with-obd-buffer-size=[size] arg to configure

22 years ago* Merged b_devel changes into b_eq (again). b_eq and b_devel have same
eeb [Fri, 13 Jun 2003 13:21:58 +0000 (13:21 +0000)]
*  Merged b_devel changes into b_eq (again).  b_eq and b_devel have same
   acceptance-small symptoms viz....

   1/ Need to sleep before lconf -d on llite client to ensure all
      connection refcounts are 0 (llmountcleanup.sh on b_eq has a 10 second
      sleep).

   2/ When running dbench 50, the OST complains about long-running
      obdfilter ops.  Eventually the MDS or OST will crash/hang.

22 years agoScript from Evan (with tweaks) for loading modules into crash, if the
adilger [Fri, 13 Jun 2003 08:39:50 +0000 (08:39 +0000)]
Script from Evan (with tweaks) for loading modules into crash, if the
"mod -S" command does not work (e.g. if the modules have been moved).

22 years ago[liblustre]: add back missing files, update test code
meijia [Fri, 13 Jun 2003 08:21:44 +0000 (08:21 +0000)]
[liblustre]: add back missing files, update test code

22 years ago[liblustre]: again add back liblustre.c
meijia [Fri, 13 Jun 2003 07:30:09 +0000 (07:30 +0000)]
[liblustre]: again add back liblustre.c

22 years agoStabalize the clump paths by clamping the number of pending sets and
zab [Fri, 13 Jun 2003 05:36:50 +0000 (05:36 +0000)]
Stabalize the clump paths by clamping the number of pending sets and
cleaning up the sleeping logic.  Also tie in ll_readpage.   Now to start
profiling and tuning this..
- record the rw mode in llwp so completion knows what to do with the pages
- introduce a simple little 'page list' data structure that uses page->private
  to chain pages off llis in fifo order
- have both read and write plists in the lli
- get rid of the old unused work_pending member
- always put the lli back on the list in _give_plist so that we don't
  lose llis
- introduce a 'new_arrival' member so lliod can only come out of l_wait_event
  when new material arrives
- pass rc into llwp completion so we can make use of it someday
- move the concurrent set checks into the set building loop, duh
- add some debugging here and there

22 years agochecking statfs cache expiration in obd_statfs
freym [Thu, 12 Jun 2003 23:03:25 +0000 (23:03 +0000)]
checking statfs cache expiration in obd_statfs

22 years agoChange ia64 to compile with -g instead of -gstabs now that we are using crash.
adilger [Thu, 12 Jun 2003 18:46:01 +0000 (18:46 +0000)]
Change ia64 to compile with -g instead of -gstabs now that we are using crash.

22 years ago* Merged in b_devel again
eeb [Thu, 12 Jun 2003 17:34:04 +0000 (17:34 +0000)]
*  Merged in b_devel again

22 years agoupdated lstat utility
girish_c [Thu, 12 Jun 2003 14:43:39 +0000 (14:43 +0000)]
updated lstat utility
Changes:
1. Summary info for client lproc tree
2. added "--all" option to print the contents of lproc

22 years ago- fix compilation of liblustre, not test yet.
meijia [Thu, 12 Jun 2003 09:36:40 +0000 (09:36 +0000)]
- fix compilation of liblustre, not test yet.

22 years agomore extN cleanup
pschwan [Thu, 12 Jun 2003 07:28:57 +0000 (07:28 +0000)]
more extN cleanup

22 years agoRemove use of local stack variables (size, ptr) for {PORTAL,OBD}_{ALLOC,FREE}().
adilger [Thu, 12 Jun 2003 05:38:42 +0000 (05:38 +0000)]
Remove use of local stack variables (size, ptr) for {PORTAL,OBD}_{ALLOC,FREE}().
I think these were intended to avoid doing multiple evals of math/function
calls for the size, but all of our usage so far is passing in a constant
or a variable that has already computed the size.

By reordering the *_FREE() macros slightly, we do not need to store the size
of the free in a local stack variable either.  We have to be careful though,
because "size" may be part of the struct we are poisoning/freeing.

22 years agoBug#1380-added to read /proc/fs/lustre*/stat in obdecho, obdfilter, llite
thantry [Wed, 11 Jun 2003 22:01:15 +0000 (22:01 +0000)]
Bug#1380-added to read /proc/fs/lustre*/stat in obdecho, obdfilter, llite

22 years ago* Merged latest b_devel
eeb [Wed, 11 Jun 2003 18:04:11 +0000 (18:04 +0000)]
*  Merged latest b_devel

*  **** NOT TESTED ****

22 years agoDisable zerocopy by default while kernel symbols get straightened out.
adilger [Wed, 11 Jun 2003 11:06:31 +0000 (11:06 +0000)]
Disable zerocopy by default while kernel symbols get straightened out.

22 years agoUse GFP_NOFS in portals, so that we don't call into the filesystem to try
adilger [Tue, 10 Jun 2003 08:58:55 +0000 (08:58 +0000)]
Use GFP_NOFS in portals, so that we don't call into the filesystem to try
and free memory when we are trying to send an RPC for a write (which is
trying to free memory).  With this change, we can now run dbench and iozone
on a single UML again.  Of importance is the addition of sk->allocation to
socknal.

22 years agoFix the socknal TCP zero-copy detection. "zercopy" does not a SOCKNAL_ZC make.
adilger [Tue, 10 Jun 2003 03:45:11 +0000 (03:45 +0000)]
Fix the socknal TCP zero-copy detection. "zercopy" does not a SOCKNAL_ZC make.

22 years agoremove 2.4.19 hp patch
wangdi [Mon, 9 Jun 2003 08:12:57 +0000 (08:12 +0000)]
remove 2.4.19 hp patch

22 years agoremove 2.4.19 hp patch
wangdi [Mon, 9 Jun 2003 07:53:43 +0000 (07:53 +0000)]
remove 2.4.19 hp patch

22 years agoAdd fixed from MDS for bug 1403 to obdfilter also.
adilger [Sun, 8 Jun 2003 20:57:44 +0000 (20:57 +0000)]
Add fixed from MDS for bug 1403 to obdfilter also.

Make mds_client_add() and filter_client_add() more closely resemble each
other (one day we may want to move the last_rcvd handling from mds/obdfilter
into some common code in obdclass or something).

22 years agoFix compile warnings on ia64 (especially the many, many in wirecheck.c).
adilger [Fri, 6 Jun 2003 21:43:36 +0000 (21:43 +0000)]
Fix compile warnings on ia64 (especially the many, many in wirecheck.c).

22 years agoFix lock ordering deadlock on the MDS when clients are connecting at the
adilger [Fri, 6 Jun 2003 21:34:09 +0000 (21:34 +0000)]
Fix lock ordering deadlock on the MDS when clients are connecting at the
same time that regular MDS operations are ongoing.

22 years agoFix braces around math, so GCC doesn't complain on ia64.
adilger [Fri, 6 Jun 2003 21:08:04 +0000 (21:08 +0000)]
Fix braces around math, so GCC doesn't complain on ia64.

22 years agob=1228
pschwan [Fri, 6 Jun 2003 19:09:42 +0000 (19:09 +0000)]
b=1228
More debugging: pass the real inode down from ll_clear_inode, so we
can determine in cancel_unused whether the lock's data pointer is wrong

22 years agoGo back to STACK_SIZE - frame_address() now that we use __builtin_dwarf_cfa().
adilger [Fri, 6 Jun 2003 17:49:55 +0000 (17:49 +0000)]
Go back to STACK_SIZE - frame_address() now that we use __builtin_dwarf_cfa().

22 years ago* fixes for multinet
eeb [Fri, 6 Jun 2003 16:02:50 +0000 (16:02 +0000)]
*  fixes for multinet

22 years ago- This brings 2.5 forward to pass most, but not all of sanity.
braam [Fri, 6 Jun 2003 07:42:46 +0000 (07:42 +0000)]
- This brings 2.5 forward to pass most, but not all of sanity.
  A few problems remain with dangling symlinks.
- Mostly this is a fairly dramatic simplification of many issues,
  maybe not possible like this in 2.4.

22 years agoQuiet compiler warnings in portals. Basically, just pass void parameters to
adilger [Thu, 5 Jun 2003 23:57:13 +0000 (23:57 +0000)]
Quiet compiler warnings in portals.  Basically, just pass void parameters to
match the function pointer declarations and cast back to the original type
inside the function.

22 years agoAllow runvmstat to log to a file for perf monitoring.
adilger [Thu, 5 Jun 2003 23:31:29 +0000 (23:31 +0000)]
Allow runvmstat to log to a file for perf monitoring.
Add a similar logging obdstat script also.

22 years ago* Added (cluster-wide) NID offset capability to qswnal
eeb [Thu, 5 Jun 2003 17:39:43 +0000 (17:39 +0000)]
*  Added (cluster-wide) NID offset capability to qswnal

*  Made lctl::mynid accept NIDs specified as a hostname string, a.b.c.d IP
   address, %i and %x (just like the routing commands).  This command now
   works with the qswnal, to tell it its cluster-wide NID offset.  Temp
   hack in lconf to actually do this function when network is 'elan' as
   well as 'tcp' or 'toe'.

*  Added lctl::shownid (no args) to print the NID

22 years agomerge from b_devel -> b_ad
rread [Thu, 5 Jun 2003 04:12:56 +0000 (04:12 +0000)]
merge from b_devel -> b_ad
  DEVEL_AD_PARENT_20030526 -> DEVEL_AD_PARENT_20030605

Large, squirrelly merge. It mounts, and acceptance_small.sh finishes
the test and then fails in cleanup, just like b_devel.

22 years agoplug into ->writepages so the vm can send us batches of pages for writeback.
zab [Wed, 4 Jun 2003 23:40:35 +0000 (23:40 +0000)]
plug into ->writepages so the vm can send us batches of pages for writeback.
In the process we make ->writepage non-blocking and hand writeback off to a
single worker thread.

we build a singly linked list of pages off of lli as writepage receives them.
when our writepage or writepages decides it wants to start io it puts the lli
into a list for the writeback thread (lliod) and wakes it.

- make llwp allocing and freeing a little more regular
-
- introduce lliod, putting its control state in the super block.
- ad ll_info2i() so that lliod can find the inode from the lli
- abstract out what 2.4 and 2.5 have to do to end writeback on a page
- consolidate initialization of lli into one place
- move set_page_clean to above the only place it was used

22 years ago__arch_ia64__ is apparently not set, so we were hitting the gcc bug
pschwan [Wed, 4 Jun 2003 23:39:18 +0000 (23:39 +0000)]
__arch_ia64__ is apparently not set, so we were hitting the gcc bug
and eating enormous amounts of stack on ia64.  Changed to __ia64__

22 years agomake the portals' definition of USERMODEHELPER blocking, too.
rread [Wed, 4 Jun 2003 22:27:19 +0000 (22:27 +0000)]
make the portals' definition of USERMODEHELPER blocking, too.

22 years agob=1400 timeout causes crash.
rread [Wed, 4 Jun 2003 22:13:21 +0000 (22:13 +0000)]
b=1400 timeout causes crash.
 vfs_fstat should call intent_release before fput, and only if f != NULL

22 years agoadd ext3-2.4-ino_t-2.4.19.pc
wangdi [Wed, 4 Jun 2003 03:05:19 +0000 (03:05 +0000)]
add ext3-2.4-ino_t-2.4.19.pc

22 years agoadd ext3-2.4-ino_t-2.4.19.patch
wangdi [Wed, 4 Jun 2003 03:01:16 +0000 (03:01 +0000)]
add ext3-2.4-ino_t-2.4.19.patch

22 years agofix a flaw in iopen-2.4.19
wangdi [Wed, 4 Jun 2003 02:56:44 +0000 (02:56 +0000)]
fix a flaw in iopen-2.4.19

22 years agoadd ext3-orhan_lock-2.4.19.pc
wangdi [Wed, 4 Jun 2003 02:41:44 +0000 (02:41 +0000)]
add ext3-orhan_lock-2.4.19.pc

22 years agoadd ext3-orphan_lock-2.4.19.patch
wangdi [Wed, 4 Jun 2003 02:39:38 +0000 (02:39 +0000)]
add ext3-orphan_lock-2.4.19.patch

22 years agoadd ext3-delete_thread-2.4.19.pc
wangdi [Wed, 4 Jun 2003 02:34:25 +0000 (02:34 +0000)]
add ext3-delete_thread-2.4.19.pc

22 years agoadd ext3-delete_thread-2.4.19.patch
wangdi [Wed, 4 Jun 2003 02:32:35 +0000 (02:32 +0000)]
add ext3-delete_thread-2.4.19.patch

22 years agoPut an obd_ucred struct in the mds_update_record struct so we can pass it
adilger [Wed, 4 Jun 2003 02:05:52 +0000 (02:05 +0000)]
Put an obd_ucred struct in the mds_update_record struct so we can pass it
directly to push_ctxt().  This saves allocating an additional obd_run_ctxt
struct on the stack in mds_reint_rec() (in one of the major stack using paths).

22 years agoMake LBUG() just a special case of LBUG_WITH_LOC().
adilger [Wed, 4 Jun 2003 00:00:30 +0000 (00:00 +0000)]
Make LBUG() just a special case of LBUG_WITH_LOC().

22 years agoRemove bogus extra arg from delete thread warning message.
adilger [Tue, 3 Jun 2003 23:27:01 +0000 (23:27 +0000)]
Remove bogus extra arg from delete thread warning message.