Whamcloud - gitweb
alex [Sat, 14 May 2005 11:51:44 +0000 (11:51 +0000)]
b=6326
- sanity tests for the bug
alex [Sat, 14 May 2005 11:46:09 +0000 (11:46 +0000)]
b=6326
- testing infrastructure to control order in which clients recover
jacob [Fri, 13 May 2005 19:42:47 +0000 (19:42 +0000)]
b=5854
fix /lib/modules/$(uname -r)/{build,source} to point to the correct
locations for 2.6 kernels
use --package arg to new-kernel-pkg if available
jacob [Fri, 13 May 2005 17:49:35 +0000 (17:49 +0000)]
b=6279
r=adilger
require CONFIG_STACK_SIZE_16KB on i386 + x86_64 non-UML
ericm [Fri, 13 May 2005 16:41:12 +0000 (16:41 +0000)]
adjust krb5 related scripts, add krb5 support in replay-ost-single,
sanityN, sanity-ldlm.
yury [Fri, 13 May 2005 15:58:50 +0000 (15:58 +0000)]
- added test_3b which emulates recursive mount. Does not pass yet.
yury [Fri, 13 May 2005 13:16:04 +0000 (13:16 +0000)]
- fixes in sanity tests about using mount command. cat /proc/mounts is used instead
yury [Fri, 13 May 2005 10:09:47 +0000 (10:09 +0000)]
fixes possible memory leak on error case path
nikita [Fri, 13 May 2005 09:34:16 +0000 (09:34 +0000)]
add b_ptlrpc_cleanup to buildcvs
adilger [Fri, 13 May 2005 07:47:14 +0000 (07:47 +0000)]
Branch HEAD
Fix duplicate filesystem feature flag (matches -fc3 kernel).
r=alex
alex [Thu, 12 May 2005 20:31:51 +0000 (20:31 +0000)]
b=6284
- for unknown yet reason, ptlrpc_client_replied() returns true while
rq_repmsg isn't allocated yet. this causes oops. the patch checks
the both and outputs debug info
ericm [Thu, 12 May 2005 20:09:56 +0000 (20:09 +0000)]
b=6277
alloc/free rpcpipe path dynamically.
alex [Thu, 12 May 2005 19:48:56 +0000 (19:48 +0000)]
b=6292
- lmv_enqueue() should re-enqueue lock if first MDS stores a name only
wangdi [Thu, 12 May 2005 15:15:43 +0000 (15:15 +0000)]
Branch: HEAD
Disable 71 in sanity.sh, because mmap code is not update on HEAD now.
yury [Thu, 12 May 2005 14:35:27 +0000 (14:35 +0000)]
- fixed sys_umount() code path. It did not release intent what caused memory leaks in sanity-gns.sh, as it uses umount for cleanups while testing GNS.
- do not free intent in GNS case if it is not lookup_it, only drop lock.
- added test_3a which checks disabling GNS mount point using chmod u-s
ericm [Thu, 12 May 2005 00:56:59 +0000 (00:56 +0000)]
again fix the intent release problem in dentry revalidate path,
simplified/rewrite some logic in ll_revalidate_it(), pls back this
out if any problem found.
nkj [Wed, 11 May 2005 15:10:24 +0000 (15:10 +0000)]
added dependency listing of IOR for ior-survey
yury [Wed, 11 May 2005 14:20:40 +0000 (14:20 +0000)]
- fixes in sanity-gns.sh
alex [Wed, 11 May 2005 11:48:14 +0000 (11:48 +0000)]
- kernel_thread() returns pid on success. ptlrpc_connect_interpret() treated
that as an error and broke the recovery state machine
yury [Wed, 11 May 2005 10:49:21 +0000 (10:49 +0000)]
- more fixes and comments about intent managing in ll_revalidate_it(). In few words, in the cases dentry is valid and no lookup will be called anymore, intent may be released by ll_intent_release() with clobering -._magic, ->it_op, etc, as it will not be used anymore. In the case revalidate failed, intent should not clobber ->magic, as it will confuse ll_frob_intent() called from ll_lookup_it(). Instead ll_drop_intent_lock() and ll_intent_free() should be used.
- in the case we found potential GNS dentry we do not cause GNS mountining by returning 0 from revalidate if GNS is disabled. This fixes possible NFS stales in some cases.
yury [Wed, 11 May 2005 10:04:43 +0000 (10:04 +0000)]
- fixes in ll_revalidate_it() about correct allocating/releasing intents.
- added comments about all the changes. Added comment for some GNS related cases.
- in the GNS case lookup intent should be released and others should be dropped and freed.
ericm [Wed, 11 May 2005 01:18:32 +0000 (01:18 +0000)]
recovery-small 21 LBUG: don't release intent for open etc. (maybe the
same problem of bug 6273?)
yury [Tue, 10 May 2005 15:14:22 +0000 (15:14 +0000)]
- removed duplicating code chunk in ll_mdc_blocking_ast() (wrong merge?)
wangdi [Mon, 9 May 2005 07:07:21 +0000 (07:07 +0000)]
Branch: HEAD
miss release parent pdir locks in mds_open
adilger [Mon, 9 May 2005 06:02:39 +0000 (06:02 +0000)]
Don't fail if Tag is not found.
yury [Sun, 8 May 2005 16:57:53 +0000 (16:57 +0000)]
- using TASK_INTERRUPTIBLE in OBD_SLEEP_ON at Andreas' suggestion.
- using IT_CHDIR in sys_chdir() vfs-intent_api-vanilla-2.6.10-fc3.patch. This fixes GNS with cd into mnt point.
- different fixes in GNS stuff.
- gns_enabled flag is moved to separate space instead of using LL_GNS_DISABLED
- removed error messages after inode_setattr(). Only comment is added.
- many changes in sanity-gns.sh, added checks for "cd" case. Deadlock tests are disabled for a while.
alex [Sat, 7 May 2005 19:34:47 +0000 (19:34 +0000)]
- reconstruct_open() should handle objects by fid properly
alex [Sat, 7 May 2005 19:21:32 +0000 (19:21 +0000)]
- minor fix
alex [Sat, 7 May 2005 19:17:45 +0000 (19:17 +0000)]
- needless debug messages are removed
alex [Sat, 7 May 2005 19:01:31 +0000 (19:01 +0000)]
b=6063
- to avoid possible lock collision during replay, we should replay all
request before any locks
alex [Sat, 7 May 2005 15:40:03 +0000 (15:40 +0000)]
b=6019
- workaround for cascading timeouts when failure of 1 server node causes
failures of unrelated connections. this caused different problems. for
example, lock cancel timeouts.
alex [Sat, 7 May 2005 12:23:17 +0000 (12:23 +0000)]
b=4920
- reconnect after recovery to all mds nodes. otherwise some mds will
miss reconnect after recovery and get to recovery timeout situation
alex [Sat, 7 May 2005 11:47:38 +0000 (11:47 +0000)]
b=6190
- deadlock in mds_open() fixed: if dentry points to another node, we
took LOOKUP lock holding UPDATE on a parent while strict resourse-
-based ordering must be used
wangdi [Sat, 7 May 2005 04:34:16 +0000 (04:34 +0000)]
Branch:HEAD
replace mds1 with SINGLEMDS according to andreas advice
ericm [Fri, 6 May 2005 19:59:07 +0000 (19:59 +0000)]
conf-sanity 15: some testbed (cnbuild) using source tree but also have
luster installed, which cause confusion of which lconf/lmc/lctl will be
used. We put $LUSTRE/utils and $LUSTRE/tests at the head of $PATH env
to ensure the tools in source tree will be searched at first. This should
not affect the tests by install, but if we got problem we'll back this
commit out.
yury [Fri, 6 May 2005 18:58:08 +0000 (18:58 +0000)]
- commiting missed fix about using symbol_get() instead of inter_module_get()
yury [Fri, 6 May 2005 18:56:26 +0000 (18:56 +0000)]
- check for GNS mount even if upcall returned to error (to not miss them)
- added IT_GETATTR to check if mount should be performed (experimental)
yury [Fri, 6 May 2005 18:32:05 +0000 (18:32 +0000)]
- fixed using of few deprected functions:
- sleep_on() is known to be racy replacing it with OBD_SLEEP_ON()
- added OBD_SLEEP_ON() which is inspired by wait_event() macro.
- inter_module_{get|put}() replaced with symbol_{get|put}()
- fixes in GNS stuff:
- ll_revalidate_it() calls missed ll_intent_release() in the case
GNS dentry is found (fix for #6176)
- added ll_gns_send_signal() which is needed to cause syscall restarting in the
case of -ERESTARTSYS is returned.
- cleanups in ll_gns_mount_object()
- ll_gns_mount_object() does not wait unconditionaly for userspace upcall,
instead it give it some time to complete (10s) and returns error if it
did not mount dentry. This fixes the case when upcall script runs endless
loop and never returns. This also fixes possible deadlock if another thread
with wait on not finished mount (this schema is disabled by and -ERESTARTSYS
is used instead. Both, waiting aproach and -ERESTARTSYS one need more
investigation).
- handling result of inode_setattr() call. Printing error message on error.
This fixes wanings that result of calling this function should be checked.
- fixed debug message about GNS inabilty to mount dentry (missed arg to CDEBUG
format string)
- cleanups in obdfilter_init()
- fixed using "mount" command in sanity-gns.sh
wangdi [Fri, 6 May 2005 14:59:21 +0000 (14:59 +0000)]
Branch: HEAD
some fix in replay-single.sh
b:6068
wangdi [Fri, 6 May 2005 13:22:07 +0000 (13:22 +0000)]
Branch: HEAD
missing up in mds_open_by_id
b:6068
jacob [Fri, 6 May 2005 13:20:57 +0000 (13:20 +0000)]
fix missing 'test'
ericm [Fri, 6 May 2005 03:46:51 +0000 (03:46 +0000)]
fix my previous careless fix
ericm [Fri, 6 May 2005 03:39:49 +0000 (03:39 +0000)]
sanity 24u: fix cross fs rename: should return error.
ericm [Fri, 6 May 2005 00:58:12 +0000 (00:58 +0000)]
add in sanity test of "rename across fs", which now cause oops anyway.
LiangZheng should already have a fix for this.
jacob [Thu, 5 May 2005 21:27:41 +0000 (21:27 +0000)]
b=3518
clean .depend files as well
nathan [Thu, 5 May 2005 20:58:52 +0000 (20:58 +0000)]
r=adilger (awhile ago)
Turn on D_CONFIG, D_CONSOLE, and D_IOCTL be default.
ericm [Thu, 5 May 2005 20:00:59 +0000 (20:00 +0000)]
conf-sanity-local: 9: the id() in python cause problem of calculating
debug mask, I'm not python export but simply remove them fixed it.
b=6104
jacob [Thu, 5 May 2005 18:18:12 +0000 (18:18 +0000)]
sync tree
ericm [Thu, 5 May 2005 05:51:17 +0000 (05:51 +0000)]
fix upcall cache: don't make acquiring entry expired too soon.
b=6244
jacob [Tue, 3 May 2005 23:27:36 +0000 (23:27 +0000)]
that's just not going to work right
jacob [Tue, 3 May 2005 22:32:46 +0000 (22:32 +0000)]
b=3031
- clean more files at clean time
jacob [Tue, 3 May 2005 22:27:54 +0000 (22:27 +0000)]
perl got away from me
jacob [Tue, 3 May 2005 22:16:17 +0000 (22:16 +0000)]
b=3031
- this is what i should have done originally
jacob [Tue, 3 May 2005 21:47:13 +0000 (21:47 +0000)]
oops
jacob [Tue, 3 May 2005 21:46:47 +0000 (21:46 +0000)]
b=3031
- remove .flags files on make clean
jacob [Tue, 3 May 2005 21:43:50 +0000 (21:43 +0000)]
b=3031
- nuke .flags files as well
jacob [Tue, 3 May 2005 21:23:14 +0000 (21:23 +0000)]
b=3418
- add dependencies on built libs
eeb [Tue, 3 May 2005 15:58:06 +0000 (15:58 +0000)]
* 6241
fixed bug in vibnal to do with mismatched RDMA fragments. Manifest by
lustre protocol errors when creating long symlinks so that the lustre
request message size didn't fit in a vibnal immediate message (4K).
* 6010
fix to prevent the LASSERT(0) in kqswnal_tx_done() if routing tables
are set up incorrectly.
jacob [Fri, 29 Apr 2005 18:48:50 +0000 (18:48 +0000)]
this got lost from b_bgl
jacob [Fri, 29 Apr 2005 18:40:13 +0000 (18:40 +0000)]
make getopt accept --with-linux, and teach the alphabet to whomever added --datestamp
liangzhen [Fri, 29 Apr 2005 07:47:10 +0000 (07:47 +0000)]
Patch for collect_pages_on_cpu() in tracefile.
(Wrong tage list is initialized)
Reviewd by Wangdi.
adilger [Thu, 28 Apr 2005 21:18:36 +0000 (21:18 +0000)]
make it clear that this is a destructive test
nathan [Thu, 28 Apr 2005 17:00:55 +0000 (17:00 +0000)]
add blank newbook template
ericm [Wed, 27 Apr 2005 10:59:28 +0000 (10:59 +0000)]
b=5538
by Nikita: fix group_info reference leak in pack_secdesc().
alex [Wed, 27 Apr 2005 03:30:42 +0000 (03:30 +0000)]
- needless CWARN() removed
alex [Wed, 27 Apr 2005 03:18:37 +0000 (03:18 +0000)]
b=6049
- handle final completion calls only
jacob [Wed, 27 Apr 2005 01:04:12 +0000 (01:04 +0000)]
b=5786
- add --disable-client and --disable-server configure options
jacob [Wed, 27 Apr 2005 00:57:09 +0000 (00:57 +0000)]
fix non-quilt build
jacob [Tue, 26 Apr 2005 23:53:19 +0000 (23:53 +0000)]
b=5782
- added --disable-server and --disable-client configure options (5782)
jacob [Tue, 26 Apr 2005 22:26:12 +0000 (22:26 +0000)]
remove leftover test
jacob [Tue, 26 Apr 2005 22:05:07 +0000 (22:05 +0000)]
add LUSTRE_VERSION to .target files
jacob [Tue, 26 Apr 2005 22:03:38 +0000 (22:03 +0000)]
b=6205
r=adilger
- move the lustre kernel modules to their own package
- lbuild now does 'make rpms' to build lustre modules, more or less
- pass --with-linux to lbuild to build only lustre rpms against an
existing kernel source tree
- drop the lustre-ldap rpm
yury [Tue, 26 Apr 2005 14:50:53 +0000 (14:50 +0000)]
- improvements in sanity-gns.sh:
- more smart way to check if fs is mounted
- in test 2g checking GNS with recursive path where last dir is not suid bit marked - should fail to mount it
yury [Tue, 26 Apr 2005 13:43:38 +0000 (13:43 +0000)]
- recursive chmod u+s for test 2g
yury [Tue, 26 Apr 2005 13:34:31 +0000 (13:34 +0000)]
- added recursive chmod for test_2g
yury [Tue, 26 Apr 2005 13:26:23 +0000 (13:26 +0000)]
- small fixes in sanity-gns.sh
yury [Tue, 26 Apr 2005 13:15:26 +0000 (13:15 +0000)]
- removed not used stuff (@lld and gns_mount_finished completion). Last is because
none of thread needs to wait for mount completion in kernel space.
- small cleanups in ll_gns_mount_object(). Some checks for input invariants are
moved after checking gns state. Fixes in comments.
- added the ability to disable GNS at all via /proc/.../gns_enabled. This is useful
for some complex tests (like recursive suid marked .mntinfo dir) at least. May be
useful for something else.
- added two new sanity tests to sanity-gns.sh:
a) test 2g: mounting /mnt/lustre/gns_entry/.mntinfo/.mntinfo/.mntinfo/.mntinfo where
only last entry is correct GNS mount object.
b) test 2h: upcall script starts mounting in background and retuns with no waiting
for mount command.
- in sanity tests disable GNS when preparing mount objects and mount points.
eeb [Tue, 26 Apr 2005 02:52:38 +0000 (02:52 +0000)]
file socknal_modparams.c was initially added on branch b_hd_newconfig.
eeb [Tue, 26 Apr 2005 02:52:34 +0000 (02:52 +0000)]
file qswnal_modparams.c was initially added on branch b_hd_newconfig.
yury [Mon, 25 Apr 2005 16:52:59 +0000 (16:52 +0000)]
- added more comments in GNS stuff about possible odd conditions
eeb [Mon, 25 Apr 2005 14:37:13 +0000 (14:37 +0000)]
* fix for 5852: vibnal HIGHMEM issues. Ensure ia64 builds OK even
when CONFIG_HIGHMEM is set.
yury [Mon, 25 Apr 2005 10:31:50 +0000 (10:31 +0000)]
- fix for #6202 accordingly to liangzhen@clusterfs.com patch
yury [Mon, 25 Apr 2005 09:13:45 +0000 (09:13 +0000)]
- fixes in filter_clear_page_cache() accordingly to green's patch.
yury [Sun, 24 Apr 2005 10:52:15 +0000 (10:52 +0000)]
- cleanups in comments, struct mtrack renamed to struct mem_track.
- making lvfs_memdbg_show() show possible memory leaks warning even if
memory allocation tracking is not enabled.
jacob [Fri, 22 Apr 2005 22:04:08 +0000 (22:04 +0000)]
If the kernel file isn't in KERNELDIR then try to wget it from our ftp server.
ericm [Fri, 22 Apr 2005 16:50:01 +0000 (16:50 +0000)]
land b_hd_sec onto HEAD:
* remote uid/gid handling
* various gss fixes
* fix broken local ACL
yury [Thu, 21 Apr 2005 15:57:40 +0000 (15:57 +0000)]
- changes in MDAPI DLD - added subchapter about nested locks issue.
- added more checks in GNS stuff:
- check for zero len mount object (mostly admin friendly thing, to show that
mount object is formed not correctly)
- set default GNS callback to "/usr/sbin/gns_upcall"
- added checks in lproc stuff:
- gns object name should be differnt than upcall
- gns timeout should be bigger than gns tick
yury [Thu, 21 Apr 2005 14:13:04 +0000 (14:13 +0000)]
- fixed missed -1 in kernel_read()
yury [Thu, 21 Apr 2005 11:39:53 +0000 (11:39 +0000)]
- GNS thread should be stopped after llite in unregistered
yury [Thu, 21 Apr 2005 11:38:17 +0000 (11:38 +0000)]
- fixes and improvements in GNS:
- fixed reading of mount object. Should be read not more than PAGE_SIZE - 1,
reserving one byte for terminating zero.
- fixed read data zero terminating. Not data[PAGE_SIZE - 1] should be zeroed
out, but rather zero should be put at the end of actual read data.
- added one more test which tests how GNS will mount dentry when there is
number of concurent thread creating entries in the same directory. test_3a
- fixes in other sanity tests. When a test creates number of concurrent threads
for checking something, it should wait their termination before getting
execution result and possibly fail. Do not remove upcalls from /tmp to have
the ability to look at them after failure.
yury [Thu, 21 Apr 2005 08:32:15 +0000 (08:32 +0000)]
- fixes and cleanups in GNS:
- mounting only on lookup() control path (revalidate() for possible mounts always return 0).
This makes it possible to use lookup_one_len() from ll_gns_mount_object(), as lookup
control path protected parent.
- on revalidate cause mount for possible mount points only if revalidate is successful
- removed not used dentry marking with MOUNT_PENDING, etc
eeb [Wed, 20 Apr 2005 19:34:52 +0000 (19:34 +0000)]
* buildcvs changes having branched portals HEAD to b_hd_newconfig
lustre b1_4 to b1_4_newconfig
jacob [Wed, 20 Apr 2005 02:39:14 +0000 (02:39 +0000)]
we don't distribute the patched kernel, kernel patch, or src.rpm, so just don't do them at all anymore
jacob [Tue, 19 Apr 2005 23:55:57 +0000 (23:55 +0000)]
add support for sles 8:
- new target for sles 8 kernel
- post(un) scripts for kernel rpms
- fixup quota build on sles 8 and sles 9
- update patches for a more recent sles 8 kernel
jacob [Tue, 19 Apr 2005 23:25:14 +0000 (23:25 +0000)]
b=b=6065
r=eeb
r=adilger
remove noops when stderr isn't present; just fail instead.
ericm [Tue, 19 Apr 2005 22:14:36 +0000 (22:14 +0000)]
from Brian: adjust default lsd location to /usr/sbin instead of /sbin/
yury [Tue, 19 Apr 2005 12:27:07 +0000 (12:27 +0000)]
- added comment in mds_open() and GNS mount points.
- cleanups in sanity-gns.sh
- added one more GNS test - mounting directory with absent mount object
yury [Tue, 19 Apr 2005 12:08:05 +0000 (12:08 +0000)]
- cosmetic changes in mdapi DLD
- more fixes in GNS stuff:
- fixes with refcount onto @dchild in ll_gns_mount_object() for error control paths
- fixing using accessing @dchild in the case of IS_ERR(dchild)
- disabled checking for GNS file in mds_open(), thus GNS mount points are also
opening by now. This is needed, because in taking lock we can't reliable to detect
GNS objects on MDS. Checking just suid bit and skip open such a dirs is not correct
as one may create structure like dir/dir/dir and mark all of them by suid.
- cleanups in sanity-gns.sh. Added few new tests for odd conditions like mount object
is symlink, dir, etc. Added test for changing mount object name to "." and "..".
Grouping tests into 2 groups:
(1) general GNS tests - mounting/umounting on timeout with different conditions
with possible deadlocks, races, etc.
(2) checking for odd conditions like recursive suid bit marked mount points, etc
yury [Mon, 18 Apr 2005 16:37:32 +0000 (16:37 +0000)]
- more fixes and improvements in GNS stuff:
- small optimization in ll_gns_wait_for_mount(), changing return code semantics to
usual 0 - ok, < 0 error.
- handled possible nested mount with returning -ERETSARTSYS and causing endless loop
of mounting.
- check if found mount object is regular file to prevent from evil cases like it is
suid bit marked directory, etc.
- fix with completting all waiters for dentry mounting in progress.
- added check for "." and ".." names in setting gns_object_name.
- fixed default gns_object_name value.
yury [Mon, 18 Apr 2005 12:40:50 +0000 (12:40 +0000)]
- fixed returning @retval in ll_lookup_it() in the case of -ERESTARTSYS