Whamcloud - gitweb
fs/lustre-release.git
21 years ago- debug.c was only used by obdclass, so I moved it there
pschwan [Thu, 5 Sep 2002 19:14:59 +0000 (19:14 +0000)]
- debug.c was only used by obdclass, so I moved it there
- some of page.c wasn't used by anything, so I removed those bits; much of
  page.c was only used in one module, so those bits moved there; two functions
  were shared, so now they're in obdclass; page.c is no more.
- stop building and distributing obdfs
- little warning and formatting fixups that I noticed along the way

21 years agoadd close_uuid
rread [Thu, 5 Sep 2002 18:33:42 +0000 (18:33 +0000)]
add close_uuid

21 years agoadd page.c symink to lov until Phil can stop the madness
rread [Thu, 5 Sep 2002 18:22:47 +0000 (18:22 +0000)]
add page.c symink to lov until Phil can stop the madness

21 years agoFixed service request buffer race
eeb [Thu, 5 Sep 2002 16:33:51 +0000 (16:33 +0000)]
Fixed service request buffer race

21 years ago- Fixed some error messages in the LOV
pschwan [Thu, 5 Sep 2002 07:05:41 +0000 (07:05 +0000)]
- Fixed some error messages in the LOV
- lov_brw was using the same callback structure for two sets of callbacks, thus
overloading the refcount and causing serious problems (sometimes failing to
free the callback data, and I could believe sometimes freeing it too soon).
Alas, this does not appear to completely fix the iozone bugs we're seeing.

21 years agonew option: --dump <file> will dump debug log to <file> during cleanup
rread [Thu, 5 Sep 2002 06:24:02 +0000 (06:24 +0000)]
new option: --dump <file> will dump debug log to <file> during cleanup

21 years ago- Rewrote connmgr_iocontrol to handle UUID specification and changing.
shaver [Thu, 5 Sep 2002 00:33:19 +0000 (00:33 +0000)]
- Rewrote connmgr_iocontrol to handle UUID specification and changing.
- connmgr now uses class_connect and class_disconnect to build dummy exports,
  so that lctl can probe before newconn.

21 years agoadded newconn to lctl
rread [Wed, 4 Sep 2002 23:30:03 +0000 (23:30 +0000)]
added newconn to lctl

21 years ago- Update jt_obd_newconn to handle the new IOC_RECOVD_NEWCONN semantics.
shaver [Wed, 4 Sep 2002 23:07:40 +0000 (23:07 +0000)]
- Update jt_obd_newconn to handle the new IOC_RECOVD_NEWCONN semantics.
- Update obdctl out of habit.

21 years agoCommit braino fix that I already fixed locally.
adilger [Wed, 4 Sep 2002 20:32:10 +0000 (20:32 +0000)]
Commit braino fix that I already fixed locally.

21 years agoGet the directory inode semaphore, just to be copasetic with VFS locking rules
adilger [Wed, 4 Sep 2002 19:38:47 +0000 (19:38 +0000)]
Get the directory inode semaphore, just to be copasetic with VFS locking rules
(we don't use this for anything other than fs-init, so it isn't a big deal).

21 years agoadd hostname to gdb script
rread [Wed, 4 Sep 2002 19:17:32 +0000 (19:17 +0000)]
add hostname to gdb script

21 years agoAdd note to --format help that the autoformat feature is currently disabled
rread [Wed, 4 Sep 2002 18:08:24 +0000 (18:08 +0000)]
Add note to --format help that the autoformat feature is currently disabled

21 years agoremove old message about autoformat
rread [Wed, 4 Sep 2002 18:04:58 +0000 (18:04 +0000)]
remove old message about autoformat

21 years agoRemove compiler warning.
adilger [Wed, 4 Sep 2002 16:30:40 +0000 (16:30 +0000)]
Remove compiler warning.

21 years agoRemove fancy but broken statistics from obdecho.
adilger [Wed, 4 Sep 2002 16:29:14 +0000 (16:29 +0000)]
Remove fancy but broken statistics from obdecho.

21 years ago- I introduced a bug in the LOV, for the case where you're writing less than
pschwan [Wed, 4 Sep 2002 15:18:31 +0000 (15:18 +0000)]
- I introduced a bug in the LOV, for the case where you're writing less than
stripe_count stripes; fixed.
- Match bulk and reply buffers by source NID, which doesn't hurt and may even
fix a bug.

21 years agoUse tex2pdf.
gord-fig [Wed, 4 Sep 2002 15:05:52 +0000 (15:05 +0000)]
Use tex2pdf.

21 years agoUpdate changelog for new release.
adilger [Wed, 4 Sep 2002 14:05:21 +0000 (14:05 +0000)]
Update changelog for new release.

21 years agoFix same problem as before - we are dereferencing a null connection...
adilger [Wed, 4 Sep 2002 11:44:00 +0000 (11:44 +0000)]
Fix same problem as before - we are dereferencing a null connection...

21 years agoMinor fixup - don't memcpy data from past end of string.
adilger [Wed, 4 Sep 2002 09:28:54 +0000 (09:28 +0000)]
Minor fixup - don't memcpy data from past end of string.
Doesn't affect anyone except me, I think.

21 years ago* stripe size in the config file was being ignored and only the default
rread [Wed, 4 Sep 2002 09:11:28 +0000 (09:11 +0000)]
* stripe size in the config file was being ignored and only the default
value of 64k was being used.  Fixed this, but check your configs, as
some of them are set to a very low stripe size (4096) for testing.

21 years agoNew improved runregression script.
adilger [Wed, 4 Sep 2002 07:19:32 +0000 (07:19 +0000)]
New improved runregression script.
It creates an OST object at the start of the test and does the write test
before the read test.  This should allow it to run against a real OST.
It also does vector I/O testing for BRW read/write tests.

21 years agothe callback data refcount should be the stripe count, but bufcount. fixed.
pschwan [Wed, 4 Sep 2002 02:55:04 +0000 (02:55 +0000)]
the callback data refcount should be the stripe count, but bufcount.  fixed.

21 years agoUse pga[i].off instead of pga[i].pg->index. Even though Andreas just fixed
pschwan [Wed, 4 Sep 2002 02:10:45 +0000 (02:10 +0000)]
Use pga[i].off instead of pga[i].pg->index.  Even though Andreas just fixed
the page index, the OSC uses .off and I'd rather behave exactly like the OSC.
Baby steps towards iozone-over-many-stripes; it writes one 192k stripe before
timing out now.

21 years ago"fix" the page->index value for direct_IO pages.
adilger [Wed, 4 Sep 2002 01:57:22 +0000 (01:57 +0000)]
"fix" the page->index value for direct_IO pages.

21 years ago- initialize LDLM client data on obdfilter, not ost, because that's where
shaver [Wed, 4 Sep 2002 01:19:14 +0000 (01:19 +0000)]
- initialize LDLM client data on obdfilter, not ost, because that's where
  the ldlm requests come from.

21 years ago- more recovery API stuff
braam [Wed, 4 Sep 2002 00:23:30 +0000 (00:23 +0000)]
- more recovery API stuff
- special API description

21 years agoFix thinko in export cleanup.
adilger [Wed, 4 Sep 2002 00:16:44 +0000 (00:16 +0000)]
Fix thinko in export cleanup.

21 years agoDisable the reply type check until bluearc has a chance to integrate that fix.
pschwan [Wed, 4 Sep 2002 00:09:00 +0000 (00:09 +0000)]
Disable the reply type check until bluearc has a chance to integrate that fix.

21 years agoInitialize the obd spinlock and import list, each time a device is attached.
adilger [Tue, 3 Sep 2002 23:53:01 +0000 (23:53 +0000)]
Initialize the obd spinlock and import list, each time a device is attached.
Do only the minimal setup of obd_dev at module init time (it is already
zero'd for us by virtue of being a global variable).

21 years ago* move recover.c from llite to ptlrpc
rread [Tue, 3 Sep 2002 23:09:48 +0000 (23:09 +0000)]
* move recover.c from llite to ptlrpc

21 years ago* fix module cleanup, previous commit broke it
rread [Tue, 3 Sep 2002 23:04:22 +0000 (23:04 +0000)]
* fix module cleanup, previous commit broke it

21 years ago- remove the PTL_RPC_TYPE_* in favour of PTL_RPC_MSG_*
pschwan [Tue, 3 Sep 2002 22:59:30 +0000 (22:59 +0000)]
- remove the PTL_RPC_TYPE_* in favour of PTL_RPC_MSG_*
- request->rq_type is used to tell the rpc subsystem which buffers to use, so
  users of ptlrpc set that
- ptlrpc will fill in request->rq_reqmsg->type or repmsg->type, so it's now
  done in exactly one place
- it was impossible to get PTL_RPC_MSG_ERR before, because we clobbered it in
  ptlrpc_reply; fixed.
- move the incoming-request type verification into the service
- add some incoming-reply type verification to ptlrpc_queue_wait.  Just in case.

21 years agoCopyright update, remove old comment
adilger [Tue, 3 Sep 2002 21:48:15 +0000 (21:48 +0000)]
Copyright update, remove old comment

21 years agoFix header problem for building on RH kernels.
adilger [Tue, 3 Sep 2002 21:47:27 +0000 (21:47 +0000)]
Fix header problem for building on RH kernels.

21 years agoreply message initialisation fix
eeb [Tue, 3 Sep 2002 21:41:26 +0000 (21:41 +0000)]
reply message initialisation fix

21 years agoldlm_handle_convert was putting the lock immediately onto the conversion queue and
pschwan [Tue, 3 Sep 2002 21:15:19 +0000 (21:15 +0000)]
ldlm_handle_convert was putting the lock immediately onto the conversion queue and
then returning with flags=0, instead of flags=LDLM_FL_BLOCK_CONV.  Thus, the client
believed that the conversion had succeeded immediately, marked the lock accordingly,
and continued on its merry way, even if it didn't actually have the lock.

21 years agoAllow obdecho to handle create and destroy and setattr methods.
adilger [Tue, 3 Sep 2002 20:19:21 +0000 (20:19 +0000)]
Allow obdecho to handle create and destroy and setattr methods.
The first two are required so that we can reasonably emulate a
BA OST, and use runregression-net.sh against a BA OST (by using
the objid returned from the create call, and destroying the objid
afterwards).

21 years ago* use specific directories when loading modules in source tree, similar
rread [Tue, 3 Sep 2002 19:39:07 +0000 (19:39 +0000)]
* use specific directories when loading modules in source tree, similar
  to common.sh  (good to have Andreas watching over my shoulder, too)
* add hostname to debug dump filename

21 years agoAllow specifying an objid for test_getattr, the same as with test_brw.
adilger [Tue, 3 Sep 2002 19:09:37 +0000 (19:09 +0000)]
Allow specifying an objid for test_getattr, the same as with test_brw.
This should allow us to work with BA OSTs.
Re-set the objid and o_valid for each test_getattr/test_brw
Use strtoull() for getting objids, to allow 64-bit objids from BA OSTs.

21 years agoAllow specifying an objid for test_getattr to be the same as test_brw.
adilger [Tue, 3 Sep 2002 19:06:30 +0000 (19:06 +0000)]
Allow specifying an objid for test_getattr to be the same as test_brw.
This should allow us to work with real BA OSTs.

21 years ago- change the prefix of bulk descriptors to bd_ and bulk pages to bp_
braam [Tue, 3 Sep 2002 13:28:43 +0000 (13:28 +0000)]
- change the prefix of bulk descriptors to bd_ and bulk pages to bp_
- documentation updates for CDR

21 years agoWARNING: we currently crash on unmount after the last phase of runtests.
shaver [Tue, 3 Sep 2002 04:06:19 +0000 (04:06 +0000)]
WARNING: we currently crash on unmount after the last phase of runtests.
Phil and I are going to debug this more tomorrow, but this changeset needed
to land.  I feel bad about this, but not as bad as I'd feel sitting with a
thousand lines of changes in my tree another day! =)

- introduced new-style obd_import structure and spread it thickly across the
  ptlrpc infrastructure.
- merged ptlrpc_prep_req and prep_req2 into one import-taking prep_req function.
- added import for ldlm to use when calling out to client
- send local-export handle info in client_obd_connect, so that said import can
  talk to a proper export during ASTs.
- new unified recovd architecture for client- and server-side recovery.
- much header frottage to hand the export -> ldlm_export_data -> import
  structure nesting.
- commented out cli_sem usage, because clients aren't the right place for that
  any more, and we probably don't need them anyway.
- connection-sharing now requires a matching (nid, uuid) pair, to avoid bogus
  sharing on the one-host, server+client case.
- chain exports on the connection for doing per-connection recovery (soon, I
  swear).
- introduced dozens, if not thousands, of bugs, no doubt.

21 years agoRemove export of non-existent function.
adilger [Mon, 2 Sep 2002 21:58:43 +0000 (21:58 +0000)]
Remove export of non-existent function.

21 years agoRemove ancient Perl cruft from obd.c.
adilger [Mon, 2 Sep 2002 21:56:21 +0000 (21:56 +0000)]
Remove ancient Perl cruft from obd.c.
Do stricter checking on all strtoul() input to avoid strange problems when
we have, say, 'w' for the object ID or whatever.

21 years agoPass UUID along when we have it.
adilger [Mon, 2 Sep 2002 21:39:05 +0000 (21:39 +0000)]
Pass UUID along when we have it.

21 years agoWe don't expect to convert from ASCII to binary UUIDs - comment out code.
adilger [Mon, 2 Sep 2002 21:38:16 +0000 (21:38 +0000)]
We don't expect to convert from ASCII to binary UUIDs - comment out code.

21 years agoMinor cleanups - use uuids in uuid2dev and uuid2obd, instead of names.
adilger [Mon, 2 Sep 2002 21:36:04 +0000 (21:36 +0000)]
Minor cleanups - use uuids in uuid2dev and uuid2obd, instead of names.

21 years agoMinor cleanups.
adilger [Mon, 2 Sep 2002 21:33:37 +0000 (21:33 +0000)]
Minor cleanups.

21 years agoMinor cleanups.
adilger [Mon, 2 Sep 2002 21:26:40 +0000 (21:26 +0000)]
Minor cleanups.

21 years agoProtect the obd export list with a spinlock (we do not sleep anywhere while
adilger [Mon, 2 Sep 2002 21:24:12 +0000 (21:24 +0000)]
Protect the obd export list with a spinlock (we do not sleep anywhere while
holding this spinlock, so I don't believe we need a semaphore).

21 years agoCheck for a few error conditions.
adilger [Mon, 2 Sep 2002 20:52:12 +0000 (20:52 +0000)]
Check for a few error conditions.
Return more useful info from the object open call.
Return the error code from the commitrw call.

21 years agoAllow chbar generation on the HOWTO.
gord-fig [Mon, 2 Sep 2002 20:22:06 +0000 (20:22 +0000)]
Allow chbar generation on the HOWTO.

21 years agoAvoid bogus typecasting.
adilger [Mon, 2 Sep 2002 19:31:43 +0000 (19:31 +0000)]
Avoid bogus typecasting.

21 years agoPass 64-bit RPC XIDs around, since this is really how big they are.
adilger [Mon, 2 Sep 2002 19:21:15 +0000 (19:21 +0000)]
Pass 64-bit RPC XIDs around, since this is really how big they are.

21 years agoDon't check bulk contents if there was an error reported on the bulk itself.
adilger [Mon, 2 Sep 2002 19:04:21 +0000 (19:04 +0000)]
Don't check bulk contents if there was an error reported on the bulk itself.

21 years agoMinor change.
adilger [Mon, 2 Sep 2002 18:32:06 +0000 (18:32 +0000)]
Minor change.

21 years ago- fix route config and clean up errors
rread [Mon, 2 Sep 2002 17:59:35 +0000 (17:59 +0000)]
- fix route config and clean up errors

21 years agoI'm a moron and spin_unlock was a no-op.
pschwan [Mon, 2 Sep 2002 17:05:10 +0000 (17:05 +0000)]
I'm a moron and spin_unlock was a no-op.

21 years agoAllow automatic specification of the RPM locations by check-build.
gord-fig [Mon, 2 Sep 2002 16:43:11 +0000 (16:43 +0000)]
Allow automatic specification of the RPM locations by check-build.

21 years agoAdd PyXML in place of the previously-removed libxml2 Requires. It's good to
pschwan [Mon, 2 Sep 2002 16:29:04 +0000 (16:29 +0000)]
Add PyXML in place of the previously-removed libxml2 Requires.  It's good to
have Andreas looking over my shoulder.

21 years agoFix eric's extremely well-spotted locking bug. It's not clear that we even
pschwan [Mon, 2 Sep 2002 16:26:21 +0000 (16:26 +0000)]
Fix eric's extremely well-spotted locking bug.  It's not clear that we even
need this lock, but I'll fix it by adding an unlock_kernel until I can ask zach.

21 years agoremove libxml2 from the Requires list
pschwan [Mon, 2 Sep 2002 15:39:19 +0000 (15:39 +0000)]
remove libxml2 from the Requires list

21 years agoJames Newsome's last dlm regression patch. I haven't reviewed any of the
pschwan [Mon, 2 Sep 2002 06:49:05 +0000 (06:49 +0000)]
James Newsome's last dlm regression patch.  I haven't reviewed any of the
changes to ldlm_test.c yet, but I want to get this into the tree before we
drift too much.

21 years agoOops, missed a file in the mdc_enqueue change
pschwan [Mon, 2 Sep 2002 06:47:13 +0000 (06:47 +0000)]
Oops, missed a file in the mdc_enqueue change

21 years ago- remove the unused 'id' from mdc_enqueue
pschwan [Mon, 2 Sep 2002 06:38:17 +0000 (06:38 +0000)]
- remove the unused 'id' from mdc_enqueue
- fix yet another LBUG related to calling ptlrpc_error after allocating a reply
  buffer

21 years agos/jt_ptl_setup/jt_ptl_network/
pschwan [Mon, 2 Sep 2002 05:56:58 +0000 (05:56 +0000)]
s/jt_ptl_setup/jt_ptl_network/

21 years agoForgot to update changelog...
braam [Fri, 30 Aug 2002 22:58:40 +0000 (22:58 +0000)]
Forgot to update changelog...

21 years ago- saw a problem with the order of how filter was initialized vs. the error
rread [Fri, 30 Aug 2002 22:13:22 +0000 (22:13 +0000)]
- saw a problem with the order of how filter was initialized vs. the error
handling code.

21 years ago* allow any host to have static routes
rread [Fri, 30 Aug 2002 17:19:48 +0000 (17:19 +0000)]
* allow any host to have static routes

21 years agoAdd assertions to {push,pop}_ctxt because we are oopsing on dev in obdfilter
adilger [Fri, 30 Aug 2002 06:03:16 +0000 (06:03 +0000)]
Add assertions to {push,pop}_ctxt because we are oopsing on dev in obdfilter
in pop_ctxt(), maybe becase saved->pwd is NULL.

21 years agoResolve my silly conflict.
gord-fig [Fri, 30 Aug 2002 05:53:24 +0000 (05:53 +0000)]
Resolve my silly conflict.

21 years agoMike's fix for the missing exp_connection oops caused us to not remove
adilger [Fri, 30 Aug 2002 05:16:03 +0000 (05:16 +0000)]
Mike's fix for the missing exp_connection oops caused us to not remove
old exports from the export list.

It also seems wrong, somehow, to be using a lock on the ptlrpc connection
to manage items that are linked to the obd_device struct.  It would
probably make more sense to protect the export list with a lock on the
obd_device.  This is doubly true, because class_new_export() is adding
the export to the obd_device without holding any lock, because there is
no ptlrpc_connection there yet.

21 years ago* sample mcr-type setups, both using a router
rread [Fri, 30 Aug 2002 02:13:48 +0000 (02:13 +0000)]
* sample mcr-type setups, both using a router

21 years ago* fix some brainos so routing config will work
rread [Fri, 30 Aug 2002 02:12:47 +0000 (02:12 +0000)]
* fix some brainos so routing config will work

21 years ago add --tcpbuf option to lconf
rread [Fri, 30 Aug 2002 01:18:25 +0000 (01:18 +0000)]
 add --tcpbuf option to lconf

21 years ago* send and recv tcp buffer size is set
rread [Thu, 29 Aug 2002 23:50:47 +0000 (23:50 +0000)]
* send and recv tcp buffer size is set
* router opens connections to gateway

21 years ago* clean up mcr config script
rread [Thu, 29 Aug 2002 19:54:51 +0000 (19:54 +0000)]
* clean up mcr config script

21 years agoAllow lmc path to be overridden.
gord-fig [Thu, 29 Aug 2002 16:59:33 +0000 (16:59 +0000)]
Allow lmc path to be overridden.

21 years agoHandle the case where the XML file has not yet been created.
adilger [Thu, 29 Aug 2002 16:33:13 +0000 (16:33 +0000)]
Handle the case where the XML file has not yet been created.

21 years ago* first cut at routing config support, pretty tacky but might just work
rread [Thu, 29 Aug 2002 12:01:53 +0000 (12:01 +0000)]
* first cut at routing config support, pretty tacky but might just work
* print more diagnostics for cleanup errors
* various cleanups

21 years ago- add initial mcr.sh for echo test on mcr cluster
rread [Thu, 29 Aug 2002 11:59:15 +0000 (11:59 +0000)]
- add initial mcr.sh for echo test on mcr cluster
- turn autoformat off in lov, use lconf --reformat
- clean up llmount.sh

21 years agoFix up the spinlock situation around rq->rq_list and the lists it lives on.
adilger [Thu, 29 Aug 2002 07:11:32 +0000 (07:11 +0000)]
Fix up the spinlock situation around rq->rq_list and the lists it lives on.
We avoid a deadlock in ptlrpc_free_committed() when it calls ptlrpc_free_req().
Likewise when ptlrpc_cleanup_client() calls ptlrpc_free_req().

A note for Mike in ll_recover() for when it is re-enabled:

   We should not hold a spinlock over such a lengthy operation.
   If necessary, drop spinlock, do operation, re-get spinlock, restart loop.
   If we need to avoid re-processing items, then delete them from the list
   as they are replayed and re-add at the tail of this list, so the next
   item to process will always be at the head of the list.

21 years ago- obsolete
rread [Thu, 29 Aug 2002 06:58:02 +0000 (06:58 +0000)]
- obsolete

21 years ago* alloc memory and copy data into type->typ_name and type->typ_ops.
rread [Thu, 29 Aug 2002 06:33:36 +0000 (06:33 +0000)]
* alloc memory and copy data into type->typ_name and type->typ_ops.
This is prevent referencing pointers into others modules memory spaces after
those modules have been unloaded.

21 years agoAdd more details to device_list output.
rread [Thu, 29 Aug 2002 06:32:07 +0000 (06:32 +0000)]
Add more details to device_list output.
Added type->typ_refcnt as the last column.
Status field (2nd column) is now:
  - unconfigured device
  AT device is attached
  UP device is setup

21 years agoAdd lconf, lmc to the RPM.
gord-fig [Thu, 29 Aug 2002 05:04:37 +0000 (05:04 +0000)]
Add lconf, lmc to the RPM.

21 years agoDon't just delete the generated PDF.
gord-fig [Thu, 29 Aug 2002 03:29:58 +0000 (03:29 +0000)]
Don't just delete the generated PDF.

21 years agoOnly overwrite the magics, not the entire page to avoid excess CPU usage.
adilger [Thu, 29 Aug 2002 01:18:27 +0000 (01:18 +0000)]
Only overwrite the magics, not the entire page to avoid excess CPU usage.

21 years agoInstall and distribute lconf and lmc.
gord-fig [Thu, 29 Aug 2002 01:08:47 +0000 (01:08 +0000)]
Install and distribute lconf and lmc.

21 years agoUse a less evil port number.
adilger [Thu, 29 Aug 2002 01:03:01 +0000 (01:03 +0000)]
Use a less evil port number.

21 years agoExit from XML configs if there is an error.
adilger [Thu, 29 Aug 2002 01:02:22 +0000 (01:02 +0000)]
Exit from XML configs if there is an error.

21 years agoUse port 988 instead of 888.
adilger [Thu, 29 Aug 2002 01:01:07 +0000 (01:01 +0000)]
Use port 988 instead of 888.

21 years agoTry and work with different versions of dbench.
adilger [Thu, 29 Aug 2002 01:00:41 +0000 (01:00 +0000)]
Try and work with different versions of dbench.

21 years agoAdd dcount assertions in obdfilter to catch any funny business (still
adilger [Thu, 29 Aug 2002 00:59:08 +0000 (00:59 +0000)]
Add dcount assertions in obdfilter to catch any funny business (still
open-unlink problems).

Only try to get a cache page if we already have a locked page.  This
is not Peter's "ignore page locks" code yet, but it's what I have in
my tree already and want to flush so I can test at LLNL.

21 years agoZap test_brw/obdecho pages we are going to read into, to ensure that we
adilger [Thu, 29 Aug 2002 00:51:54 +0000 (00:51 +0000)]
Zap test_brw/obdecho pages we are going to read into, to ensure that we
actually are getting new data therein, instead of old data.

Fix a kunmap imbalance for the non-verify case.

21 years agoAdd statfs method to LOV.
adilger [Thu, 29 Aug 2002 00:38:34 +0000 (00:38 +0000)]
Add statfs method to LOV.
Try to get valid data for getattr, if first objid fails.
Extract block count and mtime/ctime information from all objects.

21 years agoUse o_brw instead of o_getattr to distinguish between MDC and OSC operations.
adilger [Thu, 29 Aug 2002 00:35:15 +0000 (00:35 +0000)]
Use o_brw instead of o_getattr to distinguish between MDC and OSC operations.
If we ever make MDC use the OBD methods it would break.

21 years agoWe already set last_committed and last_xid for all packets, no need to do
adilger [Wed, 28 Aug 2002 23:50:06 +0000 (23:50 +0000)]
We already set last_committed and last_xid for all packets, no need to do
this for getstatus explicitly.

Store an xid in last_xid and not last_rcvd for reply packets.  Mike please
verify this.  I think it will help resolve strangeness I have observed in
the cleanup of expired requests on the clients, but that still needs some
investigation.

Add a spinlock to protect 64-bit increment of mds_last_rcvd, and also
make sure we only use each transaction value a single time.  This
would probably be a very rare issue for recovery, but may as well get
it now.