Whamcloud - gitweb
fs/lustre-release.git
22 years ago- Added an 'xid' field to struct niobuf
pschwan [Sun, 3 Mar 2002 04:34:22 +0000 (04:34 +0000)]
- Added an 'xid' field to struct niobuf
- Fixed bug in OST/OSC bulk handling by using this new field
- fixed double free in OST

22 years ago- tiny fix to ost_request.c
braam [Sun, 3 Mar 2002 04:25:40 +0000 (04:25 +0000)]
- tiny fix to ost_request.c

22 years ago- minor fixes to get mdc_readpage working over again.
braam [Sun, 3 Mar 2002 03:17:46 +0000 (03:17 +0000)]
- minor fixes to get mdc_readpage working over again.

22 years agoMerged branch 'peter' with the tip. Pre-merge tag is 't_20020302_networking'.
pschwan [Sun, 3 Mar 2002 02:29:47 +0000 (02:29 +0000)]
Merged branch 'peter' with the tip.  Pre-merge tag is 't_20020302_networking'.
Avert your eyes.

22 years agonew files - restructure RPC further
braam [Thu, 28 Feb 2002 08:49:03 +0000 (08:49 +0000)]
new files - restructure RPC further

22 years ago- added rq_type field to ptlrpc_request
pschwan [Thu, 28 Feb 2002 03:05:25 +0000 (03:05 +0000)]
- added rq_type field to ptlrpc_request
- this field must be one of PTLRPC_REQUEST, PTLRPC_REPLY, PTLRPC_BULK
- mdc_request now waits for directory pages to be sent via the bulk portal
- change MDS bulk waitq to the new style
- ptl_send_buf no longer takes the 'is_request' argument; this is encoded in
  the rq_type mentioned above
- added ptl_abort_rpc() to tear down reply structures for aborted requests.
- refactored ptl_send_buf code to rid ourselves of some of the more gross
  bulklen hacks

22 years ago-small is better for now
braam [Wed, 27 Feb 2002 22:02:48 +0000 (22:02 +0000)]
-small is better for now

22 years ago- barrier for potential cache problems
braam [Wed, 27 Feb 2002 20:57:00 +0000 (20:57 +0000)]
- barrier for potential cache problems

22 years agomds/handler.c: init the waitqueue before we call ptl_send_buf()
pschwan [Wed, 27 Feb 2002 20:06:14 +0000 (20:06 +0000)]
mds/handler.c: init the waitqueue before we call ptl_send_buf()

22 years agoyet more tracking of lost packets
braam [Wed, 27 Feb 2002 18:23:31 +0000 (18:23 +0000)]
yet more tracking of lost packets

22 years ago- add open and close to MDC interface (not used yet)
braam [Wed, 27 Feb 2002 17:49:34 +0000 (17:49 +0000)]
- add open and close to MDC interface (not used yet)
- very defensive logging to track our lost packet

22 years ago- fix connection find on OST
braam [Wed, 27 Feb 2002 10:16:11 +0000 (10:16 +0000)]
- fix connection find on OST

22 years ago- bug fix to unload modules with safe wait functions
braam [Wed, 27 Feb 2002 09:56:13 +0000 (09:56 +0000)]
- bug fix to unload modules with safe wait functions
- rewrite of ptlrpc client system
  - pivot around ptlrpc_client
  - mdc client & lustre_light uses generic routines
  - osc can do the same (tomorrow)
- different event queues in rpc.c freeing packets & getting replies
- every unexpected event or PtlError is now a BUG in rpc.c
- change gdb script output to allow reloading everything

22 years ago- New style waitqueues in ost and mds main
pschwan [Wed, 27 Feb 2002 05:41:04 +0000 (05:41 +0000)]
- New style waitqueues in ost and mds main

22 years agoNew style waitqueues for OSC
pschwan [Wed, 27 Feb 2002 05:07:30 +0000 (05:07 +0000)]
New style waitqueues for OSC

22 years ago- Don't set request->rq_repbuf until the reply has actually come in
pschwan [Wed, 27 Feb 2002 05:02:20 +0000 (05:02 +0000)]
- Don't set request->rq_repbuf until the reply has actually come in

22 years agoFix the waitqueue race in the mdc, I hope.
pschwan [Wed, 27 Feb 2002 04:58:37 +0000 (04:58 +0000)]
Fix the waitqueue race in the mdc, I hope.

22 years ago- llmount.sh creates for you a /tmp/ogdb in the hostfs, and waits for you to
pschwan [Wed, 27 Feb 2002 04:26:22 +0000 (04:26 +0000)]
- llmount.sh creates for you a /tmp/ogdb in the hostfs, and waits for you to
press enter before it continues.  This lets you set breakpoints and debug
the mount/startup process.
- Run 'mynid localhost' instead of 'mynid'

22 years agoAdd many debug messages to the RPC path
pschwan [Wed, 27 Feb 2002 04:16:27 +0000 (04:16 +0000)]
Add many debug messages to the RPC path

22 years agoAdded paths to obdctl's 'modules' command, making it immediately useful:
pschwan [Tue, 26 Feb 2002 22:32:29 +0000 (22:32 +0000)]
Added paths to obdctl's 'modules' command, making it immediately useful:

/r/usr/src/obd/utils/obdctl modules > /r/root/ogdb

22 years ago- wrong offset in commit page
braam [Mon, 25 Feb 2002 23:34:54 +0000 (23:34 +0000)]
- wrong offset in commit page

22 years agoWrong object id sent to obd - fixed
braam [Mon, 25 Feb 2002 20:45:32 +0000 (20:45 +0000)]
Wrong object id sent to obd - fixed

22 years ago- uninitialized error code
braam [Mon, 25 Feb 2002 18:57:03 +0000 (18:57 +0000)]
- uninitialized error code
- uninitialized mode of data objects
both fixed

22 years agominor changes to debug prepw and prepare page
braam [Mon, 25 Feb 2002 18:13:41 +0000 (18:13 +0000)]
minor changes to debug prepw and prepare page
remove some debug crud from the OST handler

22 years ago- add the truncate obd call
braam [Mon, 25 Feb 2002 15:49:21 +0000 (15:49 +0000)]
- add the truncate obd call

22 years agoGetting rid of tabs makes me happy, just humour me.
pschwan [Mon, 25 Feb 2002 08:00:48 +0000 (08:00 +0000)]
Getting rid of tabs makes me happy, just humour me.

22 years ago- minor debugging fixes
braam [Mon, 25 Feb 2002 07:43:46 +0000 (07:43 +0000)]
- minor debugging fixes
- write event BEFORE calling callback

22 years agoThe most tedious patch in history. Changes almost all of the 'printk's to
pschwan [Sun, 24 Feb 2002 07:09:01 +0000 (07:09 +0000)]
The most tedious patch in history.  Changes almost all of the 'printk's to
'CERROR's.  Also contains a lot of cleanup, removing things like __FUNCTION__
from CDEBUG and CERROR statements which already display that information.

22 years ago- Added DEBUG_SUBSYSTEMs
pschwan [Sun, 24 Feb 2002 06:36:26 +0000 (06:36 +0000)]
- Added DEBUG_SUBSYSTEMs
- started using CERROR, which goes both to the console and the ringbuffer,
  instead of printk

22 years agoAdded a ``modules [path]'' command to obdctl, which is complete modulo the
pschwan [Sun, 24 Feb 2002 03:57:55 +0000 (03:57 +0000)]
Added a ``modules [path]'' command to obdctl, which is complete modulo the
fact that our directory names do not currently match our module names.

The best part?  I did it with zero lines of kernel code and a function that
fits on one page.

22 years agoslightly more debugging in rpc.c - something is wrong there.
braam [Sun, 24 Feb 2002 00:01:11 +0000 (00:01 +0000)]
slightly more debugging in rpc.c - something is wrong there.

22 years ago- Converted lots of debugging printks to CDEBUGs
pschwan [Sat, 23 Feb 2002 19:06:25 +0000 (19:06 +0000)]
- Converted lots of debugging printks to CDEBUGs
- Removed duplicated debug macros and defines; they live in portals now

- All of OBD's debug information now lives in the Portals debug buffer.
Use 'ptlctl get_debug' to view the buffer.  The buffer's size is currently
hardcoded at 1 megabyte in module.c, but that's easy enough to change to a
module argument.

22 years ago MDS wasn't handling refcount on vfsmount's correctly,
braam [Fri, 22 Feb 2002 23:35:51 +0000 (23:35 +0000)]
 MDS wasn't handling refcount on vfsmount's correctly,
 resulting in oops.

22 years ago- I had set the rpc buffer size down to 4k to ease testing, the smaller
behlendo [Fri, 22 Feb 2002 02:25:59 +0000 (02:25 +0000)]
- I had set the rpc buffer size down to 4k to ease testing, the smaller
  buffer size accidentally got commited with the other changes.  This
  change just resets it to the original 64k size.

22 years ago- Added some fields for rpc double buffering
behlendo [Fri, 22 Feb 2002 02:05:30 +0000 (02:05 +0000)]
- Added some fields for rpc double buffering
- Added a NEXT_INDEX() macro
- Adjusted mds/handler.c and ost/ost_handler.c to notify
  the rpc layer when finished processing an event, this
  ensures the buffers are recycled.
- Added support for double buffering

22 years agoPass 1 of debugging and leak cleanup:
pschwan [Thu, 21 Feb 2002 23:15:18 +0000 (23:15 +0000)]
Pass 1 of debugging and leak cleanup:

- use OBD_ALLOC/OBD_FREE everywhere
- fixed a few leaks
- cleaned up a few macros
- load obdclass.o before ptlrpc.o now, since obdclass exports the memory counter
- fixed broken strdup()
- made llmount.sh a little more robust

22 years ago*** empty log message ***
braam [Thu, 21 Feb 2002 23:02:57 +0000 (23:02 +0000)]
*** empty log message ***

22 years ago- elimininate the system calls from filter obd
braam [Thu, 21 Feb 2002 23:02:32 +0000 (23:02 +0000)]
- elimininate the system calls from filter obd
- use module_init macros not old style init_module functions
- remove unused crud from lustre_light.h use the good stuff more
- move directory readpage operations to dir.c
- cleanup rw.c significantly; it's actually understandable now.
- attribute handling: indicate to setattr to call truncate on OBD or not
- add osc destroy
- fix mds_reint_unlink (checked mode wrongly)
- add ll_delete_inode (to also go to the OST to remove data object)
- add truncate.c test program to tests
- avoid using ambigous READ/WRITE constants

22 years agoUpdated 2.4.17 obd kernel patch.
behlendo [Thu, 21 Feb 2002 00:19:31 +0000 (00:19 +0000)]
Updated 2.4.17 obd kernel patch.

22 years ago*** empty log message ***
braam [Wed, 20 Feb 2002 00:56:39 +0000 (00:56 +0000)]
*** empty log message ***

22 years agoThis adds most of the metadata infrastructure: I think that all commands
braam [Tue, 19 Feb 2002 16:43:53 +0000 (16:43 +0000)]
This adds most of the metadata infrastructure: I think that all commands
(mkdir, link, symlink, rm, rmdir, mknod, mv etc) now work.  Probably the
server code is not multithread safe (particularly for renames, the others
might be OK).   I expect other fragilities and perhaps some errors with
setting {a,m,c}times, owners etc.

- add the beginning of an obd command that prints the information
  for UML gdb (class/class_obd.c)
- try to ignore errors on creating symbolic links in Makefiles
- the patch is 2.4.17 again for ext2obd, ghees, let's fix that.
- to avoid confusion, null out the buffer of result nio's in prepw
- add other reintegration functions.

22 years agoOops; I added the wrong page.c
braam [Thu, 14 Feb 2002 05:33:26 +0000 (05:33 +0000)]
Oops; I added the wrong page.c

22 years ago- changed back to using no optimization for UML modules, and no -g2
braam [Thu, 14 Feb 2002 05:31:24 +0000 (05:31 +0000)]
- changed back to using no optimization for UML modules, and no -g2
- set inode operations in object create
- s/WRITE/OBD_BRW_WRITE/ where applicable
- added much of the infrastructure for page reads/writes
- update inode->i_size on the MDS if extending the file
- removed extra mntget() which hung the loop device
- misc bugfixes
- changed notify_change to i_op->setattr; otherwise we get the mtime and ctime
  from the mds
- set reply lengths (or your replies get dropped!)

22 years agoclass/class_obd.c: small OBD_ATTACHED sanity cleanup; OBD_SET_UP fix.
braam [Wed, 13 Feb 2002 23:17:05 +0000 (23:17 +0000)]
class/class_obd.c: small OBD_ATTACHED sanity cleanup; OBD_SET_UP fix.
include/linux/lustre_net.h: set the LUSTRE_NAL back to socknal
include/linux/class_obd.h: added full license boilerplate, #define cleanup
include/linux/obd_osc: there was some confusion with obd_ost; fixed.

- added bits to structures for ost_request
- added OSC/OST portals communication
- added osc_setattr()
- added rpc_unregister_service in an attempt to get a handle on cleanup
- brought tests/ostreq.sh into the 1990s
- fixed a trivial obdctl setattr bug

22 years ago- set_page_dirty ifdef for kernels on either side of 2.4.10
braam [Wed, 13 Feb 2002 20:31:30 +0000 (20:31 +0000)]
- set_page_dirty ifdef for kernels on either side of 2.4.10

22 years agoMinor fixes to
braam [Sat, 9 Feb 2002 05:12:51 +0000 (05:12 +0000)]
Minor fixes to
 - setup scripts to do Elan stuff
 - a few changes to page cache stuff for 2.4.9 (we finally know how to #if it)
 - the usual battle the ext2obd patch
 - removed aclocal.m4 (enhances immunity from automake versions)

22 years agoNew scripts to aid in testing; on one machine you should be able to run
pschwan [Thu, 7 Feb 2002 21:24:33 +0000 (21:24 +0000)]
New scripts to aid in testing; on one machine you should be able to run

sh tests/llmount-server.sh

and on another machine:

sh tests/llmount-client.sh
mount -t lustre_light -o device=3 none /mnt/obd

and do things like:

ls -ld /mnt/obd
chmod 777 /mnt/obd

This works now--even over the socknal, on an SMP kernel (although I haven't
tested it on an SMP machine)

You'll need to edit the SERVER variable in llmount-client.sh.  If you want
to get really fancy, you'll make it a command-line option.

22 years agoMinor cleanups and debugging statements added.
braam [Wed, 6 Feb 2002 18:25:41 +0000 (18:25 +0000)]
Minor cleanups and debugging statements added.

22 years ago- zero out the request structure after allocation
braam [Fri, 1 Feb 2002 19:32:21 +0000 (19:32 +0000)]
- zero out the request structure after allocation
- do readpage via portals
- added bulk moving functionality to rpc.c
- cleaned up one small MD leak

22 years agolllocalmount: startup and mount a lustre lite FS with no networking
braam [Fri, 1 Feb 2002 18:20:06 +0000 (18:20 +0000)]
lllocalmount: startup and mount a lustre lite FS with no networking
llmountcleanup: indeed.

22 years agoMinor bugfix to allow multiple networked operations:
braam [Fri, 1 Feb 2002 18:18:06 +0000 (18:18 +0000)]
Minor bugfix to allow multiple networked operations:
- small fix to avoid crash in getattr
- properly unlink MDs and MEs after we send buffers
- fixups to the test scripts

22 years ago- new RPC infrastructure:
braam [Thu, 31 Jan 2002 21:27:21 +0000 (21:27 +0000)]
- new RPC infrastructure:
 - struct service with callback functions, buffers and Portals support
 - services are Portal aware, but the old null transport may still work.
 - lustre peer moved to Portals, as was all connection management
 -
- fixed ptl entry constants for incoming REQUESTS, REPLIES for MDS, OST etc.
- all rpc's now have a peer destination (which can be NULL)
- replies are now pre-allocated
- the XID's are used to number transactions _and_ as match entries
- packet movement differentiates Portals and non-portals case.
- reception of packets uses the new Portals callbacks
- module initialization finds peer's by UUID.  For now "mds" is a UUID.
- callbacks are used for (1) incoming requests (2) freeing outgoing packets
(3) incomding reply.  Code reuse was possible.
- adapted all the test scripts (mdc and llmount)

22 years ago- and the new files
braam [Wed, 30 Jan 2002 21:36:01 +0000 (21:36 +0000)]
- and the new files

22 years agoChanges to request processing:
braam [Wed, 30 Jan 2002 21:33:40 +0000 (21:33 +0000)]
Changes to request processing:

- all requests are now ptlrpc requests.
- better packing/unpacking once more
- new directory with Portals RPC infrastructure.
- wait and restart hooks event handlers

22 years agoChanges for file creation and small fixes.
braam [Mon, 28 Jan 2002 19:06:15 +0000 (19:06 +0000)]
Changes for file creation and small fixes.

22 years ago- more systematic unpacking/packing and handling of update records
braam [Thu, 24 Jan 2002 08:04:57 +0000 (08:04 +0000)]
- more systematic unpacking/packing and handling of update records
- create working on MDS (needs some refinements to store data object id)

22 years agoAdd missing files, cleanup .cvsignore's
braam [Wed, 23 Jan 2002 18:20:23 +0000 (18:20 +0000)]
Add missing files, cleanup .cvsignore's

22 years agoIt's been a good day: chmod/chown and friends now work for Lustre Light.
braam [Wed, 23 Jan 2002 04:59:12 +0000 (04:59 +0000)]
It's been a good day: chmod/chown and friends now work for Lustre Light.
 - most of the changes are in the packing.  It will take a little
   bit more time before this completely settles I think, but we have
   a pretty efficient system now.
 - added initial infrastructure for MDS update records.
 - the MDS server has a new routine to change the attributes: the associated
   RPC call is MDS_REINT.
 - Lustre Light is simpler than OBDFS since it doesn't have to
   change inodes when it creates them: that's part of the MDS transactions we
   send over -- removed many "change inode" and "setattr" variants.

22 years ago assorted bug fixes and a working directory readpage routine
braam [Mon, 21 Jan 2002 23:58:26 +0000 (23:58 +0000)]
 assorted bug fixes and a working directory readpage routine
 for Lustre Light.

22 years agoNasty cflag confusion in configure.ac.
braam [Mon, 21 Jan 2002 05:52:56 +0000 (05:52 +0000)]
Nasty cflag confusion in configure.ac.

22 years agodefaults back to format
braam [Mon, 21 Jan 2002 02:33:30 +0000 (02:33 +0000)]
defaults back to format

22 years agosmall fix.
braam [Mon, 21 Jan 2002 02:31:40 +0000 (02:31 +0000)]
small fix.

22 years agofix the obdfssetup script to use obdext2 (not filter) for now
braam [Mon, 21 Jan 2002 01:39:01 +0000 (01:39 +0000)]
fix the obdfssetup script to use obdext2 (not filter) for now
patch - revision one: pagecache_lock was missing. What else is missing?

22 years agocleanups coming from the new functions/macros
braam [Mon, 21 Jan 2002 01:36:04 +0000 (01:36 +0000)]
cleanups coming from the new functions/macros

22 years agoFix a broken macro.
braam [Mon, 21 Jan 2002 01:17:17 +0000 (01:17 +0000)]
Fix a broken macro.

22 years agothis minimal patch might actually work.
braam [Mon, 21 Jan 2002 00:48:30 +0000 (00:48 +0000)]
this minimal patch might actually work.
filter obd won't work: we don't need it now.

22 years agoSmall fixes to get lustre_light mount working over the request handling
braam [Sun, 20 Jan 2002 05:03:56 +0000 (05:03 +0000)]
Small fixes to get lustre_light mount working over the request handling
infrastructure.

22 years agoA few missing files...
braam [Sun, 20 Jan 2002 00:49:49 +0000 (00:49 +0000)]
A few missing files...

22 years agoSmall fixes to the request processing.
braam [Sun, 20 Jan 2002 00:45:19 +0000 (00:45 +0000)]
Small fixes to the request processing.
Split off the metadata client into a separate module, probably easier
to test  things that way.

22 years agoFix small typos in ost/osc
braam [Fri, 18 Jan 2002 06:01:30 +0000 (06:01 +0000)]
Fix small typos in ost/osc

22 years agoomitted test script
braam [Thu, 3 Jan 2002 18:39:07 +0000 (18:39 +0000)]
omitted test script

22 years agoforgotten two files
braam [Thu, 3 Jan 2002 18:37:25 +0000 (18:37 +0000)]
forgotten two files

22 years agoSmall bug fixes. OSC/OST almost working.
braam [Thu, 3 Jan 2002 17:10:42 +0000 (17:10 +0000)]
Small bug fixes.  OSC/OST almost working.

22 years agoit _compiles_ finally on 7.1 and 7.2, but the solution is not great.
braam [Sun, 30 Dec 2001 18:18:33 +0000 (18:18 +0000)]
it _compiles_ finally on 7.1 and 7.2, but the solution is not great.

22 years agoLustre light client request module
braam [Sun, 30 Dec 2001 06:43:06 +0000 (06:43 +0000)]
Lustre light client request module

22 years agoMost of the code for the OST target.
braam [Sun, 30 Dec 2001 06:17:15 +0000 (06:17 +0000)]
Most of the code for the OST target.

 lib/obd_pack.c - routines for packing/unpacking of OST requests
 lib/mds_pack.c - fixes (should be 64bit safe now)
 autogen.sh - no more Makefile.in in CVS
 tests/testreq.c tests/umreq.sh tests of the RPC infrastructure

22 years ago- fixes to the MDS: it now answers requests from the request test
braam [Sat, 29 Dec 2001 00:20:44 +0000 (00:20 +0000)]
- fixes to the MDS: it now answers requests from the request test
  program.
- beginning of OST client/server code - needs work.

22 years agoA solid part of the MDS request processing infrastructure.
braam [Fri, 28 Dec 2001 04:41:15 +0000 (04:41 +0000)]
A solid part of the MDS request processing infrastructure.
  - packing/unpacking and definition of requests
  - a server daemon
  - getattr call coded

Lustre idl file created:
  - move all interface definitions here
  - that's ioctl's, mds requests/replies & obd requests/replies

Makefiles:
  - configure.in can handle user mode linux now

utils:
  - fixed the obdctl library to support old version of readline.

22 years agosmall changes and fixes to get the filter OBD working fully.
braam [Mon, 17 Dec 2001 15:00:39 +0000 (15:00 +0000)]
small changes and fixes to get the filter OBD working fully.
It seems to work well under ext2, XFS and Reiser at the moment.

22 years ago- obd filter works over ext2, but bonnie++ is buggie
braam [Thu, 13 Dec 2001 22:21:05 +0000 (22:21 +0000)]
- obd filter works over ext2, but bonnie++ is buggie
- obd filter doesn't work over reiser
- obdctl now works and is taking over from obdcontrol

22 years agoNumerous changes:
braam [Tue, 11 Dec 2001 21:08:42 +0000 (21:08 +0000)]
Numerous changes:
  - new control utility  - no Perl
  - filter obd halfway coded ; requires new patches to kernel
  - big cleanup in the class driver with better packed ioctls
  - macros to pack other things, like network packets

22 years agoTwo small changes to support large files.
braam [Thu, 8 Nov 2001 19:54:43 +0000 (19:54 +0000)]
Two small changes to support large files.
 - internally open files with O_LARGEFILE
 - s_maxbytesize set to 64G (should do info for that).

22 years agolarge file printk cleanups
braam [Thu, 8 Nov 2001 19:11:49 +0000 (19:11 +0000)]
large file printk cleanups

22 years agofix minor kmap/kunmap mismatch.
braam [Thu, 8 Nov 2001 18:56:34 +0000 (18:56 +0000)]
fix minor kmap/kunmap mismatch.

22 years agoMinor compilation fixes for RedHat 2.4.9
braam [Tue, 6 Nov 2001 19:05:51 +0000 (19:05 +0000)]
Minor compilation fixes for RedHat 2.4.9
Fix reference decreasing bug in namei.c

22 years agominor fix to the patch
braam [Tue, 6 Nov 2001 18:42:41 +0000 (18:42 +0000)]
minor fix to the patch

22 years agocreated
braam [Tue, 6 Nov 2001 04:20:34 +0000 (04:20 +0000)]
created

22 years agoIf you free a string ... it can cost a lot of time.
braam [Mon, 5 Nov 2001 22:20:52 +0000 (22:20 +0000)]
If you free a string ... it can cost a lot of time.

22 years agomany changes: seems ok on 2.4.9
braam [Mon, 5 Nov 2001 07:35:40 +0000 (07:35 +0000)]
many changes: seems ok on 2.4.9

22 years agoBack out size/blocks changes.
gord-fig [Thu, 25 Oct 2001 16:50:17 +0000 (16:50 +0000)]
Back out size/blocks changes.

22 years agoSegregate block and size counting between obdfs and obdext2.
gord-fig [Wed, 24 Oct 2001 20:51:25 +0000 (20:51 +0000)]
Segregate block and size counting between obdfs and obdext2.

22 years agoRemove debugging clutter.
gord-fig [Wed, 24 Oct 2001 19:38:29 +0000 (19:38 +0000)]
Remove debugging clutter.

22 years agoAnother very major cleanup:
braam [Sun, 21 Oct 2001 02:52:05 +0000 (02:52 +0000)]
Another very major cleanup:
 - new lookup/new_inode code based on iget4 (saves RPC's)
 - reorganized the layout of ext2_obd.c
 - the bonnie++ bug can be fixed by patching the 2.4.3 kernel.

22 years agoAgain, only use loop device for testing.
gord-fig [Fri, 19 Oct 2001 22:54:38 +0000 (22:54 +0000)]
Again, only use loop device for testing.

22 years agoThis fixes a race between killing inodes in obdfs and in ext2obd.
braam [Fri, 19 Oct 2001 18:54:07 +0000 (18:54 +0000)]
This fixes a race between killing inodes in obdfs and in ext2obd.

When files are kept open in obdfs it is important not to propagate
the 0 link count to ext2obd since that would delete the inode prematurely.

This one seems to run dbench 50 ok.

22 years agofairly massive change to remove write_inode methods and switch
braam [Fri, 19 Oct 2001 16:06:28 +0000 (16:06 +0000)]
fairly massive change to remove write_inode methods and switch
to setattr everywhere -- dbench 50 completes, so does bonnie++.

22 years agofix statfs bug and a truncation issue.
braam [Fri, 19 Oct 2001 05:17:34 +0000 (05:17 +0000)]
fix statfs bug and a truncation issue.

22 years agofix setattr bug; check in symlink changes
braam [Fri, 19 Oct 2001 04:13:53 +0000 (04:13 +0000)]
fix setattr bug; check in symlink changes

22 years agoDon't use printk unconditionally.
gord-fig [Thu, 18 Oct 2001 23:38:29 +0000 (23:38 +0000)]
Don't use printk unconditionally.

22 years agoAdd code to help debug the rmmod hang.
gord-fig [Thu, 18 Oct 2001 22:11:07 +0000 (22:11 +0000)]
Add code to help debug the rmmod hang.