b=13181
i=alex
i=nikita

Don't take the BKL in fsfilt_ext3_setattr() for 2.6 kernels.
It causes scheduling issues when removing large files.

Use correct structure name.

b=2369
i=johann, deen

replace inode->i_size access with i_size_read/write

i=adilger, girish

Small change in ldiskfs/Makefile.in for new fiemap.h file.

Also fiemap patch chanegs the definition of ext3_prepare_callback, so effect the change in fsfilt_ext3.c.

b=12211
i=green, johann

suppress every allocation attempt message noise, and shut off a compile
warning on x86_64 in obd_alloc_fail(), due to size_t being unsigned long
there.

b=11802
i=green
i=scjody

attachment 11752 / rely on AUTOCONF_INCLUDED instead of HAVE_KERNEL_CONFIG_H

b=12606
i=shadow, nikita

Description: don't use GFP_* in generic Lustre code.
Details    : Use cfs_alloc_* functions and CFS_* flags for code portability.

b=12211
i=adilger
i=green

fix LTS AutoBuild error.

b=12211
i=adilger
i=green

Description: randomly memory allocation failure util
Details    : Make lustre randomly failed allocating memory for testing purpose.

To improve performance, disable most logging (for debug purposes)
by default.  VFSTRACE, RPCTRACE, and DLMTRACE are now off by
default, and HA includes contains fewer messages.

Original patch by adilger.

b=12417
i=scjody
i=green

b=11039
i=adilger
i=alex
i=kalpak

attachment 10893 / fsfilt_ext3: mainline extents support

Quiet verbose error message on read-only filesystem.  We will already have
been notified of this.
b=10595
i=behlendo

removed. In this case we need drop using os depended interface to
kmem_cache and use cfs_mem_cache API.

b=12764
i=green
i=wangdi

Move ldiskfs to its own package and CVS module.

b=12068
i=mjmac
i=brian

remove fs_prep_san_write operation

b=11039
i=adilger
i=nathan

Use ext3_journal_start/stop (through fsfilt_ext3_journal_start/stop())
instead of journal_start/stop.
+ Fix SLES9 build issue.

b=11743
i=alex
i=adilger

Reverve patch from bug 11743 since it breaks SLES9 build.
For the record, the revision associated to the commit was 1.176.2.14.

b=11743

Add the QUOTA mount option on the fly for kernel newer than 2.6.12.
Add a call to fsfilt_setup() in filter_common_setup().

b=12538
i=alex
i=adilger

Use ext3_journal_start/stop (through fsfilt_ext3_journal_start/stop())
instead of journal_start/stop

b=11743
i=alex
i=adilger

i=adilger
i=green

Lustre methods to set/get the inode version. Only sles10 and 2.6.18 have inode version support and older kernels may not need this support. Hence added a HAVE_DISK_INODE_VERSION in ext3_fs_i.h, so that lustre can decide whether to compile inode-version code or not without requiring a cnfigure time check.

Add uninit_groups (fast fsck) support.
b=11264
i=kalpak
i=alex

Add improved mballoc support.
i=alex (original patch)
i=adilger

Add new ldiskfs2 module based on ldiskfs.

b=11981
r=adilger
r=nathan

Move quota support from ldiskfs to lvfs.

b=11981
r=adilger
r=nathan

i=green
atomic copy remote log instead of appending records -- earlier records may
have been modfied, and record count was wrong anyhow.

assert possibly caused by uninitialized var

r=adilger
set external journal device read-only

b=11248

b=11248
r=adilger

 - very messy debug bits to check for double direntries

change CWARN to C_DEBUG

should reset the mem to 0, if not get the indicated next block.
which should fix the audit log oops.

b=7085

 - port old fixes from b1_4 to HEAD (extents related)
 - unify extents & fsfilt codebase with fsfilt

- md_close() chnaged to use mdc_op_data like other md_* functions instead of using obdo.
- fixed prototype declaration for llapi_lov_get_uuids() - bug #6900
- fixed error message in ll_gns_mount_object() - bug #7146

- a lot of fixes about memory allocation/freeing:
  - checks is memory is actually allocated by OBD_ALLOC() before sing it and do appropriate
    actions otherwise.

  - fixed many leaks on error paths.
  - fixed missed OBD_FREE(capa_hash) in capa_cache_cleanup()

- fixed few on stack allocatoins of mdc_op_data in new landed stuff.

- in smfs_mount_cache() fixes correct cleanup in the cases of memory allocation failures.
  Do mntput() for new mounted cache_mnt on error path.

- in smfs_umount_cache() nullify smb->smsi_cache_ftype and smb->smsi_ftype after freeing
  them, to prevent attempt to free again if smfs_umount_cache() is called second time.

- fixes for making liblustre at least buildable (including fixes to compilation bug found by Brian recently)