NOTE: subsytem initialize order problem is to be solved yet

  moving to the right side
- ext3_ext_find_goal() takes into account offset of a block to be allocated
  (thanks to Andreas for smart idea)

  cache-aliasing problems
- ext3_forget() should be used in conjuction with ext3_free_blocks()

NOTE: 2.4.24 + extents + batching block allocation passes data-consistency
      tests now (at least I haven't seen problems for many hours of testing)

  lots of tweaking to be done yet

  possible deadlock
- locking in extent's ioctl added
- ext3_ext_walk_space() drops bh's references in path before finding
  next extent

NOTE: dbench and syntetic tests work OK in vmware for many hours

  - fixed goal selection
  - more bits of batching allocation

b=2464

rename snapfs-core-2.4.20.patch to ext3-snapfs-2.4.20.patch, resolve some confilcts after add ext3-extents patch, some minor fix

merge up to RC_1_0_3_5:

work around lameness in mkkerneldoth

- bug in extents fixed: it could cause tree corruption
- minor bug in extents fixed: during tree growing it could possible
  to get suboptimal size for index block

1)In migration, get rid of page and buffer of src inode(fs_flushinval_pages). 2)fix some lock problems in ext3_create_indirect

- support for extents in EA
- access to extents stored in EA via ioctl

finish landing patch. (merge up to RC_1_0_3_4 of b1_0)

1 add some initial snapfs process in ext3_readsupre, 2 get rid of copy_block for reguarl file in snapfs kernel patch

update snapfs kernel patch 1)rewrite ext3_copy_blk according to the generic_file_write 2)get rid of some useless code 3)some minor fix in debug print

  vmalloc_to_page() routine
- zc patch to match 2.4.19-pre1 context

  not for emptyness
- ext3_stop_delete_thread() must not be called from ->put_super() context,
  because ->put_super() is called with sb being locked. so, that
  routine is called from ->sync_fs() context
- chaos-2.4.18 should drop sb lock before calling ->sync_fs()
- virgin 2.4.20 has no ->sync_fs(), so introduce it by delete-thread patch
- kernel_text_address patch has been added to rh-2.4.22 series

  b2326

- additional assert

NOTE: it passes sanity and single-replay

  is very simple: don't drop i_nlink to 0 in ext3_unlink() if file is
  large

NOTE: to be discussed yet

  b2505

NOTE: do not use it for a while, please. ext3-delete-thread must
      be reworked first.