ext4-documentation-update-2.patch

   1 ext4: Documention update for new ordered mode and delayed allocation
   2
   3 From: Mingming Cao <cmm@us.ibm.com>
   4
   5 Adding some documentations for delayed allocation and new ordered mode.
   6
   7 Signed-off-by: Mingming Cao <cmm@us.ibm.com>
   8 Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
   9 diff --git a/Documentation/filesystems/ext4.txt b/Documentation/filesystems/ext4.txt
  10 index 77ed8fe..e2afb6e 100644
  11 --- a/Documentation/filesystems/ext4.txt
  12 +++ b/Documentation/filesystems/ext4.txt
  13 @@ -62,7 +62,7 @@ Mailing list: linux-ext4@vger.kernel.org
  14  * extent format reduces metadata overhead (RAM, IO for access, transactions)
  15  * extent format more robust in face of on-disk corruption due to magics,
  16  * internal redunancy in tree
  17 -* improved file allocation (multi-block alloc, delayed alloc)
  18 +* improved file allocation (multi-block alloc)
  19  * fix 32000 subdirectory limit
  20  * nsec timestamps for mtime, atime, ctime, create time
  21  * inode version field on disk (NFSv4, Lustre)
  22 @@ -73,6 +73,10 @@ Mailing list: linux-ext4@vger.kernel.org
  23    flex_bg feature
  24  * large file support
  25  * Inode allocation using large virtual block groups via flex_bg
  26 +* delayed allocation
  27 +* large block (up to pagesize) support
  28 +* efficent new ordered mode in JBD2 and ext4(avoid using buffer head to force
  29 +  the ordering)
  30
  31  2.2 Candidate features for future inclusion
  32
  33 @@ -235,7 +239,9 @@ stripe=n            Number of filesystem blocks that mballoc will try
  34                         to use for allocation size and alignment. For RAID5/6
  35                         systems this should be the number of data
  36                         disks *  RAID chunk size in file system blocks.
  37 -
  38 +delalloc       (*)     Deferring block allocation until write-out time.
  39 +nodelalloc             Disable delayed allocation. Blocks are allocation
  40 +                       when data is copied from user to page cache.
  41  Data Mode
  42  =========
  43  There are 3 different data modes:
  44 @@ -249,10 +255,10 @@ typically provide the best ext4 performance.
  45
  46  * ordered mode
  47  In data=ordered mode, ext4 only officially journals metadata, but it logically
  48 -groups metadata and data blocks into a single unit called a transaction.  When
  49 -it's time to write the new metadata out to disk, the associated data blocks
  50 -are written first.  In general, this mode performs slightly slower than
  51 -writeback but significantly faster than journal mode.
  52 +groups metadata information related to data changes with the data blocks into a
  53 +single unit called a transaction.  When it's time to write the new metadata
  54 +out to disk, the associated data blocks are written first.  In general,
  55 +this mode performs slightly slower than writeback but significantly faster than journal mode.
  56
  57  * journal mode
  58  data=journal mode provides full data and metadata journaling.  All new data is
  59 @@ -260,7 +266,8 @@ written to the journal first, and then to its final location.
  60  In the event of a crash, the journal can be replayed, bringing both data and
  61  metadata into a consistent state.  This mode is the slowest except when data
  62  needs to be read from and written to disk at the same time where it
  63 -outperforms all others modes.
  64 +outperforms all others modes.  Curently ext4 does not have delayed
  65 +allocation support if this data journalling mode is selected.
  66
  67  References
  68  ==========