1. 24 4月, 2006 1 次提交
  2. 18 4月, 2006 1 次提交
    • S
      [GFS2] Use vmalloc() in dir code · fe1bdedc
      Steven Whitehouse 提交于
      When allocating memory to sort directory entries, use vmalloc()
      rather than kmalloc() since for larger directories, the required
      size can easily be graeter than the 128k maximum of kmalloc().
      
      Also adding the first steps towards getting the AOP_TRUNCATED_PAGE
      return code get in the glock code by flagging all places where we
      request a glock and we are holding a page lock.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      fe1bdedc
  3. 13 4月, 2006 1 次提交
  4. 12 4月, 2006 1 次提交
    • S
      [GFS2] Update journal accounting code. · f4154ea0
      Steven Whitehouse 提交于
      A small update to the journaling code to change the way that
      the "extra" blocks are accounted for in the journal. These are
      used at a rate of one per 503 metadata blocks or one per 251
      journaled data blocks (or just one if the total number of journaled
      blocks in the transaction is smaller). Since we are using them at
      two different rates the old method of accounting for them no longer
      works and we count them up as required.
      
      Since the "per transaction" accounting can't handle this (there is no
      fixed number of header blocks per transaction) we have to account for
      it in the general journal code. We now require that each transaction
      reserves more blocks than it actually needs to take account of the
      possible extra blocks.
      
      Also a final fix to dir.c to ensure that all ref counts are handled
      correctly.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      f4154ea0
  5. 08 4月, 2006 1 次提交
  6. 07 4月, 2006 1 次提交
    • S
      [GFS2] Fix a ref count bug and other clean ups · b09e593d
      Steven Whitehouse 提交于
      This fixes a ref count bug that sometimes showed up a umount time
      (causing it to hang) but it otherwise mostly harmless. At the same
      time there are some clean ups including making the log operations
      structures const, moving a memory allocation so that its not done
      in the fast path of checking to see if there is an outstanding
      transaction related to a particular glock.
      
      Removes the sd_log_wrap varaible which was updated, but never actually
      used anywhere. Updates the gfs2 ioctl() to run without the kernel lock
      (which it never needed anyway). Removes the "invalidate inodes" loop
      from GFS2's put_super routine. This is done in kill super anyway so
      we don't need to do it here. The loop was also bogus in that if there
      are any inodes "stuck" at this point its a bug and we need to know
      about it rather than hide it by hanging forever.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      b09e593d
  7. 31 3月, 2006 1 次提交
  8. 30 3月, 2006 1 次提交
  9. 29 3月, 2006 1 次提交
    • S
      [GFS2] Further updates to dir and logging code · 71b86f56
      Steven Whitehouse 提交于
      This reduces the size of the directory code by about 3k and gets
      readdir() to use the functions which were introduced in the previous
      directory code update.
      
      Two memory allocations are merged into one. Eliminates zeroing of some
      buffers which were never used before they were initialised by
      other data.
      
      There is still scope for further improvement in the directory code.
      
      On the logging side, a hand created mutex has been replaced by a
      standard Linux mutex in the log allocation code.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      71b86f56
  10. 21 3月, 2006 1 次提交
    • S
      [GFS2] Fix bug in directory code and tidy up · c752666c
      Steven Whitehouse 提交于
      Due to a typo, the dir leaf split operation was (for the first
      split in a directory) writing the new hash vaules at the
      wrong offset. This is now fixed.
      
      Also some other tidy ups are included:
      
       - We use GFS2's hash function for dentries (see ops_dentry.c) so that
         we don't have to keep recalculating the hash values.
       - A lot of common code is eliminated between the various directory
         lookup routines.
       - Better error checking on directory lookup (previously different
         routines checked for different errors)
       - The leaf split operation has a couple of redundant operations
         removed from it, so it should be faster.
      
      There is still further scope for further clean ups in the directory
      code, and readdir in particular could do with slimming down a bit.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      c752666c
  11. 28 2月, 2006 2 次提交
  12. 14 2月, 2006 2 次提交
  13. 08 2月, 2006 1 次提交
    • S
      [GFS2] Make journaled data files identical to normal files on disk · 18ec7d5c
      Steven Whitehouse 提交于
      This is a very large patch, with a few still to be resolved issues
      so you might want to check out the previous head of the tree since
      this is known to be unstable. Fixes for the various bugs will be
      forthcoming shortly.
      
      This patch removes the special data format which has been used
      up till now for journaled data files. Directories still retain the
      old format so that they will remain on disk compatible with earlier
      releases. As a result you can now do the following with journaled
      data files:
      
       1) mmap them
       2) export them over NFS
       3) convert to/from normal files whenever you want to (the zero length
          restriction is gone)
      
      In addition the level at which GFS' locking is done has changed for all
      files (since they all now use the page cache) such that the locking is
      done at the page cache level rather than the level of the fs operations.
      This should mean that things like loopback mounts and other things which
      touch the page cache directly should now work.
      
      Current known issues:
      
       1. There is a lock mode inversion problem related to the resource
          group hold function which needs to be resolved.
       2. Any significant amount of I/O causes an oops with an offset of hex 320
          (NULL pointer dereference) which appears to be related to a journaled data
          buffer appearing on a list where it shouldn't be.
       3. Direct I/O writes are disabled for the time being (will reappear later)
       4. There is probably a deadlock between the page lock and GFS' locks under
          certain combinations of mmap and fs operation I/O.
       5. Issue relating to ref counting on internally used inodes causes a hang
          on umount (discovered before this patch, and not fixed by it)
       6. One part of the directory metadata is different from GFS1 and will need
          to be resolved before next release.
      Signed-off-by: NSteven Whitehouse <swhiteho@redhat.com>
      18ec7d5c
  14. 30 1月, 2006 1 次提交
  15. 18 1月, 2006 2 次提交
  16. 17 1月, 2006 1 次提交