Skip to content

L
linux
Switch branch/tag
  1. 01 Mar, 2010 9 commits
    • Christoph Hellwig's avatar
      xfs: fix xfs_fsblock_t tracing · f7008d0a
      Christoph Hellwig authored 
      Using a static buffer in xfs_fmtfsblock means we can corrupt traces if
multiple CPUs hit this code path at the same.  Just remove xfs_fmtfsblock
for now and print the block number purely numerical.  If we want the
NULLFSBLOCK and NULLSTARTBLOCK formatting back the best way would be
a decoding plugin in the trace-cmd userspace command.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      f7008d0a
    • Christoph Hellwig's avatar
      xfs: fix inode pincount check in fsync · 024910cb
      Christoph Hellwig authored 
      We need to hold the ilock to check the inode pincount safely.  While
we're at it also remove the check for ip->i_itemp->ili_last_lsn, a
pinned inode always has it set.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      024910cb
    • Dave Chinner's avatar
      xfs: Non-blocking inode locking in IO completion · 77d7a0c2
      Dave Chinner authored 
      The introduction of barriers to loop devices has created a new IO
order completion dependency that XFS does not handle. The loop
device implements barriers using fsync and so turns a log IO in the
XFS filesystem on the loop device into a data IO in the backing
filesystem. That is, the completion of log IOs in the loop
filesystem are now dependent on completion of data IO in the backing
filesystem.

This can cause deadlocks when a flush daemon issues a log force with
an inode locked because the IO completion of IO on the inode is
blocked by the inode lock. This in turn prevents further data IO
completion from occuring on all XFS filesystems on that CPU (due to
the shared nature of the completion queues). This then prevents the
log IO from completing because the log is waiting for data IO
completion as well.

The fix for this new completion order dependency issue is to make
the IO completion inode locking non-blocking. If the inode lock
can't be grabbed, simply requeue the IO completion back to the work
queue so that it can be processed later. This prevents the
completion queue from being blocked and allows data IO completion on
other inodes to proceed, hence avoiding completion order dependent
deadlocks.
Signed-off-by: default avatarDave Chinner <david@fromorbit.com>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      77d7a0c2
    • Christoph Hellwig's avatar
      xfs: implement optimized fdatasync · 66d834ea
      Christoph Hellwig authored 
      Allow us to track the difference between timestamp and size updates
by using mark_inode_dirty from the I/O completion code, and checking
the VFS inode flags in xfs_file_fsync.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      66d834ea
    • Christoph Hellwig's avatar
      xfs: remove wrapper for the fsync file operation · fd3200be
      Christoph Hellwig authored 
      Currently the fsync file operation is divided into a low-level
routine doing all the work and one that implements the Linux file
operation and does minimal argument wrapping.  This is a leftover
from the days of the vnode operations layer and can be removed to
simplify the code a bit, as well as preparing for the implementation
of an optimized fdatasync which needs to look at the Linux inode
state.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      fd3200be
    • Christoph Hellwig's avatar
      xfs: remove wrappers for read/write file operations · 00258e36
      Christoph Hellwig authored 
      Currently the aio_read, aio_write, splice_read and splice_write file
operations are divided into a low-level routine doing all the work
and one that implements the Linux file operations and does minimal
argument wrapping.  This is a leftover from the days of the vnode
operations layer and can be removed to simplify the code a lot.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      00258e36
    • Christoph Hellwig's avatar
      xfs: merge xfs_lrw.c into xfs_file.c · dda35b8f
      Christoph Hellwig authored 
      Currently the code to implement the file operations is split over
two small files.  Merge the content of xfs_lrw.c into xfs_file.c to
have it in one place.  Note that I haven't done various cleanups
that are possible after this yet, they will follow in the next
patch.  Also the function xfs_dev_is_read_only which was in
xfs_lrw.c before really doesn't fit in here at all and was moved to
xfs_mount.c.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      dda35b8f
    • Christoph Hellwig's avatar
      xfs: fix dquota trace format · b262e5df
      Christoph Hellwig authored 
      The be32_to_cpu in the TP_printk output breaks automatic parsing of
the trace format by the trace-cmd tools, so we have to move it into
the TP_assign block.  While we're at it also fix the format for the
quota limits to more regular and easier parseable.
Signed-off-by: default avatarChristoph Hellwig <hch@lst.de>
Reviewed-by: default avatarDave Chinner <david@fromorbit.com>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      b262e5df
    • Eric Sandeen's avatar
      xfs: increase readdir buffer size · a9cc799e
      Eric Sandeen authored 
      While doing some testing of readdir perf a while back,
I noticed that the buffer size we're using internally is
smaller than what glibc gives us by default.  Upping this
size helped a bit, and seems safe.

glibc's __alloc_dir() does:

  const size_t default_allocation = (4 * BUFSIZ < sizeof (struct dirent64)
                                     ? sizeof (struct dirent64) : 4 * BUFSIZ);
  const size_t small_allocation = (BUFSIZ < sizeof (struct dirent64)
                                   ? sizeof (struct dirent64) : BUFSIZ);
  size_t allocation = default_allocation;
#ifdef _STATBUF_ST_BLKSIZE
  if (statp != NULL && default_allocation < statp->st_blksize)
    allocation = statp->st_blksize;
#endif

and

#define _G_BUFSIZ 8192
#define _IO_BUFSIZ _G_BUFSIZ
# define BUFSIZ _IO_BUFSIZ

so the default buffer is 4 * 8192 = 32768
(except in the unlikely case of blocks > 32k....)
Signed-off-by: default avatarEric Sandeen <sandeen@sandeen.net>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarAlex Elder <aelder@sgi.com>
      a9cc799e
  3. 26 Feb, 2010 1 commit
  5. 24 Feb, 2010 13 commits
  7. 23 Feb, 2010 17 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7