[XFS] fix, speedup and simplify atime handling let the VFS handle atime

updates and only sync back to the xfs inode when nessecary

SGI-PV: 946679
SGI-Modid: xfs-linux-melb:xfs-kern:203362a

Signed-off-by: Christoph Hellwig <hch@sgi.com>
Signed-off-by: Nathan Scott <nathans@sgi.com>

1 files changed, 8 insertions, 11 deletions

diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index 15734c54952e..688fc2cb4b8d 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -182,8 +182,8 @@ xfs_getattr(
                 break;
         }
 
-        vap->va_atime.tv_sec = ip->i_d.di_atime.t_sec;
-        vap->va_atime.tv_nsec = ip->i_d.di_atime.t_nsec;
+        /* atime is only kept uptodate in the Linux inode */
+        vap->va_atime = vp->v_inode.i_atime;
         vap->va_mtime.tv_sec = ip->i_d.di_mtime.t_sec;
         vap->va_mtime.tv_nsec = ip->i_d.di_mtime.t_nsec;
         vap->va_ctime.tv_sec = ip->i_d.di_ctime.t_sec;
@@ -982,10 +982,6 @@ xfs_readlink(
                 goto error_return;
         }
 
-        if (!(ioflags & IO_INVIS)) {
-                xfs_ichgtime(ip, XFS_ICHGTIME_ACC);
-        }
-
         /*
          * See if the symlink is stored inline.
          */
@@ -3226,7 +3222,6 @@ xfs_readdir(
         xfs_trans_t     *tp = NULL;
         int             error = 0;
         uint            lock_mode;
-        xfs_off_t       start_offset;
 
         vn_trace_entry(BHV_TO_VNODE(dir_bdp), __FUNCTION__,
                                                (inst_t *)__return_address);
@@ -3237,11 +3232,7 @@ xfs_readdir(
         }
 
         lock_mode = xfs_ilock_map_shared(dp);
-        start_offset = uiop->uio_offset;
         error = XFS_DIR_GETDENTS(dp->i_mount, tp, dp, uiop, eofp);
-        if (start_offset != uiop->uio_offset) {
-                xfs_ichgtime(dp, XFS_ICHGTIME_ACC);
-        }
         xfs_iunlock_map_shared(dp, lock_mode);
         return error;
 }
@@ -3819,6 +3810,12 @@ xfs_reclaim(
 
         ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0);
 
+        /*
+         * Make sure the atime in the XFS inode is correct before freeing the
+         * Linux inode.
+         */
+        xfs_synchronize_atime(ip);
+
         /* If we have nothing to flush with this inode then complete the
          * teardown now, otherwise break the link between the xfs inode
          * and the linux inode and clean up the xfs inode later. This