GFS2: Use ->writepages for ordered writes

Instead of using a list of buffers to write ahead of the journal
flush, this now uses a list of inodes and calls ->writepages
via filemap_fdatawrite() in order to achieve the same thing. For
most use cases this results in a shorter ordered write list,
as well as much larger i/os being issued.

The ordered write list is sorted by inode number before writing
in order to retain the disk block ordering between inodes as
per the previous code.

The previous ordered write code used to conflict in its assumptions
about how to write out the disk blocks with mpage_writepages()
so that with this updated version we can also use mpage_writepages()
for GFS2's ordered write, writepages implementation. So we will
also send larger i/os from writeback too.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

1 files changed, 18 insertions, 23 deletions

diff --git a/fs/gfs2/trans.c b/fs/gfs2/trans.c
index 14dbf6d3cdc0..88162fae27a5 100644
--- a/fs/gfs2/trans.c
+++ b/fs/gfs2/trans.c
@@ -159,7 +159,9 @@ static struct gfs2_bufdata *gfs2_alloc_bufdata(struct gfs2_glock *gl,
 }
 
 /**
- * databuf_lo_add - Add a databuf to the transaction.
+ * gfs2_trans_add_data - Add a databuf to the transaction.
+ * @gl: The inode glock associated with the buffer
+ * @bh: The buffer to add
  *
  * This is used in two distinct cases:
  * i) In ordered write mode
@@ -174,33 +176,18 @@ static struct gfs2_bufdata *gfs2_alloc_bufdata(struct gfs2_glock *gl,
  *    blocks, which isn't an enormous overhead but twice as much as
  *    for normal metadata blocks.
  */
-static void databuf_lo_add(struct gfs2_sbd *sdp, struct gfs2_bufdata *bd)
+void gfs2_trans_add_data(struct gfs2_glock *gl, struct buffer_head *bh)
 {
         struct gfs2_trans *tr = current->journal_info;
-        struct address_space *mapping = bd->bd_bh->b_page->mapping;
+        struct gfs2_sbd *sdp = gl->gl_sbd;
+        struct address_space *mapping = bh->b_page->mapping;
         struct gfs2_inode *ip = GFS2_I(mapping->host);
+        struct gfs2_bufdata *bd;
 
-        if (tr)
-                tr->tr_touched = 1;
-        if (!list_empty(&bd->bd_list))
+        if (!gfs2_is_jdata(ip)) {
+                gfs2_ordered_add_inode(ip);
                 return;
-        set_bit(GLF_LFLUSH, &bd->bd_gl->gl_flags);
-        set_bit(GLF_DIRTY, &bd->bd_gl->gl_flags);
-        if (gfs2_is_jdata(ip)) {
-                gfs2_pin(sdp, bd->bd_bh);
-                tr->tr_num_databuf_new++;
-                sdp->sd_log_num_databuf++;
-                list_add_tail(&bd->bd_list, &sdp->sd_log_le_databuf);
-        } else {
-                list_add_tail(&bd->bd_list, &sdp->sd_log_le_ordered);
         }
-}
-
-void gfs2_trans_add_data(struct gfs2_glock *gl, struct buffer_head *bh)
-{
-
-        struct gfs2_sbd *sdp = gl->gl_sbd;
-        struct gfs2_bufdata *bd;
 
         lock_buffer(bh);
         gfs2_log_lock(sdp);
@@ -214,7 +201,15 @@ void gfs2_trans_add_data(struct gfs2_glock *gl, struct buffer_head *bh)
                 gfs2_log_lock(sdp);
         }
         gfs2_assert(sdp, bd->bd_gl == gl);
-        databuf_lo_add(sdp, bd);
+        tr->tr_touched = 1;
+        if (list_empty(&bd->bd_list)) {
+                set_bit(GLF_LFLUSH, &bd->bd_gl->gl_flags);
+                set_bit(GLF_DIRTY, &bd->bd_gl->gl_flags);
+                gfs2_pin(sdp, bd->bd_bh);
+                tr->tr_num_databuf_new++;
+                sdp->sd_log_num_databuf++;
+                list_add_tail(&bd->bd_list, &sdp->sd_log_le_databuf);
+        }
         gfs2_log_unlock(sdp);
         unlock_buffer(bh);
 }