GFS2: fix jdata issues
PATCH 1/1 GFS2: fix jdata issues
This is a patch for the first three issues of RHBZ #238162
The first issue is that when you allocate a new page for a file, it will not start off uptodate. This makes sense, since you haven't written anything to that part of the file yet. Unfortunately, gfs2_pin() checks to make sure that the buffers are uptodate. The solution to this is to mark the buffers uptodate in gfs2_commit_write(), after they have been zeroed out and have the data written into them. I'm pretty confident with this fix, although it's not completely obvious that there is no problem with marking the buffers uptodate here.
The second issue is simply that you can try to pin a data buffer that is already on the incore log, and thus, already pinned. This patch checks to see if this buffer is already on the log, and exits databuf_lo_add() if it is, just like buf_lo_add() does.
The third issue is that gfs2_log_flush() doesn't do it's block accounting correctly. Both metadata and journaled data are logged, but gfs2_log_flush() only compares the number of metadata blocks with the number of blocks to commit to the ondisk journal. This patch also counts the journaled data blocks.
-------------- next part --------------
fs/gfs2/log.c +1 -1
@@ -565,7 +565,7 @@ void gfs2_log_flush(struct gfs2_sbd *sdpINIT_LIST_HEAD (&ai->ai_ail1_list);INIT_LIST_HEAD (&ai->ai_ail2_list); -gfs2_assert_withdraw (sdp, sdp->sd_log_num_buf == sdp->sd_log_commited_buf); +gfs2_assert_withdraw (sdp, sdp->sd_log_num_buf + sdp->sd_log_num_jdata == sdp->sd_log_commited_buf);gfs2_assert_withdraw (sdp, sdp->sd_log_num_revoke == sdp->sd_log_commited_revoke);
fs/gfs2/lops.c +2 -0
@@ -475,6 +475,8 @@ static void databuf_lo_add(struct gfs2_s tr->tr_num_buf++;list_add (&bd->bd_list_tr, &tr->tr_list_buf);gfs2_log_unlock (sdp); +if (!list_empty (&le->le_list)) +return ;gfs2_pin (sdp, bd->bd_bh); tr->tr_num_buf_new++; }else {
fs/gfs2/ops_address.c +2 -0
@@ -50,6 +50,8 @@ static void gfs2_page_add_databufs(struc end = start + bsize;if (end <= from || start >= to)continue ; +if (gfs2_is_jdata (ip)) +set_buffer_uptodate (bh);gfs2_trans_add_bh (ip->i_gl, bh,0 ); } }
PATCH 1/1 GFS2: fix jdata issues
Hi,
Now applied to the -nmw git tree. Thanks,
Steve.
Quoted 59 lines · Benjamin Marzinski
This is a patch for the first three issues of RHBZ #238162
The first issue is that when you allocate a new page for a file, it will not start off uptodate. This makes sense, since you haven't written anything to that part of the file yet. Unfortunately,
gfs2_pin()checks to make sure that the buffers are uptodate. The solution to this is to mark the buffers uptodate ingfs2_commit_write(), after they have been zeroed out and have the data written into them. I'm pretty confident with this fix, although it's not completely obvious that there is no problem with marking the buffers uptodate here.The second issue is simply that you can try to pin a data buffer that is already on the incore log, and thus, already pinned. This patch checks to see if this buffer is already on the log, and exits
databuf_lo_add()if it is, just likebuf_lo_add()does.The third issue is that
gfs2_log_flush()doesn't do it's block accounting correctly. Both metadata and journaled data are logged, butgfs2_log_flush()only compares the number of metadata blocks with the number of blocks to commit to the ondisk journal. This patch also counts the journaled data blocks.Signed-off-by Benjamin Marzinski <[email protected]>plain text document attachment (238162_mostly_fixed.patch)
diff -urpN --exclude='cscope.*' --exclude-from=gfs2-2.6-nmw-070530-clean/Documentation/dontdiffgfs2-2.6-nmw-070530-clean/fs/gfs2/log.cgfs2-2.6-nmw-070530/fs/gfs2/log.cfs/gfs2/log.c +1
-1@@ -565,7 +565,7 @@ void gfs2_log_flush(struct gfs2_sbd *sdpINIT_LIST_HEAD (&ai->ai_ail1_list);INIT_LIST_HEAD (&ai->ai_ail2_list); -gfs2_assert_withdraw (sdp, sdp->sd_log_num_buf == sdp->sd_log_commited_buf); +gfs2_assert_withdraw (sdp, sdp->sd_log_num_buf + sdp->sd_log_num_jdata == sdp->sd_log_commited_buf);gfs2_assert_withdraw (sdp, sdp->sd_log_num_revoke == sdp->sd_log_commited_revoke);diff -urpN --exclude='cscope.*' --exclude-from=gfs2-2.6-nmw-070530-clean/Documentation/dontdiff
gfs2-2.6-nmw-070530-clean/fs/gfs2/lops.cgfs2-2.6-nmw-070530/fs/gfs2/lops.cfs/gfs2/lops.c +2
-0@@ -475,6 +475,8 @@ static void databuf_lo_add(struct gfs2_s tr->tr_num_buf++;list_add (&bd->bd_list_tr, &tr->tr_list_buf);gfs2_log_unlock (sdp); +if (!list_empty (&le->le_list)) +return ;gfs2_pin (sdp, bd->bd_bh); tr->tr_num_buf_new++; }else {diff -urpN --exclude='cscope.*' --exclude-from=gfs2-2.6-nmw-070530-clean/Documentation/dontdiff
gfs2-2.6-nmw-070530-clean/fs/gfs2/ops_address.cgfs2-2.6-nmw-070530/fs/gfs2/ops_address.cfs/gfs2/ops_address.c +2
-0@@ -50,6 +50,8 @@ static void gfs2_page_add_databufs(struc end = start + bsize;if (end <= from || start >= to)continue ; +if (gfs2_is_jdata (ip)) +set_buffer_uptodate (bh);gfs2_trans_add_bh (ip->i_gl, bh,0 ); } }