mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Steven Whitehouse <swhiteho@redhat.com>
To: linux-kernel@vger.kernel.org, cluster-devel@redhat.com
Cc: Bob Peterson <rpeterso@redhat.com>,
	Steven Whitehouse <swhiteho@redhat.com>
Subject: [PATCH 03/32] GFS2: Combine transaction from gfs2_dir_exhash_dealloc
Date: Thu, 19 May 2011 09:46:58 +0100	[thread overview]
Message-ID: <1305794847-3291-4-git-send-email-swhiteho@redhat.com> (raw)
In-Reply-To: <1305794847-3291-1-git-send-email-swhiteho@redhat.com>

From: Bob Peterson <rpeterso@redhat.com>

At the end of function gfs2_dir_exhash_dealloc, it was setting the dinode
type to "file" to prevent directory corruption in case of a crash.
It was doing so in its own journal transaction.  This patch makes the
change occur when the last call is make to leaf_dealloc, since it needs
to rewrite the directory dinode at that time anyway.

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

diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c
index 0bb5f6b..bd57587 100644
--- a/fs/gfs2/dir.c
+++ b/fs/gfs2/dir.c
@@ -86,7 +86,7 @@ typedef int (*gfs2_dscan_t)(const struct gfs2_dirent *dent,
 			    const struct qstr *name, void *opaque);
 
 static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len,
-			u64 leaf_no);
+			u64 leaf_no, int last_dealloc);
 
 int gfs2_dir_get_new_buffer(struct gfs2_inode *ip, u64 block,
 			    struct buffer_head **bhp)
@@ -1781,10 +1781,10 @@ static int foreach_leaf(struct gfs2_inode *dip)
 	struct gfs2_leaf *leaf;
 	u32 hsize, len;
 	u32 ht_offset, lp_offset, ht_offset_cur = -1;
-	u32 index = 0;
+	u32 index = 0, next_index;
 	__be64 *lp;
 	u64 leaf_no;
-	int error = 0;
+	int error = 0, last;
 
 	hsize = 1 << dip->i_depth;
 	if (hsize * sizeof(u64) != i_size_read(&dip->i_inode)) {
@@ -1819,13 +1819,13 @@ static int foreach_leaf(struct gfs2_inode *dip)
 				goto out;
 			leaf = (struct gfs2_leaf *)bh->b_data;
 			len = 1 << (dip->i_depth - be16_to_cpu(leaf->lf_depth));
+			next_index = (index & ~(len - 1)) + len;
+			last = ((next_index >= hsize) ? 1 : 0);
 			brelse(bh);
-
-			error = leaf_dealloc(dip, index, len, leaf_no);
+			error = leaf_dealloc(dip, index, len, leaf_no, last);
 			if (error)
 				goto out;
-
-			index = (index & ~(len - 1)) + len;
+			index = next_index;
 		} else
 			index++;
 	}
@@ -1847,13 +1847,13 @@ out:
  * @index: the hash table offset in the directory
  * @len: the number of pointers to this leaf
  * @leaf_no: the leaf number
- * @data: not used
+ * last_dealloc: 1 if this is the final dealloc for the leaf, else 0
  *
  * Returns: errno
  */
 
 static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len,
-			u64 leaf_no)
+			u64 leaf_no, int last_dealloc)
 {
 	struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode);
 	struct gfs2_leaf *tmp_leaf;
@@ -1940,6 +1940,10 @@ static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len,
 		goto out_end_trans;
 
 	gfs2_trans_add_bh(dip->i_gl, dibh, 1);
+	/* On the last dealloc, make this a regular file in case we crash.
+	   (We don't want to free these blocks a second time.)  */
+	if (last_dealloc)
+		dip->i_inode.i_mode = S_IFREG;
 	gfs2_dinode_out(dip, dibh->b_data);
 	brelse(dibh);
 
@@ -1971,33 +1975,8 @@ out:
 
 int gfs2_dir_exhash_dealloc(struct gfs2_inode *dip)
 {
-	struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode);
-	struct buffer_head *bh;
-	int error;
-
 	/* Dealloc on-disk leaves to FREEMETA state */
-	error = foreach_leaf(dip);
-	if (error)
-		return error;
-
-	/* Make this a regular file in case we crash.
-	   (We don't want to free these blocks a second time.)  */
-
-	error = gfs2_trans_begin(sdp, RES_DINODE, 0);
-	if (error)
-		return error;
-
-	error = gfs2_meta_inode_buffer(dip, &bh);
-	if (!error) {
-		gfs2_trans_add_bh(dip->i_gl, bh, 1);
-		((struct gfs2_dinode *)bh->b_data)->di_mode =
-						cpu_to_be32(S_IFREG);
-		brelse(bh);
-	}
-
-	gfs2_trans_end(sdp);
-
-	return error;
+	return foreach_leaf(dip);
 }
 
 /**
-- 
1.7.4


  parent reply	other threads:[~2011-05-19  9:17 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-05-19  8:46 GFS2: Pre-pull patch posting (merge window) Steven Whitehouse
2011-05-19  8:46 ` [PATCH 01/32] GFS2: Dump better debug info if a bitmap inconsistency is detected Steven Whitehouse
2011-05-19  8:46 ` [PATCH 02/32] GFS2: remove *leaf_call_t and simplify leaf_dealloc Steven Whitehouse
2011-05-19  8:46 ` Steven Whitehouse [this message]
2011-05-19  8:46 ` [PATCH 04/32] GFS2: pass leaf_bh into leaf_dealloc Steven Whitehouse
2011-05-19  8:47 ` [PATCH 05/32] GFS2: move function foreach_leaf to gfs2_dir_exhash_dealloc Steven Whitehouse
2011-05-19  8:47 ` [PATCH 06/32] GFS2: Make ->write_inode() really write Steven Whitehouse
2011-05-19  8:47 ` [PATCH 07/32] GFS2: Use filemap_fdatawrite() to write back the AIL Steven Whitehouse
2011-05-19  8:47 ` [PATCH 08/32] GFS2: Alter point of entry to glock lru list for glocks with an address_space Steven Whitehouse
2011-05-19  8:47 ` [PATCH 09/32] GFS2: Remove unused macro Steven Whitehouse
2011-05-19  8:47 ` [PATCH 10/32] GFS2: Clean up fsync() Steven Whitehouse
2011-05-19  8:47 ` [PATCH 11/32] GFS2: Improve tracing support (adds two flags) Steven Whitehouse
2011-05-19  8:47 ` [PATCH 12/32] GFS2: Optimise glock lru and end of life inodes Steven Whitehouse
2011-05-19  8:47 ` [PATCH 13/32] GFS2: Make writeback more responsive to system conditions Steven Whitehouse
2011-05-19  8:47 ` [PATCH 14/32] GFS2: Add an AIL writeback tracepoint Steven Whitehouse
2011-05-19  8:47 ` [PATCH 15/32] GFS2: make sure fallocate bytes is a multiple of blksize Steven Whitehouse
2011-05-19  8:47 ` [PATCH 16/32] GFS2: Fix ail list traversal Steven Whitehouse
2011-05-19  8:47 ` [PATCH 17/32] GFS2: Improve bug trap code in ->releasepage() Steven Whitehouse
2011-05-19  8:47 ` [PATCH 18/32] GFS2: Double check link count under glock Steven Whitehouse
2011-05-19  8:47 ` [PATCH 19/32] GFS2: Don't use a try lock when promoting to a higher mode Steven Whitehouse
2011-05-19  8:47 ` [PATCH 20/32] GFS2: Don't use gfs2_change_nlink in link syscall Steven Whitehouse
2011-05-19  8:47 ` [PATCH 21/32] GFS2: Make gfs2_dir_del update link count when required Steven Whitehouse
2011-05-19  8:47 ` [PATCH 22/32] GFS2: When adding a new dir entry, inc link count if it is a subdir Steven Whitehouse
2011-05-19  8:47 ` [PATCH 23/32] GFS2: Remove gfs2_dinode_print() function Steven Whitehouse
2011-05-19  8:47 ` [PATCH 24/32] GFS2: Move gfs2_refresh_inode() and friends into glops.c Steven Whitehouse
2011-05-19  8:47 ` [PATCH 25/32] GFS2: Move most of the remaining inode.c into ops_inode.c Steven Whitehouse
2011-05-19  8:47 ` [PATCH 26/32] GFS2: Move final part of inode.c into super.c Steven Whitehouse
2011-05-19  8:47 ` [PATCH 27/32] GFS2: Inode.c is empty now, remove it Steven Whitehouse
2011-05-19  8:47 ` [PATCH 28/32] GFS2: Rename ops_inode.c to inode.c Steven Whitehouse
2011-05-19  8:47 ` [PATCH 29/32] GFS2: Use UUID field in generic superblock Steven Whitehouse
2011-05-19  8:47 ` [PATCH 30/32] GFS2: Clean up mkdir Steven Whitehouse
2011-05-19  8:47 ` [PATCH 31/32] GFS2: Clean up symlink creation Steven Whitehouse
2011-05-19  8:47 ` [PATCH 32/32] GFS2: Move all locking inside the inode creation function Steven Whitehouse

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1305794847-3291-4-git-send-email-swhiteho@redhat.com \
    --to=swhiteho@redhat.com \
    --cc=cluster-devel@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rpeterso@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®