Re: delalloc filesystem corruption

[Mingming Cao <cmm@us.ibm.com>]

On Tue, 2008-07-01 at 10:54 +0000, Gary Hawco wrote:
> Ted,
> 
> Thanks for your quick reply.  With the newest rc8-based snapshots I am no
> longer getting segfaults, unless you wanted me to roll back to the
> rc6-based snapshots to reproduce the segfaults.
> 
> Per your latest request I disabled delalloc, and voila, no more data
> corruption! Then I enabled delalloc and disabled mballoc and file
> corruption to /lib/rc/init.d/nettree returned. So delalloc is the culprit.
> 
> I hope this will help figure this out.
> 

Could you try the following patch? It fixes the problem of update
on-disk size too early without block allocation, the problem is
introduced unintentionally by another bug fix patch added to the patch
queue yesterday.


Signed-off-by: Mingming Cao <cmm@us.ibm.com>
---
 fs/ext4/inode.c |   32 ++++++++++++++++++++++++++++++--
 1 file changed, 30 insertions(+), 2 deletions(-)

Index: linux-2.6.26-rc8/fs/ext4/inode.c
===================================================================
--- linux-2.6.26-rc8.orig/fs/ext4/inode.c	2008-07-01 15:13:00.000000000 -0700
+++ linux-2.6.26-rc8/fs/ext4/inode.c	2008-07-01 15:34:19.000000000 -0700
@@ -1892,6 +1892,31 @@ out:
 	return ret;
 }
 
+/*
+ * Check if we should update i_disksize
+ * when write to the end of file but not require block allocation
+ */
+static int ext4_da_should_update_i_disksize(struct page *page,
+					 unsigned long offset)
+{
+	struct buffer_head *head, *bh;
+	unsigned int curr_off = 0;
+
+	head = page_buffers(page);
+	bh = head;
+	do {
+		unsigned int next_off = curr_off + bh->b_size;
+
+		if ((curr_off >= offset) &&
+		   (!buffer_mapped || (buffer_delay(bh))) {
+			return 0;
+		}
+		curr_off = next_off;
+	} while ((bh = bh->b_this_page) != head);
+
+	return 1;
+}
+
 static int ext4_da_write_end(struct file *file,
 				struct address_space *mapping,
 				loff_t pos, unsigned len, unsigned copied,
@@ -1901,6 +1926,10 @@ static int ext4_da_write_end(struct file
 	int ret = 0, ret2;
 	handle_t *handle = ext4_journal_current_handle();
 	loff_t new_i_size;
+	unsigned long start, end;
+
+	start = pos & (PAGE_CACHE_SIZE - 1);
+	end = start + copied;
 
 	/*
 	 * generic_write_end() will run mark_inode_dirty() if i_size
@@ -1910,8 +1939,7 @@ static int ext4_da_write_end(struct file
 
 	new_i_size = pos + copied;
 	if (new_i_size > EXT4_I(inode)->i_disksize)
-		if (!walk_page_buffers(NULL, page_buffers(page),
-				       0, len, NULL, ext4_bh_unmapped_or_delay)){
+		if (ext4_da_should_update_i_disksize(page, end))
 			/*
 			 * Updating i_disksize when extending file without
 			 * need block allocation



> Thanks again,
> Gary
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html