Re: What to do when the journal checksum is incorrect

[Theodore Tso <tytso@mit.edu>]

On Tue, Jun 03, 2008 at 03:52:13PM +0530, Girish Shilamkar wrote:
> I went through the code and also re-ran the e2fsprogs tests which we had
> sent upstream for journal checksum. And found that if the transaction is
> bad it is marked as info->end_transaction which indicates a bad
> transaction and is not replayed.

You're correct; my apologies.  I had misplaced or misfiled your
patches, and recreated the patch to support async checksums, and I
screwed up the patch by missing this change at the end of
do_one_pass()

	if (pass == PASS_SCAN) {
-		info->end_transaction = next_commit_ID;
+		if (!info->end_transaction)
+			info->end_transaction = next_commit_ID;
	} else {
		/* It's really bad news if different passes end up at
		 * different places (but possible due to IO errors). */

My patch to fix the buffer head leak is still relevant, though, and
I've created a patch to reflect the error up to ext4.

> > Worse yet, no indication of any problems is
> > sent back to the ext4 filesystem code.
> This definitely is not present and needs to be incorporated. 
> 

I will send the patches that I've placed into the ext4 patch queue
relevant to the journal checksum code on this thread.

	    			      	      - Ted