Re: [PATCH md ] Fix BUG in raid5 resync code.
From: Neil Brown <hidden>
Date: 2004-06-04 12:04:53
On Friday June 4, david@dgreaves.com wrote:
Hi Neil Given I'm about to resync about 400Gb of data I'd like to check this before blindly storming ahead... :) Would the following be the correct patch against 2.6.6?
Yes. Remove the "+failed". NeilBrown
quoted hunk
--- drivers/md/raid5.c~ 2004-05-30 18:38:49.000000000 +0100 +++ drivers/md/raid5.c 2004-06-04 09:06:12.000000000 +0100@@ -1037,7 +1037,7 @@ * parity, or to satisfy requests * or to load a block that is being partially written. */ - if (to_read || non_overwrite || (syncing && (uptodate+failed <disks))) { + if (to_read || non_overwrite || (syncing && (uptodate < disks))) { for (i=disks; i--;) { dev = &sh->dev[i]; if (!test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) && c