On Tuesday May 26, neilb@xxxxxxx wrote:
> >
> > All patches pass the tests of the mdadm test suite except
> > 07reshape5intr which appears fail also for the vanilla for-next tree.
>
> Hmm... I guess I should look into that!
Yep. There is a big there. Thanks for helping find it.
Here is the patch.
Thanks,
NeilBrown
Author: NeilBrown <neilb@xxxxxxx>
Date: Tue May 26 12:39:27 2009 +1000
md: raid5: avoid sector values going negative when testing reshape progress.
As sector_t in unsigned, we cannot afford to let 'safepos' etc go
negative.
So replace
a -= b;
by
a -= min(b,a);
Signed-off-by: NeilBrown <neilb@xxxxxxx>
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 6b89a7e..dce7741 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -3810,13 +3810,13 @@ static sector_t reshape_request(mddev_t *mddev, sector_t sector_nr, int *skipped
safepos = conf->reshape_safe;
sector_div(safepos, data_disks);
if (mddev->delta_disks < 0) {
- writepos -= reshape_sectors;
+ writepos -= min(reshape_sectors, writepos);
readpos += reshape_sectors;
safepos += reshape_sectors;
} else {
writepos += reshape_sectors;
- readpos -= reshape_sectors;
- safepos -= reshape_sectors;
+ readpos -= min(reshape_sectors, readpos);
+ safepos -= min(reshape_sectors, safepos);
}
/* 'writepos' is the most advanced device address we might write.
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
