On Thu, Jan 20, 2011 at 3:25 PM, NeilBrown <neilb@xxxxxxx> wrote:
> Nonetheless, I think the correct fix is to add a special case for retrying a
> zero-length.
>
> Something like this maybe?
Close, but no cigar...
Attached is a modified patch, which does the extra necessary work
(bitmap_endwrite, md_write_end) on the bio before failing it.
Does this look correct? It seems to work.
Thanks again Neil.
--
Paul
--- linux-2.6.32-71.el6.x86_64/drivers/md/raid1.c.PRISTINE 2011-01-14 14:51:56.959809669 -0500
+++ linux-2.6.32-71.el6.x86_64/drivers/md/raid1.c 2011-01-26 08:44:20.176936238 -0500
@@ -236,14 +236,18 @@ static void raid_end_bio_io(r1bio_t *r1_
/* if nobody has done the final endio yet, do it now */
if (!test_and_set_bit(R1BIO_Returned, &r1_bio->state)) {
+ int err = 0;
PRINTK(KERN_DEBUG "raid1: sync end %s on sectors %llu-%llu\n",
(bio_data_dir(bio) == WRITE) ? "write" : "read",
(unsigned long long) bio->bi_sector,
(unsigned long long) bio->bi_sector +
(bio->bi_size >> 9) - 1);
- bio_endio(bio,
- test_bit(R1BIO_Uptodate, &r1_bio->state) ? 0 : -EIO);
+ if (test_bit(R1BIO_BarrierRetry, &r1_bio->state))
+ err = -EOPNOTSUPP;
+ else if (!test_bit(R1BIO_Uptodate, &r1_bio->state))
+ err = -EIO;
+ bio_endio(bio, err);
}
free_r1bio(r1_bio);
}
@@ -1623,6 +1627,18 @@ static void raid1d(mddev_t *mddev)
*/
int i;
const bool do_sync = bio_rw_flagged(r1_bio->master_bio, BIO_RW_SYNCIO);
+ if (r1_bio->master_bio->bi_size == 0) {
+ /* cannot retry empty barriers, just fail */
+ bitmap_endwrite(r1_bio->mddev->bitmap,
+ r1_bio->sector,
+ r1_bio->sectors,
+ !test_bit(R1BIO_Degraded, &r1_bio->state),
+ test_bit(R1BIO_BehindIO, &r1_bio->state));
+ md_write_end(r1_bio->mddev);
+
+ raid_end_bio_io(r1_bio);
+ continue;
+ }
clear_bit(R1BIO_BarrierRetry, &r1_bio->state);
clear_bit(R1BIO_Barrier, &r1_bio->state);
for (i=0; i < conf->raid_disks; i++)
@@ -2139,7 +2155,7 @@ static int stop(mddev_t *mddev)
/* wait for behind writes to complete */
while (bitmap && atomic_read(&bitmap->behind_writes) > 0) {
behind_wait++;
- printk(KERN_INFO "raid1: behind writes in progress on device %s, waiting to stop (%d)\n", mdname(mddev), behind_wait);
+ printk(KERN_INFO "raid1: behind writes (%d) in progress on device %s, waiting to stop (%d)\n", atomic_read(&bitmap->behind_writes), mdname(mddev), behind_wait);
set_current_state(TASK_UNINTERRUPTIBLE);
schedule_timeout(HZ); /* wait a second */
/* need to kick something here to make sure I/O goes? */
