On mån, 2014-01-20 at 14:37 +1100, NeilBrown wrote:
>
> Thanks - that extra info is quite useful. Knowing that nothing else unusual
> is happening can be quite valuable (and I don't like to assume).
>
> I haven't found anything that would clearly cause your crash, but I have
> found something that looks wrong and conceivably could.
>
> Could you please try this patch on top of what you are currently using? By
> the look of it you get a crash at least every day, often more often. So if
> this produces a day with no crashes, that would be promising.
>
> The important aspect of the patch is that it moves the "atomic_inc" of
> "sh->count" back under the protection of ->device_lock in the case when some
> other thread might be using the same 'sh'.
I have been unable to trip this up, so this was it!
Tested-by: Ian Kumlien <ian.kuml...@gmail.com>
I hope this hits stable ASAP ;)
> Thanks,
> NeilBrown
>
>
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index 3088d3af5a89..03f82ab87d9e 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -675,8 +675,10 @@ get_active_stripe(struct r5conf *conf, sector_t sector,
> || !conf->inactive_blocked),
> *(conf->hash_locks + hash));
> conf->inactive_blocked = 0;
> - } else
> + } else {
> init_stripe(sh, sector, previous);
> + atomic_inc(&sh->count);
> + }
> } else {
> spin_lock(&conf->device_lock);
> if (atomic_read(&sh->count)) {
> @@ -695,13 +697,11 @@ get_active_stripe(struct r5conf *conf, sector_t sector,
> sh->group = NULL;
> }
> }
> + atomic_inc(&sh->count);
> spin_unlock(&conf->device_lock);
> }
> } while (sh == NULL);
>
> - if (sh)
> - atomic_inc(&sh->count);
> -
> spin_unlock_irq(conf->hash_locks + hash);
> return sh;
> }
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
