Re: [PATCH 2/2] block: fix accounting bug on cross partition merges

[Jens Axboe <jaxboe@fusionio.com>]

On 2011-01-05 15:09, Jerome Marchand wrote:
> On 01/05/2011 03:00 PM, Jens Axboe wrote:
>> On 2011-01-04 16:55, Jerome Marchand wrote:
>>> +	} else {
>>> +		part = disk_map_sector_rcu(rq->rq_disk, blk_rq_pos(rq));
>>> +		if (!kref_test_and_get(&part->ref)) {
>>> +			/*
>>> +			 * The partition is already being removed,
>>> +			 * the request will be accounted on the disk only
>>> +			 *
>>> +			 * We take a reference on disk->part0 although that
>>> +			 * partition will never be deleted, so we can treat
>>> +			 * it as any other partition.
>>> +			 */
>>> +			part = &rq->rq_disk->part0;
>>> +			kref_get(&part->ref);
>>> +		}
>>
>> This still doesn't work. So you are inside the if {} block, you know
>> what someone has dropped the last reference and the call_rcu() is being
>> scheduled. Adding a reference now will not prevent 'part' from going
>> away as soon as you do part_stat_unlock().
> 
> And what is the problem with that since we don't use 'part' (as returned
> by disk_map_sector_rcu()), but disk->part0 instead?

Ugh, I'm the one who's blind now. part0 is indeed fine, I didn't read
that carefully enough.

So I think your patch looks safe now, I don't see any holes in it.
Whether we move the kref to inside the lookup or not, that doesn't
change anything wrt using an atomic_inc_not_zero(). Can you resend 1/2
with the manual barriers removed?

-- 
Jens Axboe