From: Morten Brørup [mailto:mb@smartsharesystems.com]
Sent: Wednesday, 17 May 2023 10.38

From: Yasin CANER [mailto:yasinncaner@gmail.com]
Sent: Wednesday, 17 May 2023 10.01

Hello,

 

I don't have full knowledge of how to work rte_mempool_ops_get_count() but there is another comment about it. Maybe it relates. 

/*
 * due to race condition (access to len is not locked), the
 * total can be greater than size... so fix the result
 */

 

MB: This comment relates to the race when accessing the per-lcore cache counters.

 

MB (continued): I have added more information, regarding mempool drivers, in Bugzilla: https://bugs.dpdk.org/show_bug.cgi?id=1229

 

 

Best regards.

 

Morten Brørup <mb@smartsharesystems.com>, 16 May 2023 Sal, 19:04 tarihinde şunu yazdı:

> From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> Sent: Tuesday, 16 May 2023 17.24
>
> On Tue, 16 May 2023 13:41:46 +0000
> Yasin CANER <yasinncaner@gmail.com> wrote:
>
> > From: Yasin CANER <yasin.caner@ulakhaberlesme.com.tr>
> >
> > after a while working rte_mempool_avail_count function returns bigger
> > than mempool size that cause miscalculation rte_mempool_in_use_count.
> >
> > it helps to avoid miscalculation rte_mempool_in_use_count.
> >
> > Bugzilla ID: 1229
> >
> > Signed-off-by: Yasin CANER <yasin.caner@ulakhaberlesme.com.tr>
>
> An alternative that avoids some code duplication.
>
> diff --git a/lib/mempool/rte_mempool.c b/lib/mempool/rte_mempool.c
> index cf5dea2304a7..2406b112e7b0 100644
> --- a/lib/mempool/rte_mempool.c
> +++ b/lib/mempool/rte_mempool.c
> @@ -1010,7 +1010,7 @@ rte_mempool_avail_count(const struct rte_mempool
> *mp)
>         count = rte_mempool_ops_get_count(mp);
>
>         if (mp->cache_size == 0)
> -               return count;
> +               goto exit;

This bug can only occur here (i.e. with cache_size==0) if rte_mempool_ops_get_count() returns an incorrect value. The bug should be fixed there instead.

 

MB (continued): The bug must be in the underlying mempool driver. I took a look at the ring and stack drivers, and they seem fine.



>
>         for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++)
>                 count += mp->local_cache[lcore_id].len;
> @@ -1019,6 +1019,7 @@ rte_mempool_avail_count(const struct rte_mempool
> *mp)
>          * due to race condition (access to len is not locked), the
>          * total can be greater than size... so fix the result
>          */
> +exit:
>         if (count > mp->size)
>                 return mp->size;
>         return count;