Re: [PATCH v8 1/6] ip_frag: use a dynamic logtype

[David Marchand]

On Mon, Aug 14, 2023 at 6:31 PM Stephen Hemminger
<step...@networkplumber.org> wrote:
> @@ -52,20 +54,20 @@ rte_ip_frag_table_create(uint32_t bucket_num, uint32_t 
> bucket_entries,
>         if (rte_is_power_of_2(bucket_entries) == 0 ||
>                         nb_entries > UINT32_MAX || nb_entries == 0 ||
>                         nb_entries < max_entries) {
> -               RTE_LOG(ERR, USER1, "%s: invalid input parameter\n", 
> __func__);
> +               RTE_LOG(ERR, IPFRAG, "%s: invalid input parameter\n", 
> __func__);
>                 return NULL;
>         }
>
>         sz = sizeof (*tbl) + nb_entries * sizeof (tbl->pkt[0]);
>         if ((tbl = rte_zmalloc_socket(__func__, sz, RTE_CACHE_LINE_SIZE,
>                         socket_id)) == NULL) {
> -               RTE_LOG(ERR, USER1,
> +               RTE_LOG(ERR, IPFRAG,
>                         "%s: allocation of %zu bytes at socket %d failed 
> do\n",
>                         __func__, sz, socket_id);
>                 return NULL;
>         }
>
> -       RTE_LOG(INFO, USER1, "%s: allocated of %zu bytes at socket %d\n",
> +       RTE_LOG(INFO, IPFRAG, "%s: allocated of %zu bytes at socket %d\n",
>                 __func__, sz, socket_id);
>
>         tbl->max_cycles = max_cycles;
> --
> 2.39.2
>

Any reason not to use the IP_FRAG_LOG macro?
This is easy for me to fix when applying if you have no objection.


-- 
David Marchand