[dpdk-dev] [PATCH] mem: fix allocation failure on non-NUMA kernel

[Nicolas Dichtel]

Le 05/08/2020 à 16:53, Nick Connolly a écrit :
[snip]
>>>>> +    if (check_numa()) {
>>>>> +        ret = get_mempolicy(&cur_socket_id, NULL, 0, addr,
>>>>> +                    MPOL_F_NODE | MPOL_F_ADDR);
>>>>> +        if (ret < 0) {
>>>>> +            RTE_LOG(DEBUG, EAL, "%s(): get_mempolicy: %s\n",
>>>>> +                __func__, strerror(errno));
>>>>> +            goto mapped;
>>>>> +        } else if (cur_socket_id != socket_id) {
>>>>> +            RTE_LOG(DEBUG, EAL,
>>>>> +                    "%s(): allocation happened on wrong socket (wanted %d,
>>>>> got %d)\n",
>>>>> +                __func__, socket_id, cur_socket_id);
>>>>> +            goto mapped;
>>>>> +        }
>>>>> +    } else {
>>>>> +        if (rte_socket_count() > 1)
>>>>> +            RTE_LOG(DEBUG, EAL, "%s(): not checking socket for allocation
>>>>> (wanted %d)\n",
>>>>> +                    __func__, socket_id);
>>>> nit: maybe an higher log level like WARNING?
>>> Open to guidance here - my concern was that this is going to be generated for
>>> every call to alloc_seg() and I'm not sure what the frequency will be - I'm
>>> cautious about flooding the log with warnings under 'normal running'.  Are the
>>> implications of running on a multi socket system with NUMA support disabled in
>>> the kernel purely performance related for the DPDK or is there a functional
>>> correctness issue as well?
>> Is it really a 'normal running' to have CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES in
>> dpdk and not CONFIG_NUMA in the kernel?
> 
> I'm not an expert of DPDK, but I think it needs to be treated as 'normal
> running', for the following reasons:
> 
> 1. The existing code in eal_memalloc_alloc_seg_bulk() is designed to
>    work even if check_numa() indicates that NUMA support is not enabled:
> 
>    #ifdef RTE_EAL_NUMA_AWARE_HUGEPAGES
>    if (check_numa()) {
>             oldmask = numa_allocate_nodemask();
>             prepare_numa(&oldpolicy, oldmask, socket);
>             have_numa = true;
>         }
>    #endif
The question was not to return an error, but to display a warning. So the code
will work (after your patch), no problem.

> 2. The DPDK application could be built with
>    CONFIG_RTE_EAL_NUMA_AWARE_HUGE_PAGES and then the binary run on
>    different systems with and without NUMA support.
In a production environment, it seems odd to have a custom kernel and a generic
dpdk app, it's why I propose the log level WARNING (or NOTICE maybe?).
I let other comment about this, I don't have a strong opinion.


Regards,
Nicolas