Re: [dpdk-stable] [dpdk-dev] [PATCH v2 1/2] net/avp: fix gcc 10 maybe-uninitialized warning

[Ferruh Yigit <ferruh.yigit@intel.com>]

On 3/12/2020 3:15 PM, Kevin Traynor wrote:
> On 12/03/2020 14:31, Ferruh Yigit wrote:
>> On 3/12/2020 2:18 PM, Kevin Traynor wrote:
>>> On 12/03/2020 13:25, Ferruh Yigit wrote:
>>>> On 3/11/2020 11:32 AM, Kevin Traynor wrote:
>>>>> gcc 10.0.1 reports:
>>>>>
>>>>> ../drivers/net/avp/avp_ethdev.c: In function ‘avp_xmit_scattered_pkts’:
>>>>> ../drivers/net/avp/avp_ethdev.c:1791:24:
>>>>> warning: ‘avp_bufs[count]’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>>>>>  1791 |   tx_bufs[i] = avp_bufs[count];
>>>>>       |                ~~~~~~~~^~~~~~~
>>>>> ../drivers/net/avp/avp_ethdev.c:1791:24:
>>>>> warning: ‘avp_bufs[count]’ may be used uninitialized in this function [-Wmaybe-uninitialized]
>>>>>
>>>>> Fix by intializing the array.
>>>>>
>>>>> Fixes: 295abce2d25b ("net/avp: add packet transmit functions")
>>>>> Cc: stable@dpdk.org
>>>>>
>>>>> Signed-off-by: Kevin Traynor <ktraynor@redhat.com>
>>>>> ---
>>>>> v2: no change
>>>>>
>>>>> note, commit log violates line length but I didn't want to split warning msg.
>>>>>
>>>>> Cc: allain.legacy@windriver.com
>>>>> Cc: Steven Webster <steven.webster@windriver.com>
>>>>> Cc: Matt Peters <matt.peters@windriver.com>
>>>>> ---
>>>>>  drivers/net/avp/avp_ethdev.c | 2 +-
>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/drivers/net/avp/avp_ethdev.c b/drivers/net/avp/avp_ethdev.c
>>>>> index cd747b6be..1abe96ce5 100644
>>>>> --- a/drivers/net/avp/avp_ethdev.c
>>>>> +++ b/drivers/net/avp/avp_ethdev.c
>>>>> @@ -1695,5 +1695,5 @@ avp_xmit_scattered_pkts(void *tx_queue,
>>>>>  {
>>>>>  	struct rte_avp_desc *avp_bufs[(AVP_MAX_TX_BURST *
>>>>> -				       RTE_AVP_MAX_MBUF_SEGMENTS)];
>>>>> +				       RTE_AVP_MAX_MBUF_SEGMENTS)] = {};
>>>>>  	struct avp_queue *txq = (struct avp_queue *)tx_queue;
>>>>>  	struct rte_avp_desc *tx_bufs[AVP_MAX_TX_BURST];
>>>>>
>>>>
>>>> Isn't this a false positive, can there be any case 'avp_bufs[]' used
>>>> uninitialized? Or am I missing something.
>>>>
>>>
>>> I presume it's because it's not being initialized in the fn and there is
>>> some paths in fn's it's passed to that don't initialize it. Of course in
>>> practice with "normal" values this might not happen.
>>
>> 'avp_fifo_get(alloc_q, (void **)&avp_bufs, segments);' initializes it, and I am
>> not just talking about 'normal' case, I don't see any case that 'avp_bufs[]'
>> used uninitialized, can you see any?
>>
> 
> Well, it's initialization there is dependent on not hitting the first
> return and the loop executing.

If whole array not initialized, the next line, 'if (unlikely(n != segments))',
will catch it and function return without using 'avp_bufs[]' at all.

Anyway, as I said I can't see a case that 'avp_bufs[]' used uninitialized,
and not sure about additional zeroing out in datapath function if this is a
false positive, but if windriver guys are OK I won't object.

> 
>>>
>>>> If this is false positive, does it worth to report to issue to gcc?
>>>>
>>>
>>
>