From: Maciej Fijalkowski <maciej.fijalkowski@xxxxxxxxx>
Date: Fri, 7 Mar 2025 11:17:11 +0100
> On Wed, Mar 05, 2025 at 05:21:21PM +0100, Alexander Lobakin wrote:
>> No idea what the current barrier position was meant for. At that point,
>> nothing is read from the descriptor, only the pointer to the actual one
>> is fetched.
>> The correct barrier usage here is after the generation check, so that
>> only the first qword is read if the descriptor is not yet ready and we
>> need to stop polling. Debatable on coherent DMA as the Rx descriptor
>> size is <= cacheline size, but anyway, the current barrier position
>> only makes the codegen worse.
>
> Makes sense:
> Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@xxxxxxxxx>
>
> But you know the process... :P fixes should go to -net.
The thing's that it makes no difference for regular skb Rx, but with
ret != XDP_PASS it starts making issues. So yes, this is a fix, but
I don't think it should go separately.
>
>>
>> Fixes: 3a8845af66ed ("idpf: add RX splitq napi poll support")
>> Signed-off-by: Alexander Lobakin <aleksander.lobakin@xxxxxxxxx>
>> ---
>> drivers/net/ethernet/intel/idpf/idpf_txrx.c | 8 ++------
>> 1 file changed, 2 insertions(+), 6 deletions(-)
Thanks,
Olek
