[xdp-hints] Re: [PATCH bpf-next V1 3/7] selftests/bpf: xdp_hw_metadata track more timestamps

[Jesper Dangaard Brouer <jbrouer@redhat.com>]

On 17/03/2023 22.09, Stanislav Fomichev wrote:
>> diff --git a/tools/testing/selftests/bpf/progs/xdp_hw_metadata.c 
>> b/tools/testing/selftests/bpf/progs/xdp_hw_metadata.c
>> index 4c55b4d79d3d..f2a3b70a9882 100644
>> --- a/tools/testing/selftests/bpf/progs/xdp_hw_metadata.c
>> +++ b/tools/testing/selftests/bpf/progs/xdp_hw_metadata.c
>> @@ -69,9 +69,11 @@ int rx(struct xdp_md *ctx)
>>           return XDP_PASS;
>>       }
> 
>> -    if (!bpf_xdp_metadata_rx_timestamp(ctx, &meta->rx_timestamp))
>> -        bpf_printk("populated rx_timestamp with %llu", meta->rx_timestamp);
>> -    else
>> +    if (!bpf_xdp_metadata_rx_timestamp(ctx, &meta->rx_timestamp)) {
>> +        meta->xdp_timestamp = bpf_ktime_get_tai_ns();
>> +        bpf_printk("populated rx_timestamp with  %llu", meta->rx_timestamp);
>> +        bpf_printk("populated xdp_timestamp with %llu", meta->xdp_timestamp);
>> +    } else
>>           meta->rx_timestamp = 0; /* Used by AF_XDP as not avail signal */
> 
> Nit: curly braces around else {} block as well?

Good point, fixed in V2