Jesper Dangaard Brouer <hawk@xxxxxxxxxx> writes:
> On 31/05/2024 12.38, Sebastian Andrzej Siewior wrote:
>> On 2024-05-30 00:09:21 [+0200], Toke Høiland-Jørgensen wrote:
>>> [...]
>>>> @@ -240,12 +240,14 @@ static int cpu_map_bpf_prog_run(struct bpf_cpu_map_entry *rcpu, void **frames,
>>>> int xdp_n, struct xdp_cpumap_stats *stats,
>>>> struct list_head *list)
>>>> {
>>>> + struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx;
>>>> int nframes;
> >>
>>> I think we need to zero-initialise all the context objects we allocate
>>> on the stack.
>>>
>> Okay, I can do that.
>
> Hmm, but how will this affect performance?
My hunch would be that this would be in a cache line we're touching
anyway, so it won't make much difference? But better measure, I suppose :)
-Toke
