Re: [PATCH net-next v2 2/4] bpf: test_run: Use system page pool for XDP live frame mode

[Jesper Dangaard Brouer <hawk@kernel.org>]

On 21/02/2024 15.48, Toke Høiland-Jørgensen wrote:
> Toke Høiland-Jørgensen <toke@redhat.com> writes:
> 
>> The cookie is a random 128-bit value, which means the probability that
>> we will get accidental collisions (which would lead to recycling the
>> wrong page values and reading garbage) is on the order of 2^-128. This
>> is in the "won't happen before the heat death of the universe" range, so
>> this marking is safe for the intended usage.
> 
> Alright, got a second opinion on this from someone better at security
> than me; I'll go try out some different ideas :)

It is a general security concern for me that BPF test_run gets access to
memory used by 'system page pool', with the concern of leaking data
(from real traffic) to an attacker than can inject a BPF test_run
program via e.g. a CI pipeline.

I'm not saying we leaking data today in BPF/XDP progs, but there is a
potential, because to gain performance in XDP and page_pool we don't
clear memory to avoid cache line performance issues.
I guess today, I could BPF tail extend and read packet data from older
frames, in this way, if I get access to 'system page pool'.

--Jesper