Re: BPF memstore and bpf_validate_ext()

[Alexander Nasonov <alnsn%yandex.ru@localhost>]

Mindaugas Rasiukevicius wrote:
> Alexander Nasonov <alnsn%yandex.ru@localhost> wrote:
> > In your case BPF_COP is always the first instruction (or in the
> > first linear block). Because it's the first and it's a function
> > call, it can be moved outside of bpf program and inlined.
> > 
> > If this is the case, you don't need BPF_COP at all.
> 
> For this particular case - yes, correct.
> 
> > Except of course you may run out or memwords one day and you want
> > to have BPF_COP as a fallback.
> 
> I still need BPF_COP for other tasks (e.g. NPF_COP_TABLE).  Running out
> of words is unlikely, but COP can certainly be used to handle that too.

I wonder why do you need two different features when one is a
superset of the other if you could use BPF_COP? If the only reason
is performance, external memory is not a win-win proposition.

PS sljit has a non-standard fast call mechanism. If you're really
concerned about performance, you can generate copfuncs and I can
call them from bpfjit.

Alex