Re: [PATCH gwproxy v8 2/2] gwproxy: refactor code base to add experimental raw DNS backend

[Ahmad Gani <reyuki@gnuweeb.org>]

On Sat, Sep 6, 2025 at 11:32 AM Ahmad Gani wrote:
> On Fri, Sep 5, 2025 at 11:26 PM Alviro Iskandar Setiawan wrote:
> >> + txid = (uint16_t)rand();
> >
> > Why do you use a random number for the txid? I don't think rand()
> > guarantees uniqueness. Also, I don't find the srand() call anywhere.
> > That's equivalent to having srand(1).
> >
> > What about using a simple incrementing counter instead (which
> > guarantees the uniqueness)?
> > It will indeed reset after 65535, but that should give enough room to
> > complete prior DNS requests first. With a simple counter, the reuse of
> > an ID will always be 65535 queries away.
>
> Without calling srand, it still generates random numbers, so I thought it
> was fine. But I agree with your suggestion—an incrementing counter is
> better, because randomness doesn’t guarantee uniqueness and is more
> likely to cause collisions than a counter.

(( I'll finish this first before working on the single UDP socket model. ))

If it's incremented until it is recycled by itself, does it still need to
be atomic? I'm considering adding _Atomic because the data probably will
be accessed and incremented at the same time from multiple workers.

--
Ahmad Gani