Re: [PATCH v2 5/5] kcsan, seqlock: Fix incorrect assumption in read_seqbegin()

[Marco Elver]

On Tue, 5 Nov 2024 at 10:13, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:

> >  static inline unsigned read_seqretry(const seqlock_t *sl, unsigned start)
> >  {
> > -     /*
> > -      * Assume not nested: read_seqretry() may be called multiple times when
> > -      * completing read critical section.
> > -      */
> > -     kcsan_flat_atomic_end();
> > -
> >       return read_seqcount_retry(&sl->seqcount, start);
> >  }
>
> OK, so this takes us back to kcsan_atomic_next(KCSAN_SEQLOCK_REGION_MAX)
> and kcsan_atomic_next(0).
>
> Which I suppose is safe, except it doesn't nest properly.

Yes correct - we just give up on trying to be special here. It would
be nice to also have readers have a clear critical section, but that
seems a lot harder to enforce with a bunch of them having rather
convoluted control flow. :-/

> Anyway, these all look really nice, let me go queue them up.

Many thanks!