On Wed, Jul 17, 2024 at 05:45:23PM +0900, Dominique Martinet wrote:
> Dominique Martinet wrote on Wed, Jul 17, 2024 at 03:59:46PM +0900:
> > The only conflict with the patch was in the comment at top of the patch
> > (the commit that had changed this comment, 583c1f420173 ("bpf: Define
> > new BPF_MAP_TYPE_USER_RINGBUF map type"), has nothing to do with this
> > fix), so I went ahead with it.
> >
> > I'm not familiar with the ringbuf code but it doesn't look too wrong to
> > me at first glance; and with this all stable branches are covered.
>
> I need a bit more sleep; that obviously missed 5.15 which didn't get
> the fix backported either due to the same conflict; this commit applies
> to both branches.
>
> I've also checked something like bpftrace which uses the ring buffer for
> message passing doesn't blow up when spamming a bit on 5.10, just in
> case.
Both now queued up, thanks.
greg k-h
