On Tue, Jul 12, 2022 at 7:10 PM Stanislav Fomichev <sdf@xxxxxxxxxx> wrote:
>
> On Tue, Jul 12, 2022 at 6:39 PM Hao Luo <haoluo@xxxxxxxxxx> wrote:
> >
> > On Tue, Jul 12, 2022 at 6:26 PM Alexei Starovoitov
> > <alexei.starovoitov@xxxxxxxxx> wrote:
> > >
> > > On Tue, Jul 12, 2022 at 6:20 PM Hao Luo <haoluo@xxxxxxxxxx> wrote:
> > > >
> > > > On Tue, Jul 12, 2022 at 6:10 PM Joanne Koong <joannelkoong@xxxxxxxxx> wrote:
> > > > >
> > > > > On Tue, Jul 12, 2022 at 3:44 PM <sdf@xxxxxxxxxx> wrote:
> > > > > >
> > > > > > On 07/12, Joanne Koong wrote:
> > > > > > > This patch does two things:
> > > > > >
> > > > > > > 1. For matching against the arg type, the match should be against the
> > > > > > > base type of the arg type, since the arg type can have different
> > > > > > > bpf_type_flags set on it.
> > > > > >
> > > > > > Does this need a fixes tag? Something around the following maybe:
> > > > > >
> > > > > > Fixes: d639b9d13a39 ("bpf: Introduce composable reg, ret and arg types.")
> > > > > >
> > > > > > ?
> > > > > I will add that tag. Thanks!
> > > >
> > > > Joanne and Stan, IMO this is not necessary. I think this change is a
> > > > cleanup rather than a fix.
> > >
> > > I don't see the bug easier.
> > > The helper types that are compared directly as arg_type
> > > instead of base_type(arg_type) were all without flags so far.
> > > So I don't think the patch changes behavior or fixes anything today.
> > > It looks like a good future proofing change though.
> > > Am I missing something?
> >
> > Agree, I mean adding a fixes tag isn't necessary and the patch is
> > toward a good direction. As long as the selftests pass on this patch,
> > it looks good to me.
> >
> > Acked-by: Hao Luo <haoluo@xxxxxxxxxx>
>
> Perfect, thanks for clarifying! It wasn't clear from the description
> so I started looking for where that base_type() came from.
Great, I will leave it as is then. Thanks for the feedback.
