Re: [PATCH bpf-next v2 4/6] selftests/bpf: Null checks for link in bpf_tcp_ca

[Geliang Tang <geliang@kernel.org>]

Hi Eduard,

On Mon, 2024-07-08 at 12:42 -0700, Eduard Zingerman wrote:
> On Fri, 2024-07-05 at 10:38 +0800, Geliang Tang wrote:
> 
> [...]
> 
> I think that this patch is an improvement independent of the patch-
> set.
> Please submit it separately.
> 
> >  .../selftests/bpf/prog_tests/bpf_tcp_ca.c        | 16
> > ++++++++++++----
> 
> [...]
> 
> > @@ -489,6 +494,7 @@ static void test_mixed_links(void)
> >  	ASSERT_ERR(err, "update_map");
> >  
> >  	bpf_link__destroy(link);
> > +err:
> 
> Nit: there are two links in this test, but ASSERT_OK_PTR is added
> only
>      for a single one. Also note that bpf_link__destroy(NULL) works
>      just fine, so it is possible to initialize links as NULL and
> make
>      a jump to cleanup block w/o peeking exact position within that
> block.

Thanks for your review. I sent a set named "BPF selftests misc fixes"
yesterday to address your comments.

But reconsider it. I think here checking the first link (link_nl) is
enough. We can keep the second link as it.

I changed "BPF selftests misc fixes" as "Changes Requested".

Thanks,
-Geliang

> 
> >  	bpf_link__destroy(link_nl);
> >  	tcp_ca_update__destroy(skel);
> >  }
> 
> [...]
>