[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 3 Mar 2020 08:27:44 -0800
From: Alexei Starovoitov <alexei.starovoitov@...il.com>
To: Toke Høiland-Jørgensen <toke@...hat.com>
Cc: Daniel Borkmann <daniel@...earbox.net>,
Alexei Starovoitov <ast@...com>, bpf <bpf@...r.kernel.org>,
Network Development <netdev@...r.kernel.org>,
Andrey Ignatov <rdna@...com>
Subject: Re: [PATCH bpf] selftests/bpf: Declare bpf_log_buf variables as static
On Tue, Mar 3, 2020 at 12:10 AM Toke Høiland-Jørgensen <toke@...hat.com> wrote:
>
> Alexei Starovoitov <alexei.starovoitov@...il.com> writes:
>
> > On Mon, Mar 02, 2020 at 03:53:48PM +0100, Toke Høiland-Jørgensen wrote:
> >> The cgroup selftests did not declare the bpf_log_buf variable as static, leading
> >> to a linker error with GCC 10 (which defaults to -fno-common). Fix this by
> >> adding the missing static declarations.
> >>
> >> Fixes: 257c88559f36 ("selftests/bpf: Convert test_cgroup_attach to prog_tests")
> >> Signed-off-by: Toke Høiland-Jørgensen <toke@...hat.com>
> >
> > Applied to bpf-next.
> > It's hardly a fix. Fixes tag doesn't make it a fix in my mind.
>
> It fixes a compile error of selftests with GCC 10; how is that not a
> fix? We found it while setting up a CI test compiling Linus' tree on
> Fedora rawhide, so it does happen in the wild.
>
> > I really see no point rushing it into bpf->net->Linus's tree at this point.
>
> Well if you're not pushing any other fixes then OK, sure, no reason to
> go through the whole process just for this. But if you end up pushing
> another round of fixes anyway, please include this as well. If not, I
> guess we can wait :)
CI stands for Continuous Integration == development.
stable tree is not for development.
If you want to develop anything or accommodate the tree
for external development you need to use development tree.
Which is bpf-next.
Powered by blists - more mailing lists