[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20191011211758.GA12673@pc-63.home>
Date: Fri, 11 Oct 2019 23:17:58 +0200
From: Daniel Borkmann <daniel@...earbox.net>
To: Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc: Martin Lau <kafai@...com>, Andrii Nakryiko <andriin@...com>,
"bpf@...r.kernel.org" <bpf@...r.kernel.org>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Alexei Starovoitov <ast@...com>,
Kernel Team <Kernel-team@...com>
Subject: Re: [Potential Spoof] [PATCH bpf-next 2/2] selftests/bpf: remove
obsolete pahole/BTF support detection
On Fri, Oct 11, 2019 at 10:28:39AM -0700, Andrii Nakryiko wrote:
> On Fri, Oct 11, 2019 at 9:21 AM Martin Lau <kafai@...com> wrote:
> > On Thu, Oct 10, 2019 at 08:13:18PM -0700, Andrii Nakryiko wrote:
> > > Given lots of selftests won't work without recent enough Clang/LLVM that
> > > fully supports BTF, there is no point in maintaining outdated BTF
> > > support detection and fall-back to pahole logic. Just assume we have
> > > everything we need.
> > May be an error message to tell which llvm is needed?
>
> Not sure where we'd want this to be checked/printed. We don't do this
> today, so what I'm doing here is not really a regression.
> There is no single llvm version I'd want to pin down. For most tests
> LLVM w/ basic BTF support would be enough, for CO-RE stuff we need the
> latest Clang 10 (not yet released officially), though. So essentially
> the stance right now is that you need latest Clang built from sources
> to have all the tests compiled and I don't think it's easy to check
> for that.
At some point once bpf-gcc gets more mature, we might need something
more elaborate than just telling everyone to use latest clang/llvm
from git, but so far that's our convention we have in place today.
> > $(CPU) and $(PROBE) are no longer needed also?
>
> Good catch, removing them as well.
Ok, expecting v2 then.
Thanks,
Daniel
Powered by blists - more mailing lists