Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Sun, 28 Mar 2021 18:26:02 -0700
From:   Alexei Starovoitov <alexei.starovoitov@gmail.com>
To:     Toke =?utf-8?Q?H=C3=B8iland-J=C3=B8rgensen?= <toke@redhat.com>
Cc:     Kumar Kartikeya Dwivedi <memxor@gmail.com>, bpf@vger.kernel.org,
        brouer@redhat.com, Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Andrii Nakryiko <andrii@kernel.org>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@kernel.org>, Shuah Khan <shuah@kernel.org>,
        "David S. Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        Jesper Dangaard Brouer <hawk@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
        linux-kselftest@vger.kernel.org
Subject: Re: [PATCH bpf-next 5/5] libbpf: add selftests for TC-BPF API
Message-ID: <20210329012602.4zzysn2ewbarbn3d@ast-mbp>
References: <20210325120020.236504-1-memxor@gmail.com>
 <20210325120020.236504-6-memxor@gmail.com>
 <20210327021534.pjfjctcdczj7facs@ast-mbp>
 <87h7kwaao3.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <87h7kwaao3.fsf@toke.dk>
Precedence: bulk

On Sat, Mar 27, 2021 at 04:17:16PM +0100, Toke H?iland-J?rgensen wrote:
> Alexei Starovoitov <alexei.starovoitov@gmail.com> writes:
> 
> > On Thu, Mar 25, 2021 at 05:30:03PM +0530, Kumar Kartikeya Dwivedi wrote:
> >> This adds some basic tests for the low level bpf_tc_* API and its
> >> bpf_program__attach_tc_* wrapper on top.
> >
> > *_block() apis from patch 3 and 4 are not covered by this selftest.
> > Why were they added ? And how were they tested?
> >
> > Pls trim your cc. bpf@vger and netdev@vger would have been enough.
> >
> > My main concern with this set is that it adds netlink apis to libbpf while
> > we already agreed to split xdp manipulation pieces out of libbpf.
> > It would be odd to add tc apis now only to split them later.
> 
> We're not removing the ability to attach an XDP program via netlink from
> libxdp, though. This is the equivalent for TC: the minimum support to
> attach a program, and if you want to do more, you pull in another
> library or roll your own.
> 
> I'm fine with cutting out more stuff and making this even more minimal
> (e.g., remove the block stuff and only support attach/detach on ifaces),
> but we figured we'd err on the side of including too much and getting
> some feedback from others on which bits are the essential ones to keep,
> and which can be dropped.

This is up to you. I'm trying to understand the motivation for *_block() apis.
I'm not taking a stance for/against them.

> > I think it's better to start with new library for tc/xdp and have
> > libbpf as a dependency on that new lib.
> > For example we can add it as subdir in tools/lib/bpf/.
> 
> I agree for the higher-level stuff (though I'm not sure what that would
> be for TC), but right now TC programs are the only ones that cannot be
> attached by libbpf, which is annoying; that's what we're trying to fix.

Sure. I wasn't saying that there is no place for these APIs in libbpf+.
Just that existing libbpf is already became a kitchen sink of features
that users are not going to use like static linking.
tc-api was a straw that broke the camel's back.
I think we must move static linking and skeleton out of libbpf before
the next release.