Re: [PATCH 0/8] bpf: Add fprobe link

From: Masami Hiramatsu <mhiramat@kernel.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>,
	Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andrii@kernel.org>,
	Masami Hiramatsu <mhiramat@kernel.org>,
	Network Development <netdev@vger.kernel.org>,
	bpf <bpf@vger.kernel.org>, lkml <linux-kernel@vger.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@chromium.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Jiri Olsa <olsajiri@gmail.com>
Subject: Re: [PATCH 0/8] bpf: Add fprobe link
Date: Fri, 4 Feb 2022 09:46:19 +0900	[thread overview]
Message-ID: <20220204094619.2784e00c0b7359356458ca57@kernel.org> (raw)
In-Reply-To: <YfvvfLlM1FOTgvDm@krava>

On Thu, 3 Feb 2022 16:06:36 +0100
Jiri Olsa <jolsa@redhat.com> wrote:

> On Wed, Feb 02, 2022 at 09:30:21AM -0800, Alexei Starovoitov wrote:
> > On Wed, Feb 2, 2022 at 9:24 AM Jiri Olsa <jolsa@redhat.com> wrote:
> > >
> > > On Wed, Feb 02, 2022 at 09:09:53AM -0800, Alexei Starovoitov wrote:
> > > > On Wed, Feb 2, 2022 at 5:53 AM Jiri Olsa <jolsa@redhat.com> wrote:
> > > > >
> > > > > hi,
> > > > > this patchset adds new link type BPF_LINK_TYPE_FPROBE that attaches kprobe
> > > > > program through fprobe API [1] instroduced by Masami.
> > > >
> > > > No new prog type please.
> > > > I thought I made my reasons clear earlier.
> > > > It's a multi kprobe. Not a fprobe or any other name.
> > > > The kernel internal names should not leak into uapi.
> > > >
> > >
> > > well it's not new prog type, it's new link type that allows
> > > to attach kprobe program to multiple functions
> > >
> > > the original change used BPF_LINK_TYPE_RAW_KPROBE, which did not
> > > seem to fit anymore, so I moved to FPROBE, because that's what
> > > it is ;-)
> > 
> > Now I don't like the fprobe name even more.
> > Why invent new names? It's an ftrace interface.
> 
> how about ftrace_probe ?

I thought What Alexei pointed was that don't expose the FPROBE name
to user space. If so, I agree with that. We can continue to use
KPROBE for user space. Using fprobe is just for kernel implementation.

It means that we may better to keep simple mind model (there are only
static event or dynamic kprobe event).

> > > but if you don't want new name in uapi we could make this more
> > > obvious with link name:
> > >   BPF_LINK_TYPE_MULTI_KPROBE
> > >
> > > and bpf_attach_type:
> > >   BPF_TRACE_MULTI_KPROBE
> > 
> > I'd rather get rid of fprobe name first.
> >
> 
> Masami, any idea?

Can't we continue to use kprobe prog type for user interface
and internally, if there are multiple kprobes or kretprobes
required, switch to use fprobe?

Thank you,

> 
> thanks,
> jirka
> 

-- 
Masami Hiramatsu <mhiramat@kernel.org>