From: Steven Rostedt <rostedt@goodmis.org>
To: Nick Desaulniers <ndesaulniers@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Sami Tolvanen <samitolvanen@google.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
linux-kernel <linux-kernel@vger.kernel.org>,
Matt Mullins <mmullins@mmlx.us>, Ingo Molnar <mingo@redhat.com>,
Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Dmitry Vyukov <dvyukov@google.com>,
Martin KaFai Lau <kafai@fb.com>, Song Liu <songliubraving@fb.com>,
Yonghong Song <yhs@fb.com>, Andrii Nakryiko <andriin@fb.com>,
John Fastabend <john.fastabend@gmail.com>,
KP Singh <kpsingh@chromium.org>, netdev <netdev@vger.kernel.org>,
bpf <bpf@vger.kernel.org>, Kees Cook <keescook@chromium.org>,
Josh Poimboeuf <jpoimboe@redhat.com>,
linux-toolchains@vger.kernel.org
Subject: Re: violating function pointer signature
Date: Wed, 18 Nov 2020 12:17:30 -0500 [thread overview]
Message-ID: <20201118121730.12ee645b@gandalf.local.home> (raw)
In-Reply-To: <CAKwvOdkptuS=75WjzwOho9ZjGVHGMirEW3k3u4Ep8ya5wCNajg@mail.gmail.com>
On Wed, 18 Nov 2020 08:50:37 -0800
Nick Desaulniers <ndesaulniers@google.com> wrote:
> On Wed, Nov 18, 2020 at 5:23 AM Peter Zijlstra <peterz@infradead.org> wrote:
> >
> > On Tue, Nov 17, 2020 at 03:34:51PM -0500, Steven Rostedt wrote:
> >
> > > > > Since all tracepoints callbacks have at least one parameter (__data), we
> > > > > could declare tp_stub_func as:
> > > > >
> > > > > static void tp_stub_func(void *data, ...)
> > > > > {
> > > > > return;
> > > > > }
> > > > >
> > > > > And now C knows that tp_stub_func() can be called with one or more
> > > > > parameters, and had better be able to deal with it!
> > > >
> > > > AFAIU this won't work.
> > > >
> > > > C99 6.5.2.2 Function calls
> > > >
> > > > "If the function is defined with a type that is not compatible with the type (of the
> > > > expression) pointed to by the expression that denotes the called function, the behavior is
> > > > undefined."
> > >
> > > But is it really a problem in practice. I'm sure we could create an objtool
> > > function to check to make sure we don't break anything at build time.
> >
> > I think that as long as the function is completely empty (it never
> > touches any of the arguments) this should work in practise.
> >
> > That is:
> >
> > void tp_nop_func(void) { }
>
> or `void tp_nop_func()` if you plan to call it with different
> parameter types that are all unused in the body. If you do plan to
> use them, maybe a pointer to a tagged union would be safer?
This stub function will never use the parameters passed to it.
You can see the patch I have for the tracepoint issue here:
https://lore.kernel.org/r/20201118093405.7a6d2290@gandalf.local.home
I could change the stub from (void) to () if that would be better.
>
> >
> > can be used as an argument to any function pointer that has a void
> > return. In fact, I already do that, grep for __static_call_nop().
> >
> > I'm not sure what the LLVM-CFI crud makes of it, but that's their
> > problem.
>
> If you have instructions on how to exercise the code in question, we
> can help test it with CFI. Better to find any potential issues before
> they get committed.
If you apply the patch to the Linux kernel, and then apply:
https://lore.kernel.org/r/20201116181638.6b0de6f7@gandalf.local.home
Which will force the failed case (to use the stubs). And build and boot the
kernel with those patches applied, you can test it with:
# mount -t tracefs nodev /sys/kernel/tracing
# cd /sys/kernel/tracing
# echo 1 > events/sched/sched_switch/enable
# mkdir instances/foo
# echo 1 > instances/foo/events/sched/sched_switch/enable
# echo 0 > events/sched/sched_switch/enable
Which add two callbacks to the function array for the sched_switch
tracepoint. The remove the first one, which would add the stub instead.
-- Steve
next prev parent reply other threads:[~2020-11-18 17:17 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-16 22:51 [PATCH] tracepoint: Do not fail unregistering a probe due to memory allocation Steven Rostedt
2020-11-16 23:16 ` Steven Rostedt
2020-11-17 19:15 ` Mathieu Desnoyers
2020-11-17 19:21 ` Steven Rostedt
2020-11-17 19:47 ` Mathieu Desnoyers
2020-11-17 20:34 ` Steven Rostedt
2020-11-17 20:58 ` Steven Rostedt
2020-11-17 21:22 ` Mathieu Desnoyers
2020-11-17 22:16 ` Steven Rostedt
2020-11-17 23:08 ` Mathieu Desnoyers
2020-11-18 1:11 ` Steven Rostedt
2020-11-17 21:08 ` Mathieu Desnoyers
2020-11-18 13:21 ` violating function pointer signature Peter Zijlstra
2020-11-18 13:59 ` Florian Weimer
2020-11-18 14:12 ` Peter Zijlstra
2020-11-18 14:18 ` Florian Weimer
2020-11-18 14:34 ` [PATCH v3] tracepoint: Do not fail unregistering a probe due to memory allocation Steven Rostedt
2020-11-24 5:59 ` Matt Mullins
2020-11-18 14:22 ` violating function pointer signature Steven Rostedt
2020-11-18 19:46 ` Alexei Starovoitov
2020-11-18 20:02 ` Steven Rostedt
2020-11-18 14:02 ` Steven Rostedt
2020-11-18 16:01 ` Mathieu Desnoyers
2020-11-18 16:19 ` David Laight
2020-11-18 16:50 ` Nick Desaulniers
2020-11-18 17:17 ` Steven Rostedt [this message]
2020-11-18 18:12 ` Segher Boessenkool
2020-11-18 18:31 ` Florian Weimer
2020-11-18 18:55 ` Segher Boessenkool
2020-11-18 18:58 ` Steven Rostedt
2020-11-18 18:59 ` Steven Rostedt
2020-11-18 19:11 ` Segher Boessenkool
2020-11-18 19:33 ` Steven Rostedt
2020-11-18 19:48 ` Segher Boessenkool
2020-11-18 20:44 ` Steven Rostedt
2020-11-19 8:21 ` Peter Zijlstra
2020-11-19 8:36 ` Peter Zijlstra
2020-11-19 14:37 ` Segher Boessenkool
2020-11-19 14:59 ` Steven Rostedt
2020-11-19 16:35 ` Segher Boessenkool
2020-11-19 17:42 ` David Laight
2020-11-19 19:27 ` Segher Boessenkool
2020-11-19 17:04 ` Alexei Starovoitov
2020-11-19 17:30 ` Steven Rostedt
2020-11-20 1:31 ` Nick Desaulniers
2020-11-17 21:33 ` [PATCH] tracepoint: Do not fail unregistering a probe due to memory allocation Kees Cook
2020-11-17 22:19 ` Steven Rostedt
2020-11-17 23:12 ` Mathieu Desnoyers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20201118121730.12ee645b@gandalf.local.home \
--to=rostedt@goodmis.org \
--cc=andriin@fb.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=dvyukov@google.com \
--cc=john.fastabend@gmail.com \
--cc=jpoimboe@redhat.com \
--cc=kafai@fb.com \
--cc=keescook@chromium.org \
--cc=kpsingh@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-toolchains@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=mingo@redhat.com \
--cc=mmullins@mmlx.us \
--cc=ndesaulniers@google.com \
--cc=netdev@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=samitolvanen@google.com \
--cc=songliubraving@fb.com \
--cc=yhs@fb.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).