Re: [PATCH] x86/ibt: make is_endbr() notrace

[Peter Zijlstra <peterz@infradead.org>]

On Thu, Sep 18, 2025 at 09:32:27PM +0800, Menglong Dong wrote:
> On Thu, Sep 18, 2025 at 9:05???PM Peter Zijlstra <peterz@infradead.org> wrote:
> >
> > On Thu, Sep 18, 2025 at 08:09:39PM +0800, Menglong Dong wrote:
> > > is_endbr() is called in __ftrace_return_to_handler -> fprobe_return ->
> > > kprobe_multi_link_exit_handler -> is_endbr.
> > >
> > > It is not protected by the "bpf_prog_active", so it can't be traced by
> > > kprobe-multi, which can cause recurring and panic the kernel. Fix it by
> > > make it notrace.
> >
> > This is very much a riddle wrapped in an enigma. Notably
> > kprobe_multi_link_exit_handler() does not call is_endbr(). Nor is that
> > cryptic next line sufficient to explain why its a problem.
> >
> > I suspect the is_endbr() you did mean is the one in
> > arch_ftrace_get_symaddr(), but who knows.
> 
> Yeah, I mean
> kprobe_multi_link_exit_handler -> ftrace_get_entry_ip ->
> arch_ftrace_get_symaddr -> is_endbr
> actually. And CONFIG_X86_KERNEL_IBT is enabled of course.
> 
> >
> > Also, depending on compiler insanity, it is possible the thing
> > out-of-lines things like __is_endbr(), getting you yet another
> > __fentry__ site.
> 
> The panic happens when I run the bpf bench testing:
>   ./bench kretprobe-multi-all
> 
> And skip the "is_endbr" fix this problem.

But why does it panic? Supposedly you've done the analysis; but then
forgot to write it down?

Why is kprobe_multi_link_exit_handler() special; doesn't the issue also
exist with kprobe_multi_link_handler() ? If so, removing __fentry__
isn't going to help much, you can just stick an actual kprobe in
is_endbr(), right?