Re: [PATCH bpf-next] selftests/bpf: Filter out preempt_count_ functions from kprobe_multi bench

[Jiri Olsa <olsajiri@gmail.com>]

On Sun, Mar 19, 2023 at 09:54:56AM -0700, Alexei Starovoitov wrote:
> On Sun, Mar 19, 2023 at 03:36:57PM +0800, Yafang Shao wrote:
> > On Sat, Mar 18, 2023 at 12:52 AM Alexei Starovoitov
> > <alexei.starovoitov@gmail.com> wrote:
> > >
> > > On Fri, Mar 17, 2023 at 4:49 AM Yafang Shao <laoar.shao@gmail.com> wrote:
> > > >
> > > > It hits below warning on my test machine when running test_progs,
> > > >
> > > > [  702.223611] ------------[ cut here ]------------
> > > > [  702.224168] RCU not on for: preempt_count_sub+0x0/0xa0
> > > > [  702.224770] WARNING: CPU: 14 PID: 5267 at include/linux/trace_recursion.h:162 fprobe_handler.part.0+0x1b8/0x1c0
> > > > [  702.231740] CPU: 14 PID: 5267 Comm: main_amd64 Kdump: loaded Tainted: G           O       6.2.0+ #584
> > > > [  702.233169] RIP: 0010:fprobe_handler.part.0+0x1b8/0x1c0
> > > > [  702.241388] Call Trace:
> > > > [  702.241615]  <TASK>
> > > > [  702.241811]  fprobe_handler+0x22/0x30
> > > > [  702.242129]  0xffffffffc04710f7
> > > > [  702.242417] RIP: 0010:preempt_count_sub+0x5/0xa0
> > > > [  702.242809] Code: c8 50 68 94 42 0e b5 48 cf e9 f9 fd ff ff 0f 1f 80 00 00 00 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 e8 4b cd 38 0b <55> 8b 0d 9c d0 cf 02 48 89 e5 85 c9 75 1b 65 8b 05 be 78 f4 4a 89
> > > > [  702.244752] RSP: 0018:ffffaf6187d27f10 EFLAGS: 00000082 ORIG_RAX: 0000000000000000
> > > > [  702.245801] RAX: 000000000000000e RBX: 0000000001b6ab72 RCX: 0000000000000000
> > > > [  702.246804] RDX: 0000000000000000 RSI: ffffffffb627967d RDI: 0000000000000001
> > > > [  702.247801] RBP: ffffaf6187d27f30 R08: 0000000000000000 R09: 0000000000000000
> > > > [  702.248786] R10: 0000000000000000 R11: 0000000000000000 R12: 00000000000000ca
> > > > [  702.249782] R13: ffffaf6187d27f58 R14: 0000000000000000 R15: 0000000000000000
> > > > [  702.250785]  ? preempt_count_sub+0x5/0xa0
> > > > [  702.251540]  ? syscall_enter_from_user_mode+0x96/0xc0
> > > > [  702.252368]  ? preempt_count_sub+0x5/0xa0
> > > > [  702.253104]  ? syscall_enter_from_user_mode+0x96/0xc0
> > > > [  702.253918]  do_syscall_64+0x16/0x90
> > > > [  702.254613]  entry_SYSCALL_64_after_hwframe+0x72/0xdc
> > > > [  702.255422] RIP: 0033:0x46b793
> > > >
> > > > It's caused by bench test attaching kprobe_multi link to preempt_count_sub
> > > > function, which is not executed in rcu safe context so the kprobe handler
> > > > on top of it will trigger the rcu warning.
> > >
> > > Why is that?
> > 
> > It is caused by CONFIG_CONTEXT_TRACKING_USER=y, and it seems the
> > preempt_count_sub is executed before the RCU is watching.
> >   user_exit_irqoff
> >       if (context_tracking_enabled())  // CONFIG_CONTEXT_TRACKING_USER=y
> >           __ct_user_exit(CONTEXT_USER);
> >              ct_kernel_enter
> >                  ...
> >                  // RCU is not watching here ...
> >                  ct_kernel_enter_state(offset);
> >                  // ... but is watching here.
> > 
> > It can be reproduced with a simple bpf code as follows when
> > CONFIG_CONTEXT_TRACKING_USER=y,
> >   SEC("kprobe.multi/preempt_count_sub")
> >   int kprobe_multi_trace()
> >   {
> >       return 0;
> >   }
> > 
> > > preempt_count itself is fine.
> > > The problem is elsewhere.
> > > Since !rcu_is_watching() it some sort of idle or some other issue.
> > 
> > Not sure if we need to improve the code under
> > CONFIG_CONTEXT_TRACKING_USER=y, but it seems skipping "preempt_count_"
> > in kprobe_multi test case would be a quick fix.
> 
> It's not a fix. Only moving a goal post.
> We probably need
>         if (!rcu_is_watching())
>                 return;
> in [kf]probe handler instead.

that might also help with some of the warnings we saw from idle
functions that we skip at the moment

jirka