Re: perf, kprobes: fuzzer generates huge number of WARNings

From: Steven Rostedt
Date: Tue Jul 07 2015 - 11:10:33 EST


On Tue, 7 Jul 2015 11:06:55 -0400 (EDT)
Vince Weaver <vincent.weaver@xxxxxxxxx> wrote:

> On Tue, 7 Jul 2015, Masami Hiramatsu wrote:
>
> > > It's a Hawell machine running 4.2-rc1. I can post the .config if it's of
> > > interest.
> >
> > Yes, I'm interested in.
>
> I've put it here:
> http://web.eece.maine.edu/~vweaver/junk/haswell-config
>
> > OK, and also, if you have a chance, please run the ftracetest as below.
> >
> > $ cd tools/testing/selftest/ftrace/
> > $ sudo ./ftracetest
> >
> > This will do a series of basic tests on ftrace and report it.
>
> It won't run:
> sudo ./ftracetest
> Error: No ftrace directory found
>
> probably because I am running 4.2-rc1 and have tracefs mounted under
> /sys/kernel/tracing.

You could also just mount debugfs.

> I guess the tester needs to be updated?

Yes, it probably should be.

> I'll see if I can get debugfs mounted too, but I had changed the setup so
> perf would run.
>
>
> When I recompiled my kernel to remove the extra debug messages and run my
> fuzzer test I get a lot of these, probably because I have the kprobe
> address hard-coded into the tester.

Probably. If you recompiled the kernel, you need to find the new
address points.

-- Steve

>
> [ 290.067450] Could not insert probe at _text+1664816: -84
> [ 290.074976] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.131191] Could not insert probe at _text+1664816: -84
> [ 290.137555] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.170513] Could not insert probe at _text+1664816: -84
> [ 290.176877] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.298940] Could not insert probe at _text+1664816: -84
> [ 290.305224] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.349272] Could not insert probe at _text+1664816: -84
> [ 290.355585] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.365738] Could not insert probe at _text+1664816: -84
> [ 290.371992] Probing address(0xffffffff81196730) is not an instruction boundary.
> [ 290.383820] perf interrupt took too long (448 > 266), lowering kernel.perf_event_max_sample_rate to 468250
> [ 295.088511] perf interrupt took too long (485 > 0), lowering kernel.perf_event_max_sample_rate to 1639281051
> [ 311.415309] INFO: rcu_sched detected stalls on CPUs/tasks: { 5} (detected by 6, t=5253 jiffies, g=1818, c=1817, q=534)
> [ 311.427531] Task dump for CPU 5:
> [ 311.431631] in:imklog R running task 0 1736 1 0x00000008
> [ 311.439865] 0000000000000286 0000000100000005 0000000000000096 0000000200000000
> [ 311.448549] 0000000000000286 ffffffff81c4b2e0 ffffffff81c4b2e0 ffffffff828da234
> [ 311.457173] 0000000000001f2f 00007fa62335fe11 0000000000000071 ffffffff828da234
> [ 311.465865] Call Trace:
> [ 311.469098] [<ffffffff810c8886>] ? do_syslog+0x456/0x500
> [ 311.475518] [<ffffffff810b02a0>] ? wait_woken+0x90/0x90
> [ 311.481830] [<ffffffff81255bb4>] ? kmsg_read+0x44/0x60
> [ 311.488040] [<ffffffff816a686e>] ? mutex_unlock+0xe/0x10
> [ 311.494433] [<ffffffff81003000>] ? calibrate_delay+0x8f0/0x8f0
> [ 311.501356] [<ffffffff816a98e5>] ? retint_check+0x6/0x12
> [ 313.873152] hrtimer: interrupt took 14075878236 ns

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/