linux-kernel - intel_pstate_timer_func divide by zero oops

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 27 Mar 2013 21:49:13 -0400 (EDT)
From:	Parag Warudkar <parag.lkml@...il.com>
To:	rjw@...k.pl, cpufreq@...r.kernel.org, linux-pm@...r.kernel.org
cc:	linux-kernel@...r.kernel.org, torvalds@...ux-foundation.org
Subject: intel_pstate_timer_func divide by zero oops

I get this same oops occassionally - the machine freezes and there doesn't 
seem to be any record of the oops on disk.

I captured it on camera - 
https://lh3.googleusercontent.com/-K0lNbJrZBMQ/UVOU1vv1vvI/AAAAAAAANqI/pY92mWm3caE/s800/20130327_205245.jpg

If I am reading this right, it dies on this instruction -

   0xffffffff8145792d <+349>:   divq   0x18(%rcx)

>From the lst file that *seems* to be this inline function -

static inline void intel_pstate_calc_busy(struct cpudata *cpu,
                                        struct sample *sample)
{
        u64 core_pct;
        sample->pstate_pct_busy = 100 - div64_u64(
ffffffff8145791d:       48 8b 41 20             mov    0x20(%rcx),%rax
ffffffff81457921:       48 8d 04 80             lea    (%rax,%rax,4),%rax
ffffffff81457925:       48 8d 04 80             lea    (%rax,%rax,4),%rax
ffffffff81457929:       48 c1 e0 02             shl    $0x2,%rax
ffffffff8145792d:       48 f7 71 18             divq   0x18(%rcx)


That  is -
	sample->pstate_pct_busy = 100 - div64_u64(
					sample->idletime_us * 100,
					sample->duration_us);

So looks like sample->duration_us is 0? If so, that implies that 
ktime_us_delta(now, cpu->prev_sample) is zero. I am not entirely sure how 
to handle this case - return if sampling too early, or if there is some 
other bug making the delta calculation go poof.


Thanks,

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