linux-kernel - Re: HRTimer causing rtctest to fail

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-prev] [day] [month] [year] [list]

Message-ID: <87h8q15nlf.fsf@linux.intel.com>
Date:   Wed, 28 Feb 2018 14:02:20 +0200
From:   Felipe Balbi <felipe.balbi@...ux.intel.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Alessandro Zummo <a.zummo@...ertech.it>,
        Alexandre Belloni <alexandre.belloni@...e-electrons.com>,
        linux-kernel@...r.kernel.org, linux-rtc@...r.kernel.org
Subject: Re: HRTimer causing rtctest to fail


Hi,

Thomas Gleixner <tglx@...utronix.de> writes:
> On Wed, 28 Feb 2018, Felipe Balbi wrote:
>> Thomas Gleixner <tglx@...utronix.de> writes:
>> > All numbers in micro seconds! What happens if you load the system or
>> > restrict C-States?
>> 
>> added intel_idle.max_cstate=0 to cmdline. Still see some failures,
>> albeit a lot more rare, for example, caught one on iteration 8 of
>> 8192Hz:
>> 
>> 8192Hz:	 1 2 3 4 5 6 7
>> PIE delta error: 0.000135 should be close to 0.000122
>> 
>> That is, however, only 13 uS late. Is that within the limits of "we
>> don't quite care"?
>
> I think so. 13us (depending on the machine) can easily happen if multiple
> interrupts/timers happen at the same time or if something keeps interrupts
> disabled for a few microseconds which is certainly possible.
>
> What kind of system is this?

4 core Kabylake.

>> BTW, it took some effort to get that one failure. I'll leave rtctest
>> running in a long loop and try to figure out approximately how frequently
>> it fails, but it seems like it's within the realm of "don't care", would
>> you agree?
>
> Yes.

cool, thanks

-- 
balbi

Download attachment "signature.asc" of type "application/pgp-signature" (833 bytes)