New time code miscalculates cpu usage

New time code miscalculates cpu usage

[Olof Johansson]

Hi,

Not sure when this started happening, but I wanted to report it. I'll
start bisecting in a day or two if noone else has gotten around to
looking at it:

$ echo "int main(void) { while(1); }" > test.c ; gcc test.c
$ time ./a.out & sleep 2 ; killall a.out
real    0m2.008s
user    0m4.014s
sys     0m0.002s

Seen on POWER5 and PA6T, haven't tried anything else yet.


-Olof

Re: New time code miscalculates cpu usage

[Tony Breeds]

On Tue, Oct 16, 2007 at 03:25:25PM -0500, Olof Johansson wrote:
> Hi,
> 
> Not sure when this started happening, but I wanted to report it. I'll
> start bisecting in a day or two if noone else has gotten around to
> looking at it:
> 
> $ echo "int main(void) { while(1); }" > test.c ; gcc test.c
> $ time ./a.out & sleep 2 ; killall a.out
> real    0m2.008s
> user    0m4.014s
> sys     0m0.002s
> 
> Seen on POWER5 and PA6T, haven't tried anything else yet.

For what it's worth, this is my bug.  I suspected it, and git bisect
confirmed it.I have an ugly work around now, but hope to have something
better out tomorrow.

Thanks Olof.

Yours Tony

  linux.conf.au        http://linux.conf.au/ || http://lca2008.linux.org.au/
  Jan 28 - Feb 02 2008 The Australian Linux Technical Conference!

Re: New time code miscalculates cpu usage

[Sergei Shtylyov]

Hello.

Olof Johansson wrote:

> Not sure when this started happening, but I wanted to report it. I'll
> start bisecting in a day or two if noone else has gotten around to
> looking at it:

> $ echo "int main(void) { while(1); }" > test.c ; gcc test.c
> $ time ./a.out & sleep 2 ; killall a.out
> real    0m2.008s
> user    0m4.014s
> sys     0m0.002s
> 
> Seen on POWER5 and PA6T, haven't tried anything else yet.

    I'm not surprised -- the kernel accounts twice for each tick.

WBR, Sergei

Re: New time code miscalculates cpu usage

[Olof Johansson]

On Wed, Oct 17, 2007 at 04:07:48PM +1000, Tony Breeds wrote:
> On Tue, Oct 16, 2007 at 03:25:25PM -0500, Olof Johansson wrote:
> > Hi,
> > 
> > Not sure when this started happening, but I wanted to report it. I'll
> > start bisecting in a day or two if noone else has gotten around to
> > looking at it:
> > 
> > $ echo "int main(void) { while(1); }" > test.c ; gcc test.c
> > $ time ./a.out & sleep 2 ; killall a.out
> > real    0m2.008s
> > user    0m4.014s
> > sys     0m0.002s
> > 
> > Seen on POWER5 and PA6T, haven't tried anything else yet.
> 
> For what it's worth, this is my bug.  I suspected it, and git bisect
> confirmed it.I have an ugly work around now, but hope to have something
> better out tomorrow.

Hi Tony,

Any news?


-Olof

Re: New time code miscalculates cpu usage

[Benjamin Herrenschmidt]

On Thu, 2007-10-25 at 00:17 +0400, Sergei Shtylyov wrote:
> Hello.
> 
> Olof Johansson wrote:
> 
> > Not sure when this started happening, but I wanted to report it. I'll
> > start bisecting in a day or two if noone else has gotten around to
> > looking at it:
> 
> > $ echo "int main(void) { while(1); }" > test.c ; gcc test.c
> > $ time ./a.out & sleep 2 ; killall a.out
> > real    0m2.008s
> > user    0m4.014s
> > sys     0m0.002s
> > 
> > Seen on POWER5 and PA6T, haven't tried anything else yet.
> 
>     I'm not surprised -- the kernel accounts twice for each tick.

Your input would be much more valuable if you actually pointed out where
that happens and why since you seem to know it.

Ben.

Re: New time code miscalculates cpu usage

[Paul Mackerras]

Benjamin Herrenschmidt writes:

> Your input would be much more valuable if you actually pointed out where
> that happens and why since you seem to know it.

He did already, a couple of messages ago.

Paul.

Re: New time code miscalculates cpu usage

[Benjamin Herrenschmidt]

On Thu, 2007-10-25 at 10:19 +1000, Paul Mackerras wrote:
> Benjamin Herrenschmidt writes:
> 
> > Your input would be much more valuable if you actually pointed out where
> > that happens and why since you seem to know it.
> 
> He did already, a couple of messages ago.

Allright, I missed that.

Ben.

Re: New time code miscalculates cpu usage

[Sergei Shtylyov]

Hello.

Benjamin Herrenschmidt wrote:

>>>Not sure when this started happening, but I wanted to report it. I'll
>>>start bisecting in a day or two if noone else has gotten around to
>>>looking at it:

>>>$ echo "int main(void) { while(1); }" > test.c ; gcc test.c
>>>$ time ./a.out & sleep 2 ; killall a.out
>>>real    0m2.008s
>>>user    0m4.014s
>>>sys     0m0.002s

>>>Seen on POWER5 and PA6T, haven't tried anything else yet.

>>    I'm not surprised -- the kernel accounts twice for each tick.

> Your input would be much more valuable if you actually pointed out where
> that happens and why since you seem to know it.

   I've already pointed out the reason, yet it won't hurt to repeat indeed.
timer_interrupt() calls both account_process_time() and the hrtimers event 
handler which leads to:

- if determenistic accounting is off, account_process_time() is the same as
   update_process_times() which gets also calleed by hrtimers via the event
   handler, so we get each tick accounted for twice;

- if determenistic accounting is on, then timer_interrupt() calls
   account_process_time() and update_process_times() is still called by
   hrtimers, so each tick's gets accounted for userspace twice...

> Ben.

WBR, Sergei