* Stefano Brivio <stefano.brivio@polimi.it> wrote:

quoted text
> > Stefano, could you please try to sum up your experiences with that 
> > issue? Is it reproducable, and the 5 patches i did fix it? (if yes, 
> > could you try to re-do the mdelay verifications perhaps, to make 
> > sure it's not some other effect interacting here. In theory 
> > sched-clock scaling has no effect on udelay behavior.)
> 
> Sorry for disappearing. Anyway, yes, those patches fixed it. Precision 
> in delays isn't that good when using my crappy unstable TSC 
> (mdelay(2000) causes delays between 2 and 2.9 seconds) but it's not 
> depending on frequency changes anymore. So I'd say it's fixed, but 
> please tell me if you want me to do any other test so as to be sure it 
> is.

ok, just to make sure we are all synced up. I made 8 patches related to 
this problem category (and all the trickle effects). 3 are upstream 
already, 5 are pending for v2.6.25. One out of those 5 is an immaterial 
cleanup patch - which leaves us 4 patches to sort out.

So i'd suggest for you to try latest -git - that will tell us whether 
udelay() is acceptable on your box right now.

i've attached those 4 patches:

 x86-sched_clock-re-scheduler-fix-x86-regression-in-native-sched-clock.patch
 x86-cpu-clock-idle-event.patch
 sched-printk-recursion-fix.patch
 sched-printk-clock-fix.patch

none of them is _supposed_ to have any effect on udelay(), but the 
interactions in this area are weird.

[ note: CONFIG_PRINTK_TIME will be broken and only fixed in v2.6.25, so 
  use some other time metric for determining mdelay quality. ]

plus then there's this patch:

  http://lkml.org/lkml/2007/12/7/100

is it perhaps this one that fixed udelay for you? [ which would be much 
more expected, as this patch changes udelay ;-) ]

	Ingo