Re: [BISECTED] 2.6.35.*: horrible (exponential? >linear) slowdown to unusability (HPET)

[Nix <nix@esperi.org.uk>]

On 10 Sep 2010, Jiri Slaby spake thusly:

> On 09/10/2010 10:37 AM, Thomas Gleixner wrote:
>> HPET provides both functionalities.
>> 
>> The patch you bisected is affecting the clock events part of the
>> HPET. So yes, it's not a clock source problem.
>
> AFAIU, clocksource=jiffies shouldn't fix this issue. If that's the case
> we have another 2.6.32->2.6.33 regression (and I will report it

What? This is a 2.6.35 regression from 2.6.34 (and all previous kernels).
clocksource=jiffies has never yet been mentioned (I could try it, but
I think I know what the result would be).

2.6.32 and 2.6.33 both worked fine.

> separately). People report the system is unusable (sleepers are not
> woken), unless clocksource=jiffies, clocksource=tsc or nolapic_timer is
> used or a key pressed (i.e. some HW interrupt):
> https://bugzilla.novell.com/show_bug.cgi?id=579932

This is similar, but not identical: this particular bug is only fixed by
keypresses in the early stages (in NOHZ mode) and *interferes* with all
responsiveness including keystroke response both in NOHZ and !NOHZ mode.

It's probably another HPET timer problem, though, but that doesn't mean
it necessarily has the same cause.