Re: 2.6.10-mm2: swsusp regression [update]

[Pavel Machek]

Hi!

> > > > Thanks for pointing it out.  I have adapted this patch to -mm2, but 
> > > > unfortunately it does not fix the issue.  Still searching. ;-)
> > > 
> > > The regression is caused by the timer driver.  Obviously, turning 
> > > timer_resume() in arch/x86_64/kernel/time.c into a NOOP makes it go away.
> > > 
> > > It looks like a locking problem to me.  I'll try to find a fix, although 
> > > someone who knows more about these things would probably do it faster. :-)
> > 
> > (I do not have time right now, but...)
> > 
> > ..you might want to look at i386 time code, they have common
> > ancestor, and i386 one seems to work.
> 
> Well, I need the help of The Wise. :-)
> 
> If I comment out only the modification of jiffies in timer_resume() in 
> arch/x86_64/kernel/time.c (ie line 986), everything seems to work, but I get 
> "APIC error on CPU0: 00(00)" after device_power_up(), which seems strange to 
> me, because I boot with "noapic".  On the other hand, if it's not commented 
> out (ie jiffies _is_ modified in timer_resume()), the machine hangs solid 
> after executing device_power_up() in swsusp_suspend().

Perhaps calling handle_lost_ticks() like timer_interrupt does is right
thing to do?

> Right now I have no idea of what happens there, and it seems strange because 
> in 2.6.10-mm1 the code in time.c is the same.

Well, but in -mm1, the code is not actually called, right?

									Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/