Re: [2.6.22-rc3][ACPI?] Resume from s2r doesn't work.

[Pavel Machek <pavel@ucw.cz>]

Hi!

> > Second, you can use PM_TRACE (Documentation/power/s2ram.txt) to find the
> > place where it really fails.
> 
> First I augmented my minimal config kernel with some TRACE_RESUME()s:
> 
> --- a/kernel/power/main.c	2007-05-27 23:48:05.000000000 +0200
> +++ b/kernel/power/main.c	2007-06-03 22:28:46.000000000 +0200
> @@ -223,6 +223,7 @@
>  	pr_debug("PM: Finishing wakeup.\n");
>  	suspend_finish(state);
>   Unlock:
> +	TRACE_RESUME(error);
>  	mutex_unlock(&pm_mutex);
>  	return error;
>  }
> @@ -303,6 +304,7 @@
>  		error = enter_state(state);
>  	else
>  		error = -EINVAL;
> +	TRACE_RESUME(error);
>  	return error ? error : n;
>  }
...
> I got alternating
>   hash matches drivers/base/power/resume.c:58
> and
>   hash matches kernel/power/main.c:307
> 
> > First, you can check if the patch
> >
> > http://www.sisk.pl/kernel/hibernation_and_suspend/2.6.22-rc3/patches/20-ACPI-preserve-the-ebx-value-in-acpi_copy_wakeup_routine.patch
> 
> Then I applied this patch, but it doesn't change anything.
> 
> But either way the script never reaches "shutdown -rn now". So, it
> seems, that my laptop does a full resume every other reboot, but it
> never returns to userspace.

But it returns to kernel... can you use more trace points to figure
out where it hangs?

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html