[LAU] suspending with realtime kernel

Ken Restivo ken at restivo.org
Tue Apr 21 13:38:34 EDT 2009


On Tue, Apr 21, 2009 at 12:15:56PM +0200, Robert Persson wrote:
> Ken Restivo a ??crit :
> > In order to get suspend to play nice with RT/Ingo kernel on my EEE, I had to turn off Hyperthreading in the BIOS. Everything works fine now.
> >
> > ACPI suspend/resume is really where the rubber meets the road in terms of hardware support. If a kernel can handle suspending to RAM, pretty much it can handle anything.
> >
> > By the way, suspending to disk ("hibernate") is much easier, and there's built-in support for it in most kernels. That worked fine with my RT kernel, IIRC, it was only the suspend-to-RAM that gave me fits until I turned off hyperthreading.
> >   
> I have 3Gb of memory and hibernating takes a year and a day, much longer
> than it would take to shut down, reboot, and open all my applications
> again, so until hibernation to a compressed image becomes available
> (would it not be possible for instance to zero the unused memory and
> then pipe the used and unused memory contents through a compressor?), it
> isn't an option unfortunately.
>

Good point. You just reminded me why I blew off hibernate and instead I just use suspend-to-RAM.

> As far as I can remember, I have hyperthreading enabled in the bios, but
> it doesn't seem to be giving me any problems.
>

You're lucky. My EEE doesn't like it.

> So basically, as I said before, apart from the hibernate issue, which
> can be worked around easily, the latest rt kernel seems to be working
> like a charm. The main challenge now, will be to make jack2 survive
> through a suspend-resume cycle, which is something I never managed to do
> in the past.
> 

I doubt that would be possible.

I run a script that kills jackd and every other music app before suspending.

-ken



More information about the Linux-audio-user mailing list