Re: Suspend to RAM on a Thinkpad X41

[Jukka Ruohonen <jruohonen%iki.fi@localhost>]

On Wed, Jul 13, 2011 at 08:09:32PM +0200, Moritz Wilhelmy wrote:
> On Wed, Jul 13, 2011 at 08:16:47 +0300, Jukka Ruohonen wrote:
> > No one really is. All suspend/resume problems are generally quite difficult
> > to debug. Often even things like serial console are useless.
> 
> Out of curiosity, how are these things being fixed, then?

As this appears to be a regression, a closer time frame always helps. But
generally, when you have the hardware, it can be simple if you know the
code. But even then, often a hit and miss; things like X.org and video
drivers may be involved (let alone BIOS/ACPI...).[1]

The very first thing is to boot with the -x option (see boothowto(8)) and
see whether it panics or hangs during the pmf(9) suspend hooks. Again
generally, the greatest difficulty is in formulating a hypothesis or even
just a hint. After that it is easy.

> I don't mind learning something new, and possibly contributing to the
> project, even if it takes some time.

There are millions of lines of code in NetBSD. Thus, there is always
something to fix and improve -- a good attitude to get started.

> > Can you post the dmesg from -current?
> 
> Yes, see the MIME attachment to this mail.
> Unfortunately the 5.1-kernel does not boot anymore

A hardware regression in the stable branches is definitely a no-no, so
please send-pr(1) for this problem separately.

- Jukka.

[1] NetBSD lacks some fancy tools available in Linux that can store
    suspend/resume traces to non-volatile memory.