> J?rgen Keil wrote:
> >
> > There are two SIGSEGVs, the first one is in the nvidia module.
> >
> > Xorg tries to catch it and tries to abort the server,
> > but on the way down it gets another one, somewhere
> > in the keyboard driver. I think it tries to log the message
> > "%s: Failed to set keyboard LED's: %s\n", but one of the
> > arguments for %s must be a NULL pointer. That results
> > in a second SIGSEGV, inside a strlen(NULL) call.
> >
> > The first SIGSEGV inside the nvidia_drv.so lib should
> > be more interesting.
> >
> >
> Sorry, I missed that. I just tried this driver on a Tecra M5 and S/R works.
> However, the command I run in testing is "uadmin 3 20" which I believe
> is suspend/power-off. I'm not certain suspend to ram (uadmin 3 22) is
> tested.
uadmin 3 20: it the real S3 suspend-to-ram, where
cpu and devices are powered off; system state is saved
in ram which is still powered on
uadmin 3 22: is the same as "uadmin 3 20" but skips the
step of powering off cpu and devices; system immediately
resumes, That is, all it does is it tests driver suspend
and resume code.
> Does "uadmin 3 20" work on your system?
Nope, not yet. It does seem to reach S3 suspended
state, but is unable to resume from this state. One
box seems to hang early during wakeup from S3 suspended
state, the other one automatically powered off 2 seconds
after I start a resume be pressing the power button.
On both boxes I have to pull the power plug to get the
system back into a bootable state.
--
This message posted from opensolaris.org
