freebsd stable - May 2012 - STABLE/9 SMP ACPI suspend/resume - video mode not being restored

Hi,

I'm toying with the SMP/i386 ACPI suspend/resume patches in -9. Thanks
so much for this!

I've noticed though that the video backlight stays off after resume. A
common problem on -9, so I set hw.acpi.reset_video=1. That restores
the backlight.

However, the video mode isn't restored. I have my console set to
VGA_80x60 and the resume seems to set it up "wrong". I get half or so
of each line displayed.

A vidcontrol VGA_80x60 restores things to proper working order.

Is there a shortcoming somewhere in syscons/ACPI video restore on -9
that doesn't properly restore the configured mode?

Thanks again for all your hard work! Now that you've done that, I'll
go off and work on fixing up ath(4) suspend/resume for PCI devices. :)


Adrian

On Thu, May 24, 2012 at 08:46:03PM -0700, Adrian Chadd
wrote:
> Hi,
> 
> I'm toying with the SMP/i386 ACPI suspend/resume patches in -9. Thanks
> so much for this!
Note that enough of those patches have been committed (at least as of
r235891) that I was able to perform suspend/resume properly on my laptop
(a Dell Precision M4400; video is "NVIDIA GPU Quadro FX 770M (G96GL)",
and I use the nVidia driver (ports/x11/nvidia-driver)) without adding
more patches.
> I've noticed though that the video backlight stays off after resume. A
> common problem on -9, so I set hw.acpi.reset_video=1. That restores
> the backlight.
I do not see that behavior, and:

g1-227(8.3-S)[7] sysctl hw.acpi.reset_video
hw.acpi.reset_video: 0

(Yes, I'm running stable/8 on the present slice.  I have stable/9 on
another slice.  And the experiment I did was with a slice where I had
built stable/9 (i386) using clang.)
> However, the video mode isn't restored. I have my console set to
> VGA_80x60 and the resume seems to set it up "wrong". I get half
or so
> of each line displayed.
That is another issue that I have not observed (in my case).
> ...
> Is there a shortcoming somewhere in syscons/ACPI video restore on -9
> that doesn't properly restore the configured mode?
I believe that my experience is evidence that if such a shortciming
exists, it is not a general one.

For me, suspend/resume in stable/9 Just Works (thanks to the hard work
of others (such as iwasaki@), of course).
> Thanks again for all your hard work! Now that you've done that,
I'll
> go off and work on fixing up ath(4) suspend/resume for PCI devices. :)
Cool!  :-)

(Adrian, next BAFUG, perhaps we could compare notes/behaviors in
person, if that might be of use?)

Peace,
david
-- 
David H. Wolfskill				david@catwhisker.org
Depriving a girl or boy of an opportunity for education is evil.

See http://www.catwhisker.org/~david/publickey.gpg for my public key.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
Url :
http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20120525/9a5a562d/attachment.pgp

Hi, thanks for reporting!
> However, the video mode isn't restored. I have my console set to
> VGA_80x60 and the resume seems to set it up "wrong". I get half
or so
> of each line displayed.
> 
> A vidcontrol VGA_80x60 restores things to proper working order.
> 
> Is there a shortcoming somewhere in syscons/ACPI video restore on -9
> that doesn't properly restore the configured mode?
Do you have vesa(4) in your kernel?  It seems dev/fv/vesa.c:vesa_bios_post()
restore the mode when resuming, but it's maybe incomplete in some cases...
I think great work was done in this area, and we can improve this more.

How about switching vty to other different mode vty and switching back
in order to force changing video mode?
I think it's better than re-run vidcontrol.
> Thanks again for all your hard work! Now that you've done that,
I'll
> go off and work on fixing up ath(4) suspend/resume for PCI devices. :)
This is my pleasure :)

Thanks!