Re: Wake from sleep with nvidia-modeset

On 29/07/2021 21:30, David Wolfskill wrote:
> On Thu, Jul 29, 2021 at 06:15:19PM +0100, Graham Perrin wrote:
>> nvidia-driver-460.84 from ports, nvidia-modeset, NVIDIA Quadro K1100M
>> (GK107GLM), HP ZBook 17 G2, KDE Plasma.
>>
>> <https://bsd-hardware.info/?probe=c7fb9e9dee> …
> … head(FreeBSD 14).

True; <https://bsd-hardware.info/?probe=c7fb9e9dee#Host>

At the time of writing:

% uname -KUv
FreeBSD 14.0-CURRENT #103 main-n248269-941650aae97: Wed Jul 28 07:28:47 
BST 2021 
root@mowa219-gjp4-8570p:/usr/obj/usr/src/amd64.amd64/sys/GENERIC-NODEBUG 
1400026 1400026
%

> … possibly-salient difference is that I had been running with
>
> kern.vty="sc"
>
> (and still do, for stable/12), but since the VBE changes earlier
> this year, I let kern.vty take the default value (which I am not
> recalling at the moment) for stable/13 and head.

Without specifying kern.vty, I get vt.

With kern.vty="sc" in /boot/loader.conf, graphics failed after 
appearance of EFI framebuffer information 
<https://photos.app.goo.gl/bxQ44dqjvoBUDT2p6>, however disk activity 
continued <https://photos.app.goo.gl/VTvjpmKuyw2ksawu9> and eventually, 
through use of the power button, I could repeatedly sleep then wake the 
computer (albeit without graphics).

(Compare with vt, where things fail hard and quickly at wake time; no 
response to the power button.)

Back to vt.

I wondered whether screen.font in loader.conf might have an impact. No 
better after commenting out the line.

For 470.57.02 there's a patch and commentary at/around 
<https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=257456#c4>. I might 
try this.

Any other suggestions?

TIA