Re: Blank screen on resume

[Eric McCorkle]

I think I will give that a shot, actually...  It seems pretty unfortunate that 
we'd have to wait for full 3D graphics support on a card in order to get resume 
working.

I can't guarantee success, of course. 

On March 14, 2016 4:49:51 PM EDT, Adrian Chadd  wrote:
>On 14 March 2016 at 13:22, Eric McCorkle  wrote:
>> So no options at all until then? :(
>
>If you want to figure out what minimal set you need from the
>dri/i915kms code to restore backlight happiness, please do. :-P
>
>But yeah. We'll need to wait for the next dri sync
>
>
>-adrian
>
>>
>> On March 14, 2016 3:43:08 PM EDT, Adrian Chadd
>
>> wrote:
>>>
>>> Hiya,
>>>
>>> Yeah - the drm2 code needs to be involved in resuming, and right now
>>> there's no broadwell support in i915kms.
>>>
>>> Once that gets updated then yeah, the resume backlight will work.
>:-P
>>>
>>>
>>>
>>> -a
>>
>>
>> --
>> Sent from my Android device with K-9 Mail. Please excuse my brevity.

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

Re: Blank screen on resume

[Adrian Chadd]

On 14 March 2016 at 13:22, Eric McCorkle  wrote:
> So no options at all until then? :(

If you want to figure out what minimal set you need from the
dri/i915kms code to restore backlight happiness, please do. :-P

But yeah. We'll need to wait for the next dri sync


-adrian

>
> On March 14, 2016 3:43:08 PM EDT, Adrian Chadd 
> wrote:
>>
>> Hiya,
>>
>> Yeah - the drm2 code needs to be involved in resuming, and right now
>> there's no broadwell support in i915kms.
>>
>> Once that gets updated then yeah, the resume backlight will work. :-P
>>
>>
>>
>> -a
>
>
> --
> Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

Re: Blank screen on resume

[Eric McCorkle]

So no options at all until then? :(

On March 14, 2016 3:43:08 PM EDT, Adrian Chadd  wrote:
>Hiya,
>
>Yeah - the drm2 code needs to be involved in resuming, and right now
>there's no broadwell support in i915kms.
>
>Once that gets updated then yeah, the resume backlight will work. :-P
>
>
>
>-a

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

Re: Blank screen on resume

[Eric McCorkle]

Note: I've moved the patch and logs to a new branch: 
https://github.com/emc2/freebsd/tree/resume_blank_screen


On March 12, 2016 10:39:42 PM EST, Eric McCorkle  wrote:
>Hello everyone,
>
>It seems I'm plagued by blank screens on resume again on a different
>laptop... *sigh*
>
>I've done quite a bit of diagnostic work on this, and there seem to be
>some clues, but I'm a bit stumped at this point.  The following github
>repo has patches that add extra logging, which has proven useful in my
>attempts to diagnose the issue: https://github.com/emc2/freebsd (sorry
>for distributing patches/logs in this way, but the logs are quite big).
>Also in this repo, there's a pciconf, devinfo, and dmesg file.  The
>dmesg file shows the trace from booting a patched kernel in verbose
>mode then suspending it.
>
>There's two main curiosities in the dmesg trace.  The first is that the
>power state of all the vgapci device seems to be 0 as opposed to 3
>during resume, which was causing acpi_set_powerstate_method to skip
>resetting its power during resume.  I tried a simple patch (you can see
>it in the patch and its effects in the trace), but that didn't seem to
>work.  A second curiosity is that acpi_pwr_switch_consumer only seems
>to have an effect during resume.
>
>The only other real lead I have is that the display doesn't seem to be
>active in acpi_video.  I'm not sure if this is an artifact of the efifb
>driver.
>
>Lastly, I recall looking into vga bios calls from earlier such efforts.
>I imagine there's something similar for efifb, but I've been unable to
>track down where this is happening.  Advice or information on that
>front would be much appreciated.
>
>Best,
>Eric
>___
>freebsd-acpi@freebsd.org mailing list
>https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
>To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

Re: Blank screen on resume

[Eric McCorkle]

Laptop is the librem 13 (not helpful, I know). 

Graphics chipset is broadwell.  Cpu is i5-5200.

Haven't been able to get the motherboard model. However, I have model numbers 
from the bios. 

Bios is American Megatrends.

BIOS Version is 108.
EC version is 2.1u. 
M/B version is D. 

On March 12, 2016 11:45:00 PM EST, Adrian Chadd  wrote:
>Hi,
>
>Which laptop? Which chipset? etc.
>
>
>
>-adrian

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"

Blank screen on resume

[Eric McCorkle]

Hello everyone,

It seems I'm plagued by blank screens on resume again on a different laptop... 
*sigh*

I've done quite a bit of diagnostic work on this, and there seem to be some 
clues, but I'm a bit stumped at this point.  The following github repo has 
patches that add extra logging, which has proven useful in my attempts to 
diagnose the issue: https://github.com/emc2/freebsd (sorry for distributing 
patches/logs in this way, but the logs are quite big).  Also in this repo, 
there's a pciconf, devinfo, and dmesg file.  The dmesg file shows the trace 
from booting a patched kernel in verbose mode then suspending it.

There's two main curiosities in the dmesg trace.  The first is that the power 
state of all the vgapci device seems to be 0 as opposed to 3 during resume, 
which was causing acpi_set_powerstate_method to skip resetting its power during 
resume.  I tried a simple patch (you can see it in the patch and its effects in 
the trace), but that didn't seem to work.  A second curiosity is that 
acpi_pwr_switch_consumer only seems to have an effect during resume.

The only other real lead I have is that the display doesn't seem to be active 
in acpi_video.  I'm not sure if this is an artifact of the efifb driver.

Lastly, I recall looking into vga bios calls from earlier such efforts.  I 
imagine there's something similar for efifb, but I've been unable to track down 
where this is happening.  Advice or information on that front would be much 
appreciated.

Best,
Eric
___
freebsd-acpi@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-acpi
To unsubscribe, send any mail to "freebsd-acpi-unsubscr...@freebsd.org"