Bug 83550

Summary: [NVA5] Monitor not restored after resume [regression] [bisected]
Product: xorg Reporter: Hjulle <anka.213.bacn+freedesktop>
Component: Driver/nouveauAssignee: Nouveau Project <nouveau>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: odi
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
i915 platform: i915 features:
Description Flags
dmesg log
log from last good commit
log from first bad commit
log of last good commit
log of first bad commit
log of first bad commit none

Description Hjulle 2014-09-06 02:07:33 UTC
Created attachment 105823 [details]

I have a GT216 [GeForce 210] (rev a2). I have two monitors, one connected with DVI, the other with VGA. In revisions after 415f12efc1b only the one connected to VGA will restore after suspend.

I need to apply
[ba5e01b0d0146c2d2bb5475cf1d6efec51bf2669] drm/nv50/disp: fix a potential oops in supervisor handling
to make it restore at all in 415f12efc1b..ba5e01b0d^. Otherwise the computer completely fails to resume. (The computer wakes up but is completely unresponsive. Both monitors are off. Pressing the keyboard makes the num-lock key blink once and nothing more. This is the same behaviour it has when the computer is suspended. The only difference is that the power light is not blinking.)

xrandr seems to think that the screen was restored properly.

> lspci
01:00.0 VGA compatible controller: NVIDIA Corporation GT216 [GeForce 210] (rev a2)

> git bisect log
# first bad commit: [415f12efc1b2308411b2cbc3e82666b3db8a7758] drm/nv50/disp: start removing direct vbios parsing from supervisor
Comment 1 Hjulle 2014-09-06 02:39:14 UTC
Created attachment 105824 [details]
dmesg log
Comment 2 Hjulle 2014-09-06 02:45:15 UTC
Here is the output of xrandr after the failed resume. DVI-I-1 is not restored.

> xrandr
Screen 0: minimum 320 x 200, current 2560 x 1024, maximum 8192 x 8192
DVI-I-1 connected primary 1280x1024+0+0 (normal left inverted right x axis y axis) 376mm x 301mm
   1280x1024     60.02*+  75.02  
   1280x960      60.00  
   1152x864      75.00  
   1024x768      75.08    70.07    60.00  
   832x624       74.55  
   800x600       72.19    75.00    60.32    56.25  
   640x480       75.00    72.81    66.67    60.00  
   720x400       70.08  
VGA-1 connected 1280x1024+1280+0 (normal left inverted right x axis y axis) 337mm x 270mm
   1280x1024     60.02*+  75.02  
   1152x864      75.00  
   1024x768      75.08    70.07    60.00  
   832x624       74.55  
   800x600       72.19    75.00    60.32  
   640x480       75.00    72.81    60.00  
   720x400       70.08  
HDMI-1 disconnected (normal left inverted right x axis y axis)
Comment 3 Adam Williamson 2014-09-15 16:16:03 UTC
https://bugs.freedesktop.org/show_bug.cgi?id=80506 could very well be a case of this.
Comment 4 poma 2014-09-15 16:49:01 UTC
Created attachment 106333 [details]

cat /sys/kernel/debug/dri/0/vbios.rom > vbios.rom
Comment 5 Ortwin Glück 2014-09-19 16:50:09 UTC
Created attachment 106559 [details]
log from last good commit
Comment 6 Ortwin Glück 2014-09-19 16:51:30 UTC
Created attachment 106560 [details]
log from first bad commit

This is the debug output of a boot/suspend/resume cycle of first bad commit 415f12efc1b2308411b2cbc3e82666b3db8a7758 with nouveau.debug=trace,PDISP=spam
Comment 7 Ortwin Glück 2014-09-22 09:17:52 UTC
Created attachment 106662 [details]
log of last good commit
Comment 8 Ortwin Glück 2014-09-22 09:19:36 UTC
Created attachment 106663 [details]
log of first bad commit

Here the complete versions of the log files. gzip'ed due to size Bugzilla's restrictions.
Comment 9 Ortwin Glück 2014-09-22 09:20:06 UTC
Created attachment 106664 [details]
log of first bad commit
Comment 10 Ben Skeggs 2014-09-30 01:45:09 UTC
Can you give it a try with http://cgit.freedesktop.org/~darktama/nouveau/commit/?id=a68e95365a4248d3f58c8663e4513c79273ac704 applied?
Comment 11 poma 2014-09-30 22:17:36 UTC
3.17.0-0.rc7.git0.2.fc21.x86_64  PASSED
  patched with 
  "disp/nv50: fix dpms regression on certain boards"
Comment 12 poma 2014-10-01 23:16:30 UTC
Tested with:
Chipset: G98 (NV98)
Chipset: MCP79/MCP7A (NVAC)
Family : NV50
Comment 13 poma 2014-10-01 23:20:54 UTC
Ben, if it's late for 3.17, please push it downstream to land in Fedora 21.
Comment 14 poma 2014-10-03 18:10:03 UTC
Hjulle, this is your report so you close it when you're done with testing.
Comment 15 Ortwin Glück 2014-10-06 16:23:34 UTC
I confirm that the patch fixes the issue also for me.
Comment 16 Ortwin Glück 2014-10-06 16:25:13 UTC
(In reply to Ortwin Glück from comment #15)
> I confirm that the patch fixes the issue also for me.

Chipset: GT215 (NVA3)
Family : NV50
Comment 17 Ilia Mirkin 2015-10-22 04:26:39 UTC
Later testers indicate that the issue is fixed. OP hasn't replied, marking fixed anyways.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.