Bug 97857

Summary: card detects non-existent monitor on display port
Product: DRI Reporter: Daniel <danadam7>
Component: DRM/RadeonAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg, kernel 4.5.0-1, debug
none
xorg, kernel 4.5.0-1
none
dmesg, kernel 4.5.0-2, debug
none
xorg, kernel 4.5.0-2 none

Description Daniel 2016-09-19 00:41:16 UTC
My graphic card is:

01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Caicos [Radeon HD 6450/7450/8450 / R5 230 OEM]

I have one monitor connected to DVI port. Since kernel 4.5.0-2 my Xorg.0.log shows that something is also connected to display port, here's a diff of the log between kernel 4.5.0-1 and 4.5.0-2:

 (II) RADEON(0): EDID for output DisplayPort-0
+(II) RADEON(0): Printing probed modes for output DisplayPort-0
+(II) RADEON(0): Modeline "1024x768"x60.0   65.00  1024 1048 1184 1344  768 771 777 806 -hsync -vsync (48.4 kHz e)
+(II) RADEON(0): Modeline "800x600"x60.3   40.00  800 840 968 1056  600 601 605 628 +hsync +vsync (37.9 kHz e)
+(II) RADEON(0): Modeline "800x600"x56.2   36.00  800 824 896 1024  600 601 603 625 +hsync +vsync (35.2 kHz e)
+(II) RADEON(0): Modeline "848x480"x60.0   33.75  848 864 976 1088  480 486 494 517 +hsync +vsync (31.0 kHz e)
+(II) RADEON(0): Modeline "640x480"x59.9   25.18  640 656 752 800  480 490 492 525 -hsync -vsync (31.5 kHz e)
 (II) RADEON(0): EDID for output HDMI-0
 (II) RADEON(0): EDID for output DVI-0
...
-(II) RADEON(0): Output DisplayPort-0 disconnected
+(II) RADEON(0): Output DisplayPort-0 connected
 (II) RADEON(0): Output HDMI-0 disconnected
 (II) RADEON(0): Output DVI-0 connected
-(II) RADEON(0): Using exact sizes for initial modes
-(II) RADEON(0): Output DVI-0 using initial mode 1920x1200 +0+0
+(II) RADEON(0): Using fuzzy aspect match for initial modes
+(II) RADEON(0): Output DisplayPort-0 using initial mode 1024x768 +0+0
+(II) RADEON(0): Output DVI-0 using initial mode 1024x768 +0+0
 (II) RADEON(0): Using default gamma of (1.0, 1.0, 1.0) unless otherwise stated.
...
-(II) RADEON(0): Setting screen physical size to 508 x 317
+(II) RADEON(0): Setting screen physical size to 270 x 203

As a result my display manager (sddm) loads in 1024x768 resolution. It goes back to normal, i.e., 1920x1200 when I log in and KDE loads up.

When I explicitly disable DisplayPort by adding video=DP-1:d to kernel options, everything goes back to normal.


My original issue was with geometry of my console ( https://lists.debian.org/debian-user/2016/06/msg00851.html ), so the problem might be somewhere deeper than xorg.
Comment 1 Alex Deucher 2016-09-19 15:25:25 UTC
Please attach your full xorg log and dmesg output.
Comment 2 Daniel 2016-09-19 23:31:15 UTC
Created attachment 126639 [details]
dmesg, kernel 4.5.0-1, debug
Comment 3 Daniel 2016-09-19 23:32:00 UTC
Created attachment 126640 [details]
xorg, kernel 4.5.0-1
Comment 4 Daniel 2016-09-19 23:32:26 UTC
Created attachment 126641 [details]
dmesg, kernel 4.5.0-2, debug
Comment 5 Daniel 2016-09-19 23:32:49 UTC
Created attachment 126642 [details]
xorg, kernel 4.5.0-2
Comment 6 Michel Dänzer 2016-09-21 03:59:28 UTC
I'm seeing this as well (on mobile Kaveri), but I think it's more likely a kernel driver issue. E.g. I only noticed this a few months ago, but I can reproduce it with xf86-video-ati 7.6.1, which was released almost a year ago.

Can you try older kernels? If it doesn't happen with an older kernel, can you narrow it down to at least the kernel version which introduced it, or ideally bisect the Git commit which did?
Comment 7 Daniel 2016-09-21 11:20:52 UTC
> can you narrow it down to at least the kernel version which introduced it

But I did, in my description: "Since kernel 4.5.0-2 my Xorg.0.log shows that something is also connected to display port...".

Up to 4.5.0-1 everything was fine.
Comment 8 Michel Dänzer 2016-09-27 09:55:16 UTC
(In reply to Daniel from comment #7)
> But I did, in my description: "Since kernel 4.5.0-2 my Xorg.0.log shows that
> something is also connected to display port...".
> 
> Up to 4.5.0-1 everything was fine.

Oops, sorry, I somehow missed that before. :(

I think the most likely culprit is

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=ff0bd441bdfbfa09d05fdba9829a0401a46635c1

then, or maybe

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=1a738347df2ee4977459a8776fe2c62196bdcb1b

Can you try if reverting either of those fixes the problem?

Alex, any ideas?
Comment 9 Alex Deucher 2016-09-27 14:42:21 UTC
(In reply to Michel Dänzer from comment #8)
> (In reply to Daniel from comment #7)
> > But I did, in my description: "Since kernel 4.5.0-2 my Xorg.0.log shows that
> > something is also connected to display port...".
> > 
> > Up to 4.5.0-1 everything was fine.
> 
> Oops, sorry, I somehow missed that before. :(
> 
> I think the most likely culprit is
> 
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/
> ?id=ff0bd441bdfbfa09d05fdba9829a0401a46635c1
> 

I doubt this is the culprit.  The link training parameters aren't used until a monitor is already detected.  This wouldn't result in a fake monitor showing up.

> then, or maybe
> 
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/
> ?id=1a738347df2ee4977459a8776fe2c62196bdcb1b

That is a possibility.

> 
> Can you try if reverting either of those fixes the problem?
> 
> Alex, any ideas?

Maybe some core drm/dp change?  Any chance you could bisect?
Comment 10 Michel Dänzer 2016-09-28 02:42:56 UTC
(In reply to Alex Deucher from comment #9)
> Maybe some core drm/dp change?  Any chance you could bisect?

I can try, but probably not before next week.

FWIW, according to the attached dmesg outputs, the regression happened between upstream versions 4.5.1 and 4.5.5.
Comment 11 Michel Dänzer 2016-10-03 09:01:30 UTC
(In reply to Alex Deucher from comment #9)
> > [...] or maybe
> > 
> > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/
> > ?id=1a738347df2ee4977459a8776fe2c62196bdcb1b
> 
> That is a possibility.

Indeed, reverting just this change fixes the problem for me.

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.