Hello, I have an Intel Skull NUC, model NUC6i7KYK, with Intel video adapter 00:02.0 VGA compatible controller: Intel Corporation Iris Pro Graphics 580 (rev 09) detailed specs here: https://www.intel.it/content/www/it/it/products/docs/boards-kits/nuc/nuc-kit... It is connected to a Dell U2515H: https://www.displayspecifications.com/en/model/06291c It was initially installed with F29 (with Wayland disabled in /etc/gdm/custom.conf) and used resolution always being 2560x1440. Almost when F30 was released I upgraded and no problems until today. When powering on I see no user icons in gdm. So I switched to sddm and lxdm and there I was able to see my users login, but both trying Gnome session and Mate session it seems I don't see the top bars. While in Mate with Alt+F2 I opened a terminal window and executed xdpyinfo.
Strangely it reported a not possible resolution....
screen #0: dimensions: 5120x1440 pixels (1354x381 millimeters) resolution: 96x96 dots per inch depths (7): 24, 1, 4, 8, 15, 16, 32 root window id: 0x162 depth of root window: 24 planes
So I decided to force 2560x1440 with this file, named 00-monitor.conf and put into directory /etc/X11/xorg.conf.d https://drive.google.com/file/d/1Ml522neE-l6hsGKtGXnWHnz2cpxCsuJO/view?usp=s...
After reboot still gdm is not able to show me any user icons (perhaps it doesn't use the conf file), and again using lxdm and sddm and choosing Gnome session I don't see top bar and other things. Instead, i can use without any problem Mate and now I can see the top bar and correctly use the session. Opening xdpyinfo from a terminal I correctly get what forced: screen #0: dimensions: 2560x1440 pixels (677x381 millimeters) resolution: 96x96 dots per inch depths (7): 24, 1, 4, 8, 15, 16, 32 root window id: 0x162 depth of root window: 24 planes
Things above let me think about a problem in Gnome itself, impacting gdm too..
Yesterday evening while connected I updated as from this log, without rebooting: https://drive.google.com/file/d/10F-i8ogYAuXIouAC0o6t3DZ08yLz_q15/view?usp=s...
And I think this update generated the problem, because this evening after booting I had it. I tried to switch to a terminal and update again and I got these packages' updates: https://drive.google.com/file/d/1TUdZVlwLakIEjm8Y7ybx8m2cMhIEzlFS/view?usp=s... and reboot but still the problem is here...
Anyone experimented this?
Thanks Gianluca
I have similar problems with using my LG OLEDB6P 4K TV as a monitor. The nouveau driver seems to be confused by the EDID information. Using X it picks some resolution the monitor can't even display (invalid signal on monitor), using wayland I thought it worked, but when I examined the details it picked a strange resolution that the monitor could display, but wasn't the native 3840x2160.
When I switched to the nvidia binaries from rpmfusion it sees 3840x2160 in EDID as the best resolution and displays it no problem.
I have this feeling someone has been "improving" the EDID detection in the open source video drivers.
https://bugzilla.redhat.com/show_bug.cgi?id=1648608
https://bugzilla.redhat.com/show_bug.cgi?id=1575391
I think this started back in fedora 28, but I only recently dug into it in more detail and added info to that first bugzilla.
On 12 May 2019, at 23:47, Tom Horsley horsley1953@gmail.com wrote:
I have similar problems with using my LG OLEDB6P 4K TV as a monitor. The nouveau driver seems to be confused by the EDID information. Using X it picks some resolution the monitor can't even display (invalid signal on monitor), using wayland I thought it worked, but when I examined the details it picked a strange resolution that the monitor could display, but wasn't the native 3840x2160.
When I switched to the nvidia binaries from rpmfusion it sees 3840x2160 in EDID as the best resolution and displays it no problem.
I have this feeling someone has been "improving" the EDID detection in the open source video drivers.
https://bugzilla.redhat.com/show_bug.cgi?id=1648608
https://bugzilla.redhat.com/show_bug.cgi?id=1575391
I think this started back in fedora 28, but I only recently dug into it in more detail and added info to that first bugzilla.
There are a number of ways that I have seen EDID problems. The most popular issue the monitor's EDID is wrong. I think there are quirks tables in the kernel to fix up the well know issues (but it'd been a long time since I had to know about this stuff for work).
It used to be that you would see error reports in dmesg if the I2C bus has problems.
You can look at the details with:
$ xrandr --verboes | edid-decode
The "Detailed timing" are the monitors prefered mode I recall.
Barry
On 6/1/19 11:33 AM, Barry Scott wrote:
You can look at the details with:
$ xrandr --verboes | edid-decode
The "Detailed timing" are the monitors prefered mode I recall.
# find /sys/devices -name 'edid' /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-HDMI-A-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-VGA-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-DP-1/edid
# edid-decode < /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid Extracted contents: [...snip...] Standard timings supported: Detailed mode: Clock 69.300 MHz, 261 mm x 163 mm 1280 1306 1328 1418 hborder 0 800 802 804 814 vborder 0 -hsync -vsync VertFreq: 60 Hz, HorFreq: 48871 Hz [...snip...] Checksum: 0x2 (valid) EDID block does NOT conform to EDID 1.4! Missing name descriptor Missing monitor ranges
On 1 Jun 2019, at 19:47, Samuel Sieb samuel@sieb.net wrote:
On 6/1/19 11:33 AM, Barry Scott wrote:
You can look at the details with: $ xrandr --verboes | edid-decode The "Detailed timing" are the monitors prefered mode I recall.
# find /sys/devices -name 'edid' /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-HDMI-A-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-VGA-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-DP-1/edid
I always seem to have 0 length files in /sys for edid. That why I use the xrandr output.
# edid-decode < /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid Extracted contents: [...snip...] Standard timings supported: Detailed mode: Clock 69.300 MHz, 261 mm x 163 mm 1280 1306 1328 1418 hborder 0 800 802 804 814 vborder 0 -hsync -vsync VertFreq: 60 Hz, HorFreq: 48871 Hz [...snip...] Checksum: 0x2 (valid) EDID block does NOT conform to EDID 1.4! Missing name descriptor Missing monitor ranges
That's the problem - bad EDID.
FYI EDID Structure 1.4 recame a standard in 2006.
Barry
users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
On Sun, 2 Jun 2019 11:05:23 +0100 Barry Scott wrote:
That's the problem - bad EDID.
As near as I can tell, wayland and X treat EDID differently. Also nouveau and nvidia treat it differently.
I only get proper native resolution with nvidia and X getting EDID from my LG OLED B6P TV.
On 6/2/19 3:05 AM, Barry Scott wrote:
On 1 Jun 2019, at 19:47, Samuel Sieb samuel@sieb.net wrote: # find /sys/devices -name 'edid' /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-HDMI-A-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-VGA-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-DP-1/edid
I always seem to have 0 length files in /sys for edid. That why I use the xrandr output.
Files in /sys are not real files and they usually (always?) appear as 0 length.
# edid-decode < /sys/devices/pci0000:00/0000:00:02.0/drm/card0/card0-LVDS-1/edid Extracted contents: [...snip...] Standard timings supported: Detailed mode: Clock 69.300 MHz, 261 mm x 163 mm 1280 1306 1328 1418 hborder 0 800 802 804 814 vborder 0 -hsync -vsync VertFreq: 60 Hz, HorFreq: 48871 Hz [...snip...] Checksum: 0x2 (valid) EDID block does NOT conform to EDID 1.4! Missing name descriptor Missing monitor ranges
That's the problem - bad EDID.
I'm not the one having a problem. That's just the end of the output so I left it in.
On 2 Jun 2019, at 14:53, Tom Horsley horsley1953@gmail.com wrote:
On Sun, 2 Jun 2019 11:05:23 +0100 Barry Scott wrote:
That's the problem - bad EDID.
As near as I can tell, wayland and X treat EDID differently.
I though that the handling was the same, its in the kernel mode setting.
Also nouveau and nvidia treat it differently.
Yep coz nvidia do not use the kernel modesetting I think.
I only get proper native resolution with nvidia and X getting EDID from my LG OLED B6P TV.
I would expect the same problem with the intel driver.
Barry
On Sun, 2 Jun 2019 22:12:06 +0100 Barry wrote:
As near as I can tell, wayland and X treat EDID differently.
I though that the handling was the same, its in the kernel mode setting.
I would have thought that, but with nouveau and wayland I get visible data on the screen, but a wacky non-native resolution.
With nouveau and X, I get a blank screen because it apparently picks a setting the TV can't display at all.
Only the nvidia driver picks the right native resolution (and they all are talking to the same TV on the same hardware with, presumably, the same EDID).
-
Barry -
Barry Scott
-
Gianluca Cecchi -
Samuel Sieb -
Tom Horsley