#673669 cdrom: CRT Monitor connected DVI>VGA not detected properly - unusable - EDID not properly obtained. #673669
- Package:
- xserver-xorg
- Source:
- xorg
- Description:
- X.Org X server
- Submitter:
- doobydave
- Date:
- 2017-09-18 08:06:06 UTC
- Severity:
- important
Package: cdrom Severity: critical Justification: breaks the whole system Hi. As per title, my monitor's EDID is not read correctly, and this leads to very bad things on many linux distros. The debian live cd presents a flashing grey screen instead of the desktop. Text login works (ctrl+alt+f1), but is not displayed. Trying to launch in failsafe mode is even worse. The machine hangs with flashing scroll-lock and caps-lock lights. It's worth noting that WinXP on the same machine can use the monitor just fine. This bug is not specific to Debian (Fedora, Ubuntu and Sabayon all have certain issues) I have filed a bug with UbuntuBugs, and I have a thread at Ubuntuforums about the issue. https://bugs.launchpad.net/ubuntu/+source/read-edid/+bug/991068 ubuntuforums.org/showthread.php?t=1981935 If you'd like specific debian generated logs (rather than the Ubuntu ones available in the above links, I can get ssh installed on a live CD boot and get to them remotely (submitting this report from there now), or if you need me to install it to the hard disk (assuming it's possible), I'm happy to do so.
severity 673669 important reassign 673669 xserver-xorg thanks Reassigning to the X people, as they will be able to help more. Also reducing the inflated severity to something more reasonable. While this is undoubtedly important to you, many many other people with more mainstream hardware are using the same software are not seeing these problems.
Steve McIntyre <steve@einval.com> (20/05/2012): From the affected system, getting more info about kernel/X would be nice. You can gather that from a single script: /usr/share/bug/xserver-xorg-core/script 3>/tmp/script.log and send it to this bug report (673669@bugs.debian.org). I've seen this numerous times with bad cables, and depending on the OS, that might work out magically on some of them. You may want to double check this point. Mraw, KiBi.
xrandr: Failed to get size of gamma for output default Screen 0: minimum 320 x 240, current 1280 x 800, maximum 1280 x 800 default connected 1280x800+0+0 0mm x 0mm 1280x800 50.0* 1024x768 51.0 960x540 52.0 840x525 53.0 800x600 54.0 800x512 55.0 720x450 56.0 680x384 57.0 58.0 640x512 59.0 640x480 60.0 61.0 576x432 62.0 512x384 63.0 400x300 64.0 320x240 65.0
DAVID HAND <dooby.dave@virgin.net> (20/05/2012): Yeah (AFAICT) xrandr only knows about what X knows, which takes it from the kernel. You could try booting with something like video=$resolution to force something. That should get you a nice console. Then you can play with adding a new modeline through xrandr, then selecting it for the appropriate output. Quite a workaround, but that might help you get a working system for better use until the actual issue (if any) is found & solved. Mraw, KiBi.
Forgot to say I have a laptop with VGA out - with an nvidia go 7600 with which I could test the monitor bypassing the DVI>VGA adapter, though my hunch is that the adapter is not the problem. Not only does windows work with the same hardware configuration, but I remember having resolution issues since 2006, when the monitor was being used with a different machine using an ATI video card.
I booted my laptop from the live CD with the CRT plugged in directly, and while things are better, I'm not sure if everything is OK. I believe the nouveau driver is working, but the available resolution and refresh rates are incomplete. Looking through Xorg.0.log there are several bunches of modelines - not one group has all the correct resolutions, and many appear to be set at 0.0Hz. Also, the hex EDID seems to have many 00 in it - could be fine, just looks strange. UPDATE Just been looking at the EDID extracted from windows (using Phoenix I think). It is different. I will include the two files extracted *.dat *.raw When extracting using this software, it gave a waning about converting data from one version to another. Maybe this is where the differences came from.
(via ubuntu 10.04) is the same as in Xorg.0.log (from debian live cd on laptop) - yet gives accurate resolution choices. I'm really confused.
Would love to get to the bottom of this. Do you need any more information?
Could be kernel bug too, similar? issue was fixed recently in 3.2.x kernel (in Debian wheezy/sid). But it was in radeon driver. You could try 3.2.x kernel from squeeze-backports (for example) and see if it works there. Maybe it makes sense to file bug for kernel package and attach full dmesg from kernel boot with that report.
I'm honestly way out of my depth here. Does the bug need to be re-submitted? Cannot some of you notify the relevant people, or alter the package that it refers to?
Good day, I have an urgent and profitable business proposal for you. Kindly confirm if you received my previous mail sent. Regards, Ronney Botha.