Bug 25546

Summary:	Xorg displays a black screen with kernel > 5.2.x
Product:	Mageia	Reporter:	José Jorge <lists.jjorge>
Component:	RPM Packages	Assignee:	Thomas Backlund <tmb>
Status:	RESOLVED FIXED	QA Contact:
Severity:	critical
Priority:	Normal	CC:	jkowalzik, lists.jjorge, rolfpedersen
Version:	7
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:	kernel-5.3.2	CVE:
Status comment:
Bug Depends on:	25602
Bug Blocks:
Attachments:	dmesg of bogus boot dmesg of good boot

Description José Jorge 2019-10-10 08:19:18 CEST

Description of problem:
That's strange, this update shows an black Xorg desktop, even if the plasma desktop is launched. I hear no startup sound also, and see no errors in dmesg.
Booted two times, always the same. Went back to 5.2.16 to be able to work.


Version-Release number of selected component (if applicable):
desktop-5.3.2

How reproducible:
Always

This is a Skylake Pentium G4400 using the IGP, 2 screens 1680x1050.

Rolf Pedersen 2019-10-11 13:58:21 CEST

CC: (none) => rolfpedersen

Comment 1 José Jorge 2019-10-15 08:10:56 CEST

I have tested another system, same motherboard, only one screen, no bug.
With kernel 5.3.6, still same bug.

CC: (none) => lists.jjorge

Comment 2 José Jorge 2019-10-15 08:15:20 CEST

This looks like a gpu clock bug, if I wait ten minutes, while there is no disk activity I could see animated cursor. It is very slow, only changes the sprite every 3 seconds.

Comment 3 José Jorge 2019-10-15 08:20:05 CEST

Created attachment 11317 [details]
dmesg of bogus boot

Comment 4 José Jorge 2019-10-15 08:21:53 CEST

Created attachment 11318 [details]
dmesg of good boot

I see the drm goes much faster here, only 4/10 second in this 3 lines:

[    1.335322] [drm] Initialized i915 1.6.0 20190417 for 0000:00:02.0 on minor 0
[    1.645413] fbcon: i915drmfb (fb0) is primary device
[    1.716942] i915 0000:00:02.0: fb0: i915drmfb frame buffer device

in bad boot, it takes 2 seconds :

[    1.948457] [drm] Initialized i915 1.6.0 20190619 for 0000:00:02.0 on minor 0
[    2.248710] fbcon: i915drmfb (fb0) is primary device
[    3.908927] i915 0000:00:02.0: fb0: i915drmfb frame buffer device

José Jorge 2019-10-15 08:24:25 CEST

Summary: Xorg displays a black screen with this kernel update => Xorg displays a black screen with kernel > 5.2.x

Comment 5 Thomas Backlund 2019-10-15 09:15:36 CEST

Ah, I seem to remember something somewhere about slow gpu clocks with dual screen, but I dont remember if it was i915 or amdgpu that was about...

I need to try and find it again...

Comment 6 Thomas Backlund 2019-10-18 18:38:50 CEST

There is now a kernel-5.3.7-1.mga7 in testing, does that work any better ?

Also, there is a x11-driver-video-intel-2.99.917-57.mga7 in testing, does that make any difference ?

Comment 7 Jürgen Kowalzik 2019-10-21 21:23:34 CEST

Hi!
Kernel 5.3.6-desktop-2.mga7 makes a  ‎Intel(R) Core(TM) i5-5200U CPU @ 2.20GHz very slow.
(Avidemux makes 400 fps instead of 2500 fps or more)
Now I use 5.2.16-desktop-2.mga7 without any speed problems.

CC: (none) => jkowalzik

Comment 8 José Jorge 2019-10-22 07:52:37 CEST

(In reply to Jürgen Kowalzik from comment #7)
> Hi!
> Kernel 5.3.6-desktop-2.mga7 makes a  ‎Intel(R) Core(TM) i5-5200U CPU @
> 2.20GHz very slow.
> (Avidemux makes 400 fps instead of 2500 fps or more)
> Now I use 5.2.16-desktop-2.mga7 without any speed problems.

You must create another bug report, as this has nothing to do with a black screen. And report the '7z b' results along with 'cpupower frequency-info' for kernels 5.3.7 and 5.2.16.

Comment 9 José Jorge 2019-10-22 07:53:41 CEST

(In reply to Thomas Backlund from comment #6)
> There is now a kernel-5.3.7-1.mga7 in testing, does that work any better ?

Thanks tmb, this 5.3.7 fixes the problem. Closing, as I suppose this kernel will be validated soon?

Status: NEW => RESOLVED
Resolution: (none) => FIXED

Comment 10 Thomas Backlund 2019-10-25 21:16:29 CEST

 
kernel-5.3.7-4.mga7 is being tested / validated now in bug 25602

Depends on: (none) => 25602

Comment 11 Thomas Backlund 2019-10-29 16:09:18 CET

An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGASA-2019-0306.html