Bug 15662 - X via startx exits to black screen
Summary: X via startx exits to black screen
Status: RESOLVED WORKSFORME
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: i586 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Chris Denice
QA Contact:
URL: https://bugs.freedesktop.org/show_bug...
Whiteboard: MGA5TOO IN_ERRATA
Keywords: NEEDINFO, UPSTREAM
Depends on:
Blocks:
 
Reported: 2015-04-09 13:22 CEST by Felix Miata
Modified: 2019-03-19 05:42 CET (History)
3 users (show)

See Also:
Source RPM: xinit-1.3.4-2.mga5
CVE:
Status comment:


Attachments
lspcidrake -v from 2 rv200 systems and 1 nv11 system exiting startx to all black (8.72 KB, text/plain)
2015-05-21 12:31 CEST, Felix Miata
Details

Description Felix Miata 2015-04-09 13:22:38 CEST
Same thing reported long ago  on Fedora has not only happening in Cauldron's kernel-desktop-3.19.3 (and priors) for a long time but also just began in Tumbleweed. with 3.19.x,

Steps to Reproduce:
1.login on any vtty
2.startx to KDE4
3.log out of KDE4

Actual results
1-screen is either black or virtual black with very faint grey-white text.
2-changing to any other vtty and then back refreshes to expected screen coloring.
3-text is sprawled across the screen in a columns and width layout not matching the actual columns and rows available (line feeds in inappropriate places).

Expected results:
1-normal complement of Xorg exit messages and shell prompt in visible colors immediately on leaving KDE.
2-line feeds where they belong

mailing list thread(s):
https://ml.mageia.org/l/arc/dev/2015-02/msg00384.html
https://ml.mageia.org/l/arc/dev/2014-09/msg00218.html


Reproducible: 

Steps to Reproduce:
Comment 1 Felix Miata 2015-05-21 04:50:25 CEST
Problem continues with kernel 3.19.8 and X server 1.16.4.
Comment 2 Samuel Verschelde 2015-05-21 09:10:25 CEST
Can you also give the links to similar bug reports in other distributions?
Comment 3 Samuel Verschelde 2015-05-21 09:11:01 CEST
Also, does it happen on specific hardware (if yes please attach output of lspcidrake -v) or any?

Keywords: (none) => NEEDINFO

Comment 4 Felix Miata 2015-05-21 12:31:24 CEST
Created attachment 6599 [details]
lspcidrake -v from 2 rv200 systems and 1 nv11 system exiting startx to all black

Only other bug I filed in any tracker:
https://bugzilla.redhat.com/show_bug.cgi?id=1096487

Mailing lists have been seriously unhelpful:
http://lists.x.org/archives/xorg/2015-March/057252.html
http://lists.x.org/archives/xorg/2015-April/057314.html
http://lists.freedesktop.org/archives/systemd-devel/2015-April/030752.html
http://lists.freedesktop.org/archives/systemd-devel/2015-April/030752.html
http://www.spinics.net/lists/linux-console/msg00212.html

I've been able to reproduce this on NV11 again, by installing an nv11 gfxcard in host gx270, with which Cauldron produces this problem using rv200.

For sure this must be an upstream problem, but which upstream? Kernel? Xorg? Driver? Other?
Comment 5 Felix Miata 2015-05-22 04:44:26 CEST
Upstream bug filed: https://bugs.freedesktop.org/show_bug.cgi?id=90572

Keywords: NEEDINFO => (none)
URL: (none) => https://bugs.freedesktop.org/show_bug.cgi?id=90572

Samuel Verschelde 2015-05-22 10:04:43 CEST

Keywords: (none) => UPSTREAM

Samuel Verschelde 2015-05-22 10:05:16 CEST

Whiteboard: (none) => MGA5TOO FOR_ERRATA

Comment 6 Felix Miata 2015-05-22 10:21:19 CEST
Summary changed because this is not dependent on use of KDE as DM.

Summary: KDE run via startx exits to black screen => X via startx exits to black screen

Comment 7 Shlomi Fish 2015-06-18 10:52:16 CEST
Placed in the errata.

CC: (none) => shlomif
Whiteboard: MGA5TOO FOR_ERRATA => MGA5TOO IN_ERRATA

Comment 8 Marja Van Waes 2016-10-29 18:24:41 CEST
Is this bug still valid? (Yeah, I know nothing happened in the upstream bug report, but bug reports are sometimes forgotten or missed).

Assigning to the registered xinit maintainer, but CC'ing the kernel and drivers maintainers.

Keywords: (none) => NEEDINFO
CC: (none) => kernel, marja11
Assignee: bugsquad => eatdirt

Comment 9 Chris Denice 2017-06-12 15:47:15 CEST
I had a look, and I cannot reproduce so that's going to be difficult to fix.

That looks like a driver issues. A few ideas if that might help:

-Did you try with VESA?

-Try to remove you vga=791 at boot

-Do you have radeon-firmware installed?

Good luck :-/
Comment 10 Felix Miata 2017-06-20 09:13:50 CEST
(In reply to Chris Denice from comment #9)
> -Did you try with VESA?

VESA does not support native widescreen modes. FBDEV doesn't either.

'urpme x11-driver-video-ati' switches Xorg to using the modesetting driver, which does not help.

'urpme x11-driver-video-ati x11-driver-video-modesetting' causes Xorg to fail to start, trying to use the fbdev driver. Additionally removing the fbdev driver doesn't work either.
 
> -Try to remove you vga=791 at boot

Removing it does not help.

> -Do you have radeon-firmware installed?

Yes.

Upstream bug has been updated to show problem remains also in latest Fedora and openSUSE.
Comment 11 Felix Miata 2019-03-19 05:42:44 CET
This seems to have been fixed by last September in MGA6 (host gx27b running Trinity Desktop),
and remains so now with kernel-desktop-4.18.20.

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


Note You need to log in before you can comment on or make changes to this bug.