Bug 23740

Summary: blanks the screen in graphics mode "forever" at random moments on intel (i915 driver)
Product: Branch 5.1 Reporter: Ivan Zakharyaschev <imz>
Component: kernel-modules-drm-std-defAssignee: Michail Yakushin <silicium>
Status: NEW --- QA Contact: QA 5.1 <qa-5.1>
Severity: normal    
Priority: P3    
Version: unspecified   
Hardware: all   
OS: Linux   
Bug Depends on:    
Bug Blocks: 23850, 23739    

Description Ivan Zakharyaschev 2010-07-07 20:26:53 MSD
The following problem can also be considered as a problem of the kernel: because even after restarting X, the screen stays blank. So, this means, the kernel doesn't provide a clean state for the new, independent X process. The kernel could take care of providing a clean graphics state for new Xs.

+++ This bug was initially created as a clone of Bug #23739 +++

Created an attachment (id=4453)
/var/log/Xorg.0.log.old when this happens

# uname -a
Linux localhost.localdomain 2.6.30-std-def-alt15 #1 SMP Mon Dec 14 08:45:48 UTC 2009 i686 GNU/Linux
# lspci | fgrep -i vga
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)

At random moments (often when opening a new page in chromium-browser-6.0.401.0-alt1.r47010), the screen is blanked (more precisely, filled with one color: black, or the blue color of the Chromium's window). One can successfully switch to a linux text console. But this state of the graphics screen stays "forever" until a reboot: restarting X doesn't help.

This happens rather seldom; and is not deterministically reproducible.
Comment 1 Ivan Zakharyaschev 2010-08-05 10:53:19 MSD
(Originally posted to https://bugzilla.altlinux.org/show_bug.cgi?id=23739#c6 , but htis comments concerns 5.1 as well.)

This also happened on 2.6.32-un-def-alt5 (from 5.1) and that
xorg-drv-intel-2.4.3-alt0.M41.1 (when working in rxvt-unicode-9.02-alt1).
Comment 3 Ivan Zakharyaschev 2010-08-05 10:59:16 MSD
(In reply to comment #1)

> This also happened on 2.6.32-un-def-alt5 (from 5.1) and that

Comment 6 Ivan Zakharyaschev 2010-08-09 15:52:10 MSD
(In reply to comment #5)

> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=944001201ca0196bcdb088129e5866a9f379d08c
> .

Also, <q>Intel's Keith Packard provided a patch documenting some of the register used by Arlie's new code</q>: 

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=45503ded966c98e604c9667c0b458d40666b9ef3 .