Bug 43239

Summary:	Bad wraparound of console text after returning from graphics mode
Product:	Drivers	Reporter:	Alex (alupu01)
Component:	Console/Framebuffers	Assignee:	Alan (alan)
Status:	RESOLVED CODE_FIX
Severity:	normal	CC:	alan, jani.nikula
Priority:	P1
Hardware:	All
OS:	Linux
URL:	bugs.freedesktop.org/show_bug.cgi?id=43167
Kernel Version:	Linux 3.3.4	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	stty log

Description Alex 2012-05-12 16:18:23 UTC

(B)LFS system, i686-pc-linux-gnu, kernel-3.3.4
Xorg-7.6+ (server-1.12.1, video-intel-2.19.0)
ASUS P5E-VM HDMI, Intel E8400, 3.0 GHz, 4 GB, G35/ICH9R
Monitor (the "console"):
 Samsung SM2494, Wide (16:9), DVI input (PC HDMI->DVI cable)

-------------------------------------------------------------------------------
DEFINITIONS:
"Console Text wraparound NORMAL":
 When line length (prompt + text) reaches 80 characters
 the next character typed starts at column 1, _next_ row.

"Console Text BAD wraparound (this bug)":
 When line length (prompt + text) reaches 80 characters
 the next character typed starts at column 1, SAME row.
 The following characters typed overwrite the existing line forever.
-------------------------------------------------------------------------------
BUG, reproducible:

1. Boot to level 3 (console).  Text wraparound NORMAL.
2. Go to X11 (level 5).
3. Return to console text mode (level 3).  Text wraparound BAD (i.e., the BUG)

Work around:
After step 3. above, type the command,
'stty columns 80'
Text wraparound returns to normal.
-------------------------------------------------------------------------------
Technical details, screen snapshots, etc. and analysis can be found at

Bug 43167 "X intel driver causes wrong wraparound of console command line"
<bugs.freedesktop.org/show_bug.cgi?id=43167>

The last "official" kernel release not exhibiting the bug is 3.0.20
(in numbering sequence, NOT in time sequence).
First BAD wraparound: v3.1 (again, in numbering sequence).

A bisection of the kernel releases shows

commit 04fee895ef98ffbb91a941b53a92d6949bb6d1c4
Author: Rolf Eike Beer
Date:   Wed Jun 15 11:27:02 2011 +0200
    DRM: clean up and document parsing of video= parameter

as the point where the bug first shows up (and persists up to this writing).

Please note that "intel driver" in the bug 43167 title turned out to be wrong.
The analysts consensus seems to indicate a problem in the "vt layer" subsystem of the kernel.

Thanks,
-- Alex

Comment 1 Alan 2012-05-13 14:52:34 UTC

Please do the following 

stty -g >log
stty -a >>log

trigger the problem

stty -g >>log
stty -a >>log

and attach that to this bug. It'll show the tty state before and after the event.

Comment 2 Alex 2012-05-13 17:10:06 UTC

Created attachment 73281 [details]
stty log

Comment 3 Jani Nikula 2012-12-27 17:04:01 UTC

(In reply to comment #0)
> Technical details, screen snapshots, etc. and analysis can be found at

URL for convenience: http://bugs.freedesktop.org/show_bug.cgi?id=43167

(Please note that bug N links don't work across kernel.org and freedesktop.org bugzillas; moreover the bug numbers are independent.)

Comment 4 Jani Nikula 2012-12-27 17:08:39 UTC

(In reply to comment #2)
> Created an attachment (id=73281) [details]
> stty log

diff:

- speed 38400 baud; rows 25; columns 80; line = 0;
+ speed 38400 baud; rows 25; columns 82; line = 0;

Comment 5 Jani Nikula 2012-12-31 08:05:47 UTC

Alex, in light of [1], is this still an issue?

[1] https://bugs.freedesktop.org/show_bug.cgi?id=43167#c43

Comment 6 Alex 2012-12-31 16:08:02 UTC

(In reply to comment #5)
> Alex, in light of [1], is this still an issue?
> 
> [1] https://bugs.freedesktop.org/show_bug.cgi?id=43167#c43

Hi Jani,

No longer an issue.  This submission (43239) is actually a "spin-off" of the above (43167) submission, i.e., practically the same issue.

Thanks again and einen guten Rutsch ins neue Jahr!
-- Alex