random pixels displayed for a certain character on tty

[Uwe Geuder]

This seems to be regression in kernel 2.6.27-13. The issue also exists in 2.6.27-14. (both from intrepid-proposed)

The issue does not exist in kernels 2.6.27-7 and 2.6.27-11 (from intrepid-updates or security), when booting these old kernels in my otherwise unchanged system.

(I have not yet had time to test other kernels. I guess -12 might be interesting to find out where the problem has been really introduced and -18 might be interesting whether the issue has been solved in the meantime. Please advise, which kernel should be tested, if any)

When using a tty (Ctrl-Alt-F1 ... Ctrl-Alt-F6) the character "Latin small letter o with diaresis" unicode U+00F6 or UTF-8 0xc3b6 is displayed as random pixels.

The pixel pattern seems to be different after each boot and it's always different on each tty (tty1 ... tty6). However it does not seem to change on the same tty as long as the machine is not rebooted.

I use a Finnish keyboard (see attached console-setup file) and it also supports the non-ASCII characters

- "Latin small letter a with diaresis" Unicode U+00E4 equals UTF-8 0xc3a4 and

- "Latin small letter a with ring above" Unicode U+00E5 equals UTF-8 0xc3a5

The latter 2 characters (as well as the ASCII characters) do appear always correctly.

Without a Finnish keyboard the problem can be repeated by "cat sample.txt" (attached) (if your terminal is in utf8 mode)

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 8.10
NonfreeKernelModules: ath_hal
Package: linux-image-2.6.27-14-generic 2.6.27-14.30
ProcCmdLine: root=UUID=f6bf9b65-0cad-45bd-b3a0-a7353a3900af ro quiet splash
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games
 LANG=fi_FI.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.27-14.30-generic
SourcePackage: linux

[Martin Pitt (pitti)]

Stefan, can you please take a look at this possible regression from bug 324921? Thanks!

[Stefan Bader (smb)]

I tested on two different computers. On an self assembled Desktop and a Dell Inspiron 1521. While the Desktop does not show this corruption with any kernel, the Inspiron shows similar problems with even the 2.6.27-7 kernel (the special characters are mostly invisible but sometimes get corrupted).
When I boot with the Jaunty kernel this problem seems fixed there. One thing that seems to be different is the memory map in that case.

Uwe, I could you provide the dmesg for one of the kernels that seem to work and maybe also try a live cd with the Jaunty kernel and provide the dmesg for that as well?

[Uwe Geuder (ubuntulp-ugeuder)]

Yes, I think that should be no problem. I'll have a business trip first, so please be patient for a couple of days.

[Uwe Geuder (ubuntulp-ugeuder)]

This bug is no regression occurring around kernel 2.6.27-13. It exists (at least) in Hardy, Intrepid and also in Jaunty. I'll provide more details, log and photos of the problem soon.

But there is no reason to mark the verification of proposed Intrepid kernels 2.6.27-13 and -14 as failed because of this.

[Uwe Geuder (ubuntulp-ugeuder)]

I have create a new bug https://bugs.launchpad.net/ubuntu/+source/linux/+bug/355057. It describes the problem better (I hope) , has pictures attached and the misleading suspicion about regression is no longer mentioned. Marking this old report as a duplicate.