Bug 681825 - console handling problems running mainline < 2.6.38 kernels
Summary: console handling problems running mainline < 2.6.38 kernels
Status: RESOLVED DUPLICATE of bug 642289
Alias: None
Product: openSUSE 11.4
Classification: openSUSE
Component: Basesystem (show other bugs)
Version: Final
Hardware: x86-64 openSUSE 11.4
: P5 - None : Major (vote)
Target Milestone: ---
Assignee: Dr. Werner Fink
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-03-23 07:25 UTC by Mike Galbraith
Modified: 2011-04-07 07:23 UTC (History)
0 users

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
free sample (10.91 KB, text/plain)
2011-03-23 07:25 UTC, Mike Galbraith
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Mike Galbraith 2011-03-23 07:25:38 UTC
Created attachment 420841 [details]
free sample

Kernels < 2.6.38 do not have TIOCGDEV ioctl support, which is causing boot time console woes.  Boxen boot fine, _but_ all boot scripts do not run, eg my kbd may or may not be left as a U.S. kbd vs DE, box boots to init 5, but may or may not enter X/KDE.  What is on the console is a mess, see attachment.  Severity of mess left over depends on kernel version booted, and perhaps configuration.

To trigger: remove "quiet" from commandline, add console=ttyS0... console=tty

I have now also seen the kbd/X problem minus the console mess caused by missing TIOCGDEV support in kernel-default-2.6.38-18.1.x86_64, and I saw it in kernel-desktop-2.6.37.1-1.2.2.x86_64, so problem is not limited to TIOCGDEV, something else is going on as well.

TIOCGDEV support now being required hinders git bisections, regression testing etc, where the user will have to boot/use mainline kernels. 

'course nothing is hitting logs.

I've verified that the mess is caused by missing TIOCGDEV by adding it to a couple mainline kernels.  This report is coming from 2.6.33.8 with support added.  So far, seen nothing but clean boots with a patched kernel, BUT, as noted I have seen part of the total problem in suse kernels WITH support.

Excerpts from attachment:
showconsole: Warning: the ioctl TIOCGDEV is not known by the kernel
...
INIT: Entering runlevel: 5
blogd: can not open /dev/tty: No such device or address
...
Starting syslog services                                                                                                                             done
blogd: Can not read from fd 0: Input/output error
...
tcsetattr: Invalid argument

etc etc etc... a mess
Comment 2 Jiri Slaby 2011-04-04 14:55:43 UTC
I'm confused by this report. We always provided TIOCGDEV as a part of our kernels by patches.fixes/tiocgdev. The problem lays in userspace -> reassigning.
Comment 3 Mike Galbraith 2011-04-04 17:18:43 UTC
(In reply to comment #2)
> I'm confused by this report. We always provided TIOCGDEV as a part of our
> kernels by patches.fixes/tiocgdev. The problem lays in userspace ->
> reassigning.

Yes, this is a userspace issue new to 11.4.
Comment 4 Dr. Werner Fink 2011-04-07 07:23:39 UTC
Try update to *latest* sysvinit-tools, from changelog:

 Fri Mar 25 2011 werner@suse.de
 - Fix triggered endless loop in blogd (bnc#642289)
   * Writing on tty0 caused blogd to re-read its own messages
   * The usage of ttyname(3) on /dev/console can fail

*** This bug has been marked as a duplicate of bug 642289 ***