[oops, first posting was too large for lkml, so instead of attachments,
I'll put the relevant files online]
On 11/20/10 12:27, Randy Dunlap wrote:
> 2.6.37-rc2-git4 on x86_64, nearly allmodconfig with a several debug/test
> options disabled and a few options changed to =y so that things work
> (like serial console). Could there be some kernel options that are
> interfering here?
>
>
> Kernel config file is attached.
>
> Reproducible by running 'modules_tester' script with the ONLY environment
> variable set to "hwmon". I have not been able to reproduce this with other
> driver subsystems.
>
> Today's test failed during loading/unloaded tmp421. Previous tests have
> failed with adt7470, adm1031, and jc42. The log of modules_tester output
> is attached. (mod-hwmon-r2g4.out)
>
> (as root:)
> # ONLY=hwmon ./modules_tester 2>&1 | tee test.out
>
> modules_tester just loads and unloads modules 3 times in a row. (it is
> available from http://oss.oracle.com/~rdunlap/scripts/modules_tester)
>
>
> This scripted test always fails (WARNING: + kernel is tainted) with:
>
> WARNING: at /local/linsrc/linux-2.6.37-rc2-git4/drivers/tty/tty_io.c:1331 tty_reopen+0xa8/0xc5()
>
> The full kernel log is attached as hwmon-fail-37r2g4.txt.
>
Please see http://oss.oracle.com/~rdunlap/tty-warning/
for all of the relevant files.
--
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***
On Sat, Nov 20, 2010 at 05:05:05PM -0800, Randy Dunlap wrote:
> [oops, first posting was too large for lkml, so instead of attachments,
> I'll put the relevant files online]
>
Ah, I'm glad someone else hit this. I've been seeing it on Fedora
rawhide recently, but hadn't yet had a chance to figure out if it was my
fault (or cgroups or systemd related, since some reporters have claimed
it was caused by upgrading to it.)
regards, Kyle