Bug 1735 - Fn-F5 (video switch) hangs machine on Toshiba Satellite 1905 S303
Summary: Fn-F5 (video switch) hangs machine on Toshiba Satellite 1905 S303
Status: CLOSED PATCH_ALREADY_AVAILABLE
Alias: None
Product: Drivers
Classification: Unclassified
Component: Platform (show other bugs)
Hardware: i386 Linux
: P2 normal
Assignee: Len Brown
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-12-23 05:27 UTC by Erico Mendonca
Modified: 2010-10-08 18:16 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.4.22.x, 2.4.23.x, 2.6.x
Subsystem:
Regression: ---
Bisected commit-id:


Attachments
The DSDT I'm using (156.19 KB, text/plain)
2003-12-23 06:07 UTC, Erico Mendonca
Details
dmesg from a working kernel. (13.58 KB, text/plain)
2003-12-23 07:21 UTC, Erico Mendonca
Details
System log from a working kernel. (27.24 KB, text/plain)
2003-12-23 07:21 UTC, Erico Mendonca
Details
System log from a non-working kernel. (30.56 KB, text/plain)
2003-12-23 07:23 UTC, Erico Mendonca
Details
The original DSDT table (150.71 KB, text/plain)
2003-12-23 07:24 UTC, Erico Mendonca
Details

Description Erico Mendonca 2003-12-23 05:27:33 UTC
Distribution: Mandrake Linux 9.2
Hardware Environment: Toshiba Satellite 1905 S303 (Pentium 4 2.4Ghz, 
Software Environment:
Problem Description: 
The machine hangs as soon as Fn-F5 (switch to external video/ tv-out) is
pressed. This did not happen until kernel 2.4.21-0.13mdk (version shipped with
Mandrake Linux 9.1). I've tried compiling 3 revisions of the 2.4.22.x series,
0.13mdk, 0.18mdk and 0.25mdk. I also tried 3 revisions of the 2.4.23.x series
(currently at 2.4.23-0.1mdk), and 2.6 test7-test9, with the same results.

Following Ducrot Bruno's suggestion, I recompiled 2.4.21-0.13mdk with ACPI debug
enabled, and got a few "QUERY _Q1C" messages in the system log as I pressed
Fn-F5, suggesting that the changes in ACPI that affected my machine are in the 
code related to the EC. Unfortunately, I could not find the exact spot.

I tried enabling ACPI debug in kernel 2.4.23, but all of the ACPI modules would
fail loading (missing debug/trace symbols??).

It may be generating a kernel oops, but I couldn't find a way to catch it, since
this machine has no serial ports, and the USB serial console is only supported
on kernel 2.5.x and 2.6.x. Also trying to enable it on kernel 2.6 didn't have
any effect. I use a pl2303-compatible USB serial dongle. Enabling the KDB in
2.4.22, 2.4.23 and 2.6 also did not help, as the machine hanged solid...


I'm attaching the relevant files.


Steps to reproduce:

1. Load any kernel > 2.4.21, press Fn-F5 to switch to TV-Out;
2. Press and hold the Power button to turn off the hanged machine :)
Comment 1 Erico Mendonca 2003-12-23 06:07:02 UTC
Created attachment 1723 [details]
The DSDT I'm using
Comment 2 Erico Mendonca 2003-12-23 07:21:19 UTC
Created attachment 1724 [details]
dmesg from a working kernel.
Comment 3 Erico Mendonca 2003-12-23 07:21:52 UTC
Created attachment 1725 [details]
System log from a working kernel.
Comment 4 Erico Mendonca 2003-12-23 07:23:04 UTC
Created attachment 1726 [details]
System log from a non-working kernel.
Comment 5 Erico Mendonca 2003-12-23 07:24:04 UTC
Created attachment 1727 [details]
The original DSDT table
Comment 6 Luming Yu 2003-12-24 21:24:07 UTC
Did you try base kernel?  

You said "I tried enabling ACPI debug in kernel 2.4.23, but all of the ACPI 
modules would fail loading (missing debug/trace symbols??)." Maybe it's 
another bug, file it.

I want to see the testing result of latest base kernel.
Comment 7 Len Brown 2004-01-08 19:44:48 UTC
log.failing: 
> ACPI: Battery Slot [BAT1] (battery present) 
> ACPI: AC Adapter [ACAD] (on-line) 
> ACPI: Processor [CPU0] (supports C1 C2) 
> ACPI: Thermal Zone [THM0] (47 C) 
> ACPI: Power Button (FF) [PWRF] 
> ACPI: Lid Switch [LID] 
 
I too am confused what you mean by the ACPI modules failing to load -- 
we'd not see this if they were not loaded. 
 
The dmesg labeled for the working kernel is 2.4.23, so I think that is the failing kernel, yes? 
 
Comment 8 Luming Yu 2004-01-17 22:52:27 UTC
Please try with APIC disabled!
Comment 9 Len Brown 2004-11-15 20:10:40 UTC
better luck with linux 2.6.9? 
Comment 10 Erico Mendonca 2004-12-05 16:15:36 UTC
Sorry for the delay, this has been quite a demanding year.

Anyway, I tested it with current kernels, and it's working fine with kernel
2.6.8.1-12mdk. Kernel 2.6.9 from kernel.org also worked fine.

One problem though, which is not related to this bug and I'm reporting only for
documentation: graphical console is messed up, and simultaneous video is not
possible on Linux; to change the video back to the LCD, you must unplug the
S-Video cable. But then again, this has always worked like that with previous
kernels, and may be related to the DSDT on this machine.

I believe this can be closed now.

Note You need to log in before you can comment on or make changes to this bug.