[vdr] Re: LIRC and System-Crash

[Usenet-372112@zocki.toppoint.de (Rainer Zocholl)]

axelgruber3@gmx.de(Axel Gruber)  19.11.01 21:29

Once upon a time Axel Gruber shaped the electrons to say...


>Hi all

>I have massive Problems with my Server:

Since when?
Did it ever run?
If yes, when last OK?
When does the error occur the frist time?
How often does the error occure?
Steps to reproduce?
What did you change ? ;-)



>6* IDE HD

DMA problems?

>3* DVB-S (2new 1old)  

1 Ver 1.3 and 2 Ver 2.1 or 2 Nova? (Just to be sure)


>1* Home-MADE LIRC Device

Which of the several?
serial,parport ?


>Sometime this system crashes completly - no error-Message in SYSLOG

Checked your /etc/syslog.conf?
I know that there are systems out that do not log all....
but see below.


>The only message wich i can't understand:

>Jan 22 23:54:14 SERVER kernel: lirc_serial: AIEEEE: 0 1 32e69a96 32e69a82 = f46c 833d

>Anyone have a idea, if this is ok ?


Yes. :-)

It only shows that the system is not totally dead (crashed), 
it just "hangs" in a loop and interrupts are enabled. 
This explains, why you don't see a syslog entry!

It is able to receive erronous LIRC interrupts and log them.
This is (usually) not the cause of the crash.
To get rid of, disable "debugging" in LIRC config.


There should be a way to "escape" from such a (driver) loop
Was it "CNTL-Shift-right-SysReq" (if compilied in kernel)?

An other way would be to force an NMI and install 
a better NMI-oops handler.






HTH


Rainer




(PS: Pls tag your eMail "MIME-pq")