Re: Lockup with 2.2.3 on SMP system

gmkelly@zip.com.au
Thu, 18 Mar 1999 19:31:25 +1100

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Tigran Aivazian: "Re: [patch] /proc/<pid>/rlimit & buffer cache instrumentation"
Previous message: Jeremy Fitzhardinge: "Re: can't connect to web site"

Further on Keyboard lockups in 2.2.3. In debugging further, by running
multiple xterms watching /proc/interrupts, /proc/stat etc. it
appears that on my machine at least - the keyboard controller
chip is being placed in a "invalid state", how I don't know.
I have noticed that on reboot (by pulling the power plug on
my UPS so that system shutdown automatically, and restarts
without me touching anything) that on reboot the bios complains
saying "Keyboard failure or keyboard doesn't exist". This is
really strange, as I thought the h/w would reset this, keep
in mind that I don't remove power from the system, only trying
to narrow down the problem. Keyboard interrupts don't increment
but that may be just the result not the cause.

My h/w is SMP PPro 150Mhz, 128Mb mem, Buslogic SCSI's drives (4)
IDE drives (hda, hdc), RocketPort8, 3 x Parallel ECP ports,
4Mb Matrox Millennium, PCI NE2000 Ethernet, VideoTV Card, etc...

I do however have:
[root@bushrock /proc]# cat ./interrupts
CPU0 CPU1
0: 1634481 1637087 IO-APIC-edge timer
1: 8688 8574 IO-APIC-edge keyboard
2: 0 0 XT-PIC cascade
3: 44148 43481 IO-APIC-edge serial
5: 0 3 IO-APIC-edge soundblaster
8: 0 1 IO-APIC-edge rtc
13: 1 0 XT-PIC fpu
15: 6 1 IO-APIC-edge ide1
16: 12294 12137 IO-APIC-level BusLogic BT-946C
18: 744943 743523 IO-APIC-level bttv
19: 0 0 IO-APIC-level eth0
NMI: 0
ERR: 0

a mixture of level and edge interrupt triggering which has not
been a problem on past releases. Interesting the RTC intr
doesn't inc much. You'll also notice that I get the keyboard
hang without Ethernet activity (I disconnected network to
narrow down problem).

So, in summary I have only got 2 keyboard hangs, 1 after 4 days,
another within 12 hours of reboot, same results both times,
no log info, same reboot keyboard indications.

Hope this helps the debugging process.
Cheers, Grahame

------------------------------------------------------ Sydney Linux User Group - Coordinator & Member (SLUG) see http://www.slug.org.au for details on activities ------------------------------------------------------

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/

Next message: Tigran Aivazian: "Re: [patch] /proc/<pid>/rlimit & buffer cache instrumentation"
Previous message: Jeremy Fitzhardinge: "Re: can't connect to web site"