[linux-dvb] DEBI irq oops, crashes: kernel 2.6 & Hyperthreading unusable withDVB

[Udo Wolter <uwp@dicke-aersche.de>]

Hi !
I'm using a Hauppauge Nexus-s and try to use kernel 2.6.6.rc3 (2.4 is
not really an option anymore). I also tried 2.6.5. At first I tried
to record with vdr 1.2.6. It recorded some minutes and after 2
minutes it just stops recording. So I tried vdr 1.3.6. Now I can record
but I'm getting lots of DEBI irq oops messages. It seems the more
pressure the machine gets the more messages I have. And: the more messages
in the log the bigger the possibility of a machine crash. It just crashes,
no message on the screen or in the log. It seems to be ok when I'm
switching hyperthreading off with acpi=off. But this is not a solution.
In this case my machine runs only at 60% of the CPU power and it starts
to get problems while recording 2 channels or more  at once. BTW, I also
tried the DVB-driver (1.1.1) from the linuxtv.org page, same messages, even
more, sometimes GPIO0 irq oops messages. I can't understand why a fast machine
should have those problems while others talking about recording
3 channels at once to disk with an Athlon 1200. From my experience I can
just say: this driver doesn't work with SMP. BTW, I also tried to get
the interrupt free of other drivers:

cat /proc/interrupts
           CPU0       CPU1
  0:    9429175     226196    IO-APIC-edge  timer
  1:      20916          0    IO-APIC-edge  i8042
  2:          0          0          XT-PIC  cascade
  7:          0          0    IO-APIC-edge  parport0
  8:          1          0    IO-APIC-edge  rtc
 12:      56864          0    IO-APIC-edge  i8042
 14:     391345      33270    IO-APIC-edge  ide0
 15:         29          1    IO-APIC-edge  ide1
 16:    1533283   20198587   IO-APIC-level  saa7146 (0)
 17:      13122          0   IO-APIC-level  Intel ICH5
 18:      22248          0   IO-APIC-level  ide2, ide3, eth0
 22:          3          0   IO-APIC-level  ohci1394
 23:          0          0   IO-APIC-level  ehci_hcd
NMI:          0          0
LOC:    9655570    9655569
ERR:          0
MIS:          0

I even tried to set the smp_affinity in /proc/irq/xx that way that
IRQ16 has one CPU for itself: the DEBI messages are still there.
Maybe someone can explain those messages ? Or the GPIO message ?
This is really one of the strangest and most instable drivers I've
ever used. I cannot remember a driver which has been integrated into
the standard kernel which crashed my machine the last 2 years. Even
with experimental drivers I never had this big problems.

Any help for me ? Do I have to give more informations ?

Mermgfurt,
		Udo

-- 
Udo Wolter                      |  /"\
email:    uwp@dicke-aersche.de  |  \ / ASCII RIBBON CAMPAIGN
www:      www.dicke-aersche.de  |   X    AGAINST HTML MAIL
dark:    heaven@lutz-ziffer.de  |  / \



-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe linux-dvb" as subject.