Hello.
Often when starting my computer, vdr crashes without any specific reason and freezes the computer ! Nothing is possible/available on my comnputer ! I just have to reset it !
Looking at syslog files, I have a long list of kernel PREEMPT, the last one freezing the computer !
Most of this PREMPT finishes with the following line: Nov 3 07:18:39 localhost kernel: <1>Unable to handle kernel NULL pointer dereference at virtual address 00000020
The first PREMPT occurs after a problem on DXR3 card !
Nov 3 07:18:23 localhost vdr[3105]: dxr3: fatal: microcode upload failed: Timer expired Nov 3 07:18:23 localhost kernel: em8300: DICOM Update failed Nov 3 07:18:23 localhost kernel: em8300_video.c: Video Fifo timeout Nov 3 07:18:23 localhost vdr[2470]: dxr3: resuscitation: device failure or user initiated reset
I have a long list of call traces... It seems that problems cames from a badly managed page fault !
I don't know what to do... Does anyone have a suggestion of what I could do in order to avoid this ? What could I look in the syslog also ?
Many thanks for any help.
Michael
On Thu, 03 Nov 2005 11:37:17 +0100 michaelip@free.fr wrote:
Hello.
Often when starting my computer, vdr crashes without any specific reason and freezes the computer ! Nothing is possible/available on my comnputer ! I just have to reset it !
Looking at syslog files, I have a long list of kernel PREEMPT, the last one freezing the computer !
Most of this PREMPT finishes with the following line: Nov 3 07:18:39 localhost kernel: <1>Unable to handle kernel NULL pointer dereference at virtual address 00000020
The first PREMPT occurs after a problem on DXR3 card !
Nov 3 07:18:23 localhost vdr[3105]: dxr3: fatal: microcode upload failed: Timer expired Nov 3 07:18:23 localhost kernel: em8300: DICOM Update failed Nov 3 07:18:23 localhost kernel: em8300_video.c: Video Fifo timeout Nov 3 07:18:23 localhost vdr[2470]: dxr3: resuscitation: device failure or user initiated reset
I have a long list of call traces... It seems that problems cames from a badly managed page fault !
I don't know what to do... Does anyone have a suggestion of what I could do in order to avoid this ? What could I look in the syslog also ?
Many thanks for any help.
Michael
The PREEMPT messages suggest that you have kernel preemption enabled. That may be a bad idea with Dxr3, so try to compile a new kernel without it.
If that doesn't help, check that Dxr3 doesn't share interrupts with any other device (cat /proc/interrupts). Then you could add some delays to the scripts so that the card has enough time to initialize properly. It is best to load the dxr3 stuff in some script using modprobe with "sleep 2" after the module loading and also after the microcode setup.
-- Niko Mikkilä
michaelip@free.fr wrote:
Hello.
Often when starting my computer, vdr crashes without any specific reason and freezes the computer ! Nothing is possible/available on my comnputer ! I just have to reset it !
What version of the em8300 driver are you using? I think this issue might be fixed in the one of the recent releases from http://dxr3.sourceforge.net/download.html
A problem was fixed where the interrupt handler was installed before the card was properly initialised and a stray interrupt from a card sharing the interrupt line would cause an oops.
Disabling preempt option may help as well.
The best place to follow up on this is probably the dxr3-devel mailing list, see http://sourceforge.net/mail/?group_id=5165
Jon
Selon Niko Mikkila nm@phnet.fi:
On Thu, 03 Nov 2005 11:37:17 +0100 michaelip@free.fr wrote:
Hello.
Often when starting my computer, vdr crashes without any specific reason
and
freezes the computer ! Nothing is possible/available on my comnputer ! I
just
have to reset it !
Looking at syslog files, I have a long list of kernel PREEMPT, the last one freezing the computer !
Most of this PREMPT finishes with the following line: Nov 3 07:18:39 localhost kernel: <1>Unable to handle kernel NULL pointer dereference at virtual address 00000020
The first PREMPT occurs after a problem on DXR3 card !
Nov 3 07:18:23 localhost vdr[3105]: dxr3: fatal: microcode upload failed:
Timer
expired Nov 3 07:18:23 localhost kernel: em8300: DICOM Update failed Nov 3 07:18:23 localhost kernel: em8300_video.c: Video Fifo timeout Nov 3 07:18:23 localhost vdr[2470]: dxr3: resuscitation: device failure or
user
initiated reset
I have a long list of call traces... It seems that problems cames from a
badly
managed page fault !
I don't know what to do... Does anyone have a suggestion of what I could do
in
order to avoid this ? What could I look in the syslog also ?
Many thanks for any help.
Michael
The PREEMPT messages suggest that you have kernel preemption enabled. That may be a bad idea with Dxr3, so try to compile a new kernel without it.
If that doesn't help, check that Dxr3 doesn't share interrupts with any other device (cat /proc/interrupts). Then you could add some delays to the scripts so that the card has enough time to initialize properly. It is best to load the dxr3 stuff in some script using modprobe with "sleep 2" after the module loading and also after the microcode setup.
-- Niko Mikkilä
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Hello. I have two devices listening at the same IRQ !
Here is the output of the cat /proc/interrupts CPU0 0: 101477 XT-PIC timer 2: 0 XT-PIC cascade 4: 2086 XT-PIC lirc_serial 5: 0 XT-PIC ehci_hcd:usb1 9: 0 XT-PIC acpi 11: 1017 XT-PIC Skystar2, eth0 12: 6235 XT-PIC em8300 14: 3273 XT-PIC ide0 15: 13 XT-PIC ide1 NMI: 0 ERR: 0
The network card share the same IRQ than my DVB card ! Could I swap one of these card on IRQ 10, knowing that the network card is an intergrated one ?
Many thanks for your help.
Selon Jon Burgess jburgess@uklinux.net:
michaelip@free.fr wrote:
Hello.
Often when starting my computer, vdr crashes without any specific reason
and
freezes the computer ! Nothing is possible/available on my comnputer ! I
just
have to reset it !
What version of the em8300 driver are you using? I think this issue might be fixed in the one of the recent releases from http://dxr3.sourceforge.net/download.html
A problem was fixed where the interrupt handler was installed before the card was properly initialised and a stray interrupt from a card sharing the interrupt line would cause an oops.
Disabling preempt option may help as well.
The best place to follow up on this is probably the dxr3-devel mailing list, see http://sourceforge.net/mail/?group_id=5165
Jon
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Hello.
My em8300 driver version is the 0.15.0 at least. Maybe I have the 0.15.1 but I'm not sure, and I can't check right now ! If not, I will upgrade !
I also will remove the PREEMPT support in my kernel... Many thanks for your help.
On Fri, 04 Nov 2005 10:45:43 +0100 michaelip@free.fr wrote:
Selon Niko Mikkila nm@phnet.fi:
The PREEMPT messages suggest that you have kernel preemption enabled. That may be a bad idea with Dxr3, so try to compile a new kernel without it.
If that doesn't help, check that Dxr3 doesn't share interrupts with any other device (cat /proc/interrupts). Then you could add some delays to the scripts so that the card has enough time to initialize properly. It is best to load the dxr3 stuff in some script using modprobe with "sleep 2" after the module loading and also after the microcode setup.
-- Niko Mikkilä
Hello. I have two devices listening at the same IRQ !
Here is the output of the cat /proc/interrupts CPU0 0: 101477 XT-PIC timer 2: 0 XT-PIC cascade 4: 2086 XT-PIC lirc_serial 5: 0 XT-PIC ehci_hcd:usb1 9: 0 XT-PIC acpi 11: 1017 XT-PIC Skystar2, eth0 12: 6235 XT-PIC em8300 14: 3273 XT-PIC ide0 15: 13 XT-PIC ide1 NMI: 0 ERR: 0
The network card share the same IRQ than my DVB card ! Could I swap one of these card on IRQ 10, knowing that the network card is an intergrated one ?
This probably doesn't cause the crashes, but I think I had some problems receiving DVB in a similar IRQ setup when there was high network traffic. With some motherboards you can change the IRQ mapping in the BIOS. Otherwise you could try swapping the Skystar2 to some other PCI slot. Personally, I had to enable IO-APIC in the kernel to get enough available interrupts for all the cards. Ideally the devices should work fine with shared interrupts, but there may be bugs in drivers and the hardware that cause various problems.
-- Niko Mikkilä
-
Jon Burgess -
michaelip＠free.fr -
Niko Mikkila