[ath9k-devel] received PCI FATAL interrupt

[ath9k-devel] received PCI FATAL interrupt

[Mihai Moldovan]

Hello guys,

today I have updated to Kernel 2.6.29.1 using compat-wireless
(yesterday's snapshot) and got those messages in my kernel log ring buffer:

[   72.952758] ath9k: received PCI FATAL interrupt
[   77.087480] ath9k: received PCI FATAL interrupt
[   86.072429] ath9k: received PCI FATAL interrupt
[   86.074685] ath9k: received PCI FATAL interrupt
[   89.203901] ath9k: received PCI FATAL interrupt
[   89.448506] ath9k: received PCI FATAL interrupt
[   94.713862] ath9k: received PCI FATAL interrupt
[   94.869953] ath9k: received PCI FATAL interrupt
[  102.265172] ath9k: received PCI FATAL interrupt
[  102.281810] ath9k: received PCI FATAL interrupt
[  102.342217] ath9k: received PCI FATAL interrupt
[  102.347172] ath9k: received PCI FATAL interrupt
[  102.646480] ath9k: received PCI FATAL interrupt

I don't know where they are coming from, but I suspect those error
messages to be related to my other problems as well.

I own two Atheros cards, a PCI one (04:00.0 Network controller: Atheros
Communications Inc. AR5416 802.11abgn Wireless PCI Adapter (rev 01)) and
a PCIe one (02:00.0 Network controller: Atheros Communications Inc.
AR5418 802.11abgn Wireless PCI Express Adapter (rev 01)) - both
connected to the same PC.

As already stated, OS is Linux 2.6.29.1 with compat-wireless-2009-04-06.

Both cards (and another ethernet card) are bridged together into br0
(this piece of information is of no real importance, though.)

wlan0 and wlan1 are set up to work via hostapd into MASTER mode. (WPA1
enabled, CCMP/PSK.)

They operate on two separate, non-overlapping channels with the same ESSID.

Most of the time, everything is working fine, it seems that either ath9k
or hostapd can't handle high traffic for a long time. This particularly
means, that streaming video is impossible, because the wireless
connection will eventually go down after a few minutes. Even more
interesting in this context is, that re-starting hostapd on the router
helps solving this problem, at least temporarly.

Another interesting effect is that the signal quality seen by my ipw2200
card goes up and down, that is, is unsteady when generating high network
traffic via WLAN. This can't be normal!

This all is particularly bad, because I would like to use this box as my
wireless LAN router... but currently this is more like wireless cheese
with a few lucky packets. :(

I'm open to any kind of helping rants.

Best regards,


Mihai Moldovan


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 898 bytes
Desc: OpenPGP digital signature
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090407/9416fed9/attachment.pgp 

[ath9k-devel] received PCI FATAL interrupt

[Jason Tackaberry]

Hi,

On Tue, 2009-04-07 at 20:53 +0200, Mihai Moldovan wrote:
> today I have updated to Kernel 2.6.29.1 using compat-wireless
> (yesterday's snapshot) and got those messages in my kernel log ring buffer:
> 
> [   72.952758] ath9k: received PCI FATAL interrupt

What is the contents of /proc/interrupts?

Cheers,
Jason.

[ath9k-devel] received PCI FATAL interrupt

[Mihai Moldovan]

* On 07.04.2009 21:04, Jason Tackaberry wrote:
> Hi,
>
> On Tue, 2009-04-07 at 20:53 +0200, Mihai Moldovan wrote:
>   
>> today I have updated to Kernel 2.6.29.1 using compat-wireless
>> (yesterday's snapshot) and got those messages in my kernel log ring buffer:
>>
>> [   72.952758] ath9k: received PCI FATAL interrupt
>>     
>
> What is the contents of /proc/interrupts
Hi Jason,

actually looks fine. Relevant lines:

 19:      76501      63856      52512      64969   IO-APIC-fasteoi
uhci_hcd:usb7, ath
 21:     128257     122019     114385     128061   IO-APIC-fasteoi
uhci_hcd:usb4, ath

ath and USB should not interfere, but well...

Anyways, full content:

           CPU0       CPU1       CPU2       CPU3
  0:         46          2          2          6   IO-APIC-edge      timer
  1:          2          0          0          0   IO-APIC-edge      i8042
  4:          0          0          1          0   IO-APIC-edge
  8:       1695       1538       1484       1464   IO-APIC-edge      rtc0
  9:          0          0          0          0   IO-APIC-fasteoi   acpi
 12:          1          0          3          0   IO-APIC-edge      i8042
 16:          0          0          0          0   IO-APIC-fasteoi
uhci_hcd:usb3
 17:          0          0          0          1   IO-APIC-fasteoi
 18:          0          0          0          0   IO-APIC-fasteoi
ehci_hcd:usb1, uhci_hcd:usb5, uhci_hcd:usb8
 19:      76501      63856      52512      64969   IO-APIC-fasteoi
uhci_hcd:usb7, ath
 21:     128257     122019     114385     128061   IO-APIC-fasteoi
uhci_hcd:usb4, ath
 23:         46         36         37         36   IO-APIC-fasteoi
ehci_hcd:usb2, uhci_hcd:usb6
 24:      53885          0          0          0  HPET_MSI-edge      hpet2
 25:          0      40218          0          0  HPET_MSI-edge      hpet3
 26:          0          0      88740          0  HPET_MSI-edge      hpet4
 27:          0          0          0      82489  HPET_MSI-edge      hpet5
 32:       3515       3470       3165       2764   PCI-MSI-edge      ahci
 33:      44801      40299      38433      41709   PCI-MSI-edge      eth1
 34:      41813      45560      40506      37979   PCI-MSI-edge      eth0
NMI:          0          0          0          0   Non-maskable interrupts
LOC:         95         79         52         25   Local timer interrupts
RES:       5880       7402       5444       6837   Rescheduling interrupts
CAL:         49         73         72         58   Function call interrupts
TLB:        319        568        308        573   TLB shootdowns
TRM:          0          0          0          0   Thermal event interrupts
THR:          0          0          0          0   Threshold APIC interrupts
SPU:          0          0          0          0   Spurious interrupts
ERR:          0
MIS:          0

Best regards,


Mihai


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 898 bytes
Desc: OpenPGP digital signature
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090407/ac733e81/attachment.pgp 

[ath9k-devel] received PCI FATAL interrupt

[Mihai Moldovan]

Hi,

I've just tested the newest compat-wireless release today and the
problem still appears. :(

Any hints?

Best regards,


Mihai

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 898 bytes
Desc: OpenPGP digital signature
Url : http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20090515/21e4b51b/attachment.pgp 

[ath9k-devel] received PCI FATAL interrupt

[Luis R. Rodriguez]

On Thu, May 14, 2009 at 5:57 PM, Mihai Moldovan <ionic@ionic.de> wrote:
> Hi,
>
> I've just tested the newest compat-wireless release today and the
> problem still appears. :(
>
> Any hints?

Lets break this down into a few steps. First although you are using AP
mode please apply these patches:

http://w1.fi/p/ps/

They should apply to compat-wireless as well.

Before you try with these patches please remove one card and bootup
and see if you see this issue. If so please provide steps to
reproduce.

If you do see an issue just try the patches and try again with only
one card. If you see issues with one card and the patches applied
please provide steps to reproduce.

  Luis