Hello,
I use 2.6.8.1. On 2.6.7 things were OK. My motherboard is an Asus
P4P800, with the Intel i865PE chipset. USB stuff is compiled into the
kernel (not as modules).
When I plug in a USB device it is not recognized. It does not even
appear in lsusb. And it says that it disables IRQ 11 - which is even
NOT the IRQ used by USB!
The system works OK with USB devices that were plugged in at boot time.
But when I unplug one, wait some seconds and plug it in again, the
system does not see it anymore.
On the first plugging-in of any USB device, the following appears in
/var/log/messages :
===
Sep 19 11:49:07 localhost kernel: irq 11: nobody cared!
Sep 19 11:49:07 localhost kernel: [<c010784c>]
__report_bad_irq+0x2a/0x8b
Sep 19 11:49:07 localhost kernel: [<c0107936>] note_interrupt+0x6f/0x9f
Sep 19 11:49:07 localhost kernel: [<c0107b2c>] do_IRQ+0xe3/0xe5
Sep 19 11:49:07 localhost kernel: [<c0106104>]
common_interrupt+0x18/0x20
Sep 19 11:49:07 localhost kernel: [<c01077e2>]
handle_IRQ_event+0x24/0x64
Sep 19 11:49:07 localhost kernel: [<c0107abb>] do_IRQ+0x72/0xe5
Sep 19 11:49:07 localhost kernel: [<c0106104>]
common_interrupt+0x18/0x20
Sep 19 11:49:07 localhost kernel: [<c011d47f>] __do_softirq+0x2f/0x83
Sep 19 11:49:07 localhost kernel: [<c011d4f9>] do_softirq+0x26/0x28
Sep 19 11:49:07 localhost kernel: [<c0107b13>] do_IRQ+0xca/0xe5
Sep 19 11:49:07 localhost kernel: [<c010401e>] default_idle+0x0/0x27
Sep 19 11:49:07 localhost kernel: [<c0106104>]
common_interrupt+0x18/0x20
Sep 19 11:49:07 localhost kernel: [<c010401e>] default_idle+0x0/0x27
Sep 19 11:49:07 localhost kernel: [<c0104042>] default_idle+0x24/0x27
Sep 19 11:49:07 localhost kernel: [<c01040a9>] cpu_idle+0x2e/0x37
Sep 19 11:49:07 localhost kernel: [<c03e26ca>] start_kernel+0x17e/0x1bd
Sep 19 11:49:07 localhost kernel: [<c03e22e4>]
unknown_bootoption+0x0/0x16f
Sep 19 11:49:07 localhost kernel: handlers:
Sep 19 11:49:07 localhost kernel: [<d099d725>]
(ohci_irq_handler+0x0/0x78d [ohci1394])
Sep 19 11:49:07 localhost kernel: [<d0aa06a5>] (SkGeIsrOnePort+0x0/0x146
[sk98lin])
Sep 19 11:49:07 localhost kernel: [<d0b07796>]
(snd_intel8x0_interrupt+0x0/0x1ef [snd_intel8x0])
Sep 19 11:49:07 localhost kernel: Disabling IRQ #11
===
Here is /proc/interrupts:
===
CPU0
0: 21993633 XT-PIC timer
1: 9543 XT-PIC i8042
2: 0 XT-PIC cascade
5: 178551 XT-PIC uhci_hcd, uhci_hcd, eth0
8: 1 XT-PIC rtc
9: 0 XT-PIC acpi
10: 1864080 XT-PIC uhci_hcd, uhci_hcd, radeon@PCI:1:0:0
11: 32714 XT-PIC ohci1394, SysKonnect SK-98xx, Intel ICH5
14: 154802 XT-PIC ide0
15: 57 XT-PIC ide1
NMI: 0
ERR: 0
===
So, uhci_hcd is not on IRQ 11 at all! ohci1394 is the IEEE1394 device,
which seems to work just fine.
What can this be?..
Yours, Mikhail Ramendik
> When I plug in a USB device it is not recognized. It does not even
> appear in lsusb. And it says that it disables IRQ 11 - which is even
> NOT the IRQ used by USB!
Does it make any difference if you boot with "pci=routeirq"?
There is a known problem with USB and prism54 devices not working
after suspend/resume, and it goes away with "pci=routeirq".
If that does make a difference, can you post the whole dmesg
log with and without it?
Bjorn Helgaas wrote:
> > When I plug in a USB device it is not recognized. It does not even
> > appear in lsusb. And it says that it disables IRQ 11 - which is even
> > NOT the IRQ used by USB!
>
> Does it make any difference if you boot with "pci=routeirq"?
No. The behaviour is the same. Perhaps the message is somewhat
different, but the "IRQ 11 disabled" part is still there.
I also tried two other boots: "acpi=off", and "pci=noacpi acpi=noirq".
They also did not help at all.
I am attaching the dmesg for the normal boot. I can also send the dmesg
for any other kind of boot if necessary...
Yours, Mikhail Ramendik
On Thursday 23 September 2004 12:25 pm, Mikhail Ramendik wrote:
> Bjorn Helgaas wrote:
> > > When I plug in a USB device it is not recognized. It does not even
> > > appear in lsusb. And it says that it disables IRQ 11 - which is even
> > > NOT the IRQ used by USB!
> >
> > Does it make any difference if you boot with "pci=routeirq"?
>
> No. The behaviour is the same. Perhaps the message is somewhat
> different, but the "IRQ 11 disabled" part is still there.
Sorry, I wasn't paying enough attention. You said you were on
2.6.8.1, which doesn't have "pci=routeirq" in it. So of course
it didn't make any difference. I was thinking you were on an
-mm kernel, where there are a couple issues that can be worked
around with "pci=routeirq".