[SUCCESS REPORT / PATCH] yenta_socket IRQ woes w/ NetGear WG511T ("IRQ#11: nobody cared" either, but a different one)

[Dominique Quatravaux]

Daniel Ritz a écrit :

>hi dominique
>
>thanks for the report. the problem is the patch there is a quick hack which
>is wrong and breaks other stuff. could you instead try with the attached one?
>it contains all my current changes (which i really need to split up soon).
>  
>
I have mixed success with your patch. The dreaded "IRQ 11: nobody cared"
is still there, and getting the wifi card to work still requires irqpoll
on the kernel command-line.

The good news, however is that I could not get the card to hang even
under stress (flood ping, multiple reboots and inserts-removes) despite
that "nobody cared" message in dmesg (see below). Well, except under one
particular circumstance: yanking the card off while it's busy tuning
itself into the (WEPped) WLAN causes nice kernel oopses, see pictures at
http://kilimandjaro.dyndns.org/~dom/nobody-cared/. They are only barely
legible, but I am not transcribing them to the list since from my
limited understanding the stack traces are madwifi-only so they do not
seem related to the problem at hand.

The laptop is only loaned to me, and I have to give it back tonight to
his rightful owner, sorry.

dmesg excerpt (note that irqpoll is mentioned even though I *did* set it
on the cmdline):

irq 11: nobody cared (try booting with the "irqpoll" option.
 [<c013813a>] __report_bad_irq+0x2a/0x90
 [<c0138239>] note_interrupt+0x79/0x100
 [<c0137ada>] __do_IRQ+0x13a/0x160
 [<c0105519>] do_IRQ+0x19/0x30
 [<c0103aae>] common_interrupt+0x1a/0x20
 [<c0137950>] handle_IRQ_event+0x20/0x70
 [<c0138270>] note_interrupt+0xb0/0x100
 [<c0137a7b>] __do_IRQ+0xdb/0x160
 [<c0105519>] do_IRQ+0x19/0x30
 [<c0103aae>] common_interrupt+0x1a/0x20
 [<c011e570>] __do_softirq+0x30/0x90
 [<c011e5f6>] do_softirq+0x26/0x30
 [<c0137924>] irq_exit+0x34/0x40
 [<c010551e>] do_IRQ+0x1e/0x30
 [<c0103aae>] common_interrupt+0x1a/0x20
 [<ce82ab06>] acpi_processor_idle+0x123/0x260 [processor]
 [<c01010ca>] cpu_idle+0x3a/0x60
 [<c03587f2>] start_kernel+0x172/0x1b0
 [<c03583b0>] unknown_bootoption+0x0/0x1b0
handlers:
[<ce99f7e0>] (yenta_interrupt+0x0/0xb0 [yenta_socket])
[<cea25c30>] (usb_hcd_irq+0x0/0x70 [usbcore])
[<cea25c30>] (usb_hcd_irq+0x0/0x70 [usbcore])
Disabling IRQ #11


Regards,

-- 
<< Tout n'y est pas parfait, mais on y honore certainement les jardiniers >>

			Dominique Quatravaux <dom at kilimandjaro.dyndns.org>