Re: [PATCH] xirc2ps_cs irq return fix

Jeff Garzik (jgarzik@pobox.com)
Mon, 26 May 2003 23:49:18 -0400


Alan Cox wrote:
> On Llu, 2003-05-26 at 02:00, Zwane Mwaikambo wrote:
>
>>My interpretation of it is the PCMCIA controller was triggering interrupts
>>on exit and the link handler for the card was still installed even after
>>the netdevice was down.
>
>
> This is exactly what will happen all the time on PCMCIA devices. The
> edge triggered interrupt will cause an IRQ to float around every remove
> of the device on most hardware
>
> The fix is basically correct, although the odd floating IRQ ought to be
> cleaned up by the heuristics being fixed in the core IRQ disaster
> detector

huh?

If the fix is correct, we need to do a bombing run that adds the
following code to each driver,

if (!netif_device_present(dev))
return IRQ_HANDLED;

but of course the irq _wasn't_ handled by the driver. Silly, isn't it?

I still maintain it needs to be fixed elsewhere. Touching every driver
for this is just not scalable, in addition to pointing the finger at the
pcmcia core instead of individual drivers.

Jeff

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/