On Wed, 28 Jan 2004 09:31:12 -0800
David Brownell <[EMAIL PROTECTED]> wrote:
> Stephen Hemminger wrote:
> > USB hci interface's probably will not work with shared IRQ without
> > this patch. The hcd core always returns IRQ_HANDLED.
>
> Actually they do work, but whatever corner cases motivated the
> irqreturn_t change clearly won't. (Mostly detecting IRQ lines
> that are wedged.)
The problem with returning handled is that the next irq in the chain
doesn't get called. None of my machines share USB irq, but some
share IRQ between network and serial. And if the network card
always returns handled then the serial doesn't work; therefore
I notice when other drivers have the same problem.
>
> > The fix is to change the irq hook to return a value and propagate
> > result back out.
>
> Looks more or less right to me. BK history says that Linus just
> did the quick'n'dirty patch in this part of his irqreturn_t fix;
> he didn't update usbcore, and nobody went back to it.
>
> Though I do wonder about the OHCI code returning IRQ_NONE for
> the case of a cardbus eject ("device removed"). That doesn't
> seem right, and it doesn't match your change to the EHCI driver.
I am not really sure what the correct response is in the device
removed case. Some drivers return NONE and some return handled.
> I think this'd be OK for Greg to merge if you changed that bit.
>
> - Dave
>
>
--
Stephen Hemminger mailto:[EMAIL PROTECTED]
Open Source Development Lab http://developer.osdl.org/shemminger
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
