Re: [PATCH] uio: User IRQ Mode

["Magnus Damm" <magnus.damm@gmail.com>]

On Fri, Jul 4, 2008 at 4:55 AM, Hans J. Koch <hjk@linutronix.de> wrote:
> On Thu, Jul 03, 2008 at 10:23:16PM +0900, Paul Mundt wrote:
>> On Thu, Jul 03, 2008 at 02:45:05PM +0200, Hans J. Koch wrote:
>> > On Thu, Jul 03, 2008 at 09:10:19AM +0200, Uwe Kleine-K??nig wrote:
>> > >     Moreover this might prevent some bugs.  (And obviously this
>> > >     function has the potential to have a buggy implementation as
>> > >     the comment by Alan Cox shows.)
>> >
>> > For me, this shows two things:
>> >
>> > - I never ever had to use disable_irq_nosync() in any UIO driver yet,
>> >   otherwise I would have noticed.
>> >
>> > - Magnus turned in a patch that he never tested.
>> >
>> Note that the deadlock in question is in relation to SMP, it's true that
>> the patch was never tested in an SMP environment and only on UP, but it
>> certainly was tested.
>
> Ok, so I take back that accusation ;-)

That's good, thank you. In the future it would be even better if you
didn't accuse to begin with, since that will only heat up the
discussion.

> Nonetheless, the patch changes a UIO core file, and everything in there
> should work in all situations on all architectures unless there is a
> _very_ good reason to do something different. This not only applies to
> SMP issues but also to the limitation to non-shared interrupts.

I will resolve the SMP issue and repost, no problem.

>> The vast majority of driver writers don't have a
>> need to use disable_irq_nosync(), so whether you've had to use it or not
>> is largely irrelevant to the conversation at hand ;-)
>
> Sure ;-) I merely wanted to point out that this is an unusual way to
> handle an interrupt.

Grep shows that there is nothing unusual about it.

damm@rx1 ~/git/linux-2.6 $ find drivers/ | xargs grep -m 1 -e
[[:blank:]]disable_irq\( -e [[:blank:]]disable_irq_nosync\( | wc -l
105

/ magnus