atom feed45 messages in org.kernel.vger.linux-usbRe: PATCH: 2.6.26-rc8: Fix IRQF_DISAB...
FromSent OnAttachments
Stefan BeckerJun 22, 2008 9:55 am 
Rene HermanJun 22, 2008 10:42 am 
Alan SternJun 22, 2008 12:31 pm 
Stefan BeckerJun 23, 2008 8:52 am 
Alan SternJun 23, 2008 11:09 am 
Stefan BeckerJun 24, 2008 11:41 am 
Alan SternJun 24, 2008 2:15 pm 
Stefan BeckerJun 25, 2008 8:51 am 
Alan SternJun 25, 2008 11:38 am 
Stefan BeckerJun 25, 2008 11:31 pm 
Alan SternJun 26, 2008 7:25 am 
Stefan BeckerJun 26, 2008 3:07 pm.bz2
David BrownellJun 27, 2008 9:07 am 
Alan SternJun 27, 2008 9:09 am 
Stefan BeckerJun 28, 2008 7:31 am 
Stefan BeckerJun 28, 2008 7:36 am.bz2
Stefan BeckerJun 28, 2008 8:38 am.bz2
Alan SternJun 28, 2008 9:52 am 
Becker Stefan (Nokia-D/Salo)Jun 28, 2008 12:34 pm 
David BrownellJun 28, 2008 12:50 pm 
Stefan BeckerJun 29, 2008 7:57 am 
David BrownellJun 29, 2008 8:09 pm 
Stefan BeckerJun 29, 2008 10:22 pm 
Stefan BeckerJun 30, 2008 2:34 am 
David BrownellJun 30, 2008 4:14 am 
Alan CoxJun 30, 2008 7:26 am 
Henrique de Moraes HolschuhJun 30, 2008 7:27 am 
Alan SternJun 30, 2008 7:37 am 
Stefan BeckerJun 30, 2008 11:52 am 
Alan SternJun 30, 2008 12:35 pm 
David BrownellJun 30, 2008 12:57 pm 
David BrownellJun 30, 2008 1:30 pm 
Stefan BeckerJun 30, 2008 2:26 pm 
Alan SternJun 30, 2008 2:28 pm 
David BrownellJun 30, 2008 2:48 pm 
Alan SternJul 1, 2008 7:10 am 
Leonardo ChiquittoJul 1, 2008 7:19 am 
Stefan BeckerJul 1, 2008 9:18 am 
David BrownellJul 1, 2008 9:22 am 
Greg KHJul 1, 2008 11:25 am 
Alan SternJul 1, 2008 11:59 am 
Greg KHJul 1, 2008 12:13 pm 
Stefan BeckerJul 1, 2008 12:14 pm 
David BrownellJul 1, 2008 12:21 pm 
Greg KHJul 1, 2008 12:50 pm 
Subject:Re: PATCH: 2.6.26-rc8: Fix IRQF_DISABLED for shared interrupts
From:David Brownell (davi@public.gmane.org)
Date:Jun 30, 2008 12:57:25 pm
List:org.kernel.vger.linux-usb

On Monday 30 June 2008, David Brownell wrote:

On Monday 30 June 2008, Stefan Becker wrote:

It seems IRQF_SHARED | IRQF_DISABLED has already been discussed several times on LKML:

Given that, I'm surprised that nobody has added a warning that prints when those two flags are both passed to request_irq().

Well, here's a fix for that little problem.

Notice the rude interaction with LOCKDEP too. If you used that, you'd never have seen the behavior you saw. And if you did use that, with non-IRQF_DISABLED interrupt handlers, you'd wrongly believe some IRQ code paths always ran with IRQs disabled...

- Dave

===== CUT HERE We periodically have problems that get tracked down to the IRQ framework not respecting IRQF_DISABLED for some shared IRQ cases. Linus views this as "will not fix", but we're still left with the bugs caused by this misbehavior.

This patch adds a nag message in request_irq(), so that drivers can fix their IRQ handlers to avoid this problem.

Note that developers will never see the relevant bugs when they run with LOCKDEP, so it's no wonder these bugs are hard to find. (That also means LOCKDEP will be missing some IRQ-related bugs involving IRQ handlers that don't set IRQF_DISABLED.)

Signed-off-by: David Brownell
<dbrownell-Rn4VEauK+AKRv+LV9M@public.gmane.org>

--- kernel/irq/manage.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)

--- a/kernel/irq/manage.c 2008-06-30 12:28:58.000000000 -0700 +++ b/kernel/irq/manage.c 2008-06-30 12:46:54.000000000 -0700 @@ -539,6 +539,18 @@ int request_irq(unsigned int irq, irq_ha struct irqaction *action; int retval;

+ /* + * handle_IRQ_event() always ignores IRQF_DISABLED except for + * the _first_ irqaction (sigh). That can cause oopsing, but + * the behavior is classified as "will not fix" so we need to + * start nudging drivers away from using that idiom. + */ + if ((irqflags & (IRQF_SHARED|IRQF_DISABLED)) + == (IRQF_SHARED|IRQF_DISABLED)) + pr_warning("IRQ %d/%s: IRQF_DISABLED is not " + "guaranteed on shared IRQs\n", + irq, devname); + #ifdef CONFIG_LOCKDEP /* * Lockdep wants atomic interrupt handlers: