[RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[Adrian Bunk]

If #includ'ing interrupt.h should be enough for getting the prototype of 
e.g. enable_irq() on all architectures, we need this patch.


Signed-off-by: Adrian Bunk <bunk@stusta.de>

--- linux-2.6.13-rc6-mm1-modular/include/linux/interrupt.h.old	2005-08-22 23:44:42.000000000 +0200
+++ linux-2.6.13-rc6-mm1-modular/include/linux/interrupt.h	2005-08-22 23:45:04.000000000 +0200
@@ -12,6 +12,7 @@
 #include <asm/atomic.h>
 #include <asm/ptrace.h>
 #include <asm/system.h>
+#include <asm/irq.h>
 
 /*
  * For 2.4.x compatibility, 2.4.x can use

Re: [RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[Christoph Hellwig]

On Wed, Aug 24, 2005 at 10:57:50AM +0200, Adrian Bunk wrote:
> If #includ'ing interrupt.h should be enough for getting the prototype of 
> e.g. enable_irq() on all architectures, we need this patch.

Per defintion you need to include <asm/irq.h> right now.  I'd like to change
that to <linux/interrupt.h>, but not my including <asm/irq.h> there.
We should just make the prototypes in <linux/interrupt.h> unconditional
and get rid of the macro/inline tricks some architectures do, these calls
aren't exactly fastpathes where that matters.

Re: [RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[Adrian Bunk]

On Wed, Aug 24, 2005 at 10:22:50AM +0100, Christoph Hellwig wrote:
> On Wed, Aug 24, 2005 at 10:57:50AM +0200, Adrian Bunk wrote:
> > If #includ'ing interrupt.h should be enough for getting the prototype of 
> > e.g. enable_irq() on all architectures, we need this patch.
> 
> Per defintion you need to include <asm/irq.h> right now.  I'd like to change
> that to <linux/interrupt.h>, but not my including <asm/irq.h> there.
> We should just make the prototypes in <linux/interrupt.h> unconditional
> and get rid of the macro/inline tricks some architectures do, these calls
> aren't exactly fastpathes where that matters.

Looking at 2.6.13-rc6-mm2, the only architectures with own enable_irq() 
implementations are m68knommu and sparc.

On m68knommu, enable_irq() does nothing unless a hook is used that has 
no in-kernel users.

The 32bit sparc arch seems to be the only arch doing funky things.

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: [RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[Thomas Gleixner]

On Wed, 2005-08-24 at 12:08 +0200, Adrian Bunk wrote:

> Looking at 2.6.13-rc6-mm2, the only architectures with own enable_irq() 
> implementations are m68knommu and sparc.

You missed ARM.

tglx

Re: [RFC: 2.6 patch] #include <asm/irq.h> in interrupt.h

[Adrian Bunk]

On Wed, Aug 24, 2005 at 10:44:55AM +0000, Thomas Gleixner wrote:
> On Wed, 2005-08-24 at 12:08 +0200, Adrian Bunk wrote:
> 
> > Looking at 2.6.13-rc6-mm2, the only architectures with own enable_irq() 
> > implementations are m68knommu and sparc.
> 
> You missed ARM.

Yes and no.

Yes, because you are right that I missed architectures.

No, because I was only looking at what would be required to 
unconditionally offer the enable_irq() prototyp in interrupt.h .

> tglx

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed