[PATCH] powerpc-genirq: Use generic_handle_irq()

[PATCH] powerpc-genirq: Use generic_handle_irq()

[Benjamin Herrenschmidt]

This patch updates the ppc and powerpc architectures to use the new
generic_handle_irq() so that interrupt controllers can be ported to the
new genirq layer.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

Patch depends on Ingo/Thomas genirq patchset.

Index: linux-work/arch/powerpc/kernel/irq.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/irq.c	2006-06-05 15:29:37.000000000 +1000
+++ linux-work/arch/powerpc/kernel/irq.c	2006-06-05 15:50:53.000000000 +1000
@@ -219,15 +219,19 @@ void do_IRQ(struct pt_regs *regs)
 		curtp = current_thread_info();
 		irqtp = hardirq_ctx[smp_processor_id()];
 		if (curtp != irqtp) {
+			struct irq_desc *desc = irq_desc + irq;
+			void *handler = desc->handle_irq;
+			if (handler == NULL)
+				handler = &__do_IRQ;
 			irqtp->task = curtp->task;
 			irqtp->flags = 0;
-			call___do_IRQ(irq, regs, irqtp);
+			call_handle_irq(irq, desc, regs, irqtp, handler);
 			irqtp->task = NULL;
 			if (irqtp->flags)
 				set_bits(irqtp->flags, &curtp->flags);
 		} else
 #endif
-			__do_IRQ(irq, regs);
+			generic_handle_irq(irq, regs);
 	} else if (irq != -2)
 		/* That's not SMP safe ... but who cares ? */
 		ppc_spurious_interrupts++;
Index: linux-work/arch/powerpc/kernel/misc_64.S
===================================================================
--- linux-work.orig/arch/powerpc/kernel/misc_64.S	2006-06-05 15:29:37.000000000 +1000
+++ linux-work/arch/powerpc/kernel/misc_64.S	2006-06-05 15:38:01.000000000 +1000
@@ -89,12 +89,14 @@ _GLOBAL(call_do_softirq)
 	mtlr	r0
 	blr
 
-_GLOBAL(call___do_IRQ)
+_GLOBAL(call_handle_irq)
+	ld	r8,0(r7)
 	mflr	r0
 	std	r0,16(r1)
-	stdu	r1,THREAD_SIZE-112(r5)
-	mr	r1,r5
-	bl	.__do_IRQ
+	mtctr	r8
+	stdu	r1,THREAD_SIZE-112(r6)
+	mr	r1,r6
+	bctrl
 	ld	r1,0(r1)
 	ld	r0,16(r1)
 	mtlr	r0
Index: linux-work/include/asm-powerpc/irq.h
===================================================================
--- linux-work.orig/include/asm-powerpc/irq.h	2006-06-05 15:29:37.000000000 +1000
+++ linux-work/include/asm-powerpc/irq.h	2006-06-05 15:50:53.000000000 +1000
@@ -494,8 +494,8 @@ extern struct thread_info *softirq_ctx[N
 
 extern void irq_ctx_init(void);
 extern void call_do_softirq(struct thread_info *tp);
-extern int call___do_IRQ(int irq, struct pt_regs *regs,
-		struct thread_info *tp);
+extern int call_handle_irq(int irq, void *p1, void *p2,
+			   struct thread_info *tp, void *func);
 
 #else
 #define irq_ctx_init()

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[Josh Boyer]

On Tue, 2006-06-13 at 13:47 +1000, Benjamin Herrenschmidt wrote:
> This patch updates the ppc and powerpc architectures to use the new
> generic_handle_irq() so that interrupt controllers can be ported to the
> new genirq layer.

It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
somewhere else?

And yes, arch/ppc still exists.  We can debate that later ;).

josh

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[David Woodhouse]

On Tue, 2006-06-13 at 05:51 -0500, Josh Boyer wrote:
> It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
> somewhere else?
> 
> And yes, arch/ppc still exists.  We can debate that later ;).

Did a maintainer step forward for it yet?

-- 
dwmw2

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[Josh Boyer]

On Tue, 2006-06-13 at 13:39 +0100, David Woodhouse wrote:
> On Tue, 2006-06-13 at 05:51 -0500, Josh Boyer wrote:
> > It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
> > somewhere else?
> > 
> > And yes, arch/ppc still exists.  We can debate that later ;).
> 
> Did a maintainer step forward for it yet?

For arch/ppc?  Not that I'm aware of.  But that wasn't particularly my
point.  The email says "This patch updates the ppc and powerpc
architectures..." when it doesn't.

I think most agree that arch/ppc needs to die via a complete migration
to arch/powerpc.  But patches shouldn't claim to do something they
don't :).

josh

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[Benjamin Herrenschmidt]

On Tue, 2006-06-13 at 05:51 -0500, Josh Boyer wrote:
> On Tue, 2006-06-13 at 13:47 +1000, Benjamin Herrenschmidt wrote:
> > This patch updates the ppc and powerpc architectures to use the new
> > generic_handle_irq() so that interrupt controllers can be ported to the
> > new genirq layer.
> 
> It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
> somewhere else?
> 
> And yes, arch/ppc still exists.  We can debate that later ;).

arch/ppc uses arch/powerpc/kernel/irq.c

Ben.

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[Benjamin Herrenschmidt]

On Tue, 2006-06-13 at 07:46 -0500, Josh Boyer wrote:
> On Tue, 2006-06-13 at 13:39 +0100, David Woodhouse wrote:
> > On Tue, 2006-06-13 at 05:51 -0500, Josh Boyer wrote:
> > > It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
> > > somewhere else?
> > > 
> > > And yes, arch/ppc still exists.  We can debate that later ;).
> > 
> > Did a maintainer step forward for it yet?
> 
> For arch/ppc?  Not that I'm aware of.  But that wasn't particularly my
> point.  The email says "This patch updates the ppc and powerpc
> architectures..." when it doesn't.

It does, check again :)

Ben.

Re: [PATCH] powerpc-genirq: Use generic_handle_irq()

[Josh Boyer]

On Wed, 2006-06-14 at 09:20 +1000, Benjamin Herrenschmidt wrote:
> On Tue, 2006-06-13 at 07:46 -0500, Josh Boyer wrote:
> > On Tue, 2006-06-13 at 13:39 +0100, David Woodhouse wrote:
> > > On Tue, 2006-06-13 at 05:51 -0500, Josh Boyer wrote:
> > > > It does?  I only see arch/powerpc in the patch.  Is arch/ppc hiding
> > > > somewhere else?
> > > > 
> > > > And yes, arch/ppc still exists.  We can debate that later ;).
> > > 
> > > Did a maintainer step forward for it yet?
> > 
> > For arch/ppc?  Not that I'm aware of.  But that wasn't particularly my
> > point.  The email says "This patch updates the ppc and powerpc
> > architectures..." when it doesn't.
> 
> It does, check again :)

Ugh.  Today is not my day apparently.  Sorry for the noise.

josh