Fix (workaround?) for BCM Bigsur

Fix (workaround?) for BCM Bigsur

[Thiemo Seufer]

Hello All,

the appended patchlet allows to boot current HEAD on a BCM1480.
Without it the kernel dies in an unhandled interrupt loop.


Thiemo


Signed-off-by: Thiemo Seufer <ths@networkno.de>
---
diff --git a/arch/mips/sibyte/bcm1480/time.c b/arch/mips/sibyte/bcm1480/time.c
index 610f025..9344c7b 100644
--- a/arch/mips/sibyte/bcm1480/time.c
+++ b/arch/mips/sibyte/bcm1480/time.c
@@ -142,7 +142,6 @@ void __cpuinit sb1480_clockevent_init(void)
 			R_BCM1480_IMR_INTERRUPT_MAP_BASE_H) + (irq << 3)));
 
 	bcm1480_unmask_irq(cpu, irq);
-	bcm1480_steal_irq(irq);
 
 	action->handler	= sibyte_counter_handler;
 	action->flags	= IRQF_DISABLED | IRQF_PERCPU;

Re: Fix (workaround?) for BCM Bigsur

[Ralf Baechle]

On Sun, Oct 28, 2007 at 08:22:49PM +0000, Thiemo Seufer wrote:

> Hello All,
> 
> the appended patchlet allows to boot current HEAD on a BCM1480.
> Without it the kernel dies in an unhandled interrupt loop.

I think it's a bugfix.  The whole {sb1250,bcm1250}_steal_irq() irq seems
to be a hack papering needed only to paper over other implementation
issues.

  Ralf

Re: Fix (workaround?) for BCM Bigsur

[Ralf Baechle]

On Sun, Oct 28, 2007 at 08:22:49PM +0000, Thiemo Seufer wrote:

> the appended patchlet allows to boot current HEAD on a BCM1480.
> Without it the kernel dies in an unhandled interrupt loop.

I went a little further in 3b4137342e088f03e0f2c44608b832af7a8af814 but
basically the functional change is the same as your patch.  It also
applies the same to the sb1250 code.

Thanks,

  Ralf

Re: Fix (workaround?) for BCM Bigsur

[Thiemo Seufer]

Ralf Baechle wrote:
> On Sun, Oct 28, 2007 at 08:22:49PM +0000, Thiemo Seufer wrote:
> 
> > Hello All,
> > 
> > the appended patchlet allows to boot current HEAD on a BCM1480.
> > Without it the kernel dies in an unhandled interrupt loop.
> 
> I think it's a bugfix.  The whole {sb1250,bcm1250}_steal_irq() irq seems
> to be a hack papering needed only to paper over other implementation
> issues.

FYI, I removed all the *_steal_irq stuff in my tree, the resulting
kernel works fine on a BCM1480.


Thiemo

Re: Fix (workaround?) for BCM Bigsur

[Ralf Baechle]

On Mon, Oct 29, 2007 at 11:44:06AM +0000, Thiemo Seufer wrote:

> > > the appended patchlet allows to boot current HEAD on a BCM1480.
> > > Without it the kernel dies in an unhandled interrupt loop.
> > 
> > I think it's a bugfix.  The whole {sb1250,bcm1250}_steal_irq() irq seems
> > to be a hack papering needed only to paper over other implementation
> > issues.
> 
> FYI, I removed all the *_steal_irq stuff in my tree, the resulting
> kernel works fine on a BCM1480.

Thanks.  I already have f514a361491aea38160f2c9eb4f7494f06b8263a which
deletes the thing in my tree and queued for Linus.

  Ralf