Re: G5 troubles booting powerpc-git (July 6)

On Fri, 07 Jul 2006 18:27:04 +1000
Benjamin Herrenschmidt [off-list ref] wrote:

On Fri, 2006-07-07 at 01:23 -0700, Andrew Morton wrote:

quoted

On Fri, 07 Jul 2006 08:56:27 +1000
Benjamin Herrenschmidt [off-list ref] wrote:

quoted

If you look at arch/powerpc/kernel/irq.c from line 479:

	/* Check if mapping already exist, if it does, call
	 * host->ops->map() to update the flags
	 */
	virq = irq_find_mapping(host, hwirq);
	if (virq != IRQ_NONE) {
		pr_debug("irq: -> existing mapping on virq %d\n", virq);
		host->ops->map(host, virq, hwirq, flags);
		return virq;
	}

What if you comment out the host->ops->map(...) call in there ? Does it
help ?

No, there's no change in behaviour.

Doh ! Ok, looks like a different issue. Can you try this patch ? Please,
do _not_ apply it to your tree as, as it is it will break non-powermac.
It's some work I'm doing to clean up a couple of rough edges in my irq
rework and fix a little misdesign.

If it doesn't fix it, then it's indeed a completely different issue that I'll have to track down tomorrow
hopefully.

Needs this to compile:

diff -puN arch/powerpc/platforms/pseries/ras.c~a-fix arch/powerpc/platforms/pseries/ras.c

--- a/arch/powerpc/platforms/pseries/ras.c~a-fix
+++ a/arch/powerpc/platforms/pseries/ras.c

@@ -93,8 +93,7 @@ static void request_ras_irqs(struct devi
 		for (i = 0; i < opicplen; i++) {
 			if (count > 15)
 				break;
-			virqs[count] = irq_create_mapping(NULL, *(opicprop++),
-							 IRQ_TYPE_NONE);
+			virqs[count] = irq_create_mapping(NULL, *(opicprop++));
 			if (virqs[count] == NO_IRQ)
 				printk(KERN_ERR "Unable to allocate interrupt "
 				       "number for %s\n", np->full_name);

diff -puN arch/powerpc/platforms/pseries/xics.c~a-fix arch/powerpc/platforms/pseries/xics.c

--- a/arch/powerpc/platforms/pseries/xics.c~a-fix
+++ a/arch/powerpc/platforms/pseries/xics.c

@@ -757,7 +757,7 @@ void xics_request_IPIs(void)
 {
 	unsigned int ipi;
 
-	ipi = irq_create_mapping(xics_host, XICS_IPI, 0);
+	ipi = irq_create_mapping(xics_host, XICS_IPI);
 	BUG_ON(ipi == NO_IRQ);
 
 	/*

diff -puN drivers/char/hvsi.c~a-fix drivers/char/hvsi.c

--- a/drivers/char/hvsi.c~a-fix
+++ a/drivers/char/hvsi.c

@@ -1299,7 +1299,7 @@ static int __init hvsi_console_init(void
 		hp->inbuf_end = hp->inbuf;
 		hp->state = HVSI_CLOSED;
 		hp->vtermno = *vtermno;
-		hp->virq = irq_create_mapping(NULL, irq[0], 0);
+		hp->virq = irq_create_mapping(NULL, irq[0]);
 		if (hp->virq == NO_IRQ) {
 			printk(KERN_ERR "%s: couldn't create irq mapping for 0x%x\n",
 				__FUNCTION__, irq[0]);

_

But it still doesn't help.


These:

arch/powerpc/sysdev/i8259.c:222: warning: initialization from incompatible pointer type
arch/powerpc/platforms/pseries/xics.c:555: warning: initialization from incompatible pointer type
arch/powerpc/platforms/pseries/xics.c:561: warning: initialization from incompatible pointer type

look like super-serious box-killers.  Hope I'm not using either of those.