[PATCH] powerpc: fix trigger handling in the new irq code

This patch slightly reworks the new irq code to fix a small design error.  I
removed the passing of the trigger to the map() calls entirely, it was not a
good idea to have one call do two different things.  It also fixes a couple of
corner cases.

Mapping a linux virtual irq to a physical irq now does only that.  Setting the
trigger is a different action which has a different call.

The main changes are:

- I no longer call host->ops->map() for an already mapped irq, I just return
  the virtual number that was already mapped.  It was called before to give an
  opportunity to change the trigger, but that was causing issues as that could
  happen while the interrupt was in use by a device, and because of the
  trigger change, map would potentially muck around with things in a racy way.
   That was causing much burden on a given's controller implementation of
  map() to get it right.  This is much simpler now.  map() is only called on
  the initial mapping of an irq, meaning that you know that this irq is _not_
  being used.  You can initialize the hardware if you want (though you don't
  have to).

- Controllers that can handle different type of triggers (level/edge/etc...)
  now implement the standard irq_chip->set_type() call as defined by the
  generic code.  That means that you can use the standard set_irq_type() to
  configure an irq line manually if you wish or (though I don't like that
  interface), pass explicit trigger flags to request_irq() as defined by the
  generic kernel interfaces.  Also, using those interfaces guarantees that
  your controller set_type callback is called with the descriptor lock held,
  thus providing locking against activity on the same interrupt (including
  mask/unmask/etc...) automatically.  A result is that, for example, MPIC's
  own map() implementation calls irq_set_type(NONE) to configure the hardware
  to the default triggers.

- To allow the above, the irq_map array entry for the new mapped interrupt
  is now set before map() callback is called for the controller.

- The irq_create_of_mapping() (also used by irq_of_parse_and_map()) function
  for mapping interrupts from the device-tree now also call the separate
  set_irq_type(), and only does so if there is a change in the trigger type.

- While I was at it, I changed pci_read_irq_line() (which is the helper I
  would expect most archs to use in their pcibios_fixup() to get the PCI
  interrupt routing from the device tree) to also handle a fallback when the
  DT mapping fails consisting of reading the PCI_INTERRUPT_PIN to know wether
  the device has an interrupt at all, and the the PCI_INTERRUPT_LINE to get an
  interrupt number from the device.  That number is then mapped using the
  default controller, and the trigger is set to level low.  That default
  behaviour works for several platforms that don't have a proper interrupt
  tree like Pegasos.  If it doesn't work for your platform, then either
  provide a proper interrupt tree from the firmware so that fallback isn't
  needed, or don't call pci_read_irq_line()

- Add back a bit that got dropped by my main rework patch for properly
  clearing pending IPIs on pSeries when using a kexec

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

8 files changed, 74 insertions, 68 deletions

diff --git a/arch/powerpc/platforms/cell/interrupt.c b/arch/powerpc/platforms/cell/interrupt.c
index 9d5da78..d7bbb61 100644
--- a/arch/powerpc/platforms/cell/interrupt.c
+++ b/arch/powerpc/platforms/cell/interrupt.c
@@ -159,7 +159,7 @@ static void iic_request_ipi(int ipi, const char *name)
 		if (iic_hosts[node] == NULL)
 			continue;
 		virq = irq_create_mapping(iic_hosts[node],
-					  iic_ipi_to_irq(ipi), 0);
+					  iic_ipi_to_irq(ipi));
 		if (virq == NO_IRQ) {
 			printk(KERN_ERR
 			       "iic: failed to map IPI %s on node %d\n",
@@ -197,7 +197,7 @@ static int iic_host_match(struct irq_host *h, struct device_node *node)
 }
 
 static int iic_host_map(struct irq_host *h, unsigned int virq,
-			irq_hw_number_t hw, unsigned int flags)
+			irq_hw_number_t hw)
 {
 	if (hw < IIC_IRQ_IPI0)
 		set_irq_chip_and_handler(virq, &iic_chip, handle_fasteoi_irq);
diff --git a/arch/powerpc/platforms/cell/spider-pic.c b/arch/powerpc/platforms/cell/spider-pic.c
index ae7ef88..15217bb 100644
--- a/arch/powerpc/platforms/cell/spider-pic.c
+++ b/arch/powerpc/platforms/cell/spider-pic.c
@@ -85,9 +85,6 @@ static void spider_unmask_irq(unsigned int virq)
 	struct spider_pic *pic = spider_virq_to_pic(virq);
 	void __iomem *cfg = spider_get_irq_config(pic, irq_map[virq].hwirq);
 
-	/* We use no locking as we should be covered by the descriptor lock
-	 * for access to invidual source configuration registers
-	 */
 	out_be32(cfg, in_be32(cfg) | 0x30000000u);
 }
 
@@ -96,9 +93,6 @@ static void spider_mask_irq(unsigned int virq)
 	struct spider_pic *pic = spider_virq_to_pic(virq);
 	void __iomem *cfg = spider_get_irq_config(pic, irq_map[virq].hwirq);
 
-	/* We use no locking as we should be covered by the descriptor lock
-	 * for access to invidual source configuration registers
-	 */
 	out_be32(cfg, in_be32(cfg) & ~0x30000000u);
 }
 
@@ -120,26 +114,14 @@ static void spider_ack_irq(unsigned int virq)
 	out_be32(pic->regs + TIR_EDC, 0x100 | (src & 0xf));
 }
 
-static struct irq_chip spider_pic = {
-	.typename = " SPIDER   ",
-	.unmask = spider_unmask_irq,
-	.mask = spider_mask_irq,
-	.ack = spider_ack_irq,
-};
-
-static int spider_host_match(struct irq_host *h, struct device_node *node)
-{
-	struct spider_pic *pic = h->host_data;
-	return node == pic->of_node;
-}
-
-static int spider_host_map(struct irq_host *h, unsigned int virq,
-			irq_hw_number_t hw, unsigned int flags)
+static int spider_set_irq_type(unsigned int virq, unsigned int type)
 {
-	unsigned int sense = flags & IRQ_TYPE_SENSE_MASK;
-	struct spider_pic *pic = h->host_data;
+	unsigned int sense = type & IRQ_TYPE_SENSE_MASK;
+	struct spider_pic *pic = spider_virq_to_pic(virq);
+	unsigned int hw = irq_map[virq].hwirq;
 	void __iomem *cfg = spider_get_irq_config(pic, hw);
-	int level = 0;
+	struct irq_desc *desc = get_irq_desc(virq);
+	u32 old_mask;
 	u32 ic;
 
 	/* Note that only level high is supported for most interrupts */
@@ -157,29 +139,57 @@ static int spider_host_map(struct irq_host *h, unsigned int virq,
 		break;
 	case IRQ_TYPE_LEVEL_LOW:
 		ic = 0x0;
-		level = 1;
 		break;
 	case IRQ_TYPE_LEVEL_HIGH:
 	case IRQ_TYPE_NONE:
 		ic = 0x1;
-		level = 1;
 		break;
 	default:
 		return -EINVAL;
 	}
 
+	/* Update irq_desc */
+	desc->status &= ~(IRQ_TYPE_SENSE_MASK | IRQ_LEVEL);
+	desc->status |= type & IRQ_TYPE_SENSE_MASK;
+	if (type & (IRQ_TYPE_LEVEL_HIGH | IRQ_TYPE_LEVEL_LOW))
+		desc->status |= IRQ_LEVEL;
+
 	/* Configure the source. One gross hack that was there before and
 	 * that I've kept around is the priority to the BE which I set to
 	 * be the same as the interrupt source number. I don't know wether
 	 * that's supposed to make any kind of sense however, we'll have to
 	 * decide that, but for now, I'm not changing the behaviour.
 	 */
-	out_be32(cfg, (ic << 24) | (0x7 << 16) | (pic->node_id << 4) | 0xe);
+	old_mask = in_be32(cfg) & 0x30000000u;
+	out_be32(cfg, old_mask | (ic << 24) | (0x7 << 16) |
+		 (pic->node_id << 4) | 0xe);
 	out_be32(cfg + 4, (0x2 << 16) | (hw & 0xff));
 
-	if (level)
-		get_irq_desc(virq)->status |= IRQ_LEVEL;
+	return 0;
+}
+
+static struct irq_chip spider_pic = {
+	.typename = " SPIDER   ",
+	.unmask = spider_unmask_irq,
+	.mask = spider_mask_irq,
+	.ack = spider_ack_irq,
+	.set_type = spider_set_irq_type,
+};
+
+static int spider_host_match(struct irq_host *h, struct device_node *node)
+{
+	struct spider_pic *pic = h->host_data;
+	return node == pic->of_node;
+}
+
+static int spider_host_map(struct irq_host *h, unsigned int virq,
+			irq_hw_number_t hw)
+{
 	set_irq_chip_and_handler(virq, &spider_pic, handle_level_irq);
+
+	/* Set default irq type */
+	set_irq_type(virq, IRQ_TYPE_NONE);
+
 	return 0;
 }
 
@@ -283,7 +293,7 @@ static unsigned int __init spider_find_cascade_and_node(struct spider_pic *pic)
 	if (iic_host == NULL)
 		return NO_IRQ;
 	/* Manufacture an IIC interrupt number of class 2 */
-	virq = irq_create_mapping(iic_host, 0x20 | unit, 0);
+	virq = irq_create_mapping(iic_host, 0x20 | unit);
 	if (virq == NO_IRQ)
 		printk(KERN_ERR "spider_pic: failed to map cascade !");
 	return virq;
diff --git a/arch/powerpc/platforms/cell/spu_base.c b/arch/powerpc/platforms/cell/spu_base.c
index 5d2313a..d06042d 100644
--- a/arch/powerpc/platforms/cell/spu_base.c
+++ b/arch/powerpc/platforms/cell/spu_base.c
@@ -583,9 +583,9 @@ static int __init spu_map_interrupts(struct spu *spu, struct device_node *np)
 	spu->isrc = isrc = tmp[0];
 
 	/* Now map interrupts of all 3 classes */
-	spu->irqs[0] = irq_create_mapping(host, 0x00 | isrc, 0);
-	spu->irqs[1] = irq_create_mapping(host, 0x10 | isrc, 0);
-	spu->irqs[2] = irq_create_mapping(host, 0x20 | isrc, 0);
+	spu->irqs[0] = irq_create_mapping(host, 0x00 | isrc);
+	spu->irqs[1] = irq_create_mapping(host, 0x10 | isrc);
+	spu->irqs[2] = irq_create_mapping(host, 0x20 | isrc);
 
 	/* Right now, we only fail if class 2 failed */
 	return spu->irqs[2] == NO_IRQ ? -EINVAL : 0;
diff --git a/arch/powerpc/platforms/iseries/irq.c b/arch/powerpc/platforms/iseries/irq.c
index 2275e64..e324468 100644
--- a/arch/powerpc/platforms/iseries/irq.c
+++ b/arch/powerpc/platforms/iseries/irq.c
@@ -300,7 +300,7 @@ int __init iSeries_allocate_IRQ(HvBusNumber bus,
 	realirq = (((((sub_bus << 8) + (bus - 1)) << 3) + (idsel - 1)) << 3)
 		+ function;
 
-	return irq_create_mapping(NULL, realirq, IRQ_TYPE_NONE);
+	return irq_create_mapping(NULL, realirq);
 }
 
 #endif /* CONFIG_PCI */
@@ -341,7 +341,7 @@ unsigned int iSeries_get_irq(struct pt_regs *regs)
 }
 
 static int iseries_irq_host_map(struct irq_host *h, unsigned int virq,
-				irq_hw_number_t hw, unsigned int flags)
+				irq_hw_number_t hw)
 {
 	set_irq_chip_and_handler(virq, &iseries_pic, handle_fasteoi_irq);
 
diff --git a/arch/powerpc/platforms/powermac/pci.c b/arch/powerpc/platforms/powermac/pci.c
index 556b349..205d044 100644
--- a/arch/powerpc/platforms/powermac/pci.c
+++ b/arch/powerpc/platforms/powermac/pci.c
@@ -16,6 +16,7 @@
 #include <linux/string.h>
 #include <linux/init.h>
 #include <linux/bootmem.h>
+#include <linux/irq.h>
 
 #include <asm/sections.h>
 #include <asm/io.h>
@@ -24,10 +25,7 @@
 #include <asm/machdep.h>
 #include <asm/pmac_feature.h>
 #include <asm/grackle.h>
-#ifdef CONFIG_PPC64
-//#include <asm/iommu.h>
 #include <asm/ppc-pci.h>
-#endif
 
 #undef DEBUG
 
@@ -46,7 +44,6 @@ static int has_uninorth;
 static struct pci_controller *u3_agp;
 static struct pci_controller *u4_pcie;
 static struct pci_controller *u3_ht;
-#define has_second_ohare 0
 #else
 static int has_second_ohare;
 #endif /* CONFIG_PPC64 */
@@ -993,6 +990,7 @@ void __init pmac_pcibios_fixup(void)
 		/* Read interrupt from the device-tree */
 		pci_read_irq_line(dev);
 
+#ifdef CONFIG_PPC32
 		/* Fixup interrupt for the modem/ethernet combo controller.
 		 * on machines with a second ohare chip.
 		 * The number in the device tree (27) is bogus (correct for
@@ -1002,8 +1000,11 @@ void __init pmac_pcibios_fixup(void)
 		 */
 		if (has_second_ohare &&
 		    dev->vendor == PCI_VENDOR_ID_DEC &&
-		    dev->device == PCI_DEVICE_ID_DEC_TULIP_PLUS)
-			dev->irq = irq_create_mapping(NULL, 60, 0);
+		    dev->device == PCI_DEVICE_ID_DEC_TULIP_PLUS) {
+			dev->irq = irq_create_mapping(NULL, 60);
+			set_irq_type(dev->irq, IRQ_TYPE_LEVEL_LOW);
+		}
+#endif /* CONFIG_PPC32 */
 	}
 }
 
diff --git a/arch/powerpc/platforms/powermac/pic.c b/arch/powerpc/platforms/powermac/pic.c
index 3d328bc..060789e 100644
--- a/arch/powerpc/platforms/powermac/pic.c
+++ b/arch/powerpc/platforms/powermac/pic.c
@@ -291,7 +291,7 @@ static int pmac_pic_host_match(struct irq_host *h, struct device_node *node)
 }
 
 static int pmac_pic_host_map(struct irq_host *h, unsigned int virq,
-			     irq_hw_number_t hw, unsigned int flags)
+			     irq_hw_number_t hw)
 {
 	struct irq_desc *desc = get_irq_desc(virq);
 	int level;
@@ -318,6 +318,7 @@ static int pmac_pic_host_xlate(struct irq_host *h, struct device_node *ct,
 			       unsigned int *out_flags)
 
 {
+	*out_flags = IRQ_TYPE_NONE;
 	*out_hwirq = *intspec;
 	return 0;
 }
@@ -434,7 +435,7 @@ static void __init pmac_pic_probe_oldstyle(void)
 
 	printk(KERN_INFO "irq: System has %d possible interrupts\n", max_irqs);
 #ifdef CONFIG_XMON
-	setup_irq(irq_create_mapping(NULL, 20, 0), &xmon_action);
+	setup_irq(irq_create_mapping(NULL, 20), &xmon_action);
 #endif
 }
 #endif /* CONFIG_PPC32 */
@@ -579,9 +580,10 @@ void __init pmac_pic_init(void)
 		flags |= OF_IMAP_OLDWORLD_MAC;
 	if (get_property(of_chosen, "linux,bootx", NULL) != NULL)
 		flags |= OF_IMAP_NO_PHANDLE;
-	of_irq_map_init(flags);
 #endif /* CONFIG_PPC_32 */
 
+	of_irq_map_init(flags);
+
 	/* We first try to detect Apple's new Core99 chipset, since mac-io
 	 * is quite different on those machines and contains an IBM MPIC2.
 	 */
diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index 9df783088..c7ffde1 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c
@@ -93,8 +93,7 @@ static void request_ras_irqs(struct device_node *np,
 		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 --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c
index 716972a..2d0da6f 100644
--- a/arch/powerpc/platforms/pseries/xics.c
+++ b/arch/powerpc/platforms/pseries/xics.c
@@ -502,16 +502,9 @@ static int xics_host_match(struct irq_host *h, struct device_node *node)
 }
 
 static int xics_host_map_direct(struct irq_host *h, unsigned int virq,
-				irq_hw_number_t hw, unsigned int flags)
+				irq_hw_number_t hw)
 {
-	unsigned int sense = flags & IRQ_TYPE_SENSE_MASK;
-
-	pr_debug("xics: map_direct virq %d, hwirq 0x%lx, flags: 0x%x\n",
-		 virq, hw, flags);
-
-	if (sense && sense != IRQ_TYPE_LEVEL_LOW)
-		printk(KERN_WARNING "xics: using unsupported sense 0x%x"
-		       " for irq %d (h: 0x%lx)\n", flags, virq, hw);
+	pr_debug("xics: map_direct virq %d, hwirq 0x%lx\n", virq, hw);
 
 	get_irq_desc(virq)->status |= IRQ_LEVEL;
 	set_irq_chip_and_handler(virq, &xics_pic_direct, handle_fasteoi_irq);
@@ -519,16 +512,9 @@ static int xics_host_map_direct(struct irq_host *h, unsigned int virq,
 }
 
 static int xics_host_map_lpar(struct irq_host *h, unsigned int virq,
-			      irq_hw_number_t hw, unsigned int flags)
+			      irq_hw_number_t hw)
 {
-	unsigned int sense = flags & IRQ_TYPE_SENSE_MASK;
-
-	pr_debug("xics: map_lpar virq %d, hwirq 0x%lx, flags: 0x%x\n",
-		 virq, hw, flags);
-
-	if (sense && sense != IRQ_TYPE_LEVEL_LOW)
-		printk(KERN_WARNING "xics: using unsupported sense 0x%x"
-		       " for irq %d (h: 0x%lx)\n", flags, virq, hw);
+	pr_debug("xics: map_direct virq %d, hwirq 0x%lx\n", virq, hw);
 
 	get_irq_desc(virq)->status |= IRQ_LEVEL;
 	set_irq_chip_and_handler(virq, &xics_pic_lpar, handle_fasteoi_irq);
@@ -757,7 +743,7 @@ 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);
 
 	/*
@@ -783,6 +769,14 @@ void xics_teardown_cpu(int secondary)
 	xics_set_cpu_priority(cpu, 0);
 
 	/*
+	 * Clear IPI
+	 */
+	if (firmware_has_feature(FW_FEATURE_LPAR))
+		lpar_qirr_info(cpu, 0xff);
+	else
+		direct_qirr_info(cpu, 0xff);
+
+	/*
 	 * we need to EOI the IPI if we got here from kexec down IPI
 	 *
 	 * probably need to check all the other interrupts too
@@ -795,7 +789,7 @@ void xics_teardown_cpu(int secondary)
 		return;
 	desc = get_irq_desc(ipi);
 	if (desc->chip && desc->chip->eoi)
-		desc->chip->eoi(XICS_IPI);
+		desc->chip->eoi(ipi);
 
 	/*
 	 * Some machines need to have at least one cpu in the GIQ,