Skip to contentSkip to conversation

GIC OF bindings

PATCH In review Rob Herring 3 participants 8 messages Sep 29, 2011 – Sep 30, 2011

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover GIC OF bindings

From: Rob Herring <[email protected]>

With recent patches for VIC device tree support and GIC MULTI_IRQ support, it is evident that the irq domain support should be moved into the core GIC code and not be dependent on the DT initialization. So I've split things up into 2 patches separating irq_domain and device-tree code.

This should only change the internals of the GIC code. The binding is unchanged from the previous version. Yeah!

The full series is available here. This includes Russell's devel-stable and for-next branches and Nico's vmalloc branch.

git://git.jdl.com/software/linux-3.0.git gic

Rob

Rob Herring (2):
ARM: gic: add irq_domain support
ARM: gic: add OF based initialization

 Documentation/devicetree/bindings/arm/gic.txt |   55 ++++++++++
 arch/arm/common/Kconfig                       |    1 +
 arch/arm/common/gic.c                         |  144 +++++++++++++++++++------
 arch/arm/include/asm/hardware/gic.h           |    4 +
 4 files changed, 171 insertions(+), 33 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/gic.txt
Signature · 1 line
1.7.5.4
#1 lore raw

PATCH 1/2 ARM: gic: add irq_domain support

From: Rob Herring <[email protected]>

Convert the gic interrupt controller to use irq domains in preparation for device-tree binding and MULTI_IRQ. This allows for translation between GIC interrupt IDs and Linux irq numbers.

The meaning of irq_offset has changed. It now is just the number of skipped GIC interrupt IDs for the controller. It will be 16 for primary GIC and 32 for secondary GICs.

---
 arch/arm/common/Kconfig             |    1 +
 arch/arm/common/gic.c               |   80 ++++++++++++++++++++--------------
 arch/arm/include/asm/hardware/gic.h |    3 +
 3 files changed, 51 insertions(+), 33 deletions(-)
arch/arm/common/Kconfig +1 -0
index 4b71766..74df9ca 100644
@@ -1,4 +1,5 @@
 config ARM_GIC
+	select IRQ_DOMAIN
 	bool
 
 config ARM_VIC
arch/arm/common/gic.c +47 -33

157 lines not on this page · raw message

arch/arm/include/asm/hardware/gic.h +3 -0
index c562705..4893b10 100644
@@ -33,6 +33,8 @@
 #define GIC_DIST_SOFTINT		0xf00
 
 #ifndef __ASSEMBLY__
+#include <linux/irqdomain.h>
+
 extern void __iomem *gic_cpu_base_addr;
 extern struct irq_chip gic_arch_extn;
 
@@ -53,6 +55,7 @@ struct gic_chip_data {
 	u32 __percpu *saved_ppi_enable;
 	u32 __percpu *saved_ppi_conf;
 #endif
+	struct irq_domain domain;
 	unsigned int gic_irqs;
 };
 #endif
Signature · 1 line
1.7.5.4
#2 lore raw

Re: [PATCH 1/2] ARM: gic: add irq_domain support

Quoted 9 lines · Rob Herring

From: Rob Herring <[email protected]>

Convert the gic interrupt controller to use irq domains in preparation for device-tree binding and MULTI_IRQ. This allows for translation between GIC interrupt IDs and Linux irq numbers.

The meaning of irq_offset has changed. It now is just the number of skipped GIC interrupt IDs for the controller. It will be 16 for primary GIC and 32 for secondary GICs.

[...]

arch/arm/common/gic.c static inline unsigned int gic_irq(struct irq_data *d) · line 85

Quoted 7 lines

/*

arch/arm/common/gic.c +1 -1
@@ -81,7 +82,7 @@ static inline unsigned int gic_irq(struct irq_data *d)
  */
 static void gic_mask_irq(struct irq_data *d)
 {
-	u32 mask = 1 << (d->irq % 32);
+	u32 mask = 1 << (gic_irq(d) % 32);

This can probably change simply to d->hwirq if irq_offset is eliminated as I describe below.

Quoted 14 lines
arch/arm/common/gic.c +0 -0
 void __init gic_init(unsigned int gic_nr, unsigned int irq_start,
 	void __iomem *dist_base, void __iomem *cpu_base)
 {
arch/arm/common/gic.c +3 -1
 	struct gic_chip_data *gic;
+	struct irq_domain *domain;
+	int gic_irqs;
 
 	BUG_ON(gic_nr >= MAX_GIC_NR);
 
 	gic = &gic_data[gic_nr];
+	domain = &gic->domain;
 	gic->dist_base = dist_base;
 	gic->cpu_base = cpu_base;
-	gic->irq_offset = (irq_start - 1) & ~31;

arch/arm/common/gic.c static void __init gic_pm_init(struct gic_chip_data *gic) · line 546

Quoted 10 lines
arch/arm/common/gic.c +7 -1
 
-	if (gic_nr == 0)
+	/*
+	 * For primary GICs, skip over SGIs.
+	 * For secondary GICs, skip over PPIs, too.
+	 */
+	if (gic_nr == 0) {
 		gic_cpu_base_addr = cpu_base;
+		gic->irq_offset = 16;
+		irq_start = (irq_start & ~31) + 16;

With the switch to irq_domain, there should no longer be any need for a ~31 mask on the irq_start number. Yes, you'll want to make sure that it doesn't allocate below irq 16, but the driver should completely use the irq_domain to manage the mapping from linux-irq number to hwirq number. The ~31 mask appears to have been an optimization to quickly calculate hwirq number from the linux one, but that value is now found in irq_data->hwirq.

irq_offset could almost be dropped entirely outside of probe because it no longer matters. Instead, populate the to_irq() hook so that the hwirq value gets set correctly from the outset.

It would be better in general to have a consistent view of what the hwirq number means regardless of if it is a root GIC.

g.

#3 lore raw

Re: [PATCH 1/2] ARM: gic: add irq_domain support

Quoted 49 lines · Grant Likely
Quoted 9 lines · Rob Herring

From: Rob Herring <[email protected]>

Convert the gic interrupt controller to use irq domains in preparation for device-tree binding and MULTI_IRQ. This allows for translation between GIC interrupt IDs and Linux irq numbers.

The meaning of irq_offset has changed. It now is just the number of skipped GIC interrupt IDs for the controller. It will be 16 for primary GIC and 32 for secondary GICs.

[...]

arch/arm/common/gic.c static inline unsigned int gic_irq(struct irq_data *d) · line 85

Quoted 7 lines

/*

arch/arm/common/gic.c +1 -1
@@ -81,7 +82,7 @@ static inline unsigned int gic_irq(struct irq_data *d)
  */
 static void gic_mask_irq(struct irq_data *d)
 {
-	u32 mask = 1 << (d->irq % 32);
+	u32 mask = 1 << (gic_irq(d) % 32);

This can probably change simply to d->hwirq if irq_offset is eliminated as I describe below.

Quoted 14 lines
arch/arm/common/gic.c +0 -0
 void __init gic_init(unsigned int gic_nr, unsigned int irq_start,
 	void __iomem *dist_base, void __iomem *cpu_base)
 {
arch/arm/common/gic.c +3 -1
 	struct gic_chip_data *gic;
+	struct irq_domain *domain;
+	int gic_irqs;
 
 	BUG_ON(gic_nr >= MAX_GIC_NR);
 
 	gic = &gic_data[gic_nr];
+	domain = &gic->domain;
 	gic->dist_base = dist_base;
 	gic->cpu_base = cpu_base;
-	gic->irq_offset = (irq_start - 1) & ~31;

arch/arm/common/gic.c static void __init gic_pm_init(struct gic_chip_data *gic) · line 546

Quoted 10 lines
arch/arm/common/gic.c +7 -1
 
-	if (gic_nr == 0)
+	/*
+	 * For primary GICs, skip over SGIs.
+	 * For secondary GICs, skip over PPIs, too.
+	 */
+	if (gic_nr == 0) {
 		gic_cpu_base_addr = cpu_base;
+		gic->irq_offset = 16;
+		irq_start = (irq_start & ~31) + 16;
Quoted 7 lines

With the switch to irq_domain, there should no longer be any need for a ~31 mask on the irq_start number. Yes, you'll want to make sure that it doesn't allocate below irq 16, but the driver should completely use the irq_domain to manage the mapping from linux-irq number to hwirq number. The ~31 mask appears to have been an optimization to quickly calculate hwirq number from the linux one, but that value is now found in irq_data->hwirq.

I started out exactly this way removing irq_offset. The problem is the core irq domain code assumes that hwirq starts at 0, but for the gic it starts at 16 or 32.

It could be fixed in the domain core, but that would effectively mean moving irq_offset into the core.

The masking here is mainly to not break exynos. Everyone else passes in a value of 16-31 which I could just hardcode to 16. So when exynos gets all the interrupts converted over to domains, this can be removed. I looked at it briefly and it doesn't look trivial.

Quoted 7 lines

irq_offset could almost be dropped entirely outside of probe because it no longer matters. Instead, populate the to_irq() hook so that the hwirq value gets set correctly from the outset.

It would be better in general to have a consistent view of what the hwirq number means regardless of if it is a root GIC.

Agreed.

Rob

#4 lore raw

Re: [PATCH 1/2] ARM: gic: add irq_domain support

Quoted 58 lines · Rob Herring
Quoted 49 lines · Grant Likely
Quoted 9 lines · Rob Herring

From: Rob Herring <[email protected]>

Convert the gic interrupt controller to use irq domains in preparation for device-tree binding and MULTI_IRQ. This allows for translation between GIC interrupt IDs and Linux irq numbers.

The meaning of irq_offset has changed. It now is just the number of skipped GIC interrupt IDs for the controller. It will be 16 for primary GIC and 32 for secondary GICs.

[...]

arch/arm/common/gic.c static inline unsigned int gic_irq(struct irq_data *d) · line 85

Quoted 7 lines

/*

arch/arm/common/gic.c +1 -1
@@ -81,7 +82,7 @@ static inline unsigned int gic_irq(struct irq_data *d)
  */
 static void gic_mask_irq(struct irq_data *d)
 {
-	u32 mask = 1 << (d->irq % 32);
+	u32 mask = 1 << (gic_irq(d) % 32);

This can probably change simply to d->hwirq if irq_offset is eliminated as I describe below.

Quoted 14 lines
arch/arm/common/gic.c +0 -0
 void __init gic_init(unsigned int gic_nr, unsigned int irq_start,
 	void __iomem *dist_base, void __iomem *cpu_base)
 {
arch/arm/common/gic.c +3 -1
 	struct gic_chip_data *gic;
+	struct irq_domain *domain;
+	int gic_irqs;
 
 	BUG_ON(gic_nr >= MAX_GIC_NR);
 
 	gic = &gic_data[gic_nr];
+	domain = &gic->domain;
 	gic->dist_base = dist_base;
 	gic->cpu_base = cpu_base;
-	gic->irq_offset = (irq_start - 1) & ~31;

arch/arm/common/gic.c static void __init gic_pm_init(struct gic_chip_data *gic) · line 546

Quoted 10 lines
arch/arm/common/gic.c +7 -1
 
-	if (gic_nr == 0)
+	/*
+	 * For primary GICs, skip over SGIs.
+	 * For secondary GICs, skip over PPIs, too.
+	 */
+	if (gic_nr == 0) {
 		gic_cpu_base_addr = cpu_base;
+		gic->irq_offset = 16;
+		irq_start = (irq_start & ~31) + 16;
Quoted 7 lines

With the switch to irq_domain, there should no longer be any need for a ~31 mask on the irq_start number. Yes, you'll want to make sure that it doesn't allocate below irq 16, but the driver should completely use the irq_domain to manage the mapping from linux-irq number to hwirq number. The ~31 mask appears to have been an optimization to quickly calculate hwirq number from the linux one, but that value is now found in irq_data->hwirq.

Quoted 6 lines

I started out exactly this way removing irq_offset. The problem is the core irq domain code assumes that hwirq starts at 0, but for the gic it starts at 16 or 32.

It could be fixed in the domain core, but that would effectively mean moving irq_offset into the core.

Fix irq_domain. If it cannot handle this trivial case, then it won't be able to handle the complex powerpc use cases.

g.

#5 lore raw

PATCH 2/2 ARM: gic: add OF based initialization

In reply to Rob Herring

From: Rob Herring <[email protected]>

This adds ARM gic interrupt controller initialization using device tree data.

The initialization function is intended to be called by of_irq_init function like this:

const static struct of_device_id irq_match[] = {
	{ .compatible = "arm,cortex-a9-gic", .data = gic_of_init, },
	{}
};

static void __init init_irqs(void)

{
	of_irq_init(irq_match);
}
---
 Documentation/devicetree/bindings/arm/gic.txt |   55 +++++++++++++++++++++
 arch/arm/common/gic.c                         |   64 +++++++++++++++++++++++++
 arch/arm/include/asm/hardware/gic.h           |    1 +
 3 files changed, 120 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/gic.txt
Documentation/devicetree/bindings/arm/gic.txt new file +55 -0
new file mode 100644
index 0000000..52916b4
@@ -0,0 +1,55 @@
+* ARM Generic Interrupt Controller
+
+ARM SMP cores are often associated with a GIC, providing per processor
+interrupts (PPI), shared processor interrupts (SPI) and software
+generated interrupts (SGI).
+
+Primary GIC is attached directly to the CPU and typically has PPIs and SGIs.
+Secondary GICs are cascaded into the upward interrupt controller and do not
+have PPIs or SGIs.
+
+Main node required properties:
+
+- compatible : should be one of:
+	"arm,cortex-a9-gic"
+	"arm,arm11mp-gic"
+- interrupt-controller : Identifies the node as an interrupt controller
+- #interrupt-cells : Specifies the number of cells needed to encode an
+  interrupt source.  The type shall be a <u32> and the value shall be 3.
+
+  The 1st cell is the interrupt type; 0 for SPI interrupts, 1 for PPI
+  interrupts.
+
+  The 2nd cell contains the interrupt number for the interrupt type.
+  SPI interrupts are in the range [0-987].  PPI interrupts are in the
+  range [0-15].
+
+  The 3rd cell is the flags, encoded as follows:
+	bits[3:0] trigger type and level flags.
+		1 = low-to-high edge triggered
+		2 = high-to-low edge triggered
+		4 = active high level-sensitive
+		8 = active low level-sensitive
+	bits[15:8] PPI interrupt cpu mask.  Each bit corresponds to each of
+	the 8 possible cpus attached to the GIC.  A bit set to '1' indicated
+	the interrupt is wired to that CPU.  Only valid for PPI interrupts.
+
+- reg : Specifies base physical address(s) and size of the GIC registers. The
+  first region is the GIC distributor register base and size. The 2nd region is
+  the GIC cpu interface register base and size.
+
+Optional
+- interrupts	: Interrupt source of the parent interrupt controller. Only
+  present on secondary GICs.
+
+Example:
+
+	intc: interrupt-controller at fff11000 {
+		compatible = "arm,cortex-a9-gic";
+		#interrupt-cells = <3>;
+		#address-cells = <1>;
+		interrupt-controller;
+		reg = <0xfff11000 0x1000>,
+		      <0xfff10100 0x100>;
+	};
+
arch/arm/common/gic.c +64 -0

80 lines not on this page · raw message

arch/arm/include/asm/hardware/gic.h +1 -0
index 4893b10..a343c96 100644
@@ -39,6 +39,7 @@ extern void __iomem *gic_cpu_base_addr;
 extern struct irq_chip gic_arch_extn;
 
 void gic_init(unsigned int, unsigned int, void __iomem *, void __iomem *);
+int gic_of_init(struct device_node *node, struct device_node *parent);
 void gic_secondary_init(unsigned int);
 void gic_cascade_irq(unsigned int gic_nr, unsigned int irq);
 void gic_raise_softirq(const struct cpumask *mask, unsigned int irq);
Signature · 1 line
1.7.5.4
#6 lore raw

Re: [PATCH 2/2] ARM: gic: add OF based initialization

Quoted 13 lines · Rob Herring

From: Rob Herring <[email protected]>

This adds ARM gic interrupt controller initialization using device tree data.

The initialization function is intended to be called by of_irq_init function like this:

const static struct of_device_id irq_match[] = {
	{ .compatible = "arm,cortex-a9-gic", .data = gic_of_init, },
	{}
};
Quoted 6 lines

static void __init init_irqs(void)

{
	of_irq_init(irq_match);
}

Looks good.

g.

Reviewed-by on the v1 series
#7 lore raw
In reply to Rob Herring
Quoted 7 lines · Rob Herring

From: Rob Herring <[email protected]>

With recent patches for VIC device tree support and GIC MULTI_IRQ support, it is evident that the irq domain support should be moved into the core GIC code and not be dependent on the DT initialization. So I've split things up into 2 patches separating irq_domain and device-tree code.

Quoted 5 lines

This should only change the internals of the GIC code. The binding is unchanged from the previous version. Yeah!

The full series is available here. This includes Russell's devel-stable and for-next branches and Nico's vmalloc branch.

Looks good to me!