Skip to contentSkip to conversation

ARM: Common GIC entry macro code V2

PATCH New Magnus Damm 5 participants 9 messages Nov 11, 2010 – Nov 12, 2010

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

---

Changes since V1:
- dropped msm as requested by Abhijeet, thank you.

This patch is almost identical to 6284/1 in the RMKs patch tracker. The main difference is that this patch requires each platform to specify base address using the macro get_irqnr_preable. This patch includes tegra, but drops omap support due to multi-omap complexity.

Thanks to Shilimkar, Santosh and Srinidhi Kasagar for feedback.

 arch/arm/include/asm/hardware/entry-macro-gic.S   |   68 +++++++++++++++++++++
 arch/arm/mach-cns3xxx/include/mach/entry-macro.S  |   61 ------------------
 arch/arm/mach-realview/include/mach/entry-macro.S |   60 ------------------
 arch/arm/mach-tegra/include/mach/entry-macro.S    |   64 -------------------
 arch/arm/mach-ux500/include/mach/entry-macro.S    |   67 --------------------
 arch/arm/mach-vexpress/include/mach/entry-macro.S |   57 -----------------
 6 files changed, 73 insertions(+), 304 deletions(-)
arch/arm/include/asm/hardware/entry-macro-gic.S new file +68 -0
@@ -0,0 +1,68 @@
+/*
+ * arch/arm/include/asm/hardware/entry-macro-gic.S
+ *
+ * Low-level IRQ helper macros for GIC
+ *
+ * This file is licensed under  the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <asm/hardware/gic.h>
+
+/*
+ * The interrupt numbering scheme is defined in the
+ * interrupt controller spec.  To wit:
+ *
+ * Interrupts 0-15 are IPI
+ * 16-28 are reserved
+ * 29-31 are local.  We allow 30 to be used for the watchdog.
+ * 32-1020 are global
+ * 1021-1022 are reserved
+ * 1023 is "spurious" (no interrupt)
+ *
+ * For now, we ignore all local interrupts so only return an interrupt if it's
+ * between 30 and 1020.  The test_for_ipi routine below will pick up on IPIs.
+ *
+ * A simple read from the controller will tell us the number of the highest
+ * priority enabled interrupt.  We then just need to check whether it is in the
+ * valid range for an IRQ (30-1020 inclusive).
+ */
+
+	.macro  get_irqnr_and_base, irqnr, irqstat, base, tmp
+
+	ldr     \irqstat, [\base, #GIC_CPU_INTACK]
+	/* bits 12-10 = src CPU, 9-0 = int # */
+
+	ldr	\tmp, =1021
+	bic     \irqnr, \irqstat, #0x1c00
+	cmp     \irqnr, #29
+	cmpcc	\irqnr, \irqnr
+	cmpne	\irqnr, \tmp
+	cmpcs	\irqnr, \irqnr
+	.endm
+
+/* We assume that irqstat (the raw value of the IRQ acknowledge
+ * register) is preserved from the macro above.
+ * If there is an IPI, we immediately signal end of interrupt on the
+ * controller, since this requires the original irqstat value which
+ * we won't easily be able to recreate later.
+ */
+
+	.macro test_for_ipi, irqnr, irqstat, base, tmp
+	bic	\irqnr, \irqstat, #0x1c00
+	cmp	\irqnr, #16
+	strcc	\irqstat, [\base, #GIC_CPU_EOI]
+	cmpcs	\irqnr, \irqnr
+	.endm
+
+/* As above, this assumes that irqstat and base are preserved.. */
+
+	.macro test_for_ltirq, irqnr, irqstat, base, tmp
+	bic	\irqnr, \irqstat, #0x1c00
+	mov 	\tmp, #0
+	cmp	\irqnr, #29
+	moveq	\tmp, #1
+	streq	\irqstat, [\base, #GIC_CPU_EOI]
+	cmp	\tmp, #0
+	.endm
arch/arm/mach-cns3xxx/include/mach/entry-macro.S +1 -60

70 lines not on this page · raw message

arch/arm/mach-realview/include/mach/entry-macro.S +1 -59

69 lines not on this page · raw message

arch/arm/mach-tegra/include/mach/entry-macro.S +1 -63

76 lines not on this page · raw message

arch/arm/mach-ux500/include/mach/entry-macro.S +1 -66

76 lines not on this page · raw message

arch/arm/mach-vexpress/include/mach/entry-macro.S +1 -56

63 lines not on this page · raw message

#1 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Hi,

Can you include the "arch/arm/mach-s5pv310/include/mach/entry-macro.S" also?

Thank you,
Kyungmin Park

Quoted 496 lines · Magnus Damm

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

---

?Changes since V1:
?- dropped msm as requested by Abhijeet, thank you.

?This patch is almost identical to 6284/1 in the RMKs patch ?tracker. The main difference is that this patch requires ?each platform to specify base address using the macro ?get_irqnr_preable. This patch includes tegra, but drops ?omap support due to multi-omap complexity.

?Thanks to Shilimkar, Santosh and Srinidhi Kasagar for feedback.

?arch/arm/include/asm/hardware/entry-macro-gic.S ? | ? 68 +++++++++++++++++++++
?arch/arm/mach-cns3xxx/include/mach/entry-macro.S ?| ? 61 ------------------
?arch/arm/mach-realview/include/mach/entry-macro.S | ? 60 ------------------
?arch/arm/mach-tegra/include/mach/entry-macro.S ? ?| ? 64 -------------------
?arch/arm/mach-ux500/include/mach/entry-macro.S ? ?| ? 67 --------------------
?arch/arm/mach-vexpress/include/mach/entry-macro.S | ? 57 -----------------

?6 files changed, 73 insertions(+), 304 deletions(-)

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? ?2010-11-10 16:06:26.000000000 +0900 new file +68 -0
@@ -0,0 +1,68 @@
+/*
+ * arch/arm/include/asm/hardware/entry-macro-gic.S
+ *
+ * Low-level IRQ helper macros for GIC
+ *
+ * This file is licensed under ?the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <asm/hardware/gic.h>
+
+/*
+ * The interrupt numbering scheme is defined in the
+ * interrupt controller spec. ?To wit:
+ *
+ * Interrupts 0-15 are IPI
+ * 16-28 are reserved
+ * 29-31 are local. ?We allow 30 to be used for the watchdog.
+ * 32-1020 are global
+ * 1021-1022 are reserved
+ * 1023 is "spurious" (no interrupt)
+ *
+ * For now, we ignore all local interrupts so only return an interrupt if it's
+ * between 30 and 1020. ?The test_for_ipi routine below will pick up on IPIs.
+ *
+ * A simple read from the controller will tell us the number of the highest
+ * priority enabled interrupt. ?We then just need to check whether it is in the
+ * valid range for an IRQ (30-1020 inclusive).
+ */
+
+ ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
+
+ ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK]
+ ? ? ? /* bits 12-10 = src CPU, 9-0 = int # */
+
+ ? ? ? ldr ? ? \tmp, =1021
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? cmpcc ? \irqnr, \irqnr
+ ? ? ? cmpne ? \irqnr, \tmp
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* We assume that irqstat (the raw value of the IRQ acknowledge
+ * register) is preserved from the macro above.
+ * If there is an IPI, we immediately signal end of interrupt on the
+ * controller, since this requires the original irqstat value which
+ * we won't easily be able to recreate later.
+ */
+
+ ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #16
+ ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* As above, this assumes that irqstat and base are preserved.. */
+
+ ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? mov ? ? \tmp, #0
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? moveq ? \tmp, #1
+ ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmp ? ? \tmp, #0
+ ? ? ? .endm
arch/arm/mach-cns3xxx/include/mach/entry-macro.S ? ? ? 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-cns3xxx/include/mach/entry-macro.S +0 -0
@@ -9,7 +9,7 @@

?*/

?#include <mach/hardware.h>

arch/arm/mach-cns3xxx/include/mach/entry-macro.S ? ? ? 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-cns3xxx/include/mach/entry-macro.S +1 -1
-#include <asm/hardware/gic.h>
+#include <asm/hardware/entry-macro-gic.S>

? ? ? ? ? ? ? ?.macro ?disable_fiq
? ? ? ? ? ? ? ?.endm

arch/arm/mach-cns3xxx/include/mach/entry-macro.S ? ? ? 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-cns3xxx/include/mach/entry-macro.S +0 -0
@@ -21,62 +21,3 @@
 

? ? ? ? ? ? ? ?.macro ?arch_ret_to_user, tmp1, tmp2 ? ? ? ? ? ? ? ?.endm

arch/arm/mach-cns3xxx/include/mach/entry-macro.S ? ? ? 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-cns3xxx/include/mach/entry-macro.S +0 -59
-
- ? ? ? ? ? ? ? /*
- ? ? ? ? ? ? ? ?* The interrupt numbering scheme is defined in the
- ? ? ? ? ? ? ? ?* interrupt controller spec. ?To wit:
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* Interrupts 0-15 are IPI
- ? ? ? ? ? ? ? ?* 16-28 are reserved
- ? ? ? ? ? ? ? ?* 29-31 are local. ?We allow 30 to be used for the watchdog.
- ? ? ? ? ? ? ? ?* 32-1020 are global
- ? ? ? ? ? ? ? ?* 1021-1022 are reserved
- ? ? ? ? ? ? ? ?* 1023 is "spurious" (no interrupt)
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* For now, we ignore all local interrupts so only return an interrupt if it's
- ? ? ? ? ? ? ? ?* between 30 and 1020. ?The test_for_ipi routine below will pick up on IPIs.
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* A simple read from the controller will tell us the number of the highest
- ? ? ? ? ? ? ? ? * priority enabled interrupt. ?We then just need to check whether it is in the
- ? ? ? ? ? ? ? ?* valid range for an IRQ (30-1020 inclusive).
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
-
- ? ? ? ? ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK] /* bits 12-10 = src CPU, 9-0 = int # */
-
- ? ? ? ? ? ? ? ldr ? ? \tmp, =1021
-
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
-
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? cmpcc ? \irqnr, \irqnr
- ? ? ? ? ? ? ? cmpne ? \irqnr, \tmp
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
-
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* We assume that irqstat (the raw value of the IRQ acknowledge
- ? ? ? ? ? ? ? ?* register) is preserved from the macro above.
- ? ? ? ? ? ? ? ?* If there is an IPI, we immediately signal end of interrupt on the
- ? ? ? ? ? ? ? ?* controller, since this requires the original irqstat value which
- ? ? ? ? ? ? ? ?* we won't easily be able to recreate later.
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #16
- ? ? ? ? ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* As above, this assumes that irqstat and base are preserved.. */
-
- ? ? ? ? ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? mov ? ? \tmp, #0
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? moveq ? \tmp, #1
- ? ? ? ? ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmp ? ? \tmp, #0
- ? ? ? ? ? ? ? .endm
arch/arm/mach-realview/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-realview/include/mach/entry-macro.S +0 -0
@@ -8,7 +8,7 @@

?* warranty of any kind, whether express or implied. ?*/
?#include <mach/hardware.h>

arch/arm/mach-realview/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-realview/include/mach/entry-macro.S +1 -1
-#include <asm/hardware/gic.h>
+#include <asm/hardware/entry-macro-gic.S>

? ? ? ? ? ? ? ?.macro ?disable_fiq
? ? ? ? ? ? ? ?.endm

arch/arm/mach-realview/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-realview/include/mach/entry-macro.S +0 -0
@@ -21,61 +21,3 @@

? ? ? ? ? ? ? ?.macro ?arch_ret_to_user, tmp1, tmp2 ? ? ? ? ? ? ? ?.endm

- ? ? ? ? ? ? ? /*
- ? ? ? ? ? ? ? ?* The interrupt numbering scheme is defined in the
- ? ? ? ? ? ? ? ?* interrupt controller spec. ?To wit:
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* Interrupts 0-15 are IPI
- ? ? ? ? ? ? ? ?* 16-28 are reserved
- ? ? ? ? ? ? ? ?* 29-31 are local. ?We allow 30 to be used for the watchdog.
- ? ? ? ? ? ? ? ?* 32-1020 are global
- ? ? ? ? ? ? ? ?* 1021-1022 are reserved
- ? ? ? ? ? ? ? ?* 1023 is "spurious" (no interrupt)
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* For now, we ignore all local interrupts so only return an interrupt if it's
- ? ? ? ? ? ? ? ?* between 30 and 1020. ?The test_for_ipi routine below will pick up on IPIs.
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* A simple read from the controller will tell us the number of the highest
- ? ? ? ? ? ? ? ? * priority enabled interrupt. ?We then just need to check whether it is in the
- ? ? ? ? ? ? ? ?* valid range for an IRQ (30-1020 inclusive).
- ? ? ? ? ? ? ? ?*/

arch/arm/mach-realview/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-realview/include/mach/entry-macro.S +0 -40
-
- ? ? ? ? ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
-
- ? ? ? ? ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK] /* bits 12-10 = src CPU, 9-0 = int # */
-
- ? ? ? ? ? ? ? ldr ? ? \tmp, =1021
-
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
-
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? cmpcc ? \irqnr, \irqnr
- ? ? ? ? ? ? ? cmpne ? \irqnr, \tmp
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
-
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* We assume that irqstat (the raw value of the IRQ acknowledge
- ? ? ? ? ? ? ? ?* register) is preserved from the macro above.
- ? ? ? ? ? ? ? ?* If there is an IPI, we immediately signal end of interrupt on the
- ? ? ? ? ? ? ? ?* controller, since this requires the original irqstat value which
- ? ? ? ? ? ? ? ?* we won't easily be able to recreate later.
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #16
- ? ? ? ? ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* As above, this assumes that irqstat and base are preserved.. */
-
- ? ? ? ? ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? mov ? ? \tmp, #0
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? moveq ? \tmp, #1
- ? ? ? ? ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmp ? ? \tmp, #0
- ? ? ? ? ? ? ? .endm
arch/arm/mach-tegra/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-tegra/include/mach/entry-macro.S +1 -1
@@ -17,7 +17,7 @@
 
 ?#if defined(CONFIG_ARM_GIC)
 
-#include <asm/hardware/gic.h>
+#include <asm/hardware/entry-macro-gic.S>
 

? ? ? ?/* Uses the GIC interrupt controller built into the cpu */ ?#define ICTRL_BASE (IO_CPU_VIRT + 0x100)

arch/arm/mach-tegra/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-tegra/include/mach/entry-macro.S +0 -0
@@ -32,68 +32,6 @@
 

? ? ? ?.macro ?arch_ret_to_user, tmp1, tmp2
? ? ? ?.endm

arch/arm/mach-tegra/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-tegra/include/mach/entry-macro.S +0 -62
-
- ? ? ? /*
- ? ? ? ?* The interrupt numbering scheme is defined in the
- ? ? ? ?* interrupt controller spec. ?To wit:
- ? ? ? ?*
- ? ? ? ?* Interrupts 0-15 are IPI
- ? ? ? ?* 16-28 are reserved
- ? ? ? ?* 29-31 are local. ?We allow 30 to be used for the watchdog.
- ? ? ? ?* 32-1020 are global
- ? ? ? ?* 1021-1022 are reserved
- ? ? ? ?* 1023 is "spurious" (no interrupt)
- ? ? ? ?*
- ? ? ? ?* For now, we ignore all local interrupts so only return an interrupt
- ? ? ? ?* if it's between 30 and 1020. ?The test_for_ipi routine below will
- ? ? ? ?* pick up on IPIs.
- ? ? ? ?*
- ? ? ? ?* A simple read from the controller will tell us the number of the
- ? ? ? ?* highest priority enabled interrupt. ?We then just need to check
- ? ? ? ?* whether it is in the valid range for an IRQ (30-1020 inclusive).
- ? ? ? ?*/
-
- ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
-
- ? ? ? /* bits 12-10 = src CPU, 9-0 = int # */
- ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK]
-
- ? ? ? ldr ? ? ? ? ? ? \tmp, =1021
-
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
-
- ? ? ? cmp ? ? \irqnr, #29
- ? ? ? cmpcc ? \irqnr, \irqnr
- ? ? ? cmpne ? \irqnr, \tmp
- ? ? ? cmpcs ? \irqnr, \irqnr
-
- ? ? ? .endm
-
- ? ? ? /* We assume that irqstat (the raw value of the IRQ acknowledge
- ? ? ? ?* register) is preserved from the macro above.
- ? ? ? ?* If there is an IPI, we immediately signal end of interrupt on the
- ? ? ? ?* controller, since this requires the original irqstat value which
- ? ? ? ?* we won't easily be able to recreate later.
- ? ? ? ?*/
-
- ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? cmp ? ? \irqnr, #16
- ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? .endm
-
- ? ? ? /* As above, this assumes that irqstat and base are preserved.. */
-
- ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? mov ? ? \tmp, #0
- ? ? ? cmp ? ? \irqnr, #29
- ? ? ? moveq ? \tmp, #1
- ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? cmp ? ? \tmp, #0
- ? ? ? .endm
-

?#else
? ? ? ?/* legacy interrupt controller for AP16 */ ? ? ? ?.macro ?disable_fiq

arch/arm/mach-ux500/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-ux500/include/mach/entry-macro.S +0 -0
@@ -11,7 +11,7 @@

?* warranty of any kind, whether express or implied. ?*/
?#include <mach/hardware.h>

arch/arm/mach-ux500/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-ux500/include/mach/entry-macro.S +1 -1
-#include <asm/hardware/gic.h>
+#include <asm/hardware/entry-macro-gic.S>

? ? ? ? ? ? ? ?.macro ?disable_fiq
? ? ? ? ? ? ? ?.endm

arch/arm/mach-ux500/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-ux500/include/mach/entry-macro.S +0 -0
@@ -22,68 +22,3 @@
 

? ? ? ? ? ? ? ?.macro ?arch_ret_to_user, tmp1, tmp2 ? ? ? ? ? ? ? ?.endm

arch/arm/mach-ux500/include/mach/entry-macro.S 2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-ux500/include/mach/entry-macro.S +0 -65
-
- ? ? ? ? ? ? ? /*
- ? ? ? ? ? ? ? ?* The interrupt numbering scheme is defined in the
- ? ? ? ? ? ? ? ?* interrupt controller spec. ?To wit:
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* Interrupts 0-15 are IPI
- ? ? ? ? ? ? ? ?* 16-28 are reserved
- ? ? ? ? ? ? ? ?* 29-31 are local. ?We allow 30 to be used for the watchdog.
- ? ? ? ? ? ? ? ?* 32-1020 are global
- ? ? ? ? ? ? ? ?* 1021-1022 are reserved
- ? ? ? ? ? ? ? ?* 1023 is "spurious" (no interrupt)
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* For now, we ignore all local interrupts so only return an
- ? ? ? ? ? ? ? ?* interrupt if it's between 30 and 1020. The test_for_ipi
- ? ? ? ? ? ? ? ?* routine below will pick up on IPIs.
- ? ? ? ? ? ? ? ?*
- ? ? ? ? ? ? ? ?* A simple read from the controller will tell us the number
- ? ? ? ? ? ? ? ?* of the highest priority enabled interrupt. We then just
- ? ? ? ? ? ? ? ?* need to check whether it is in the valid range for an
- ? ? ? ? ? ? ? ?* IRQ (30-1020 inclusive).
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
-
- ? ? ? ? ? ? ? /* bits 12-10 = src CPU, 9-0 = int # */
- ? ? ? ? ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK]
-
- ? ? ? ? ? ? ? ldr ? ? \tmp, =1021
-
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
-
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? cmpcc ? \irqnr, \irqnr
- ? ? ? ? ? ? ? cmpne ? \irqnr, \tmp
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
-
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* We assume that irqstat (the raw value of the IRQ
- ? ? ? ? ? ? ? ?* acknowledge register) is preserved from the macro above.
- ? ? ? ? ? ? ? ?* If there is an IPI, we immediately signal end of
- ? ? ? ? ? ? ? ?* interrupt on the controller, since this requires the
- ? ? ? ? ? ? ? ?* original irqstat value which we won't easily be able
- ? ? ? ? ? ? ? ?* to recreate later.
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #16
- ? ? ? ? ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? ? ? ? ? .endm
-
- ? ? ? ? ? ? ? /* As above, this assumes that irqstat and base
- ? ? ? ? ? ? ? ?* are preserved..
- ? ? ? ? ? ? ? ?*/
-
- ? ? ? ? ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
- ? ? ? ? ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? ? ? ? ? mov ? ? \tmp, #0
- ? ? ? ? ? ? ? cmp ? ? \irqnr, #29
- ? ? ? ? ? ? ? moveq ? \tmp, #1
- ? ? ? ? ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? ? ? ? ? cmp ? ? \tmp, #0
- ? ? ? ? ? ? ? .endm
arch/arm/mach-vexpress/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-vexpress/include/mach/entry-macro.S +1 -1
@@ -1,4 +1,4 @@
-#include <asm/hardware/gic.h>
+#include <asm/hardware/entry-macro-gic.S>
 

? ? ? ?.macro ?disable_fiq
? ? ? ?.endm

arch/arm/mach-vexpress/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-vexpress/include/mach/entry-macro.S +0 -0
@@ -10,58 +10,3 @@
 

? ? ? ?.macro ?arch_ret_to_user, tmp1, tmp2
? ? ? ?.endm

arch/arm/mach-vexpress/include/mach/entry-macro.S ? ? ?2010-11-10 16:06:25.000000000 +0900 renamed from arch/arm/mach-vexpress/include/mach/entry-macro.S +0 -55
-
- ? ? ? /*
- ? ? ? ?* The interrupt numbering scheme is defined in the
- ? ? ? ?* interrupt controller spec. ?To wit:
- ? ? ? ?*
- ? ? ? ?* Interrupts 0-15 are IPI
- ? ? ? ?* 16-28 are reserved
- ? ? ? ?* 29-31 are local. ?We allow 30 to be used for the watchdog.
- ? ? ? ?* 32-1020 are global
- ? ? ? ?* 1021-1022 are reserved
- ? ? ? ?* 1023 is "spurious" (no interrupt)
- ? ? ? ?*
- ? ? ? ?* For now, we ignore all local interrupts so only return an interrupt if it's
- ? ? ? ?* between 30 and 1020. ?The test_for_ipi routine below will pick up on IPIs.
- ? ? ? ?*
- ? ? ? ?* A simple read from the controller will tell us the number of the highest
- ? ? ? ?* priority enabled interrupt. ?We then just need to check whether it is in the
- ? ? ? ?* valid range for an IRQ (30-1020 inclusive).
- ? ? ? ?*/
-
- ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
- ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK] /* bits 12-10 = src CPU, 9-0 = int # */
- ? ? ? ldr ? ? \tmp, =1021
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? cmp ? ? \irqnr, #29
- ? ? ? cmpcc ? \irqnr, \irqnr
- ? ? ? cmpne ? \irqnr, \tmp
- ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? .endm
-
- ? ? ? /* We assume that irqstat (the raw value of the IRQ acknowledge
- ? ? ? ?* register) is preserved from the macro above.
- ? ? ? ?* If there is an IPI, we immediately signal end of interrupt on the
- ? ? ? ?* controller, since this requires the original irqstat value which
- ? ? ? ?* we won't easily be able to recreate later.
- ? ? ? ?*/
-
- ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? cmp ? ? \irqnr, #16
- ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? cmpcs ? \irqnr, \irqnr
- ? ? ? .endm
-
- ? ? ? /* As above, this assumes that irqstat and base are preserved.. */
-
- ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
- ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
- ? ? ? mov ? ? \tmp, #0
- ? ? ? cmp ? ? \irqnr, #29
- ? ? ? moveq ? \tmp, #1
- ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
- ? ? ? cmp ? ? \tmp, #0
- ? ? ? .endm
-

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

#2 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Quoted 3 lines · Kyungmin Park

Hi,

Hi,

Quoted 1 line

Can you include the "arch/arm/mach-s5pv310/include/mach/entry-macro.S"

also?

Quoted 1 line

In conclusion, can't it now.
Because according to current interrupt number structure of Samsung SoCs, irqnr macro need to add S5P_IRQ_OFFSET, 32.

Or have to add #ifdef in there...But don't like it.

Quoted 1 line

On Thu, Nov 11, 2010 at 4:49 PM, Magnus Damm <[email protected]>

Quoted 35 lines · wrote:
Quoted 35 lines

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

---

?Changes since V1:
?- dropped msm as requested by Abhijeet, thank you.

?This patch is almost identical to 6284/1 in the RMKs patch ?tracker. The main difference is that this patch requires ?each platform to specify base address using the macro ?get_irqnr_preable. This patch includes tegra, but drops ?omap support due to multi-omap complexity.

Quoted 5 lines
Quoted 3 lines

?Thanks to Shilimkar, Santosh and Srinidhi Kasagar for feedback.

?arch/arm/include/asm/hardware/entry-macro-gic.S ? | ? 68

+++++++++++++++++++++

Quoted 1 line

?arch/arm/mach-cns3xxx/include/mach/entry-macro.S ?| ? 61

-----------------

Quoted 2 lines

-

Quoted 1 line

?arch/arm/mach-realview/include/mach/entry-macro.S | ? 60

-----------------

Quoted 2 lines

-

Quoted 1 line

?arch/arm/mach-tegra/include/mach/entry-macro.S ? ?| ? 64

-----------------

Quoted 2 lines

--

Quoted 1 line

?arch/arm/mach-ux500/include/mach/entry-macro.S ? ?| ? 67

-----------------

Quoted 2 lines

---

Quoted 1 line

?arch/arm/mach-vexpress/include/mach/entry-macro.S | ? 57

-----------------

Quoted 4 lines

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ?

Quoted 4 lines

?6 files changed, 73 insertions(+), 304 deletions(-)

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +0 -0

?2010-11-10

Quoted 18 lines

16:06:26.000000000 +0900

Quoted 17 lines
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +16 -0
@@ -0,0 +1,68 @@
+/*
+ * arch/arm/include/asm/hardware/entry-macro-gic.S
+ *
+ * Low-level IRQ helper macros for GIC
+ *
+ * This file is licensed under ?the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <asm/hardware/gic.h>
+
+/*
+ * The interrupt numbering scheme is defined in the
+ * interrupt controller spec. ?To wit:
+ *
Quoted 8 lines

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? line 24

Quoted 8 lines
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +8 -0
+ * Interrupts 0-15 are IPI
+ * 16-28 are reserved
+ * 29-31 are local. ?We allow 30 to be used for the watchdog.
+ * 32-1020 are global
+ * 1021-1022 are reserved
+ * 1023 is "spurious" (no interrupt)
+ *
+ * For now, we ignore all local interrupts so only return an interrupt

if

Quoted 7 lines

it's

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? line 25

Quoted 1 line
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +1 -0
+ * between 30 and 1020. ?The test_for_ipi routine below will pick up on

IPIs.

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? line 27

Quoted 2 lines
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +2 -0
+ *
+ * A simple read from the controller will tell us the number of the

highest

Quoted 1 line
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +1 -0
+ * priority enabled interrupt. ?We then just need to check whether it

is

Quoted 33 lines

in the

Quoted 32 lines
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +32 -0
+ * valid range for an IRQ (30-1020 inclusive).
+ */
+
+ ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
+
+ ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK]
+ ? ? ? /* bits 12-10 = src CPU, 9-0 = int # */
+
+ ? ? ? ldr ? ? \tmp, =1021
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? cmpcc ? \irqnr, \irqnr
+ ? ? ? cmpne ? \irqnr, \tmp
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* We assume that irqstat (the raw value of the IRQ acknowledge
+ * register) is preserved from the macro above.
+ * If there is an IPI, we immediately signal end of interrupt on the
+ * controller, since this requires the original irqstat value which
+ * we won't easily be able to recreate later.
+ */
+
+ ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #16
+ ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* As above, this assumes that irqstat and base are preserved.. */
+
Quoted 8 lines

arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? line 68

Quoted 8 lines
arch/arm/include/asm/hardware/entry-macro-gic.S ? ? ? new file +8 -0
+ ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? mov ? ? \tmp, #0
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? moveq ? \tmp, #1
+ ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmp ? ? \tmp, #0
+ ? ? ? .endm

(snip)

Thanks.

Best regards,
Kgene.
--
Kukjin Kim <[email protected]>, Senior Engineer, SW Solution Development Team, Samsung Electronics Co., Ltd.

#3 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Quoted 6 lines · Kukjin Kim
Quoted 3 lines · Kyungmin Park

Hi,

Hi,

Quoted 8 lines
Quoted 1 line

Can you include the "arch/arm/mach-s5pv310/include/mach/entry-macro.S"

also?

Quoted 1 line

In conclusion, can't it now.
Because according to current interrupt number structure of Samsung SoCs, irqnr macro need to add S5P_IRQ_OFFSET, 32.

Or have to add #ifdef in there...But don't like it.

I wonder why Samsung goes another way, use the common codes. if required, modify to use the common codes.

Thank you,
Kyungmin Park

Quoted 149 lines
Quoted 2 lines

On Thu, Nov 11, 2010 at 4:49 PM, Magnus Damm <[email protected]>

Quoted 35 lines · wrote:
Quoted 35 lines

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

---

?Changes since V1:
?- dropped msm as requested by Abhijeet, thank you.

?This patch is almost identical to 6284/1 in the RMKs patch ?tracker. The main difference is that this patch requires ?each platform to specify base address using the macro ?get_irqnr_preable. This patch includes tegra, but drops ?omap support due to multi-omap complexity.

Quoted 5 lines
Quoted 3 lines

?Thanks to Shilimkar, Santosh and Srinidhi Kasagar for feedback.

?arch/arm/include/asm/hardware/entry-macro-gic.S ? | ? 68

+++++++++++++++++++++

Quoted 1 line

?arch/arm/mach-cns3xxx/include/mach/entry-macro.S ?| ? 61

-----------------

Quoted 2 lines

-

Quoted 1 line

?arch/arm/mach-realview/include/mach/entry-macro.S | ? 60

-----------------

Quoted 2 lines

-

Quoted 1 line

?arch/arm/mach-tegra/include/mach/entry-macro.S ? ?| ? 64

-----------------

Quoted 2 lines

--

Quoted 1 line

?arch/arm/mach-ux500/include/mach/entry-macro.S ? ?| ? 67

-----------------

Quoted 2 lines

---

Quoted 1 line

?arch/arm/mach-vexpress/include/mach/entry-macro.S | ? 57

-----------------

Quoted 4 lines

arch/arm/include/asm/hardware/entry-macro-gic.S

Quoted 4 lines

?6 files changed, 73 insertions(+), 304 deletions(-)

arch/arm/include/asm/hardware/entry-macro-gic.S new file +0 -0

?2010-11-10

Quoted 18 lines

16:06:26.000000000 +0900

Quoted 17 lines
arch/arm/include/asm/hardware/entry-macro-gic.S new file +16 -0
@@ -0,0 +1,68 @@
+/*
+ * arch/arm/include/asm/hardware/entry-macro-gic.S
+ *
+ * Low-level IRQ helper macros for GIC
+ *
+ * This file is licensed under ?the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <asm/hardware/gic.h>
+
+/*
+ * The interrupt numbering scheme is defined in the
+ * interrupt controller spec. ?To wit:
+ *
Quoted 8 lines

arch/arm/include/asm/hardware/entry-macro-gic.S

Quoted 8 lines
arch/arm/include/asm/hardware/entry-macro-gic.S new file +8 -0
+ * Interrupts 0-15 are IPI
+ * 16-28 are reserved
+ * 29-31 are local. ?We allow 30 to be used for the watchdog.
+ * 32-1020 are global
+ * 1021-1022 are reserved
+ * 1023 is "spurious" (no interrupt)
+ *
+ * For now, we ignore all local interrupts so only return an interrupt

if

Quoted 7 lines

it's

arch/arm/include/asm/hardware/entry-macro-gic.S line 25

Quoted 1 line
arch/arm/include/asm/hardware/entry-macro-gic.S new file +1 -0
+ * between 30 and 1020. ?The test_for_ipi routine below will pick up on

IPIs.

arch/arm/include/asm/hardware/entry-macro-gic.S line 27

Quoted 2 lines
arch/arm/include/asm/hardware/entry-macro-gic.S new file +2 -0
+ *
+ * A simple read from the controller will tell us the number of the

highest

Quoted 1 line
arch/arm/include/asm/hardware/entry-macro-gic.S new file +1 -0
+ * priority enabled interrupt. ?We then just need to check whether it

is

Quoted 33 lines

in the

Quoted 32 lines
arch/arm/include/asm/hardware/entry-macro-gic.S new file +32 -0
+ * valid range for an IRQ (30-1020 inclusive).
+ */
+
+ ? ? ? .macro ?get_irqnr_and_base, irqnr, irqstat, base, tmp
+
+ ? ? ? ldr ? ? \irqstat, [\base, #GIC_CPU_INTACK]
+ ? ? ? /* bits 12-10 = src CPU, 9-0 = int # */
+
+ ? ? ? ldr ? ? \tmp, =1021
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? cmpcc ? \irqnr, \irqnr
+ ? ? ? cmpne ? \irqnr, \tmp
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* We assume that irqstat (the raw value of the IRQ acknowledge
+ * register) is preserved from the macro above.
+ * If there is an IPI, we immediately signal end of interrupt on the
+ * controller, since this requires the original irqstat value which
+ * we won't easily be able to recreate later.
+ */
+
+ ? ? ? .macro test_for_ipi, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? cmp ? ? \irqnr, #16
+ ? ? ? strcc ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmpcs ? \irqnr, \irqnr
+ ? ? ? .endm
+
+/* As above, this assumes that irqstat and base are preserved.. */
+
Quoted 8 lines

arch/arm/include/asm/hardware/entry-macro-gic.S line 68

Quoted 8 lines
arch/arm/include/asm/hardware/entry-macro-gic.S new file +8 -0
+ ? ? ? .macro test_for_ltirq, irqnr, irqstat, base, tmp
+ ? ? ? bic ? ? \irqnr, \irqstat, #0x1c00
+ ? ? ? mov ? ? \tmp, #0
+ ? ? ? cmp ? ? \irqnr, #29
+ ? ? ? moveq ? \tmp, #1
+ ? ? ? streq ? \irqstat, [\base, #GIC_CPU_EOI]
+ ? ? ? cmp ? ? \tmp, #0
+ ? ? ? .endm

(snip)

Thanks.

Best regards,
Kgene.
--
Kukjin Kim <[email protected]>, Senior Engineer, SW Solution Development Team, Samsung Electronics Co., Ltd.

#4 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Quoted 6 lines · Kyungmin Park
Quoted 5 lines · Kukjin Kim
Quoted 2 lines · Kyungmin Park

Can you include the "arch/arm/mach-s5pv310/include/mach/entry-macro.S"

also?

Quoted 1 line
Quoted 8 lines
Quoted 5 lines

In conclusion, can't it now.
Because according to current interrupt number structure of Samsung SoCs, irqnr macro need to add S5P_IRQ_OFFSET, 32.

Or have to add #ifdef in there...But don't like it.

I wonder why Samsung goes another way, use the common codes. if required, modify to use the common codes.

Yes, it would be nice if Samsung could use the common code as well. I will however ignore this for now to focus on getting something merged that most people can use. The Samsung bits can be dealt with incrementally.

Thanks,

/ magnus

#5 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

In reply to Magnus Damm
Quoted 15 lines · Magnus Damm

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

Quoted 6 lines

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

I'm a little disappoint you didn't CC someone MSM related on this. Your patch is small enough that you can CC everyone effected. Here's get_maintainer.pl output on it,

Russell King <[email protected]> Anton Vorontsov <[email protected]> Srinidhi Kasagar <[email protected]> Linus Walleij <[email protected]> Alessandro Rubini <[email protected]> linux-arm-kernel at lists.infradead.org linux-kernel at vger.kernel.org

At least those people should be getting this. Also, why can't you break this up per sub-architecture ?

Daniel

Signature · 2 lines
Sent by a consultant of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.
#6 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Quoted 27 lines · Daniel Walker
Quoted 15 lines · Magnus Damm

From: Magnus Damm <[email protected]>

This patch is the identical GIC demux implementation merge V2. Instead of implementing same code over and over simply share it in entry-macro-gic.S. The shared code is based on the realview implementation.

Each GIC demux instance still has to setup the base address of the controller using the get_irqnr_preamble macro. The rest of the GIC specific code can be shared.

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

Quoted 6 lines

The s5pv310 subarch is excluded due to the special "addne" instruction in the get_irqnr_and_base macro.

The msm subarch is excluded as well since it treats PPIs differently than other platforms and needs a different version of the get_irqnr_and_base macro.

I'm a little disappoint you didn't CC someone MSM related on this. Your patch is small enough that you can CC everyone effected. Here's get_maintainer.pl output on it,

Quoted 7 lines

Russell King <[email protected]> Anton Vorontsov <[email protected]> Srinidhi Kasagar <[email protected]> Linus Walleij <[email protected]> Alessandro Rubini <[email protected]> linux-arm-kernel at lists.infradead.org linux-kernel at vger.kernel.org

Hi Daniel,

Sorry to disappoint you, will use get_maintainer.pl and include people from former versions when I send out next version.

Quoted 2 lines

At least those people should be getting this. Also, why can't you break this up per sub-architecture ?

maintainer
#7 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

In reply to Magnus Damm

Hi Magnus,

Quoted 4 lines · Magnus Damm

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

Hmm test_for_ipi and test_for_ltirq are needed for the SMP_ON_UP case, so I've moved those in to the MULTI_OMAP2 section.

Maybe merge the following into your original patch?

Regards,

Tony

From: Tony Lindgren <[email protected]>
Date: Thu, 11 Nov 2010 16:19:44 -0800
Subject: [PATCH] ARM: Common GIC entry macro for omap

Common GIC entry macro for omap

arch/arm/mach-omap2/include/mach/entry-macro.S +31 -61
index 06e64e1..2e358df 100644
@@ -105,6 +105,35 @@ omap_irq_base:	.word	0
 9999:
 		.endm
 
+#ifdef CONFIG_SMP
+		/* We assume that irqstat (the raw value of the IRQ acknowledge
+		 * register) is preserved from the macro above.
+		 * If there is an IPI, we immediately signal end of interrupt
+		 * on the controller, since this requires the original irqstat
+		 * value which we won't easily be able to recreate later.
+		 */
+
+		.macro test_for_ipi, irqnr, irqstat, base, tmp
+		bic	\irqnr, \irqstat, #0x1c00
+		cmp	\irqnr, #16
+		it	cc
+		strcc	\irqstat, [\base, #GIC_CPU_EOI]
+		it	cs
+		cmpcs	\irqnr, \irqnr
+		.endm
+
+		/* As above, this assumes that irqstat and base are preserved */
+
+		.macro test_for_ltirq, irqnr, irqstat, base, tmp
+		bic	\irqnr, \irqstat, #0x1c00
+		mov 	\tmp, #0
+		cmp	\irqnr, #29
+		itt	eq
+		moveq	\tmp, #1
+		streq	\irqstat, [\base, #GIC_CPU_EOI]
+		cmp	\tmp, #0
+		.endm
+#endif	/* CONFIG_SMP */
 
 #else	/* MULTI_OMAP2 */
 
@@ -141,74 +170,15 @@ omap_irq_base:	.word	0
 
 
 #ifdef CONFIG_ARCH_OMAP4
+#include <asm/hardware/entry-macro-gic.S>
 
 		.macro  get_irqnr_preamble, base, tmp
 		ldr     \base, =OMAP4_IRQ_BASE
 		.endm
 
-		/*
-		 * The interrupt numbering scheme is defined in the
-		 * interrupt controller spec.  To wit:
-		 *
-		 * Interrupts 0-15 are IPI
-		 * 16-28 are reserved
-		 * 29-31 are local.  We allow 30 to be used for the watchdog.
-		 * 32-1020 are global
-		 * 1021-1022 are reserved
-		 * 1023 is "spurious" (no interrupt)
-		 *
-		 * For now, we ignore all local interrupts so only return an
-		 * interrupt if it's between 30 and 1020.  The test_for_ipi
-		 * routine below will pick up on IPIs.
-		 * A simple read from the controller will tell us the number
-		 * of the highest priority enabled interrupt.
-		 * We then just need to check whether it is in the
-		 * valid range for an IRQ (30-1020 inclusive).
-		 */
-		.macro  get_irqnr_and_base, irqnr, irqstat, base, tmp
-		ldr     \irqstat, [\base, #GIC_CPU_INTACK]
-
-		ldr     \tmp, =1021
-
-		bic     \irqnr, \irqstat, #0x1c00
-
-		cmp     \irqnr, #29
-		cmpcc   \irqnr, \irqnr
-		cmpne   \irqnr, \tmp
-		cmpcs   \irqnr, \irqnr
-		.endm
 #endif
-#endif	/* MULTI_OMAP2 */
-
-#ifdef CONFIG_SMP
-		/* We assume that irqstat (the raw value of the IRQ acknowledge
-		 * register) is preserved from the macro above.
-		 * If there is an IPI, we immediately signal end of interrupt
-		 * on the controller, since this requires the original irqstat
-		 * value which we won't easily be able to recreate later.
-		 */
 
-		.macro test_for_ipi, irqnr, irqstat, base, tmp
-		bic	\irqnr, \irqstat, #0x1c00
-		cmp	\irqnr, #16
-		it	cc
-		strcc	\irqstat, [\base, #GIC_CPU_EOI]
-		it	cs
-		cmpcs	\irqnr, \irqnr
-		.endm
-
-		/* As above, this assumes that irqstat and base are preserved */
-
-		.macro test_for_ltirq, irqnr, irqstat, base, tmp
-		bic	\irqnr, \irqstat, #0x1c00
-		mov 	\tmp, #0
-		cmp	\irqnr, #29
-		itt	eq
-		moveq	\tmp, #1
-		streq	\irqstat, [\base, #GIC_CPU_EOI]
-		cmp	\tmp, #0
-		.endm
-#endif	/* CONFIG_SMP */
+#endif	/* MULTI_OMAP2 */
 
 		.macro	irq_prio_table
 		.endm
#8 lore raw

PATCH 1/1 ARM: Common GIC entry macro code V2

Hi Tony,

Quoted 11 lines · Tony Lindgren

Hi Magnus,

Quoted 4 lines · Magnus Damm

The omap code is excluded due to complexity and lack of hardware - the CONFIG_SMP macros are used for multi-omap and the omap4 case - this makes is difficult to share code.

Hmm test_for_ipi and test_for_ltirq are needed for the SMP_ON_UP case, so I've moved those in to the MULTI_OMAP2 section.

Looking good!

Quoted 1 line

Maybe merge the following into your original patch?

Sure, I'll include your code as a broken out patch in V3.

Thank you!

/ magnus