Skip to contentSkip to conversation

Cleanup MSM_GPIOMUX and add DT support for gpio-msm

PATCH Superseded Rohit Vaswani 5 participants 11 messages May 23, 2013 – May 31, 2013

Subsystems (5)

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

v2 cover Cleanup MSM_GPIOMUX and add DT support for gpio-msm

Hi Linus,
Could this series go through David's tree or is there a better way to do this? It would be great if I can have you ack for the gpio patch.

Thanks,
Rohit

Rohit Vaswani (3):
ARM: msm: Remove gpiomux-v2 and re-organize MSM_GPIOMUX configs
ARM: msm: Remove unused and unmapped MSM_TLMM_BASE for 8x60
gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

 .../devicetree/bindings/gpio/gpio-msm.txt          |   26 +++
 arch/arm/boot/dts/msm8660-surf.dts                 |   11 ++
 arch/arm/boot/dts/msm8960-cdp.dts                  |   11 ++
 arch/arm/mach-msm/Kconfig                          |   13 +-
 arch/arm/mach-msm/Makefile                         |    6 +-
 arch/arm/mach-msm/gpiomux-8x60.c                   |   19 --
 arch/arm/mach-msm/gpiomux-v2.c                     |   25 ---
 arch/arm/mach-msm/gpiomux-v2.h                     |   61 -------
 arch/arm/mach-msm/gpiomux.c                        |   15 ++
 arch/arm/mach-msm/gpiomux.h                        |    5 -
 arch/arm/mach-msm/include/mach/msm_iomap-8x60.h    |    4 -
 drivers/gpio/Kconfig                               |    2 +-
 drivers/gpio/gpio-msm-v2.c                         |  173 +++++++++++++-------
 13 files changed, 181 insertions(+), 190 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/gpio-msm.txt
 delete mode 100644 arch/arm/mach-msm/gpiomux-8x60.c
 delete mode 100644 arch/arm/mach-msm/gpiomux-v2.c
 delete mode 100644 arch/arm/mach-msm/gpiomux-v2.h
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation
#1 lore raw

PATCH v2 1/3 ARM: msm: Remove gpiomux-v2 and re-organize MSM_GPIOMUX configs

Remove gpiomux-v2 as it's not being used and make way for future improvements.

---
 arch/arm/mach-msm/Kconfig        |   13 +++-----
 arch/arm/mach-msm/Makefile       |    6 +--
 arch/arm/mach-msm/gpiomux-8x60.c |   19 ------------
 arch/arm/mach-msm/gpiomux-v2.c   |   25 ---------------
 arch/arm/mach-msm/gpiomux-v2.h   |   61 --------------------------------------
 arch/arm/mach-msm/gpiomux.c      |   15 +++++++++
 arch/arm/mach-msm/gpiomux.h      |    5 ---
 drivers/gpio/gpio-msm-v2.c       |    5 +--
 8 files changed, 24 insertions(+), 125 deletions(-)
 delete mode 100644 arch/arm/mach-msm/gpiomux-8x60.c
 delete mode 100644 arch/arm/mach-msm/gpiomux-v2.c
 delete mode 100644 arch/arm/mach-msm/gpiomux-v2.h
arch/arm/mach-msm/Kconfig +5 -8
index fceb093..614e41e 100644
@@ -48,9 +48,7 @@ config ARCH_MSM8X60
 	select CPU_V7
 	select GPIO_MSM_V2
 	select HAVE_SMP
-	select MSM_GPIOMUX
 	select MSM_SCM if SMP
-	select MSM_V2_TLMM
 	select USE_OF
 
 config ARCH_MSM8960
@@ -58,9 +56,8 @@ config ARCH_MSM8960
 	select ARM_GIC
 	select CPU_V7
 	select HAVE_SMP
-	select MSM_GPIOMUX
+	select GPIO_MSM_V2
 	select MSM_SCM if SMP
-	select MSM_V2_TLMM
 	select USE_OF
 
 config MSM_HAS_DEBUG_UART_HS
@@ -124,10 +121,10 @@ config MSM_SMD
 	bool
 
 config MSM_GPIOMUX
-	bool
-
-config MSM_V2_TLMM
-	bool
+	depends on !(ARCH_MSM8X60 || ARCH_MSM8960)
+	bool "MSM V1 TLMM GPIOMUX architecture"
+	help
+	  Support for MSM V1 TLMM GPIOMUX architecture.
 
 config MSM_SCM
 	bool
arch/arm/mach-msm/Makefile +2 -4
index 17519fa..1a26d04 100644
@@ -27,7 +27,5 @@ obj-$(CONFIG_ARCH_MSM7X30) += board-msm7x30.o devices-msm7x30.o
 obj-$(CONFIG_ARCH_QSD8X50) += board-qsd8x50.o devices-qsd8x50.o
 obj-$(CONFIG_ARCH_MSM8X60) += board-dt-8660.o
 obj-$(CONFIG_ARCH_MSM8960) += board-dt-8960.o
-
-obj-$(CONFIG_ARCH_MSM7X30) += gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_MSM8X60) += gpiomux-8x60.o gpiomux-v2.o gpiomux.o
+obj-$(CONFIG_MSM_GPIOMUX) += gpiomux.o
+obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o
arch/arm/mach-msm/gpiomux-8x60.c deleted +0 -19
deleted file mode 100644
index 7b380b3..0000000
@@ -1,19 +0,0 @@
-/* Copyright (c) 2010, Code Aurora Forum. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 and
- * only version 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
- * 02110-1301, USA.
- */
-#include "gpiomux.h"
-
-struct msm_gpiomux_config msm_gpiomux_configs[GPIOMUX_NGPIOS] = {};
arch/arm/mach-msm/gpiomux-v2.c deleted +0 -25
deleted file mode 100644
index 273396d..0000000
@@ -1,25 +0,0 @@
-/* Copyright (c) 2010, Code Aurora Forum. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 and
- * only version 2 as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
- * 02110-1301, USA.
- */
-#include <linux/io.h>
-#include <mach/msm_iomap.h>
-#include "gpiomux.h"
-
-void __msm_gpiomux_write(unsigned gpio, gpiomux_config_t val)
-{
-	writel(val & ~GPIOMUX_CTL_MASK,
-	       MSM_TLMM_BASE + 0x1000 + (0x10 * gpio));
-}
arch/arm/mach-msm/gpiomux-v2.h deleted +0 -61

61 lines not on this page · raw message

arch/arm/mach-msm/gpiomux.c +15 -0
index 53af21a..2b8e2d2 100644
@@ -17,9 +17,24 @@
 #include <linux/module.h>
 #include <linux/spinlock.h>
 #include "gpiomux.h"
+#include "proc_comm.h"
 
 static DEFINE_SPINLOCK(gpiomux_lock);
 
+static void __msm_gpiomux_write(unsigned gpio, gpiomux_config_t val)
+{
+	unsigned tlmm_config  = (val & ~GPIOMUX_CTL_MASK) |
+				((gpio & 0x3ff) << 4);
+	unsigned tlmm_disable = 0;
+	int rc;
+
+	rc = msm_proc_comm(PCOM_RPC_GPIO_TLMM_CONFIG_EX,
+			   &tlmm_config, &tlmm_disable);
+	if (rc)
+		pr_err("%s: unexpected proc_comm failure %d: %08x %08x\n",
+		       __func__, rc, tlmm_config, tlmm_disable);
+}
+
 int msm_gpiomux_write(unsigned gpio,
 		      gpiomux_config_t active,
 		      gpiomux_config_t suspended)
arch/arm/mach-msm/gpiomux.h +0 -5
index 00459f6..8e82f41 100644
@@ -20,12 +20,7 @@
 #include <linux/bitops.h>
 #include <linux/errno.h>
 #include <mach/msm_gpiomux.h>
-
-#if defined(CONFIG_MSM_V2_TLMM)
-#include "gpiomux-v2.h"
-#else
 #include "gpiomux-v1.h"
-#endif
 
 /**
  * struct msm_gpiomux_config: gpiomux settings for one gpio line.
drivers/gpio/gpio-msm-v2.c +2 -3
index dd2edde..75cc821 100644
@@ -29,7 +29,6 @@
 #include <linux/platform_device.h>
 #include <linux/spinlock.h>
 
-#include <mach/msm_gpiomux.h>
 #include <mach/msm_iomap.h>
 
 /* Bits of interest in the GPIO_IN_OUT register.
@@ -159,12 +158,12 @@ static int msm_gpio_direction_output(struct gpio_chip *chip,
 
 static int msm_gpio_request(struct gpio_chip *chip, unsigned offset)
 {
-	return msm_gpiomux_get(chip->base + offset);
+	return 0;
 }
 
 static void msm_gpio_free(struct gpio_chip *chip, unsigned offset)
 {
-	msm_gpiomux_put(chip->base + offset);
+	return;
 }
 
 static int msm_gpio_to_irq(struct gpio_chip *chip, unsigned offset)
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation
#2 lore raw

Re: [PATCH 1/3] ARM: msm: Remove gpiomux-v2 and re-organize MSM_GPIOMUX configs

Quoted 26 lines · Rohit Vaswani
arch/arm/mach-msm/Kconfig +1 -4
index fceb093..614e41e 100644
@@ -48,9 +48,7 @@ config ARCH_MSM8X60
 	select CPU_V7
 	select GPIO_MSM_V2
 	select HAVE_SMP
-	select MSM_GPIOMUX
 	select MSM_SCM if SMP
-	select MSM_V2_TLMM
 	select USE_OF
 
 config ARCH_MSM8960
@@ -58,9 +56,8 @@ config ARCH_MSM8960
 	select ARM_GIC
 	select CPU_V7
 	select HAVE_SMP
-	select MSM_GPIOMUX
+	select GPIO_MSM_V2
 	select MSM_SCM if SMP
-	select MSM_V2_TLMM
 	select USE_OF
 
 config MSM_HAS_DEBUG_UART_HS

arch/arm/mach-msm/Kconfig config MSM_SMD · line 124

Quoted 8 lines
arch/arm/mach-msm/Kconfig +1 -4
@@ -124,10 +121,10 @@ config MSM_SMD
 	bool
 
 config MSM_GPIOMUX
-	bool
-
-config MSM_V2_TLMM
-	bool
+	depends on !(ARCH_MSM8X60 || ARCH_MSM8960)

This list will keep getting bigger over time. Please just keep it a hidden option and don't select it on 8x60/8960.

Quoted 11 lines
arch/arm/mach-msm/Kconfig +3 -0
+	bool "MSM V1 TLMM GPIOMUX architecture"
+	help
+	  Support for MSM V1 TLMM GPIOMUX architecture.
 
 config MSM_SCM
 	bool
arch/arm/mach-msm/Makefile +0 -0
index 17519fa..1a26d04 100644

arch/arm/mach-msm/Makefile obj-$(CONFIG_ARCH_MSM7X30) += board-msm7x30.o devices-msm7x30.o · line 31

Quoted 9 lines
arch/arm/mach-msm/Makefile +2 -4
@@ -27,7 +27,5 @@ obj-$(CONFIG_ARCH_MSM7X30) += board-msm7x30.o devices-msm7x30.o
 obj-$(CONFIG_ARCH_QSD8X50) += board-qsd8x50.o devices-qsd8x50.o
 obj-$(CONFIG_ARCH_MSM8X60) += board-dt-8660.o
 obj-$(CONFIG_ARCH_MSM8960) += board-dt-8960.o
-
-obj-$(CONFIG_ARCH_MSM7X30) += gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_MSM8X60) += gpiomux-8x60.o gpiomux-v2.o gpiomux.o
+obj-$(CONFIG_MSM_GPIOMUX) += gpiomux.o
+obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o

Did you mean to delete gpiomux-v1.c as well?

Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by The Linux Foundation
#3 lore raw

Re: [PATCH 1/3] ARM: msm: Remove gpiomux-v2 and re-organize MSM_GPIOMUX configs

Quoted 29 lines · Stephen Boyd
Quoted 28 lines · Rohit Vaswani
arch/arm/mach-msm/Kconfig +1 -4
index fceb093..614e41e 100644
@@ -48,9 +48,7 @@ config ARCH_MSM8X60
  	select CPU_V7
  	select GPIO_MSM_V2
  	select HAVE_SMP
-	select MSM_GPIOMUX
  	select MSM_SCM if SMP
-	select MSM_V2_TLMM
  	select USE_OF
 
  config ARCH_MSM8960
@@ -58,9 +56,8 @@ config ARCH_MSM8960
  	select ARM_GIC
  	select CPU_V7
  	select HAVE_SMP
-	select MSM_GPIOMUX
+	select GPIO_MSM_V2
  	select MSM_SCM if SMP
-	select MSM_V2_TLMM
  	select USE_OF
 
  config MSM_HAS_DEBUG_UART_HS
@@ -124,10 +121,10 @@ config MSM_SMD
  	bool
 
Quoted 8 lines

arch/arm/mach-msm/Kconfig config MSM_SMD · line 124

Quoted 6 lines
arch/arm/mach-msm/Kconfig +1 -4
  config MSM_GPIOMUX
-	bool
-
-config MSM_V2_TLMM
-	bool
+	depends on !(ARCH_MSM8X60 || ARCH_MSM8960)

This list will keep getting bigger over time. Please just keep it a hidden option and don't select it on 8x60/8960.

This is fixed in my forth coming patches, so hoping this is okay for now ?

Quoted 14 lines
Quoted 14 lines
arch/arm/mach-msm/Kconfig +3 -0
+	bool "MSM V1 TLMM GPIOMUX architecture"
+	help
+	  Support for MSM V1 TLMM GPIOMUX architecture.
 
  config MSM_SCM
  	bool
arch/arm/mach-msm/Makefile +0 -0
index 17519fa..1a26d04 100644
@@ -27,7 +27,5 @@ obj-$(CONFIG_ARCH_MSM7X30) += board-msm7x30.o devices-msm7x30.o
  obj-$(CONFIG_ARCH_QSD8X50) += board-qsd8x50.o devices-qsd8x50.o
  obj-$(CONFIG_ARCH_MSM8X60) += board-dt-8660.o
Quoted 8 lines

arch/arm/mach-msm/Makefile obj-$(CONFIG_ARCH_MSM7X30) += board-msm7x30.o devices-msm7x30.o · line 31

Quoted 7 lines
arch/arm/mach-msm/Makefile +2 -4
  obj-$(CONFIG_ARCH_MSM8960) += board-dt-8960.o
-
-obj-$(CONFIG_ARCH_MSM7X30) += gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o gpiomux-v1.o gpiomux.o
-obj-$(CONFIG_ARCH_MSM8X60) += gpiomux-8x60.o gpiomux-v2.o gpiomux.o
+obj-$(CONFIG_MSM_GPIOMUX) += gpiomux.o
+obj-$(CONFIG_ARCH_QSD8X50) += gpiomux-8x50.o

Did you mean to delete gpiomux-v1.c as well?

Yes. There was only 1 function (__msm_gpiomux_write) which was selected based on v1/v2. Since v2 doesn't exist anymore, there is no need for a separate file for v1.

Quoted 1 line

Thanks,
Rohit Vaswani

Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
#4 lore raw

PATCH v2 2/3 ARM: msm: Remove unused and unmapped MSM_TLMM_BASE for 8x60

In reply to Rohit Vaswani

MSM_TLMM_BASE is currently not mapped by 8x60. Remove it.

---
 arch/arm/mach-msm/include/mach/msm_iomap-8x60.h |    4 ----
 1 files changed, 0 insertions(+), 4 deletions(-)
arch/arm/mach-msm/include/mach/msm_iomap-8x60.h +0 -4
index 199372e..cf24b5c 100644
@@ -41,10 +41,6 @@
 #define MSM8X60_QGIC_CPU_PHYS	0x02081000
 #define MSM8X60_QGIC_CPU_SIZE	SZ_4K
 
-#define MSM_TLMM_BASE		IOMEM(0xF0004000)
-#define MSM_TLMM_PHYS		0x00800000
-#define MSM_TLMM_SIZE		SZ_16K
-
 #define MSM8X60_TMR_PHYS	0x02000000
 #define MSM8X60_TMR_SIZE	SZ_4K
 
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation
#5 lore raw

PATCH v2 3/3 gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

In reply to Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

---
 .../devicetree/bindings/gpio/gpio-msm.txt          |   26 +++
 arch/arm/boot/dts/msm8660-surf.dts                 |   11 ++
 arch/arm/boot/dts/msm8960-cdp.dts                  |   11 ++
 drivers/gpio/Kconfig                               |    2 +-
 drivers/gpio/gpio-msm-v2.c                         |  168 +++++++++++++-------
 5 files changed, 157 insertions(+), 61 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/gpio-msm.txt
Documentation/devicetree/bindings/gpio/gpio-msm.txt new file +26 -0
new file mode 100644
index 0000000..ac20e68
@@ -0,0 +1,26 @@
+MSM GPIO controller bindings
+
+Required properties:
+- compatible:
+  - "qcom,msm-gpio" for MSM controllers
+- #gpio-cells : Should be two.
+  - first cell is the pin number
+  - second cell is used to specify optional parameters (unused)
+- gpio-controller : Marks the device node as a GPIO controller.
+- #interrupt-cells : Should be 2.
+- interrupt-controller: Mark the device node as an interrupt controller
+- interrupts : Specify the TLMM summary interrupt number
+- ngpio : Specify the number of MSM GPIOs
+
+Example:
+
+	msmgpio: gpio@fd510000 {
+		compatible = "qcom,msm-gpio";
+		gpio-controller;
+		#gpio-cells = <2>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+		reg = <0xfd510000 0x4000>;
+		interrupts = <0 208 0>;
+		ngpio = <150>;
+	};
arch/arm/boot/dts/msm8660-surf.dts +11 -0
index 9bf49b3..8931906 100644
@@ -26,6 +26,17 @@
 		cpu-offset = <0x40000>;
 	};
 
+	msmgpio: gpio@800000 {
+		compatible = "qcom,msm-gpio";
+		reg = <0x00800000 0x1000>;
+		gpio-controller;
+		#gpio-cells = <2>;
+		ngpio = <173>;
+		interrupts = <0 32 0x4>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+	};
+
 	serial@19c400000 {
 		compatible = "qcom,msm-hsuart", "qcom,msm-uart";
 		reg = <0x19c40000 0x1000>,
arch/arm/boot/dts/msm8960-cdp.dts +11 -0
index 2e4d87a..52fe253 100644
@@ -26,6 +26,17 @@
 		cpu-offset = <0x80000>;
 	};
 
+	msmgpio: gpio@fd510000 {
+		compatible = "qcom,msm-gpio";
+		gpio-controller;
+		#gpio-cells = <2>;
+		ngpio = <150>;
+		interrupts = <0 32 0x4>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+		reg = <0xfd510000 0x4000>;
+	};
+
 	serial@19c400000 {
 		compatible = "qcom,msm-hsuart", "qcom,msm-uart";
 		reg = <0x16440000 0x1000>,
drivers/gpio/Kconfig +1 -1
index 87d5670..9ef3b04 100644
@@ -165,7 +165,7 @@ config GPIO_MSM_V1
 
 config GPIO_MSM_V2
 	tristate "Qualcomm MSM GPIO v2"
-	depends on GPIOLIB && ARCH_MSM
+	depends on GPIOLIB && OF
 	help
 	  Say yes here to support the GPIO interface on ARM v7 based
 	  Qualcomm MSM chips.  Most of the pins on the MSM can be
drivers/gpio/gpio-msm-v2.c +108 -60

263 lines not on this page · raw message

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

Re: [PATCHv2 3/3] gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

Hi Rohit,

Thanks for the new version!

I have few more comments below.

Quoted 14 lines · Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

---
 .../devicetree/bindings/gpio/gpio-msm.txt          |   26 +++
 arch/arm/boot/dts/msm8660-surf.dts                 |   11 ++
 arch/arm/boot/dts/msm8960-cdp.dts                  |   11 ++
 drivers/gpio/Kconfig                               |    2 +-
 drivers/gpio/gpio-msm-v2.c                         |  168 +++++++++++++-------
 5 files changed, 157 insertions(+), 61 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/gpio-msm.txt

<cut>

drivers/gpio/gpio-msm-v2.c static void msm_gpio_free(struct gpio_chip *chip, unsigned offset) · line 178

Quoted 5 lines

static int msm_gpio_to_irq(struct gpio_chip *chip, unsigned offset) {

drivers/gpio/gpio-msm-v2.c +2 -1
-	return MSM_GPIO_TO_INT(chip->base + offset);
+	struct msm_gpio_dev *g_dev = to_msm_gpio_dev(chip);
+	struct irq_domain *domain = g_dev->domain;

Could you add blank line here.

drivers/gpio/gpio-msm-v2.c static void msm_gpio_free(struct gpio_chip *chip, unsigned offset) · line 184

Quoted 7 lines
drivers/gpio/gpio-msm-v2.c +2 -1
+	return irq_create_mapping(domain, offset);
 }
 
 static inline int msm_irq_to_gpio(struct gpio_chip *chip, unsigned irq)
 {
-	return irq - MSM_GPIO_TO_INT(chip->base);
+	struct irq_data *irq_data = irq_get_irq_data(irq);

Blank line please.

Quoted 25 lines
drivers/gpio/gpio-msm-v2.c +3 -3
+	return irq_data->hwirq;
 }
 
 static struct msm_gpio_dev msm_gpio = {
 	.gpio_chip = {
 		.base             = 0,
-		.ngpio            = NR_GPIO_IRQS,
 		.direction_input  = msm_gpio_direction_input,
 		.direction_output = msm_gpio_direction_output,
 		.get              = msm_gpio_get,
@@ -226,9 +234,9 @@ static void msm_gpio_update_dual_edge_pos(unsigned gpio)
 		if (intstat || val == val2)
 			return;
 	} while (loop_limit-- > 0);
-	pr_err("dual-edge irq failed to stabilize, "
+	pr_err("%s: dual-edge irq failed to stabilize, "
 	       "interrupts dropped. %#08x != %#08x\n",
-	       val, val2);
+	       __func__, val, val2);
 }
 
 static void msm_gpio_irq_ack(struct irq_data *d)
@@ -312,10 +320,11 @@ static void msm_summary_irq_handler(unsigned int irq, struct irq_desc *desc)
 {
 	unsigned long i;

drivers/gpio/gpio-msm-v2.c static void msm_summary_irq_handler(unsigned int irq, struct irq_desc *desc) · line 330

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +2 -1
 	struct irq_chip *chip = irq_desc_get_chip(desc);
+	int ngpio = msm_gpio.gpio_chip.ngpio;
 
 	chained_irq_enter(chip, desc);
 
-	for_each_set_bit(i, msm_gpio.enabled_irqs, NR_GPIO_IRQS) {
+	for_each_set_bit(i, msm_gpio.enabled_irqs, ngpio) {
 		if (readl(GPIO_INTR_STATUS(i)) & BIT(INTR_STATUS))
 			generic_handle_irq(msm_gpio_to_irq(&msm_gpio.gpio_chip,
 							   i));

As you decided to call irq_create_mapping() from msm_gpio_to_irq, then in irq handler you must call irq_find_mapping() as stated in IRQ-domain.txt and probably check return value.

<cut>

drivers/gpio/gpio-msm-v2.c static struct irq_chip msm_gpio_irq_chip = { · line 390

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +10 -0
+static int msm_gpio_irqdomain_init(struct device_node *node, int ngpio)
+{
+	msm_gpio.domain = irq_domain_add_linear(node, ngpio,
+			&msm_gpio_irq_domain_ops, &msm_gpio);
+	if (!msm_gpio.domain)
+		return -ENOMEM;
+
+	return 0;
+}
+

This function seems meaningless, could you call irq_domain_add_linear from .probe directly?

- Stan

#7 lore raw

Re: [PATCHv2 3/3] gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

In reply to Rohit Vaswani
Quoted 5 lines · Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

Hi Rohit,

Comments below on this driver...

Quoted 113 lines
---
 .../devicetree/bindings/gpio/gpio-msm.txt          |   26 +++
 arch/arm/boot/dts/msm8660-surf.dts                 |   11 ++
 arch/arm/boot/dts/msm8960-cdp.dts                  |   11 ++
 drivers/gpio/Kconfig                               |    2 +-
 drivers/gpio/gpio-msm-v2.c                         |  168 +++++++++++++-------
 5 files changed, 157 insertions(+), 61 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/gpio/gpio-msm.txt
Documentation/devicetree/bindings/gpio/gpio-msm.txt new file +26 -0
new file mode 100644
index 0000000..ac20e68
@@ -0,0 +1,26 @@
+MSM GPIO controller bindings
+
+Required properties:
+- compatible:
+  - "qcom,msm-gpio" for MSM controllers
+- #gpio-cells : Should be two.
+  - first cell is the pin number
+  - second cell is used to specify optional parameters (unused)
+- gpio-controller : Marks the device node as a GPIO controller.
+- #interrupt-cells : Should be 2.
+- interrupt-controller: Mark the device node as an interrupt controller
+- interrupts : Specify the TLMM summary interrupt number
+- ngpio : Specify the number of MSM GPIOs
+
+Example:
+
+	msmgpio: gpio@fd510000 {
+		compatible = "qcom,msm-gpio";
+		gpio-controller;
+		#gpio-cells = <2>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+		reg = <0xfd510000 0x4000>;
+		interrupts = <0 208 0>;
+		ngpio = <150>;
+	};
arch/arm/boot/dts/msm8660-surf.dts +11 -0
index 9bf49b3..8931906 100644
@@ -26,6 +26,17 @@
 		cpu-offset = <0x40000>;
 	};
 
+	msmgpio: gpio@800000 {
+		compatible = "qcom,msm-gpio";
+		reg = <0x00800000 0x1000>;
+		gpio-controller;
+		#gpio-cells = <2>;
+		ngpio = <173>;
+		interrupts = <0 32 0x4>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+	};
+
 	serial@19c400000 {
 		compatible = "qcom,msm-hsuart", "qcom,msm-uart";
 		reg = <0x19c40000 0x1000>,
arch/arm/boot/dts/msm8960-cdp.dts +11 -0
index 2e4d87a..52fe253 100644
@@ -26,6 +26,17 @@
 		cpu-offset = <0x80000>;
 	};
 
+	msmgpio: gpio@fd510000 {
+		compatible = "qcom,msm-gpio";
+		gpio-controller;
+		#gpio-cells = <2>;
+		ngpio = <150>;
+		interrupts = <0 32 0x4>;
+		interrupt-controller;
+		#interrupt-cells = <2>;
+		reg = <0xfd510000 0x4000>;
+	};
+
 	serial@19c400000 {
 		compatible = "qcom,msm-hsuart", "qcom,msm-uart";
 		reg = <0x16440000 0x1000>,
drivers/gpio/Kconfig +1 -1
index 87d5670..9ef3b04 100644
@@ -165,7 +165,7 @@ config GPIO_MSM_V1
 
 config GPIO_MSM_V2
 	tristate "Qualcomm MSM GPIO v2"
-	depends on GPIOLIB && ARCH_MSM
+	depends on GPIOLIB && OF
 	help
 	  Say yes here to support the GPIO interface on ARM v7 based
 	  Qualcomm MSM chips.  Most of the pins on the MSM can be
drivers/gpio/gpio-msm-v2.c +1 -0
index 75cc821..025c11a 100644
@@ -19,17 +19,22 @@
 
 #include <linux/bitmap.h>
 #include <linux/bitops.h>
+#include <linux/err.h>
 #include <linux/gpio.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/irqchip/chained_irq.h>
 #include <linux/irq.h>

drivers/gpio/gpio-msm-v2.c line 37

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +5 -1
+#include <linux/irqdomain.h>
 #include <linux/module.h>
+#include <linux/of_address.h>
 #include <linux/platform_device.h>
 #include <linux/spinlock.h>
+#include <linux/slab.h>
 
-#include <mach/msm_iomap.h>
+static int summary_irq;
+void __iomem *msm_tlmm_base;

Making this a global means there can only be one instance of this chip in a system. That may not be a concern for you, but I generally recommend to driver authors not to use any global variables if possible. Instead, things like base address should be stored in the msm_gpio_dev structure.

Quoted 21 lines
drivers/gpio/gpio-msm-v2.c +5 -5
 /* Bits of interest in the GPIO_IN_OUT register.
  */
@@ -77,11 +82,11 @@ enum {
 };
 
 
-#define GPIO_INTR_CFG_SU(gpio)    (MSM_TLMM_BASE + 0x0400 + (0x04 * (gpio)))
-#define GPIO_CONFIG(gpio)         (MSM_TLMM_BASE + 0x1000 + (0x10 * (gpio)))
-#define GPIO_IN_OUT(gpio)         (MSM_TLMM_BASE + 0x1004 + (0x10 * (gpio)))
-#define GPIO_INTR_CFG(gpio)       (MSM_TLMM_BASE + 0x1008 + (0x10 * (gpio)))
-#define GPIO_INTR_STATUS(gpio)    (MSM_TLMM_BASE + 0x100c + (0x10 * (gpio)))
+#define GPIO_INTR_CFG_SU(gpio)    (msm_tlmm_base + 0x0400 + (0x04 * (gpio)))
+#define GPIO_CONFIG(gpio)         (msm_tlmm_base + 0x1000 + (0x10 * (gpio)))
+#define GPIO_IN_OUT(gpio)         (msm_tlmm_base + 0x1004 + (0x10 * (gpio)))
+#define GPIO_INTR_CFG(gpio)       (msm_tlmm_base + 0x1008 + (0x10 * (gpio)))
+#define GPIO_INTR_STATUS(gpio)    (msm_tlmm_base + 0x100c + (0x10 * (gpio)))
 
 /**
  * struct msm_gpio_dev: the MSM8660 SoC GPIO device structure

drivers/gpio/gpio-msm-v2.c enum { · line 111

Quoted 9 lines
drivers/gpio/gpio-msm-v2.c +3 -3
@@ -101,9 +106,10 @@ enum {
  */
 struct msm_gpio_dev {
 	struct gpio_chip gpio_chip;
-	DECLARE_BITMAP(enabled_irqs, NR_GPIO_IRQS);
-	DECLARE_BITMAP(wake_irqs, NR_GPIO_IRQS);
-	DECLARE_BITMAP(dual_edge_irqs, NR_GPIO_IRQS);
+	unsigned long *enabled_irqs;
+	unsigned long *wake_irqs;
+	unsigned long *dual_edge_irqs;

See my comments below, but because these are merely bitmaps they are pretty small arrays. 3 arrays of <173 bits. That's merely 72 bytes. It is more expensive to dynamically allocate that much space.

Quoted 97 lines
drivers/gpio/gpio-msm-v2.c +31 -12
+	struct irq_domain *domain;
 };
 
 static DEFINE_SPINLOCK(tlmm_lock);
@@ -168,18 +174,20 @@ static void msm_gpio_free(struct gpio_chip *chip, unsigned offset)
 
 static int msm_gpio_to_irq(struct gpio_chip *chip, unsigned offset)
 {
-	return MSM_GPIO_TO_INT(chip->base + offset);
+	struct msm_gpio_dev *g_dev = to_msm_gpio_dev(chip);
+	struct irq_domain *domain = g_dev->domain;
+	return irq_create_mapping(domain, offset);
 }
 
 static inline int msm_irq_to_gpio(struct gpio_chip *chip, unsigned irq)
 {
-	return irq - MSM_GPIO_TO_INT(chip->base);
+	struct irq_data *irq_data = irq_get_irq_data(irq);
+	return irq_data->hwirq;
 }
 
 static struct msm_gpio_dev msm_gpio = {
 	.gpio_chip = {
 		.base             = 0,
-		.ngpio            = NR_GPIO_IRQS,
 		.direction_input  = msm_gpio_direction_input,
 		.direction_output = msm_gpio_direction_output,
 		.get              = msm_gpio_get,
@@ -226,9 +234,9 @@ static void msm_gpio_update_dual_edge_pos(unsigned gpio)
 		if (intstat || val == val2)
 			return;
 	} while (loop_limit-- > 0);
-	pr_err("dual-edge irq failed to stabilize, "
+	pr_err("%s: dual-edge irq failed to stabilize, "
 	       "interrupts dropped. %#08x != %#08x\n",
-	       val, val2);
+	       __func__, val, val2);
 }
 
 static void msm_gpio_irq_ack(struct irq_data *d)
@@ -312,10 +320,11 @@ static void msm_summary_irq_handler(unsigned int irq, struct irq_desc *desc)
 {
 	unsigned long i;
 	struct irq_chip *chip = irq_desc_get_chip(desc);
+	int ngpio = msm_gpio.gpio_chip.ngpio;
 
 	chained_irq_enter(chip, desc);
 
-	for_each_set_bit(i, msm_gpio.enabled_irqs, NR_GPIO_IRQS) {
+	for_each_set_bit(i, msm_gpio.enabled_irqs, ngpio) {
 		if (readl(GPIO_INTR_STATUS(i)) & BIT(INTR_STATUS))
 			generic_handle_irq(msm_gpio_to_irq(&msm_gpio.gpio_chip,
 							   i));
@@ -327,15 +336,16 @@ static void msm_summary_irq_handler(unsigned int irq, struct irq_desc *desc)
 static int msm_gpio_irq_set_wake(struct irq_data *d, unsigned int on)
 {
 	int gpio = msm_irq_to_gpio(&msm_gpio.gpio_chip, d->irq);
+	int ngpio = msm_gpio.gpio_chip.ngpio;
 
 	if (on) {
-		if (bitmap_empty(msm_gpio.wake_irqs, NR_GPIO_IRQS))
-			irq_set_irq_wake(TLMM_SCSS_SUMMARY_IRQ, 1);
+		if (bitmap_empty(msm_gpio.wake_irqs, ngpio))
+			irq_set_irq_wake(summary_irq, 1);
 		set_bit(gpio, msm_gpio.wake_irqs);
 	} else {
 		clear_bit(gpio, msm_gpio.wake_irqs);
-		if (bitmap_empty(msm_gpio.wake_irqs, NR_GPIO_IRQS))
-			irq_set_irq_wake(TLMM_SCSS_SUMMARY_IRQ, 0);
+		if (bitmap_empty(msm_gpio.wake_irqs, ngpio))
+			irq_set_irq_wake(summary_irq, 0);
 	}
 
 	return 0;
@@ -350,30 +360,93 @@ static struct irq_chip msm_gpio_irq_chip = {
 	.irq_set_wake	= msm_gpio_irq_set_wake,
 };
 
-static int msm_gpio_probe(struct platform_device *dev)
+static struct lock_class_key msm_gpio_lock_class;
+
+static int msm_gpio_irq_domain_map(struct irq_domain *d, unsigned int irq,
+				   irq_hw_number_t hwirq)
 {
-	int i, irq, ret;
+	irq_set_lockdep_class(irq, &msm_gpio_lock_class);
+	irq_set_chip_and_handler(irq, &msm_gpio_irq_chip,
+			handle_level_irq);
+	set_irq_flags(irq, IRQF_VALID);
+
+	return 0;
+}
+
+static const struct irq_domain_ops msm_gpio_irq_domain_ops = {
+	.xlate = irq_domain_xlate_twocell,
+	.map = msm_gpio_irq_domain_map,
+};

drivers/gpio/gpio-msm-v2.c static struct irq_chip msm_gpio_irq_chip = { · line 389

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +10 -0
+
+static int msm_gpio_irqdomain_init(struct device_node *node, int ngpio)
+{
+	msm_gpio.domain = irq_domain_add_linear(node, ngpio,
+			&msm_gpio_irq_domain_ops, &msm_gpio);
+	if (!msm_gpio.domain)
+		return -ENOMEM;
+
+	return 0;
+}

No need to split this into a separate function. All irq_domain_add_linear() directly please.

Quoted 19 lines
drivers/gpio/gpio-msm-v2.c +19 -0
+
+static int msm_gpio_probe(struct platform_device *pdev)
+{
+	int ret, ngpio;
+	struct resource *res;
+
+	msm_gpio.gpio_chip.label = pdev->name;
+	msm_gpio.gpio_chip.dev = &pdev->dev;
+	if (!of_property_read_u32(pdev->dev.of_node, "ngpio", &ngpio)) {
+		dev_err(&pdev->dev, "%s: ngpio property missing\n", __func__);
+		return -EINVAL;
+	}
+
+	msm_gpio.gpio_chip.ngpio = ngpio;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	msm_tlmm_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msm_tlmm_base))
+		return PTR_ERR(msm_tlmm_base);

drivers/gpio/gpio-msm-v2.c static struct irq_chip msm_gpio_irq_chip = { · line 418

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +10 -0
+
+	msm_gpio.enabled_irqs = devm_kzalloc(&pdev->dev,
+						sizeof(unsigned long) * ngpio,
+						GFP_KERNEL);
+	msm_gpio.wake_irqs = devm_kzalloc(&pdev->dev,
+						sizeof(unsigned long) * ngpio,
+						GFP_KERNEL);
+	msm_gpio.dual_edge_irqs = devm_kzalloc(&pdev->dev,
+						sizeof(unsigned long) * ngpio,
+						GFP_KERNEL);

It's expensive to call devm_kzalloc multiple times when you don't need to. do this instead:

	msm_gpio.enabled_irqs = devm_kzalloc(&pdev->dev,
					sizeof(unsigned long) * ngpio * 3,
					GFP_KERNEL);
	msm_gpio.wake_irqs = &msm_gpio.enable_irqs[ngpio];
	msm_gpio.dual_edge_irqs = &msm_gpio.enable_irqs[ngpio*2];

(I wrote the above comment before looking closely at the bitmaps, so the above is irrelevant if you revert to the static bitmap table.)

drivers/gpio/gpio-msm-v2.c static struct irq_chip msm_gpio_irq_chip = { · line 421

Quoted 3 lines
drivers/gpio/gpio-msm-v2.c +3 -0
+	bitmap_zero(msm_gpio.enabled_irqs, ngpio);
+	bitmap_zero(msm_gpio.wake_irqs, ngpio);
+	bitmap_zero(msm_gpio.dual_edge_irqs, ngpio);

Umm... this looks wrong. You've allocated an array with one unsigned long for each gpio, but then are using it as a bitmask (an array of one /bit/ for each gpio) meaning most of the array is unused.

Besides, if you've kzalloced the memory, then the bitmap is already zeroed. :)

What is the maximum number of gpios that this controller can have? From the dts fragments above it looks to be 173, which would be 5 banks of 32 bits each.

Quoted 44 lines
drivers/gpio/gpio-msm-v2.c +19 -12
-	bitmap_zero(msm_gpio.enabled_irqs, NR_GPIO_IRQS);
-	bitmap_zero(msm_gpio.wake_irqs, NR_GPIO_IRQS);
-	bitmap_zero(msm_gpio.dual_edge_irqs, NR_GPIO_IRQS);
-	msm_gpio.gpio_chip.label = dev->name;
 	ret = gpiochip_add(&msm_gpio.gpio_chip);
-	if (ret < 0)
+	if (ret < 0) {
+		dev_err(&pdev->dev, "gpiochip_add failed with error %d\n", ret);
 		return ret;
+	}
 
-	for (i = 0; i < msm_gpio.gpio_chip.ngpio; ++i) {
-		irq = msm_gpio_to_irq(&msm_gpio.gpio_chip, i);
-		irq_set_chip_and_handler(irq, &msm_gpio_irq_chip,
-					 handle_level_irq);
-		set_irq_flags(irq, IRQF_VALID);
+	summary_irq = platform_get_irq(pdev, 0);
+	if (summary_irq < 0) {
+		dev_err(&pdev->dev, "No Summary irq defined for msmgpio\n");
+		return summary_irq;
 	}
 
-	irq_set_chained_handler(TLMM_SCSS_SUMMARY_IRQ,
-				msm_summary_irq_handler);
+	ret = msm_gpio_irqdomain_init(pdev->dev.of_node,
+					msm_gpio.gpio_chip.ngpio);
+	if (ret)
+		return ret;
+
+	irq_set_chained_handler(summary_irq, msm_summary_irq_handler);
+
 	return 0;
 }
 
+static struct of_device_id msm_gpio_of_match[] = {
+	{ .compatible = "qcom,msm-gpio", },
+	{ },
+};
+
 static int msm_gpio_remove(struct platform_device *dev)
 {
 	int ret = gpiochip_remove(&msm_gpio.gpio_chip);

drivers/gpio/gpio-msm-v2.c static struct platform_driver msm_gpio_driver = { · line 468

Quoted 4 lines
drivers/gpio/gpio-msm-v2.c +1 -0
@@ -392,36 +465,11 @@ static struct platform_driver msm_gpio_driver = {
 	.driver = {
 		.name = "msmgpio",
 		.owner = THIS_MODULE,
+		.of_match_table = of_match_ptr(msm_gpio_of_match),

You don't need the of_match_ptr() macro if msm_gpio_of_match[] is unconditionally present (which in this patch it is)

Quoted 38 lines
drivers/gpio/gpio-msm-v2.c +1 -27
 	},
 };
 
-static struct platform_device msm_device_gpio = {
-	.name = "msmgpio",
-	.id   = -1,
-};
-
-static int __init msm_gpio_init(void)
-{
-	int rc;
-
-	rc = platform_driver_register(&msm_gpio_driver);
-	if (!rc) {
-		rc = platform_device_register(&msm_device_gpio);
-		if (rc)
-			platform_driver_unregister(&msm_gpio_driver);
-	}
-
-	return rc;
-}
-
-static void __exit msm_gpio_exit(void)
-{
-	platform_device_unregister(&msm_device_gpio);
-	platform_driver_unregister(&msm_gpio_driver);
-}
-
-postcore_initcall(msm_gpio_init);
-module_exit(msm_gpio_exit);
+module_platform_driver(msm_gpio_driver)
 
 MODULE_AUTHOR("Gregory Bean <[email protected]>");
 MODULE_DESCRIPTION("Driver for Qualcomm MSM TLMMv2 SoC GPIOs");
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation
Signature · 2 lines
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies, Ltd.
#8 lore raw

Re: [PATCHv2 3/3] gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

Quoted 9 lines · Grant Likely
Quoted 5 lines · Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

Hi Rohit,

Comments below on this driver...

Thanks Grant for your comments. Have addressed some of these in v3. Will fix the bitmap allocation and other comments in v4.

<snip>

Thanks,
Rohit Vaswani

Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
#9 lore raw

Re: [PATCHv2 3/3] gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

In reply to Rohit Vaswani
Quoted 3 lines · Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

Besides the other comments, I have one important here.

Quoted 13 lines
drivers/gpio/gpio-msm-v2.c +0 -13
-static int __init msm_gpio_init(void)
-{
-       int rc;
-
-       rc = platform_driver_register(&msm_gpio_driver);
-       if (!rc) {
-               rc = platform_device_register(&msm_device_gpio);
-               if (rc)
-                       platform_driver_unregister(&msm_gpio_driver);
-       }
-
-       return rc;
-}

drivers/gpio/gpio-msm-v2.c static struct platform_driver msm_gpio_driver = { · line 472

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +1 -9
-
-static void __exit msm_gpio_exit(void)
-{
-       platform_device_unregister(&msm_device_gpio);
-       platform_driver_unregister(&msm_gpio_driver);
-}
-
-postcore_initcall(msm_gpio_init);
-module_exit(msm_gpio_exit);
+module_platform_driver(msm_gpio_driver)

It's a really big mistake to do this.
GPIO nowadays is used in many hardware subsystems to support early stages of booting and initializing.

postcore_initcall is early enough to support mostly anything that hardware wants from GPIO (for example, ACPI events, though it seems not a case here).

Please, do not do such changes without good argumets.

--
With Best Regards,
Andy Shevchenko

#10 lore raw

Re: [PATCHv2 3/3] gpio: msm: Add device tree and irqdomain support for gpio-msm-v2

Quoted 35 lines · Andy Shevchenko
Quoted 3 lines · Rohit Vaswani

This cleans up the gpio-msm-v2 driver of all the global define usage. The number of gpios are now defined in the device tree. This enables adding irqdomain support as well.

Besides the other comments, I have one important here.

Quoted 13 lines
drivers/gpio/gpio-msm-v2.c +0 -13
-static int __init msm_gpio_init(void)
-{
-       int rc;
-
-       rc = platform_driver_register(&msm_gpio_driver);
-       if (!rc) {
-               rc = platform_device_register(&msm_device_gpio);
-               if (rc)
-                       platform_driver_unregister(&msm_gpio_driver);
-       }
-
-       return rc;
-}

drivers/gpio/gpio-msm-v2.c static struct platform_driver msm_gpio_driver = { · line 472

Quoted 10 lines
drivers/gpio/gpio-msm-v2.c +1 -9
-
-static void __exit msm_gpio_exit(void)
-{
-       platform_device_unregister(&msm_device_gpio);
-       platform_driver_unregister(&msm_gpio_driver);
-}
-
-postcore_initcall(msm_gpio_init);
-module_exit(msm_gpio_exit);
+module_platform_driver(msm_gpio_driver)

It's a really big mistake to do this.
GPIO nowadays is used in many hardware subsystems to support early stages of booting and initializing.

Quoted 3 lines

postcore_initcall is early enough to support mostly anything that hardware wants from GPIO (for example, ACPI events, though it seems not a case here).

Actually, we're going the other way in the kernel. I do agree that for existing drivers you need to be careful and make sure that switching the initcall level won't cause any breakage with existing systems, but new platform support must not rely on the initcall order for setting up gpios. That's what deferred probe is intended to solve.

g.