Skip to contentSkip to conversation

Remove GENERIC_GPIO from architecture code

RFC In review Alexandre Courbot 8 participants 37 messages Mar 12, 2013 – Mar 14, 2013 1 message not in the archive

Subsystems (12)
lore
Share & tools
lore
#0 lore raw

cover Remove GENERIC_GPIO from architecture code

This series makes sure the GENERIC_GPIO option can only be set through GPIOLIB (and not by individual architectures), as a first step towards its removal.

It is targeted at getting feedback from architecture maintainers (and complains if this is deemed too bold a move - however, I hope the rationale behind this will be convincing).

AFAICT every GPIO driver that implement the generic GPIO API support gpiolib at least optionally, and the overwhelming majority actually requires it. Using the generic GPIO API alone has become a marginal practice, and the benefit of retaining this option is rather uncertain when compared to the confusion and complexity it induces:
* More and more GPIO features are built around gpiolib. The sysfs interface is one of them ; the future gpiod API is another one. Platforms that only implement the generic GPIO API cannot benefit from these and the split between those who only implement the generic GPIO API and full-fledge users of gpiolib is becoming wider and wider.
* Having two layers of GPIO support is confusing to both GPIO users and providers, and it is easy to e.g. only depend on GENERIC_GPIO when one actually needs gpiolib. This series actually fixes a few of these cases.
* Simplicity and consistency is always a good thing - features in the kernel are typically implemented through well-defined frameworks, and similarly GPIO support could be consolidated around gpiolib.

Arguments against using gpiolib:
* Memory footprint of gpiolib. According to my tests a gpiolib with 256 GPIOs induces an overhead of ~15KB, which sounds rather reasonable.
* Performance. gpiolib introduces another layer of indirection compared to drivers that only implement the generic GPIO API. However, a fast path is available to platforms for which GPIO performance matters through the implementation of custom gpio_set_value() and gpio_get_value() functions which test for a given GPIO range and shortcut gpiolib.

For most platforms, this change should be a no-op. However I would like to make sure that everyone is ok with it and that nothing gets broken, as the effect of changing configuration options are sometimes difficult to predict.

Alexandre Courbot (17):
arm: remove unneeded select GENERIC_GPIO
arm: remove redundant GENERIC_GPIO selection
arm: plat-orion: use GPIO driver on CONFIG_GPIOLIB
mips: remove redundant GENERIC_GPIO select
mips: loongson: use GPIO driver on CONFIG_GPIOLIB
mips: txx9: change GENERIC_GPIO to GPIOLIB
unicore32: remove unneeded select GENERIC_GPIO
powerpc: remove redundant GENERIC_GPIO selection
sh: replace CONFIG_GENERIC_GPIO by CONFIG_GPIOLIB
xtensa: remove explicit selection of GENERIC_GPIO
mips: alchemy: require gpiolib
mips: pnx833x: remove requirement for GENERIC_GPIO
avr32: default GENERIC_GPIO to false
m68k: coldfire: use gpiolib
avr32: default GENERIC_GPIO to false
openrisc: default GENERIC_GPIO to false
unicore32: default GENERIC_GPIO to false

 arch/arm/Kconfig                     | 2 --
 arch/arm/plat-orion/Makefile         | 2 +-
 arch/avr32/Kconfig                   | 2 +-
 arch/blackfin/Kconfig                | 2 +-
 arch/m68k/Kconfig.cpu                | 3 +--
 arch/mips/Kconfig                    | 7 +------
 arch/mips/loongson/common/Makefile   | 2 +-
 arch/mips/txx9/generic/setup.c       | 2 +-
 arch/openrisc/Kconfig                | 2 +-
 arch/powerpc/platforms/40x/Kconfig   | 1 -
 arch/powerpc/platforms/44x/Kconfig   | 1 -
 arch/powerpc/platforms/85xx/Kconfig  | 1 -
 arch/powerpc/platforms/86xx/Kconfig  | 3 ---
 arch/powerpc/platforms/8xx/Kconfig   | 1 -
 arch/powerpc/platforms/Kconfig       | 4 ----
 arch/sh/boards/mach-sdk7786/Makefile | 2 +-
 arch/sh/boards/mach-x3proto/Makefile | 2 +-
 arch/sh/kernel/cpu/sh2a/Makefile     | 2 +-
 arch/sh/kernel/cpu/sh3/Makefile      | 2 +-
 arch/sh/kernel/cpu/sh4a/Makefile     | 2 +-
 arch/unicore32/Kconfig               | 3 +--
 arch/xtensa/configs/iss_defconfig    | 1 -
 arch/xtensa/configs/s6105_defconfig  | 1 -
 23 files changed, 14 insertions(+), 36 deletions(-)
Signature · 1 line
1.8.1.5
#1 lore raw

RFC 1/17 arm: remove unneeded select GENERIC_GPIO

Remove the selection of GENERIC_GPIO for S3C24XX and S5PC100 in preparation of the removal of this option.

ARCH_REQUIRE_GPIOLIB (which also selects GENERIC_GPIO) is already selected whenever these machines are chosen anyway, so this patch should not have any side-effect.

---
 arch/arm/Kconfig | 2 --
 1 file changed, 2 deletions(-)
arch/arm/Kconfig +0 -2
index ad06749..26b6ddd 100644
@@ -773,7 +773,6 @@ config ARCH_S3C24XX
 	select CLKDEV_LOOKUP
 	select CLKSRC_MMIO
 	select GENERIC_CLOCKEVENTS
-	select GENERIC_GPIO
 	select HAVE_CLK
 	select HAVE_S3C2410_I2C if I2C
 	select HAVE_S3C2410_WATCHDOG if WATCHDOG
@@ -832,7 +831,6 @@ config ARCH_S5PC100
 	select CLKSRC_MMIO
 	select CPU_V7
 	select GENERIC_CLOCKEVENTS
-	select GENERIC_GPIO
 	select HAVE_CLK
 	select HAVE_S3C2410_I2C if I2C
 	select HAVE_S3C2410_WATCHDOG if WATCHDOG
Signature · 1 line
1.8.1.5
#2 lore raw

RFC 2/17 arm: remove redundant GENERIC_GPIO selection

In reply to Alexandre Courbot

GENERIC_GPIO was selected next to ARCH_REQUIRE_GPIOLIB, which itself selects GENERIC_GPIO.

---
 arch/mips/Kconfig | 2 --
 1 file changed, 2 deletions(-)
arch/mips/Kconfig +0 -2
index 32eb3d6..b7b7228 100644
@@ -225,7 +225,6 @@ config MACH_JZ4740
 	select SYS_SUPPORTS_ZBOOT_UART16550
 	select DMA_NONCOHERENT
 	select IRQ_CPU
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	select SYS_HAS_EARLY_PRINTK
 	select HAVE_PWM
@@ -935,7 +934,6 @@ config CSRC_SB1250
 	bool
 
 config GPIO_TXX9
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	bool
 
Signature · 1 line
1.8.1.5
#3 lore raw

RFC 3/17 arm: plat-orion: use GPIO driver on CONFIG_GPIOLIB

In reply to Alexandre Courbot

The GPIO driver uses gpiolib, thus it should be compiled when CONFIG_GPIOLIB is defined and not only CONFIG_GENERIC_GPIO

---
 arch/arm/plat-orion/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/plat-orion/Makefile +1 -1
index a82cecb..bbc0d6b 100644
@@ -5,6 +5,6 @@ ccflags-$(CONFIG_ARCH_MULTIPLATFORM) := -I$(srctree)/$(src)/include
 
 obj-y                             += addr-map.o
 
-orion-gpio-$(CONFIG_GENERIC_GPIO) += gpio.o
+orion-gpio-$(CONFIG_GPIOLIB)      += gpio.o
 obj-$(CONFIG_PLAT_ORION_LEGACY)   += irq.o pcie.o time.o common.o mpp.o
 obj-$(CONFIG_PLAT_ORION_LEGACY)   += $(orion-gpio-y)
Signature · 1 line
1.8.1.5
#4 lore raw

RFC 4/17 mips: remove redundant GENERIC_GPIO select

In reply to Alexandre Courbot
---
 arch/mips/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/mips/Kconfig +0 -1
index b7b7228..ae70ab1 100644
@@ -1199,7 +1199,6 @@ config CPU_LOONGSON2F
 	bool "Loongson 2F"
 	depends on SYS_HAS_CPU_LOONGSON2F
 	select CPU_LOONGSON2
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	help
 	  The Loongson 2F processor implements the MIPS III instruction set
Signature · 1 line
1.8.1.5
#5 lore raw

RFC 5/17 mips: loongson: use GPIO driver on CONFIG_GPIOLIB

In reply to Alexandre Courbot

The GPIO driver uses gpiolib, thus it should be compiled when CONFIG_GPIOLIB is defined and not only CONFIG_GENERIC_GPIO.

---
 arch/mips/loongson/common/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/mips/loongson/common/Makefile +1 -1
index e526488..4c57b3e 100644
@@ -4,7 +4,7 @@
 
 obj-y += setup.o init.o cmdline.o env.o time.o reset.o irq.o \
     pci.o bonito-irq.o mem.o machtype.o platform.o
-obj-$(CONFIG_GENERIC_GPIO) += gpio.o
+obj-$(CONFIG_GPIOLIB) += gpio.o
 
 #
 # Serial port support
Signature · 1 line
1.8.1.5
#6 lore raw

RFC 6/17 mips: txx9: change GENERIC_GPIO to GPIOLIB

In reply to Alexandre Courbot

The GPIO driver for txx9 requires gpiolib, so this is more accurate and let us get rid of one mention of GENERIC_GPIO which is due to disappear.

---
 arch/mips/txx9/generic/setup.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/mips/txx9/generic/setup.c +1 -1
index 5524f2c..5364aab 100644
@@ -118,7 +118,7 @@ EXPORT_SYMBOL(clk_put);
 
 /* GPIO support */
 
-#ifdef CONFIG_GENERIC_GPIO
+#ifdef CONFIG_GPIOLIB
 int gpio_to_irq(unsigned gpio)
 {
 	return -EINVAL;
Signature · 1 line
1.8.1.5
#7 lore raw

RFC 7/17 unicore32: remove unneeded select GENERIC_GPIO

In reply to Alexandre Courbot

PUV3_GPIO depends on ARCH_PUV3 which requires GPIOLIB and thus already selects GENERIC_GPIO

---
 arch/unicore32/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/unicore32/Kconfig +0 -1
index dc50b15..a889117 100644
@@ -219,7 +219,6 @@ if ARCH_PUV3
 config PUV3_GPIO
 	bool
 	depends on !ARCH_FPGA
-	select GENERIC_GPIO
 	select GPIO_SYSFS
 	default y
 
Signature · 1 line
1.8.1.5
#8 lore raw

RFC 8/17 powerpc: remove redundant GENERIC_GPIO selection

In reply to Alexandre Courbot

GENERIC_GPIO was explicitly selected along with ARCH_REQUIRE_GPIOLIB, which itself selects GENERIC_GPIO.

---
 arch/powerpc/platforms/40x/Kconfig  | 1 -
 arch/powerpc/platforms/44x/Kconfig  | 1 -
 arch/powerpc/platforms/85xx/Kconfig | 1 -
 arch/powerpc/platforms/86xx/Kconfig | 3 ---
 arch/powerpc/platforms/8xx/Kconfig  | 1 -
 arch/powerpc/platforms/Kconfig      | 4 ----
 6 files changed, 11 deletions(-)
arch/powerpc/platforms/40x/Kconfig +0 -1
index a392d12..d0b46a2 100644
@@ -145,7 +145,6 @@ config PPC4xx_GPIO
 	bool "PPC4xx GPIO support"
 	depends on 40x
 	select ARCH_REQUIRE_GPIOLIB
-	select GENERIC_GPIO
 	help
 	  Enable gpiolib support for ppc40x based boards
 
arch/powerpc/platforms/44x/Kconfig +0 -1
index 0effe9f..50306b8 100644
@@ -248,7 +248,6 @@ config PPC4xx_GPIO
 	bool "PPC4xx GPIO support"
 	depends on 44x
 	select ARCH_REQUIRE_GPIOLIB
-	select GENERIC_GPIO
 	help
 	  Enable gpiolib support for ppc440 based boards
 
arch/powerpc/platforms/85xx/Kconfig +0 -1
index a0dcd57..78549b6 100644
@@ -203,7 +203,6 @@ config GE_IMP3A
 	select DEFAULT_UIMAGE
 	select SWIOTLB
 	select MMIO_NVRAM
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	select GE_FPGA
 	help
arch/powerpc/platforms/86xx/Kconfig +0 -3
index 7a6279e..1afd1e4 100644
@@ -37,7 +37,6 @@ config GEF_PPC9A
 	bool "GE PPC9A"
 	select DEFAULT_UIMAGE
 	select MMIO_NVRAM
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	select GE_FPGA
 	help
@@ -47,7 +46,6 @@ config GEF_SBC310
 	bool "GE SBC310"
 	select DEFAULT_UIMAGE
 	select MMIO_NVRAM
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	select GE_FPGA
 	help
@@ -57,7 +55,6 @@ config GEF_SBC610
 	bool "GE SBC610"
 	select DEFAULT_UIMAGE
 	select MMIO_NVRAM
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	select GE_FPGA
 	select HAS_RAPIDIO
arch/powerpc/platforms/8xx/Kconfig +0 -1
index 1fb0b3c..8dec3c0 100644
@@ -114,7 +114,6 @@ config 8xx_COPYBACK
 
 config 8xx_GPIO
 	bool "GPIO API Support"
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	help
 	  Saying Y here will cause the ports on an MPC8xx processor to be used
arch/powerpc/platforms/Kconfig +0 -4
index 52de8bc..15095d5 100644
@@ -303,7 +303,6 @@ config QUICC_ENGINE
 config QE_GPIO
 	bool "QE GPIO support"
 	depends on QUICC_ENGINE
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	help
 	  Say Y here if you're going to use hardware that connects to the
@@ -316,7 +315,6 @@ config CPM2
 	select PPC_LIB_RHEAP
 	select PPC_PCI_CHOICE
 	select ARCH_REQUIRE_GPIOLIB
-	select GENERIC_GPIO
 	help
 	  The CPM2 (Communications Processor Module) is a coprocessor on
 	  embedded CPUs made by Freescale.  Selecting this option means that
@@ -355,7 +353,6 @@ config OF_RTC
 config SIMPLE_GPIO
 	bool "Support for simple, memory-mapped GPIO controllers"
 	depends on PPC
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	help
 	  Say Y here to support simple, memory-mapped GPIO controllers.
@@ -366,7 +363,6 @@ config SIMPLE_GPIO
 config MCU_MPC8349EMITX
 	bool "MPC8349E-mITX MCU driver"
 	depends on I2C=y && PPC_83xx
-	select GENERIC_GPIO
 	select ARCH_REQUIRE_GPIOLIB
 	help
 	  Say Y here to enable soft power-off functionality on the Freescale
Signature · 1 line
1.8.1.5
#9 lore raw

RFC 9/17 sh: replace CONFIG_GENERIC_GPIO by CONFIG_GPIOLIB

In reply to Alexandre Courbot

SH GPIO drivers all use gpiolib and CONFIG_GENERIC_GPIO is only selected through CONFIG_GPIOLIB, yet some compilation units depended on CONFIG_GENERIC_GPIO. Make them depend on CONFIG_GPIOLIB instead since it is more accurate and prepares us for the future removal of CONFIG_GENERIC_GPIO.

---
 arch/sh/boards/mach-sdk7786/Makefile | 2 +-
 arch/sh/boards/mach-x3proto/Makefile | 2 +-
 arch/sh/kernel/cpu/sh2a/Makefile     | 2 +-
 arch/sh/kernel/cpu/sh3/Makefile      | 2 +-
 arch/sh/kernel/cpu/sh4a/Makefile     | 2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)
arch/sh/boards/mach-sdk7786/Makefile +1 -1
index 8ae56e9..45d32e3 100644
@@ -1,4 +1,4 @@
 obj-y	:= fpga.o irq.o nmi.o setup.o
 
-obj-$(CONFIG_GENERIC_GPIO)	+= gpio.o
+obj-$(CONFIG_GPIOLIB)		+= gpio.o
 obj-$(CONFIG_HAVE_SRAM_POOL)	+= sram.o
arch/sh/boards/mach-x3proto/Makefile +1 -1
index 708c21c..0cbe3d0 100644
@@ -1,3 +1,3 @@
 obj-y += setup.o ilsel.o
 
-obj-$(CONFIG_GENERIC_GPIO)	+= gpio.o
+obj-$(CONFIG_GPIOLIB)		+= gpio.o
arch/sh/kernel/cpu/sh2a/Makefile +1 -1
index 130984c..92f0da4 100644
@@ -21,5 +21,5 @@ pinmux-$(CONFIG_CPU_SUBTYPE_SH7203)	:= pinmux-sh7203.o
 pinmux-$(CONFIG_CPU_SUBTYPE_SH7264)	:= pinmux-sh7264.o
 pinmux-$(CONFIG_CPU_SUBTYPE_SH7269)	:= pinmux-sh7269.o
 
-obj-$(CONFIG_GENERIC_GPIO)		+= $(pinmux-y)
+obj-$(CONFIG_GPIOLIB)			+= $(pinmux-y)
 obj-$(CONFIG_HAVE_HW_BREAKPOINT)	+= ubc.o
arch/sh/kernel/cpu/sh3/Makefile +1 -1
index 6f13f33..d3634ae 100644
@@ -30,4 +30,4 @@ clock-$(CONFIG_CPU_SUBTYPE_SH7712)	:= clock-sh7712.o
 pinmux-$(CONFIG_CPU_SUBTYPE_SH7720)	:= pinmux-sh7720.o
 
 obj-y	+= $(clock-y)
-obj-$(CONFIG_GENERIC_GPIO)	+= $(pinmux-y)
+obj-$(CONFIG_GPIOLIB)			+= $(pinmux-y)
arch/sh/kernel/cpu/sh4a/Makefile +1 -1
index 8fc6ec2..0705df7 100644
@@ -47,6 +47,6 @@ pinmux-$(CONFIG_CPU_SUBTYPE_SHX3)	:= pinmux-shx3.o
 
 obj-y					+= $(clock-y)
 obj-$(CONFIG_SMP)			+= $(smp-y)
-obj-$(CONFIG_GENERIC_GPIO)		+= $(pinmux-y)
+obj-$(CONFIG_GPIOLIB)			+= $(pinmux-y)
 obj-$(CONFIG_PERF_EVENTS)		+= perf_event.o
 obj-$(CONFIG_HAVE_HW_BREAKPOINT)	+= ubc.o
Signature · 1 line
1.8.1.5
#10 lore raw

Re: [RFC 09/17] sh: replace CONFIG_GENERIC_GPIO by CONFIG_GPIOLIB

Quoted 7 lines · Alexandre Courbot

SH GPIO drivers all use gpiolib and CONFIG_GENERIC_GPIO is only selected through CONFIG_GPIOLIB, yet some compilation units depended on CONFIG_GENERIC_GPIO. Make them depend on CONFIG_GPIOLIB instead since it is more accurate and prepares us for the future removal of CONFIG_GENERIC_GPIO.

Note that the bulk of the GENERIC_GPIO use for SH has shifted to drivers/pinctrl/sh-pfc/. If GPIOLIB is forced then a good chunk of the Kconfig/Makefile bits there ought to be refactored too.

#11 lore raw

Re: [RFC 09/17] sh: replace CONFIG_GENERIC_GPIO by CONFIG_GPIOLIB

Quoted 7 lines · Paul Mundt
Quoted 6 lines · Alexandre Courbot

SH GPIO drivers all use gpiolib and CONFIG_GENERIC_GPIO is only selected through CONFIG_GPIOLIB, yet some compilation units depended on CONFIG_GENERIC_GPIO. Make them depend on CONFIG_GPIOLIB instead since it is more accurate and prepares us for the future removal of CONFIG_GENERIC_GPIO.

Quoted 5 lines
Quoted 1 line

Note that the bulk of the GENERIC_GPIO use for SH has shifted to drivers/pinctrl/sh-pfc/. If GPIOLIB is forced then a good chunk of the Kconfig/Makefile bits there ought to be refactored too.

Yes, that is the case for many other drivers actually. This series just makes sure that GENERIC_GPIO is set through GPIOLIB *only*, thus making both options equivalent. If nobody shoots me for this, I will refactor individual drivers and ultimately remove GENERIC_GPIO completely.

Alex.

Acked-by
#12 lore raw

Re: [RFC 09/17] sh: replace CONFIG_GENERIC_GPIO by CONFIG_GPIOLIB

Quoted 14 lines · Alex Courbot
Quoted 7 lines · Paul Mundt
Quoted 6 lines · Alexandre Courbot

SH GPIO drivers all use gpiolib and CONFIG_GENERIC_GPIO is only selected through CONFIG_GPIOLIB, yet some compilation units depended on CONFIG_GENERIC_GPIO. Make them depend on CONFIG_GPIOLIB instead since it is more accurate and prepares us for the future removal of CONFIG_GENERIC_GPIO.

Quoted 5 lines
Quoted 1 line

Note that the bulk of the GENERIC_GPIO use for SH has shifted to drivers/pinctrl/sh-pfc/. If GPIOLIB is forced then a good chunk of the Kconfig/Makefile bits there ought to be refactored too.

Quoted 6 lines

Yes, that is the case for many other drivers actually. This series just makes sure that GENERIC_GPIO is set through GPIOLIB *only*, thus making both options equivalent. If nobody shoots me for this, I will refactor individual drivers and ultimately remove GENERIC_GPIO completely.

Ok, just wanted to make sure you were aware of this. I don't have any strong opinions on this one way or the other, the entire GPIO Kconfig mess is too ugly for words, so any effort to clean that up is probably worth the added bloat.

#13 lore raw

RFC 10/17 xtensa: remove explicit selection of GENERIC_GPIO

In reply to Alexandre Courbot

The Xtensa GPIO driver uses gpiolib and cannot compile with only GENERIC_GPIO.

---
 arch/xtensa/configs/iss_defconfig   | 1 -
 arch/xtensa/configs/s6105_defconfig | 1 -
 2 files changed, 2 deletions(-)
arch/xtensa/configs/iss_defconfig +0 -1
index ddab37b..77c52f8 100644
@@ -10,7 +10,6 @@ CONFIG_RWSEM_XCHGADD_ALGORITHM=y
 CONFIG_GENERIC_FIND_NEXT_BIT=y
 CONFIG_GENERIC_HWEIGHT=y
 CONFIG_GENERIC_HARDIRQS=y
-CONFIG_GENERIC_GPIO=y
 # CONFIG_ARCH_HAS_ILOG2_U32 is not set
 # CONFIG_ARCH_HAS_ILOG2_U64 is not set
 CONFIG_NO_IOPORT=y
arch/xtensa/configs/s6105_defconfig +0 -1
index eaf1b8f..4799c6a 100644
@@ -10,7 +10,6 @@ CONFIG_RWSEM_XCHGADD_ALGORITHM=y
 CONFIG_GENERIC_FIND_NEXT_BIT=y
 CONFIG_GENERIC_HWEIGHT=y
 CONFIG_GENERIC_HARDIRQS=y
-CONFIG_GENERIC_GPIO=y
 # CONFIG_ARCH_HAS_ILOG2_U32 is not set
 # CONFIG_ARCH_HAS_ILOG2_U64 is not set
 CONFIG_NO_IOPORT=y
Signature · 1 line
1.8.1.5
#14 lore raw

RFC 11/17 mips: alchemy: require gpiolib

In reply to Alexandre Courbot

Require gpiolib instead of just GENERIC_GPIO.

---
 arch/mips/Kconfig | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
arch/mips/Kconfig +1 -2
index ae70ab1..5cfa818 100644
@@ -61,8 +61,7 @@ config MIPS_ALCHEMY
 	select SYS_HAS_CPU_MIPS32_R1
 	select SYS_SUPPORTS_32BIT_KERNEL
 	select SYS_SUPPORTS_APM_EMULATION
-	select GENERIC_GPIO
-	select ARCH_WANT_OPTIONAL_GPIOLIB
+	select ARCH_REQUIRE_GPIOLIB
 	select SYS_SUPPORTS_ZBOOT
 	select USB_ARCH_HAS_OHCI
 	select USB_ARCH_HAS_EHCI
Signature · 1 line
1.8.1.5
#15 lore raw

RFC 12/17 mips: pnx833x: remove requirement for GENERIC_GPIO

In reply to Alexandre Courbot

pnx833x does not seem to use the generic gpio API.

---
 arch/mips/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/mips/Kconfig +0 -1
index 5cfa818..c1c6549 100644
@@ -1107,7 +1107,6 @@ config SOC_PNX833X
 	select SYS_SUPPORTS_32BIT_KERNEL
 	select SYS_SUPPORTS_LITTLE_ENDIAN
 	select SYS_SUPPORTS_BIG_ENDIAN
-	select GENERIC_GPIO
 	select CPU_MIPSR2_IRQ_VI
 
 config SOC_PNX8335
Signature · 1 line
1.8.1.5
#16 lore raw

RFC 13/17 avr32: default GENERIC_GPIO to false

In reply to Alexandre Courbot

This will help removing the GENERIC_GPIO option.

GPIO access is done through gpiolib, which selects GENERIC_GPIO anyway.

---
 arch/avr32/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/avr32/Kconfig +1 -1
index 9b89257..85c377e 100644
@@ -27,7 +27,7 @@ config AVR32
 	  http://avr32linux.org/.
 
 config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
 config STACKTRACE_SUPPORT
 	def_bool y
Signature · 1 line
1.8.1.5
#17 lore raw

RFC 14/17 m68k: coldfire: use gpiolib

In reply to Alexandre Courbot

Force use of gpiolib for Coldfire, as a step towards the deprecation of GENERIC_GPIO.

---
 arch/m68k/Kconfig.cpu | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
arch/m68k/Kconfig.cpu +1 -2
index b1cfff8..d266787 100644
@@ -22,8 +22,7 @@ config M68KCLASSIC
 
 config COLDFIRE
 	bool "Coldfire CPU family support"
-	select GENERIC_GPIO
-	select ARCH_WANT_OPTIONAL_GPIOLIB
+	select ARCH_REQUIRE_GPIOLIB
 	select ARCH_HAVE_CUSTOM_GPIO_H
 	select CPU_HAS_NO_BITFIELDS
 	select CPU_HAS_NO_MULDIV64
Signature · 1 line
1.8.1.5
#18 lore raw

RFC 15/17 avr32: default GENERIC_GPIO to false

In reply to Alexandre Courbot
---
 arch/blackfin/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/blackfin/Kconfig +1 -1
index 600494c..b947e26 100644
@@ -56,7 +56,7 @@ config ZONE_DMA
 	def_bool y
 
 config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
 config FORCE_MAX_ZONEORDER
 	int
Signature · 1 line
1.8.1.5
#19 lore raw

Re: [RFC 15/17] avr32: default GENERIC_GPIO to false

Around Tue 12 Mar 2013 19:12:28 +0900 or thereabout, Alexandre Courbot wrote:

$(Subject) is wrong, you state avr32, but it is for blackfin.

Quoted 18 lines
---
 arch/blackfin/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/blackfin/Kconfig +1 -1
index 600494c..b947e26 100644
@@ -56,7 +56,7 @@ config ZONE_DMA
 	def_bool y
 
 config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
 config FORCE_MAX_ZONEORDER
 	int
Signature · 2 lines
mvh
Hans-Christian Egtvedt
#20 lore raw

Re: [RFC 15/17] avr32: default GENERIC_GPIO to false

Quoted 3 lines · Hans-Christian Egtvedt

Around Tue 12 Mar 2013 19:12:28 +0900 or thereabout, Alexandre Courbot wrote:

$(Subject) is wrong, you state avr32, but it is for blackfin.

Argh, apologies. That's very wrong, especially since blackfin is one of the archs most affected by this. Thanks for catching this, I hope the maintainers will notice it too.

Alex.

#21 lore raw

RFC 16/17 openrisc: default GENERIC_GPIO to false

In reply to Alexandre Courbot

This is one step towards the removal of the GENERIC_GPIO option. OpenRISC mandates the use of GPIOLIB, which enables GENERIC_GPIO anyway, so this patch should be a no-op.

---
 arch/openrisc/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/openrisc/Kconfig +1 -1
index 28df179..5366bbf 100644
@@ -46,7 +46,7 @@ config NO_IOPORT
 	def_bool y
 
 config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
 config TRACE_IRQFLAGS_SUPPORT
         def_bool y
Signature · 1 line
1.8.1.5
maintainer Acked-by on v1 16/17
#22 lore raw

Re: [RFC 16/17] openrisc: default GENERIC_GPIO to false

Quoted 3 lines · Alexandre Courbot

This is one step towards the removal of the GENERIC_GPIO option. OpenRISC mandates the use of GPIOLIB, which enables GENERIC_GPIO anyway, so this patch should be a no-op.

Quoted 18 lines
---
  arch/openrisc/Kconfig | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
arch/openrisc/Kconfig +1 -1
index 28df179..5366bbf 100644
@@ -46,7 +46,7 @@ config NO_IOPORT
  	def_bool y
 
  config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
  config TRACE_IRQFLAGS_SUPPORT
          def_bool y
#23 lore raw

RFC 17/17 unicore32: default GENERIC_GPIO to false

In reply to Alexandre Courbot

GPIO implementation for unicore32 uses gpiolib, which sets GENERIC_GPIO when selected.

---
 arch/unicore32/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/unicore32/Kconfig +1 -1
index a889117..9bb1180 100644
@@ -24,7 +24,7 @@ config UNICORE32
 	  Please see web page at <http://www.pkunity.com/>.
 
 config GENERIC_GPIO
-	def_bool y
+	def_bool n
 
 config GENERIC_CSUM
 	def_bool y
Signature · 1 line
1.8.1.5
#24 lore raw

Re: [RFC 17/17] unicore32: default GENERIC_GPIO to false

arch/unicore32/Kconfig config UNICORE32 · line 28

Quoted 7 lines · Alexandre Courbot
arch/unicore32/Kconfig +1 -1
@@ -24,7 +24,7 @@ config UNICORE32
          Please see web page at <http://www.pkunity.com/>.
 
 config GENERIC_GPIO
-       def_bool y
+       def_bool n
 

"def_bool n" is the same as "bool", but in general the latter syntax is preferred.

Arnd

A message that is not in the archive (sent off-list or not yet received): CAAVeFuLJ0VvdOkqyNaVCbFHE-x==OYFC_=HEF2YLqaK6mM+rBw@mail.gmail.com

#26 lore raw

Re: [RFC 17/17] unicore32: default GENERIC_GPIO to false

Quoted 6 lines · Alexandre Courbot

Thanks, I will fix this in the next version.

Btw, if this (or rather the next iteration) gets merged, may I suggest to do it through the GPIO tree instead of having each architecture picking the patches of relevance. A considerable number of other patches that depend on this will follow and this will ensure they are picked in the right order.

Sounds reasonable, but I would also suggest you prepare a branch that can get merged into the GPIO, ARM, PowerPC and MIPS trees to avoid merge conflicts in that case, since these all will have a number of changes conflicting with yours.

In the end, I would not care which tree it gets merged through, as long as all have the same version of your patches.

Arnd

#27 lore raw

Re: [RFC 17/17] unicore32: default GENERIC_GPIO to false

In reply to Arnd Bergmann
Quoted 10 lines · Arnd Bergmann

arch/unicore32/Kconfig config UNICORE32 · line 28

Quoted 7 lines · Alexandre Courbot
arch/unicore32/Kconfig +1 -1
@@ -24,7 +24,7 @@ config UNICORE32
          Please see web page at <http://www.pkunity.com/>.
 
 config GENERIC_GPIO
-       def_bool y
+       def_bool n
 

"def_bool n" is the same as "bool", but in general the latter syntax is preferred.

I have fixed this for unicore32, but a few other architectures use the "def_bool n" syntax consistently across their Kconfig, so for consistency reasons it is maybe preferable to leave it that way for them.

Alex.

#28 lore raw

Re: [RFC 17/17] unicore32: default GENERIC_GPIO to false

Quoted 4 lines · Alexandre Courbot

I have fixed this for unicore32, but a few other architectures use the "def_bool n" syntax consistently across their Kconfig, so for consistency reasons it is maybe preferable to leave it that way for them.

Ok, fair enough.

Arnd

Acked-by on the v1 series
#29 lore raw
In reply to Alexandre Courbot
Quoted 4 lines · Alexandre Courbot

This series makes sure the GENERIC_GPIO option can only be set through GPIOLIB (and not by individual architectures), as a first step towards its removal.

Nice series,

Quoted 3 lines

For most platforms, this change should be a no-op. However I would like to make sure that everyone is ok with it and that nothing gets broken, as the effect of changing configuration options are sometimes difficult to predict.

It's probably worth pointing out which platforms actually change with this. For all I can tell, the only user-visible change is that m68k now cannot turn off gpiolib any more.

Arnd

#30 lore raw
Quoted 3 lines · Arnd Bergmann

Nice series,

Thanks!

Quoted 7 lines
Quoted 3 lines

For most platforms, this change should be a no-op. However I would like to make sure that everyone is ok with it and that nothing gets broken, as the effect of changing configuration options are sometimes difficult to predict.

It's probably worth pointing out which platforms actually change with this. For all I can tell, the only user-visible change is that m68k now cannot turn off gpiolib any more.

That's the one we are sure is affected indeed. I am also a little bit worried about blackfin, since it used to compile with GENERIC_GPIO and patch #15 removes that. However it seems that blackfin's GPIO driver is unconditionally compiled, so the whole architecture might actually need ARCH_REQUIRE_GPIOLIB now.

There are also some more subtle changes in mips and arm where I rely on indirect selection of gpiolib to assume that GENERIC_GPIO will remain true - it's a bit of a headache, actually. Anyway the arch maintainers know better what they really need, so I prefer not to assume too much on my own.

Alex.

#32 lore raw
Quoted 9 lines · Linus Walleij
Quoted 2 lines · Alexandre Courbot

This series makes sure the GENERIC_GPIO option can only be set through GPIOLIB (and not by individual architectures), as a first step towards its removal.

Nice!

I bet something will break, anyway: no pain no gain.

Since this is a pretty invasive patch series that touches a lot of architectures, we should get it into a separate branch that can get pulled into linux-next.

g.

#33 lore raw
Quoted 6 lines · Grant Likely
Quoted 5 lines · Linus Walleij
Quoted 2 lines · Alexandre Courbot

This series makes sure the GENERIC_GPIO option can only be set through GPIOLIB (and not by individual architectures), as a first step towards its removal.

Quoted 8 lines
Quoted 4 lines

Nice!

I bet something will break, anyway: no pain no gain.

Since this is a pretty invasive patch series that touches a lot of architectures, we should get it into a separate branch that can get pulled into linux-next.

Good point. Alexandre, do you have a git repo to attain this?

Yours,
Linus Walleij

#34 lore raw
Quoted 10 lines · Linus Walleij
Quoted 9 lines · Grant Likely
Quoted 8 lines · Linus Walleij
Quoted 2 lines · Alexandre Courbot

This series makes sure the GENERIC_GPIO option can only be set through GPIOLIB (and not by individual architectures), as a first step towards its removal.

Nice!

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

I bet something will break, anyway: no pain no gain.

Since this is a pretty invasive patch series that touches a lot of architectures, we should get it into a separate branch that can get pulled into linux-next.

Good point. Alexandre, do you have a git repo to attain this?

Would be reasonable indeed. I can publish the branch on Github, but (sorry for being clueless) what's the procedure to add a new branch to linux-next? Shall I write to Stephen Rothwell directly?

Alex.

#35 lore raw
Quoted 1 line · Alexandre Courbot

Would be reasonable indeed. I can publish the branch on Github,

That'll work.

For submitting the pull request to Torvalds in the end you will need to create a signed tag. Do you have a signed GPG key? Else it's about time to create a key and start collecting signatures... then you can also get a repo at kernel.org in the future.

Quoted 3 lines

but
(sorry for being clueless) what's the procedure to add a new branch to linux-next? Shall I write to Stephen Rothwell directly?

Yes. Join the linux-next mailing list too.

Yours,
Linus Walleij

#36 lore raw
Quoted 4 lines · Linus Walleij

For submitting the pull request to Torvalds in the end you will need to create a signed tag. Do you have a signed GPG key? Else it's about time to create a key and start collecting signatures... then you can also get a repo at kernel.org in the future.

I have one or two signatures of kernel devs on my GPG key, but I thought that maybe this set could be pulled by you or Grant once it has undergone enough testing - it's not like I intend to push regularly on this branch.

If I understand correctly the reason for having this in its own branch is to be able remove it quickly would it attempt to screw the dog.

Quoted 4 lines
Quoted 2 lines

(sorry for being clueless) what's the procedure to add a new branch to linux-next? Shall I write to Stephen Rothwell directly?

Yes. Join the linux-next mailing list too.

Will do that, thanks. First of all let me send the relevant patches again to blackfin and m68k and get some feedback from them, since they are the most affected.

Alex.

#37 lore raw
Quoted 8 lines · Alexandre Courbot
Quoted 4 lines · Linus Walleij

For submitting the pull request to Torvalds in the end you will need to create a signed tag. Do you have a signed GPG key? Else it's about time to create a key and start collecting signatures... then you can also get a repo at kernel.org in the future.

Quoted 7 lines

I have one or two signatures of kernel devs on my GPG key, but I thought that maybe this set could be pulled by you or Grant once it has undergone enough testing - it's not like I intend to push regularly on this branch.

If I understand correctly the reason for having this in its own branch is to be able remove it quickly would it attempt to screw the dog.

The main reason I think is so that branch can be pulled by other people that don't want to pull in all of the other gpio changes.

It could still just be another branch in the existing gpio tree, at least once everyone is happy with the state of the patches.

Arnd