MIPS: Lantiq: switch vmmc to use gpiod API

PATCH Applied Dmitry Torokhov 2 participants 8 messages Sep 23, 2022 – Sep 27, 2022

Subsystems (2)

MIPS Lantiq MIPS architecture

lore

Share & tools

lore

Dmitry Torokhov

author

2022-09-23 04:55 UTC #0 lore raw

PATCH 1/1 MIPS: Lantiq: switch vmmc to use gpiod API

This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.

Signed-off-by Dmitry Torokhov <[email protected]>

---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

arch/mips/lantiq/xway/vmmc.c +13 -9

index 7a14da8d9d15..a1947306ac18 100644
@@ -4,9 +4,10 @@
  *  Copyright (C) 2012 John Crispin <[email protected]>
  */
 
+#include <linux/err.h>
 #include <linux/export.h>
+#include <linux/gpio/consumer.h>
 #include <linux/of_platform.h>
-#include <linux/of_gpio.h>
 #include <linux/dma-mapping.h>
 
 #include <lantiq_soc.h>

@@ -25,6 +26,7 @@ EXPORT_SYMBOL(ltq_get_cp1_base);
 static int vmmc_probe(struct platform_device *pdev)
 {
 #define CP1_SIZE       (1 << 20)
+	struct gpio_desc *gpio;
 	int gpio_count;
 	dma_addr_t dma;

@@ -32,16 +34,18 @@ static int vmmc_probe(struct platform_device *pdev)
 		(void *) CPHYSADDR(dma_alloc_coherent(&pdev->dev, CP1_SIZE,
 						    &dma, GFP_KERNEL));
 
-	gpio_count = of_gpio_count(pdev->dev.of_node);
+	gpio_count = gpiod_count(&pdev->dev, NULL);
 	while (gpio_count > 0) {
-		enum of_gpio_flags flags;
-		int gpio = of_get_gpio_flags(pdev->dev.of_node,
-					     --gpio_count, &flags);
-		if (gpio_request(gpio, "vmmc-relay"))
+		gpio = devm_gpiod_get_index(&pdev->dev,
+					    NULL, --gpio_count, GPIOD_OUT_HIGH);
+		if (IS_ERR(gpio)) {
+			dev_err(&pdev->dev,
+				"failed to request GPIO idx %d: %d\n",
+				gpio_count, PTR_ERR(gpio);
 			continue;
-		dev_info(&pdev->dev, "requested GPIO %d\n", gpio);
-		gpio_direction_output(gpio,
-				      (flags & OF_GPIO_ACTIVE_LOW) ? (0) : (1));
+		}
+
+		gpio_consumer_set_name(gpio, "vmmc-relay");
 	}
 
 	dev_info(&pdev->dev, "reserved %dMB at 0x%p", CP1_SIZE >> 20, cp1_base);

Signature · 5 lines

2.37.3.998.g577e59143f-goog


-- 
Dmitry

Thomas Bogendoerfer

maintainer Applied

2022-09-24 12:06 UTC #1 lore raw

Quoted 7 lines · Dmitry Torokhov

This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

applied to mips-next.

Thomas.

Signature · 2 lines

Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Dmitry Torokhov

author

2022-09-27 04:56 UTC #2 lore raw

Hi Thomas,

Quoted 10 lines · Thomas Bogendoerfer

Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)
applied to mips-next.

My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.

Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.

Thanks.

Signature · 12 lines

Dmitry


MIPS: Lantiq: vmmc: fix compile break introduced by gpiod patch

"MIPS: Lantiq: switch vmmc to use gpiod API" patch introduced compile
errors, this patch fixes them.

Signed-off-by: Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

arch/mips/lantiq/xway/vmmc.c +2 -2

index a1947306ac18..b6b0d9884b88 100644
@@ -41,11 +41,11 @@ static int vmmc_probe(struct platform_device *pdev)
 		if (IS_ERR(gpio)) {
 			dev_err(&pdev->dev,
 				"failed to request GPIO idx %d: %d\n",
-				gpio_count, PTR_ERR(gpio);
+				gpio_count, PTR_ERR(gpio));
 			continue;
 		}
 
-		gpio_consumer_set_name(gpio, "vmmc-relay");
+		gpiod_set_consumer_name(gpio, "vmmc-relay");
 	}
 
 	dev_info(&pdev->dev, "reserved %dMB at 0x%p", CP1_SIZE >> 20, cp1_base);

Signature · 1 line

2.38.0.rc1.362.ged0d419d3c-goog

Thomas Bogendoerfer

maintainer

2022-09-27 07:50 UTC #3 lore raw

Quoted 12 lines · Dmitry Torokhov

Hi Thomas,
Quoted 9 lines · Thomas Bogendoerfer
Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

Quoted 7 lines

Quoted 1 line
applied to mips-next.
My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.
Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.

I need a fixup, but this one still fails in my build:

/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c: In function ‘vmmc_probe’:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:43:5: error: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Werror=format=]
"failed to request GPIO idx %d: %d\n",
^
/local/tbogendoerfer/korg/linux/include/linux/dev_printk.h:110:16: note: in definition of macro ‘dev_printk_index_wrap’

   _p_func(dev, fmt, ##__VA_ARGS__);   \
                ^~~
/local/tbogendoerfer/korg/linux/include/linux/dev_printk.h:144:49: note: in expansion of macro ‘dev_fmt’
  dev_printk_index_wrap(_dev_err, KERN_ERR, dev, dev_fmt(fmt), ##__VA_ARGS__)
                                                 ^~~~~~~
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:42:4: note: in expansion of macro ‘dev_err’
    dev_err(&pdev->dev,
    ^~~~~~~

Thomas.

Signature · 2 lines

Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Dmitry Torokhov

author

2022-09-27 08:13 UTC #4 lore raw

Quoted 21 lines · Thomas Bogendoerfer

Quoted 12 lines · Dmitry Torokhov
Hi Thomas,
Quoted 9 lines · Thomas Bogendoerfer
Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)
Quoted 7 lines
Quoted 1 line
applied to mips-next.
My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.
Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.

Quoted 6 lines

I need a fixup, but this one still fails in my build:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c: In function ‘vmmc_probe’:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:43:5: error: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Werror=format=]
"failed to request GPIO idx %d: %d\n",
^

I see, I did not realize PTR_ERR() is actually long. I guess I can introduce a temp variable and use PTR_ERR_OR_ZERO(), but there are a lot of places in the kernel that use %d and PTR_ERR(). I wonder why we can't define PTR_ERR() as (int)(long)ptr or something.

What compiler/version are you using for your builds?

Thanks.

Signature · 1 line

Dmitry

Thomas Bogendoerfer

maintainer

2022-09-27 08:29 UTC #5 lore raw

Quoted 24 lines · Dmitry Torokhov

Quoted 23 lines · Thomas Bogendoerfer
Quoted 12 lines · Dmitry Torokhov
Hi Thomas,
Quoted 9 lines · Thomas Bogendoerfer
Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)
Quoted 7 lines
Quoted 1 line
applied to mips-next.
My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.
Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.
I need a fixup, but this one still fails in my build:

Quoted 8 lines

Quoted 4 lines
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c: In function ‘vmmc_probe’:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:43:5: error: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Werror=format=]
"failed to request GPIO idx %d: %d\n",
^
I see, I did not realize PTR_ERR() is actually long. I guess I can introduce a temp variable and use PTR_ERR_OR_ZERO(), but there are a lot of places in the kernel that use %d and PTR_ERR(). I wonder why we can't define PTR_ERR() as (int)(long)ptr or something.
What compiler/version are you using for your builds?

it's rather old:

gcc version 6.1.1 20160621 (Red Hat Cross 6.1.1-2) (GCC)

Thomas.

Signature · 2 lines

Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]

Dmitry Torokhov

author

2022-09-27 15:29 UTC #6 lore raw

Quoted 30 lines · Thomas Bogendoerfer

Quoted 29 lines · Dmitry Torokhov
Quoted 27 lines · Thomas Bogendoerfer
Quoted 12 lines · Dmitry Torokhov
Hi Thomas,
Quoted 9 lines · Thomas Bogendoerfer
Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)
Quoted 7 lines
Quoted 1 line
applied to mips-next.
My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.
Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.
I need a fixup, but this one still fails in my build:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c: In function ‘vmmc_probe’:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:43:5: error: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Werror=format=]
"failed to request GPIO idx %d: %d\n",
^

Quoted 7 lines

Quoted 3 lines
I see, I did not realize PTR_ERR() is actually long. I guess I can introduce a temp variable and use PTR_ERR_OR_ZERO(), but there are a lot of places in the kernel that use %d and PTR_ERR(). I wonder why we can't define PTR_ERR() as (int)(long)ptr or something.
What compiler/version are you using for your builds?
it's rather old:
gcc version 6.1.1 20160621 (Red Hat Cross 6.1.1-2) (GCC)

OK, I tried the below with gcc 12.1.0 cross-compiler, hopefully this does not trip on 6.1.1.

Thanks.

Signature · 11 lines

Dmitry

MIPS: Lantiq: vmmc: fix compile break introduced by gpiod patch

"MIPS: Lantiq: switch vmmc to use gpiod API" patch introduced compile
errors, this patch fixes them.

Signed-off-by: Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

arch/mips/lantiq/xway/vmmc.c +5 -3

index a1947306ac18..2796e87dfcae 100644
@@ -29,6 +29,7 @@ static int vmmc_probe(struct platform_device *pdev)
 	struct gpio_desc *gpio;
 	int gpio_count;
 	dma_addr_t dma;
+	int error;
 
 	cp1_base =
 		(void *) CPHYSADDR(dma_alloc_coherent(&pdev->dev, CP1_SIZE,

@@ -38,14 +39,15 @@ static int vmmc_probe(struct platform_device *pdev)
 	while (gpio_count > 0) {
 		gpio = devm_gpiod_get_index(&pdev->dev,
 					    NULL, --gpio_count, GPIOD_OUT_HIGH);
-		if (IS_ERR(gpio)) {
+		error = PTR_ERR_OR_ZERO(gpio);
+		if (error) {
 			dev_err(&pdev->dev,
 				"failed to request GPIO idx %d: %d\n",
-				gpio_count, PTR_ERR(gpio);
+				gpio_count, error);
 			continue;
 		}
 
-		gpio_consumer_set_name(gpio, "vmmc-relay");
+		gpiod_set_consumer_name(gpio, "vmmc-relay");
 	}
 
 	dev_info(&pdev->dev, "reserved %dMB at 0x%p", CP1_SIZE >> 20, cp1_base);

Signature · 1 line

2.38.0.rc1.362.ged0d419d3c-goog

Thomas Bogendoerfer

maintainer

2022-09-27 20:14 UTC #7 lore raw

Quoted 33 lines · Dmitry Torokhov

Quoted 32 lines · Thomas Bogendoerfer
Quoted 31 lines · Dmitry Torokhov
Quoted 21 lines · Thomas Bogendoerfer
Quoted 12 lines · Dmitry Torokhov
Hi Thomas,
Quoted 9 lines · Thomas Bogendoerfer
Quoted 7 lines · Dmitry Torokhov
This switches vmmc to use gpiod API instead of OF-specific legacy gpio API that we want to stop exporting from gpiolib.
Signed-off-by Dmitry Torokhov <[email protected]>
---
 arch/mips/lantiq/xway/vmmc.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)
Quoted 7 lines
Quoted 1 line
applied to mips-next.
My apologies, I screwed up. I thought this patch passed 0day before I sent it to you, but apparently it has not.
Here is a fixup (actually cross-compiled this time), or I can send a v2 incorporating it into the original change.
Quoted 6 lines
I need a fixup, but this one still fails in my build:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c: In function ‘vmmc_probe’:
/local/tbogendoerfer/korg/linux/arch/mips/lantiq/xway/vmmc.c:43:5: error: format ‘%d’ expects argument of type ‘int’, but argument 4 has type ‘long int’ [-Werror=format=]
"failed to request GPIO idx %d: %d\n",
^
I see, I did not realize PTR_ERR() is actually long. I guess I can introduce a temp variable and use PTR_ERR_OR_ZERO(), but there are a lot of places in the kernel that use %d and PTR_ERR(). I wonder why we can't define PTR_ERR() as (int)(long)ptr or something.

Quoted 8 lines

Quoted 5 lines
Quoted 1 line
What compiler/version are you using for your builds?
it's rather old:
gcc version 6.1.1 20160621 (Red Hat Cross 6.1.1-2) (GCC)
OK, I tried the below with gcc 12.1.0 cross-compiler, hopefully this does not trip on 6.1.1.

works with 6.1.1 as well. Applied to mips-next.

Thomas.

Signature · 2 lines

Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]