[PATCH 0/2] soc: samsung: be a module!

Hi,

A question - what is the convention for naming modules: underscores '_' or
hyphens '-'?

The C-unit file already uses a hyphen "exynos-chipid.c", so I could
create a module named:
1. exynos_chipid
2. soc-exynos-chipid
3. exynos-chip-id

The Linux device name is "exynos-chipid".

Best regards,
Krzysztof


Krzysztof Kozlowski (2):
  soc: export soc_device_to_device symbol
  soc: samsung: exynos-chipid: convert to a module

 drivers/base/soc.c                   |  1 +
 drivers/soc/samsung/Kconfig          |  3 ++-
 drivers/soc/samsung/Makefile         |  3 ++-
 drivers/soc/samsung/exynos-chipid.c  | 11 ++++++++++-
 drivers/soc/samsung/exynos5422-asv.c |  1 +
 5 files changed, 16 insertions(+), 3 deletions(-)

-- 
2.30.2


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

In case if soc-bus drivers are modules soc_device_to_device() has to be
exported.  Since it is trivial, export it as non-GPL.

Signed-off-by: Krzysztof Kozlowski <redacted>
---
 drivers/base/soc.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/base/soc.c b/drivers/base/soc.c
index 0af5363a582c..ac8db303daa1 100644
--- a/drivers/base/soc.c
+++ b/drivers/base/soc.c

@@ -41,6 +41,7 @@ struct device *soc_device_to_device(struct soc_device *soc_dev)
 {
 	return &soc_dev->dev;
 }
+EXPORT_SYMBOL(soc_device_to_device);
 
 static umode_t soc_attribute_mode(struct kobject *kobj,
 				struct attribute *attr,

-- 
2.30.2


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

Exynos ChipID and ASV (Adaptive Supply Voltage) driver is not essential
to system boot and it can successfully be built and loaded as module.

This makes core kernel image smaller and reduces the memory footprint
when multi-platform kernel is booted on non-Exynos board.  Usually it is
also distro-friendly.

Add multiple authors of the driver since its conversion from
mach-exynos, ordered alphabetically by first name.

Signed-off-by: Krzysztof Kozlowski <redacted>
---
 drivers/soc/samsung/Kconfig          |  3 ++-
 drivers/soc/samsung/Makefile         |  3 ++-
 drivers/soc/samsung/exynos-chipid.c  | 11 ++++++++++-
 drivers/soc/samsung/exynos5422-asv.c |  1 +
 4 files changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/soc/samsung/Kconfig b/drivers/soc/samsung/Kconfig
index 1f643c0f5c93..fe139f26d093 100644
--- a/drivers/soc/samsung/Kconfig
+++ b/drivers/soc/samsung/Kconfig

@@ -13,13 +13,14 @@ config EXYNOS_ASV_ARM
 	depends on EXYNOS_CHIPID
 
 config EXYNOS_CHIPID
-	bool "Exynos ChipID controller and ASV driver" if COMPILE_TEST
+	tristate "Exynos ChipID controller and ASV driver" if COMPILE_TEST
 	depends on ARCH_EXYNOS || COMPILE_TEST
 	select EXYNOS_ASV_ARM if ARM && ARCH_EXYNOS
 	select MFD_SYSCON
 	select SOC_BUS
 	help
 	  Support for Samsung Exynos SoC ChipID and Adaptive Supply Voltage.
+	  This driver can also be built as module (exynos_chipid).
 
 config EXYNOS_PMU
 	bool "Exynos PMU controller driver" if COMPILE_TEST

diff --git a/drivers/soc/samsung/Makefile b/drivers/soc/samsung/Makefile
index 0c523a8de4eb..2ae4bea804cf 100644
--- a/drivers/soc/samsung/Makefile
+++ b/drivers/soc/samsung/Makefile

@@ -1,8 +1,9 @@
 # SPDX-License-Identifier: GPL-2.0
 
 obj-$(CONFIG_EXYNOS_ASV_ARM)	+= exynos5422-asv.o
+obj-$(CONFIG_EXYNOS_CHIPID)	+= exynos_chipid.o
+exynos_chipid-y			+= exynos-chipid.o exynos-asv.o
 
-obj-$(CONFIG_EXYNOS_CHIPID)	+= exynos-chipid.o exynos-asv.o
 obj-$(CONFIG_EXYNOS_PMU)	+= exynos-pmu.o
 
 obj-$(CONFIG_EXYNOS_PMU_ARM_DRIVERS)	+= exynos3250-pmu.o exynos4-pmu.o \

diff --git a/drivers/soc/samsung/exynos-chipid.c b/drivers/soc/samsung/exynos-chipid.c
index 5c1d0f97f766..5e8c957231ff 100644
--- a/drivers/soc/samsung/exynos-chipid.c
+++ b/drivers/soc/samsung/exynos-chipid.c

@@ -15,6 +15,7 @@
 #include <linux/device.h>
 #include <linux/errno.h>
 #include <linux/mfd/syscon.h>
+#include <linux/module.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/regmap.h>

@@ -129,6 +130,7 @@ static const struct of_device_id exynos_chipid_of_device_ids[] = {
 	{ .compatible = "samsung,exynos4210-chipid" },
 	{}
 };
+MODULE_DEVICE_TABLE(of, exynos_chipid_of_device_ids);
 
 static struct platform_driver exynos_chipid_driver = {
 	.driver = {

@@ -138,4 +140,11 @@ static struct platform_driver exynos_chipid_driver = {
 	.probe	= exynos_chipid_probe,
 	.remove	= exynos_chipid_remove,
 };
-builtin_platform_driver(exynos_chipid_driver);
+module_platform_driver(exynos_chipid_driver);
+
+MODULE_DESCRIPTION("Samsung Exynos ChipID controller and ASV driver");
+MODULE_AUTHOR("Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>");
+MODULE_AUTHOR("Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>");
+MODULE_AUTHOR("Pankaj Dubey <pankaj.dubey@samsung.com>");
+MODULE_AUTHOR("Sylwester Nawrocki <s.nawrocki@samsung.com>");
+MODULE_LICENSE("GPL");

diff --git a/drivers/soc/samsung/exynos5422-asv.c b/drivers/soc/samsung/exynos5422-asv.c
index ca409a976e34..475ae5276529 100644
--- a/drivers/soc/samsung/exynos5422-asv.c
+++ b/drivers/soc/samsung/exynos5422-asv.c

@@ -503,3 +503,4 @@ int exynos5422_asv_init(struct exynos_asv *asv)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(exynos5422_asv_init);

-- 
2.30.2


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

On Fri, Sep 17, 2021 at 07:51:33PM +0200, Krzysztof Kozlowski wrote:

I thought we were getting rid of the use of this function, why export it
and not just fix it up instead?

Or am I confused about some other soc device abuse...

thanks,

greg k-h

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

On 18/09/2021 09:02, Greg Kroah-Hartman wrote:

On Fri, Sep 17, 2021 at 07:51:33PM +0200, Krzysztof Kozlowski wrote:

quoted

In case if soc-bus drivers are modules soc_device_to_device() has to be
exported.  Since it is trivial, export it as non-GPL.

Signed-off-by: Krzysztof Kozlowski <redacted>
---
 drivers/base/soc.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/base/soc.c b/drivers/base/soc.c
index 0af5363a582c..ac8db303daa1 100644
--- a/drivers/base/soc.c
+++ b/drivers/base/soc.c

@@ -41,6 +41,7 @@ struct device *soc_device_to_device(struct soc_device *soc_dev)
 {
 	return &soc_dev->dev;
 }
+EXPORT_SYMBOL(soc_device_to_device);

I thought we were getting rid of the use of this function, why export it
and not just fix it up instead?

Or am I confused about some other soc device abuse...

I was not aware of discussion about soc_device_to_device(). Thanks for
pointing out the issue. I googled a little and found previous talks:
https://lore.kernel.org/lkml/20191111052741.GB3176397@kroah.com/

I can easily get rid of soc_device_to_device() in my driver, so first
patch won't be needed.

Best regards,
Krzysztof

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

On Sun, Sep 19, 2021 at 11:12:29AM +0200, Krzysztof Kozlowski wrote:

On 18/09/2021 09:02, Greg Kroah-Hartman wrote:

quoted

On Fri, Sep 17, 2021 at 07:51:33PM +0200, Krzysztof Kozlowski wrote:

quoted

In case if soc-bus drivers are modules soc_device_to_device() has to be
exported.  Since it is trivial, export it as non-GPL.

Signed-off-by: Krzysztof Kozlowski <redacted>
---
 drivers/base/soc.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/base/soc.c b/drivers/base/soc.c
index 0af5363a582c..ac8db303daa1 100644
--- a/drivers/base/soc.c
+++ b/drivers/base/soc.c

@@ -41,6 +41,7 @@ struct device *soc_device_to_device(struct soc_device *soc_dev)
 {
 	return &soc_dev->dev;
 }
+EXPORT_SYMBOL(soc_device_to_device);

I thought we were getting rid of the use of this function, why export it
and not just fix it up instead?

Or am I confused about some other soc device abuse...

I was not aware of discussion about soc_device_to_device(). Thanks for
pointing out the issue. I googled a little and found previous talks:
https://lore.kernel.org/lkml/20191111052741.GB3176397@kroah.com/

I can easily get rid of soc_device_to_device() in my driver, so first
patch won't be needed.

Wonderful, thanks for fixing that up.

greg k-h

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

On Fri, 17 Sept 2021 at 20:51, Krzysztof Kozlowski
[off-list ref] wrote:

Exynos ChipID and ASV (Adaptive Supply Voltage) driver is not essential
to system boot and it can successfully be built and loaded as module.

This makes core kernel image smaller and reduces the memory footprint
when multi-platform kernel is booted on non-Exynos board.  Usually it is
also distro-friendly.

Add multiple authors of the driver since its conversion from
mach-exynos, ordered alphabetically by first name.

Signed-off-by: Krzysztof Kozlowski <redacted>
---

Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>

Just curious, how exactly ASV is related to chip-id driver?

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