Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1371451 > unrolled thread

[PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices Laxman Dewangan <ldewangan@nvidia.com> - 2016-04-05 14:00 +0200
    Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and  mfd_release_devices Lee Jones <lee.jones@linaro.org> - 2016-04-07 12:50 +0200
      Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices Laxman Dewangan <ldewangan@nvidia.com> - 2016-04-07 13:00 +0200
        Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and  mfd_release_devices Lee Jones <lee.jones@linaro.org> - 2016-04-07 13:50 +0200
          Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices Laxman Dewangan <ldewangan@nvidia.com> - 2016-04-07 14:00 +0200

#1371451 — [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices

Add device managed APIs devm_mfd_add_devices() and
devm_mfd_remove_devices() for the APIs mfd_add_devices()
and mfd_remove_devices().

This helps in reducing code in error path and sometimes
removal of .remove callback for driver unbind.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
 drivers/mfd/mfd-core.c   | 62 ++++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/core.h |  7 ++++++
 2 files changed, 69 insertions(+)

diff --git a/drivers/mfd/mfd-core.c b/drivers/mfd/mfd-core.c
index 409da01..145712e 100644
--- a/drivers/mfd/mfd-core.c
+++ b/drivers/mfd/mfd-core.c
@@ -334,6 +334,68 @@ void mfd_remove_devices(struct device *parent)
 }
 EXPORT_SYMBOL(mfd_remove_devices);
 
+static void devm_mfd_dev_release(struct device *dev, void *res)
+{
+	mfd_remove_devices(dev);
+}
+
+/**
+ * devm_mfd_add_devices - Resource managed version of mfd_add_devices()
+ *
+ * This returns the 0 on success otherwise error number in the failure.
+ * The allocated mfd devices will automatically be released when the
+ * device is unbound.
+ */
+int devm_mfd_add_devices(struct device *dev, int id,
+		    const struct mfd_cell *cells, int n_devs,
+		    struct resource *mem_base,
+		    int irq_base, struct irq_domain *domain)
+{
+	struct device **ptr;
+	int ret;
+
+	ptr = devres_alloc(devm_mfd_dev_release, sizeof(*ptr), GFP_KERNEL);
+	if (!ptr)
+		return -ENOMEM;
+
+	ret = mfd_add_devices(dev, id, cells, n_devs, mem_base,
+			      irq_base, domain);
+	if (!ret) {
+		*ptr = dev;
+		devres_add(dev, ptr);
+	} else {
+		devres_free(ptr);
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL(devm_mfd_add_devices);
+
+static int devm_mfd_devs_match(struct device *dev, void *res, void *data)
+{
+	struct device **r = res;
+
+	if (WARN_ON(!r || !*r))
+		return 0;
+
+	return *r == data;
+}
+
+/**
+ * devm_mfd_remove_device - Resource managed version of mfd_remove_devices()
+ * @dev: Device for which which resource was allocated.
+ *
+ * Remove all mfd devices added on the device.
+ * Normally this function will not need to be called and the resource
+ * management code will ensure that the resource is freed.
+ */
+void devm_mfd_remove_devices(struct device *dev)
+{
+	WARN_ON(devres_release(dev, devm_mfd_dev_release,
+			       devm_mfd_devs_match, dev));
+}
+EXPORT_SYMBOL(devm_mfd_remove_devices);
+
 int mfd_clone_cell(const char *cell, const char **clones, size_t n_clones)
 {
 	struct mfd_cell cell_entry;
diff --git a/include/linux/mfd/core.h b/include/linux/mfd/core.h
index bc6f7e0..d658d7f 100644
--- a/include/linux/mfd/core.h
+++ b/include/linux/mfd/core.h
@@ -131,4 +131,11 @@ static inline int mfd_add_hotplug_devices(struct device *parent,
 
 extern void mfd_remove_devices(struct device *parent);
 
+extern int devm_mfd_add_devices(struct device *dev, int id,
+			   const struct mfd_cell *cells, int n_devs,
+			   struct resource *mem_base,
+			   int irq_base, struct irq_domain *irq_domain);
+
+extern void devm_mfd_remove_devices(struct device *dev);
+
 #endif
-- 
2.1.4

[toc] | [next] | [standalone]

#1373267 — Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices

On Tue, 05 Apr 2016, Laxman Dewangan wrote:

> Add device managed APIs devm_mfd_add_devices() and
> devm_mfd_remove_devices() for the APIs mfd_add_devices()
> and mfd_remove_devices().

Nit: Line wrap after "devm_mfd_remove_devices()" instead.

> This helps in reducing code in error path and sometimes
> removal of .remove callback for driver unbind.

s/for/during/

> Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
> ---
>  drivers/mfd/mfd-core.c   | 62 ++++++++++++++++++++++++++++++++++++++++++++++++
>  include/linux/mfd/core.h |  7 ++++++
>  2 files changed, 69 insertions(+)
> 
> diff --git a/drivers/mfd/mfd-core.c b/drivers/mfd/mfd-core.c
> index 409da01..145712e 100644
> --- a/drivers/mfd/mfd-core.c
> +++ b/drivers/mfd/mfd-core.c
> @@ -334,6 +334,68 @@ void mfd_remove_devices(struct device *parent)
>  }
>  EXPORT_SYMBOL(mfd_remove_devices);
>  
> +static void devm_mfd_dev_release(struct device *dev, void *res)
> +{
> +	mfd_remove_devices(dev);
> +}
> +
> +/**
> + * devm_mfd_add_devices - Resource managed version of mfd_add_devices()
> + *
> + * This returns the 0 on success otherwise error number in the failure.

"Returns 0 on success or an appropriate negative error number on failure."

> + * The allocated mfd devices will automatically be released when the

s/mfd/MFD/

> + * device is unbound.
> + */
> +int devm_mfd_add_devices(struct device *dev, int id,
> +		    const struct mfd_cell *cells, int n_devs,
> +		    struct resource *mem_base,
> +		    int irq_base, struct irq_domain *domain)
> +{
> +	struct device **ptr;
> +	int ret;
> +
> +	ptr = devres_alloc(devm_mfd_dev_release, sizeof(*ptr), GFP_KERNEL);
> +	if (!ptr)
> +		return -ENOMEM;
> +
> +	ret = mfd_add_devices(dev, id, cells, n_devs, mem_base,
> +			      irq_base, domain);
> +	if (!ret) {
> +		*ptr = dev;
> +		devres_add(dev, ptr);
> +	} else {
> +		devres_free(ptr);
> +	}

Switch these round.  If you encounter a problem, free and return.  If
not, skip the error handling and add the device outside of the if().

> +	return ret;
> +}
> +EXPORT_SYMBOL(devm_mfd_add_devices);
> +
> +static int devm_mfd_devs_match(struct device *dev, void *res, void *data)
> +{
> +	struct device **r = res;
> +
> +	if (WARN_ON(!r || !*r))
> +		return 0;
> +
> +	return *r == data;
> +}
> +
> +/**
> + * devm_mfd_remove_device - Resource managed version of mfd_remove_devices()
> + * @dev: Device for which which resource was allocated.

Did you proof read your own writing?  Pleaes re-word.

> + * Remove all mfd devices added on the device.

s/mfd/MFD/

'D' already means devices, so here you are saying "devices devices".
Please re-word.  Besides, you need to be more specific as to which
"devices on the devices" you are detailing, since this sentence
doesn't really make a great deal of sense.

> + * Normally this function will not need to be called and the resource
> + * management code will ensure that the resource is freed.

Then what is the purpose of providing it?  Do you have a user?

> + */
> +void devm_mfd_remove_devices(struct device *dev)
> +{
> +	WARN_ON(devres_release(dev, devm_mfd_dev_release,
> +			       devm_mfd_devs_match, dev));
> +}
> +EXPORT_SYMBOL(devm_mfd_remove_devices);
> +
>  int mfd_clone_cell(const char *cell, const char **clones, size_t n_clones)
>  {
>  	struct mfd_cell cell_entry;
> diff --git a/include/linux/mfd/core.h b/include/linux/mfd/core.h
> index bc6f7e0..d658d7f 100644
> --- a/include/linux/mfd/core.h
> +++ b/include/linux/mfd/core.h
> @@ -131,4 +131,11 @@ static inline int mfd_add_hotplug_devices(struct device *parent,
>  
>  extern void mfd_remove_devices(struct device *parent);
>  
> +extern int devm_mfd_add_devices(struct device *dev, int id,
> +			   const struct mfd_cell *cells, int n_devs,
> +			   struct resource *mem_base,
> +			   int irq_base, struct irq_domain *irq_domain);
> +
> +extern void devm_mfd_remove_devices(struct device *dev);
> +
>  #endif

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

[toc] | [prev] | [next] | [standalone]

#1373271

Hi Lee,
Thanks for review.
I will send another patch with incorporating your comments.


On Thursday 07 April 2016 04:14 PM, Lee Jones wrote:
> On Tue, 05 Apr 2016, Laxman Dewangan wrote:
>
> +	if (!ret) {
> +		*ptr = dev;
> +		devres_add(dev, ptr);
> +	} else {
> +		devres_free(ptr);
> +	}
> Switch these round.  If you encounter a problem, free and return.  If
> not, skip the error handling and add the device outside of the if().

Like below?

if (ret) {
      devres_free(ptr);
      return ret;
}

*ptr = dev;
devres_add(dev, ptr);

return ret;

>> + * Remove all mfd devices added on the device.
> s/mfd/MFD/
>
> 'D' already means devices, so here you are saying "devices devices".
> Please re-word.  Besides, you need to be more specific as to which
> "devices on the devices" you are detailing, since this sentence
> doesn't really make a great deal of sense.
Wanted to say
Remove all devices added by mfd_add_devices()  from parent device.


>> + * Normally this function will not need to be called and the resource
>> + * management code will ensure that the resource is freed.
> Then what is the purpose of providing it?  Do you have a user?

To have pair of release. I have not seen the usage of most of 
devm_*_release() function other than devm_kfree().

[toc] | [prev] | [next] | [standalone]

#1373300 — Re: [PATCH 01/20] mfd: Add devm_ apis for mfd_add_devices and mfd_release_devices

On Thu, 07 Apr 2016, Laxman Dewangan wrote:

> Hi Lee,
> Thanks for review.
> I will send another patch with incorporating your comments.
> 
> 
> On Thursday 07 April 2016 04:14 PM, Lee Jones wrote:
> >On Tue, 05 Apr 2016, Laxman Dewangan wrote:
> >
> >+	if (!ret) {
> >+		*ptr = dev;
> >+		devres_add(dev, ptr);
> >+	} else {
> >+		devres_free(ptr);
> >+	}
> >Switch these round.  If you encounter a problem, free and return.  If
> >not, skip the error handling and add the device outside of the if().
> 
> Like below?
> 
> if (ret) {
>      devres_free(ptr);
>      return ret;
> }
> 
> *ptr = dev;
> devres_add(dev, ptr);
> 
> return ret;

This is more in line with what I expect, yes.

> >>+ * Remove all mfd devices added on the device.
> >s/mfd/MFD/
> >
> >'D' already means devices, so here you are saying "devices devices".
> >Please re-word.  Besides, you need to be more specific as to which
> >"devices on the devices" you are detailing, since this sentence
> >doesn't really make a great deal of sense.
> Wanted to say
> Remove all devices added by mfd_add_devices()  from parent device.

Remove all chlid-devices?

> >>+ * Normally this function will not need to be called and the resource
> >>+ * management code will ensure that the resource is freed.
> >Then what is the purpose of providing it?  Do you have a user?
> 
> To have pair of release. I have not seen the usage of most of
> devm_*_release() function other than devm_kfree().

Unless you have a need or a user, I would omit this for now.

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

[toc] | [prev] | [next] | [standalone]

#1373312

On Thursday 07 April 2016 05:12 PM, Lee Jones wrote:
> On Thu, 07 Apr 2016, Laxman Dewangan wrote:
>
> + * Normally this function will not need to be called and the resource
> + * management code will ensure that the resource is freed.
>>> Then what is the purpose of providing it?  Do you have a user?
>> To have pair of release. I have not seen the usage of most of
>> devm_*_release() function other than devm_kfree().
> Unless you have a need or a user, I would omit this for now.
>
OK, I will remove it.

BTW, I have collected all acks and reviewed by which I got from this series.
So I can send this series with all acks/reviewed by.

If you want me to provide the git location for these patchs then I can 
do it also. For this I will need base from where I need to take branch.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web