version 2: - simplify function prototype to only keep device as parameter - rebased on v4.10
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus. This also could make drivers more robust in case that probe failed after calling of_platform_populate().
Benjamin Gaignard (2): of: add devm_ functions for populate and depopulate drm: sti: make driver use devm_of_platform_populate()
drivers/gpu/drm/sti/sti_drv.c | 3 +- drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 3 files changed, 83 insertions(+), 2 deletions(-)
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org --- version 2: - simplify function prototype to only keep device as parameter
drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 2 files changed, 82 insertions(+)
diff --git a/drivers/of/platform.c b/drivers/of/platform.c index b8064bc..f5bbb50 100644 --- a/drivers/of/platform.c +++ b/drivers/of/platform.c @@ -571,6 +571,77 @@ void of_platform_depopulate(struct device *parent) } EXPORT_SYMBOL_GPL(of_platform_depopulate);
+static void devm_of_platform_populate_release(struct device *dev, void *res) +{ + of_platform_depopulate(*(struct device **)res); +} + +/** + * devm_of_platform_populate() - Populate platform_devices from device tree data + * @dev: device that requested to populate from device tree data + * + * Similar to of_platform_populate(), but will automatically call + * of_platform_depopulate() when the device is unbound from the bus. + * + * Returns 0 on success, < 0 on failure. + */ +int devm_of_platform_populate(struct device *dev) +{ + struct device **ptr; + int ret; + + if (!dev) + return -EINVAL; + + ptr = devres_alloc(devm_of_platform_populate_release, + sizeof(*ptr), GFP_KERNEL); + if (!ptr) + return -ENOMEM; + + ret = of_platform_populate(dev->of_node, NULL, NULL, dev); + if (ret) { + devres_free(ptr); + } else { + *ptr = dev; + devres_add(dev, ptr); + } + + return ret; +} +EXPORT_SYMBOL_GPL(devm_of_platform_populate); + +static int devm_of_platform_match(struct device *dev, void *res, void *data) +{ + struct device **ptr = res; + + if (!ptr) { + WARN_ON(!ptr); + return 0; + } + + return *ptr == data; +} + +/** + * devm_of_platform_depopulate() - Remove devices populated from device tree + * @dev: device that requested to depopulate from device tree data + * + * Complementary to devm_of_platform_populate(), this function removes children + * of the given device (and, recurrently, their children) that have been + * created from their respective device tree nodes (and only those, + * leaving others - eg. manually created - unharmed). + */ +void devm_of_platform_depopulate(struct device *dev) +{ + int ret; + + ret = devres_release(dev, devm_of_platform_populate_release, + devm_of_platform_match, dev); + + WARN_ON(ret); +} +EXPORT_SYMBOL_GPL(devm_of_platform_depopulate); + #ifdef CONFIG_OF_DYNAMIC static int of_platform_notify(struct notifier_block *nb, unsigned long action, void *arg) diff --git a/include/linux/of_platform.h b/include/linux/of_platform.h index 956a100..dc8224a 100644 --- a/include/linux/of_platform.h +++ b/include/linux/of_platform.h @@ -76,6 +76,10 @@ extern int of_platform_default_populate(struct device_node *root, const struct of_dev_auxdata *lookup, struct device *parent); extern void of_platform_depopulate(struct device *parent); + +extern int devm_of_platform_populate(struct device *dev); + +extern void devm_of_platform_depopulate(struct device *dev); #else static inline int of_platform_populate(struct device_node *root, const struct of_device_id *matches, @@ -91,6 +95,13 @@ static inline int of_platform_default_populate(struct device_node *root, return -ENODEV; } static inline void of_platform_depopulate(struct device *parent) { } + +static inline int devm_of_platform_populate(struct device *dev) +{ + return -ENODEV; +} + +static inline void devm_of_platform_depopulate(struct device *dev) { } #endif
#if defined(CONFIG_OF_DYNAMIC) && defined(CONFIG_OF_ADDRESS)
On Fri, Feb 24, 2017 at 10:14 AM, Benjamin Gaignard benjamin.gaignard@linaro.org wrote:
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org
version 2:
- simplify function prototype to only keep device as parameter
drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 2 files changed, 82 insertions(+)
For this and patch 2:
Acked-by: Rob Herring robh@kernel.org
Rob
On Fri, Feb 24, 2017 at 10:31:25AM -0600, Rob Herring wrote:
On Fri, Feb 24, 2017 at 10:14 AM, Benjamin Gaignard benjamin.gaignard@linaro.org wrote:
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org
version 2:
- simplify function prototype to only keep device as parameter
drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 2 files changed, 82 insertions(+)
For this and patch 2:
Acked-by: Rob Herring robh@kernel.org
Is this an ack for merging both through drm-misc, or should we do a topic-branch dance here?
/me not sure anymore what we discussed here already
Cheers, Daniel
On Sun, Feb 26, 2017 at 2:11 PM, Daniel Vetter daniel@ffwll.ch wrote:
On Fri, Feb 24, 2017 at 10:31:25AM -0600, Rob Herring wrote:
On Fri, Feb 24, 2017 at 10:14 AM, Benjamin Gaignard benjamin.gaignard@linaro.org wrote:
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org
version 2:
- simplify function prototype to only keep device as parameter
drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 2 files changed, 82 insertions(+)
For this and patch 2:
Acked-by: Rob Herring robh@kernel.org
Is this an ack for merging both through drm-misc, or should we do a topic-branch dance here?
You can apply it directly.
Rob
On Mon, Feb 27, 2017 at 07:30:24AM -0600, Rob Herring wrote:
On Sun, Feb 26, 2017 at 2:11 PM, Daniel Vetter daniel@ffwll.ch wrote:
On Fri, Feb 24, 2017 at 10:31:25AM -0600, Rob Herring wrote:
On Fri, Feb 24, 2017 at 10:14 AM, Benjamin Gaignard benjamin.gaignard@linaro.org wrote:
Lots of calls to of_platform_populate() are not unbalanced by a call to of_platform_depopulate(). This create issues while drivers are bind/unbind.
In way to solve those issues is to add devm_of_platform_populate() which will call of_platform_depopulate() when the device is unbound from the bus.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org
version 2:
- simplify function prototype to only keep device as parameter
drivers/of/platform.c | 71 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/of_platform.h | 11 +++++++ 2 files changed, 82 insertions(+)
For this and patch 2:
Acked-by: Rob Herring robh@kernel.org
Is this an ack for merging both through drm-misc, or should we do a topic-branch dance here?
You can apply it directly.
Ok, this seems like a pretty cool idea, so applied both to drm-misc for 4.12.
Cheers, Daniel
This make sure that of_platform_depopulate() is called if an error occur in probe after populating the date from the device tree.
Signed-off-by: Benjamin Gaignard benjamin.gaignard@linaro.org --- version 2: - use simplified devm_of_platform_populate function prototype
drivers/gpu/drm/sti/sti_drv.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/sti/sti_drv.c b/drivers/gpu/drm/sti/sti_drv.c index ff71e25..b26b1bc 100644 --- a/drivers/gpu/drm/sti/sti_drv.c +++ b/drivers/gpu/drm/sti/sti_drv.c @@ -438,7 +438,7 @@ static int sti_platform_probe(struct platform_device *pdev)
dma_set_coherent_mask(dev, DMA_BIT_MASK(32));
- of_platform_populate(node, NULL, NULL, dev); + devm_of_platform_populate(dev);
child_np = of_get_next_available_child(node, NULL);
@@ -454,7 +454,6 @@ static int sti_platform_probe(struct platform_device *pdev) static int sti_platform_remove(struct platform_device *pdev) { component_master_del(&pdev->dev, &sti_ops); - of_platform_depopulate(&pdev->dev);
return 0; }
linaro-kernel@lists.linaro.org
-
Benjamin Gaignard -
Daniel Vetter -
Rob Herring