[PATCH] media: rc: Pospone ir raw decoders loading until really needed

[PATCH] media: rc: Pospone ir raw decoders loading until really needed

[Ezequiel Garcia]

This changes rc_core to not load the IR decoders at load time,
postponing it to load only if a RC_DRIVER_IR_RAW device is registered
via rc_register_device.

Signed-off-by: Ezequiel Garcia <elezegarcia@gmail.com>
---
 drivers/media/rc/rc-main.c |    8 ++++++--
 include/media/rc-core.h    |    2 ++
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c
index f6a930b..adf4a99 100644
--- a/drivers/media/rc/rc-main.c
+++ b/drivers/media/rc/rc-main.c
@@ -1103,6 +1103,12 @@ int rc_register_device(struct rc_dev *dev)
 	kfree(path);
 
 	if (dev->driver_type == RC_DRIVER_IR_RAW) {
+		/* Load raw decoders, if they aren't already */
+		if (dev->raw_init) {
+			IR_dprintk(1, "Loading raw decoders\n");
+			ir_raw_init();
+			dev->raw_init = true;
+		}
 		rc = ir_raw_event_register(dev);
 		if (rc < 0)
 			goto out_input;
@@ -1176,8 +1182,6 @@ static int __init rc_core_init(void)
 		return rc;
 	}
 
-	/* Initialize/load the decoders/keymap code that will be used */
-	ir_raw_init();
 	rc_map_register(&empty_map);
 
 	return 0;
diff --git a/include/media/rc-core.h b/include/media/rc-core.h
index b0c494a..71c99c1 100644
--- a/include/media/rc-core.h
+++ b/include/media/rc-core.h
@@ -47,6 +47,7 @@ enum rc_driver_type {
  *	anyone can call show_protocols or store_protocols
  * @devno: unique remote control device number
  * @raw: additional data for raw pulse/space devices
+ * @raw_init: used to load raw decoders modules if needed
  * @input_dev: the input child device used to communicate events to userspace
  * @driver_type: specifies if protocol decoding is done in hardware or software
  * @idle: used to keep track of RX state
@@ -95,6 +96,7 @@ struct rc_dev {
 	struct mutex			lock;
 	unsigned long			devno;
 	struct ir_raw_event_ctrl	*raw;
+	bool				raw_init;
 	struct input_dev		*input_dev;
 	enum rc_driver_type		driver_type;
 	bool				idle;
-- 
1.7.3.4

Re: [PATCH] media: rc: Pospone ir raw decoders loading until really needed

[Mauro Carvalho Chehab]

Em 15-03-2012 16:39, Ezequiel Garcia escreveu:
> This changes rc_core to not load the IR decoders at load time,
> postponing it to load only if a RC_DRIVER_IR_RAW device is registered
> via rc_register_device.
> 
> Signed-off-by: Ezequiel Garcia <elezegarcia@gmail.com>
> ---
>  drivers/media/rc/rc-main.c |    8 ++++++--
>  include/media/rc-core.h    |    2 ++
>  2 files changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c
> index f6a930b..adf4a99 100644
> --- a/drivers/media/rc/rc-main.c
> +++ b/drivers/media/rc/rc-main.c
> @@ -1103,6 +1103,12 @@ int rc_register_device(struct rc_dev *dev)
>  	kfree(path);
>  
>  	if (dev->driver_type == RC_DRIVER_IR_RAW) {
> +		/* Load raw decoders, if they aren't already */
> +		if (dev->raw_init) {

The logic here seems to be inverted. it should be, instead !dev->raw_init.

> +			IR_dprintk(1, "Loading raw decoders\n");
> +			ir_raw_init();
> +			dev->raw_init = true;
> +		}
>  		rc = ir_raw_event_register(dev);
>  		if (rc < 0)
>  			goto out_input;
> @@ -1176,8 +1182,6 @@ static int __init rc_core_init(void)
>  		return rc;
>  	}
>  
> -	/* Initialize/load the decoders/keymap code that will be used */
> -	ir_raw_init();
>  	rc_map_register(&empty_map);
>  
>  	return 0;
> diff --git a/include/media/rc-core.h b/include/media/rc-core.h
> index b0c494a..71c99c1 100644
> --- a/include/media/rc-core.h
> +++ b/include/media/rc-core.h
> @@ -47,6 +47,7 @@ enum rc_driver_type {
>   *	anyone can call show_protocols or store_protocols
>   * @devno: unique remote control device number
>   * @raw: additional data for raw pulse/space devices
> + * @raw_init: used to load raw decoders modules if needed
>   * @input_dev: the input child device used to communicate events to userspace
>   * @driver_type: specifies if protocol decoding is done in hardware or software
>   * @idle: used to keep track of RX state
> @@ -95,6 +96,7 @@ struct rc_dev {
>  	struct mutex			lock;
>  	unsigned long			devno;
>  	struct ir_raw_event_ctrl	*raw;
> +	bool				raw_init;
>  	struct input_dev		*input_dev;
>  	enum rc_driver_type		driver_type;
>  	bool				idle;

Re: [PATCH] media: rc: Pospone ir raw decoders loading until really needed

[Jarod Wilson]

On Thu, Mar 15, 2012 at 04:39:02PM -0300, Ezequiel Garcia wrote:
> This changes rc_core to not load the IR decoders at load time,
> postponing it to load only if a RC_DRIVER_IR_RAW device is registered
> via rc_register_device.
> 
> Signed-off-by: Ezequiel Garcia <elezegarcia@gmail.com>
> ---
>  drivers/media/rc/rc-main.c |    8 ++++++--
>  include/media/rc-core.h    |    2 ++
>  2 files changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c
> index f6a930b..adf4a99 100644
> --- a/drivers/media/rc/rc-main.c
> +++ b/drivers/media/rc/rc-main.c
> @@ -1103,6 +1103,12 @@ int rc_register_device(struct rc_dev *dev)
>  	kfree(path);
>  
>  	if (dev->driver_type == RC_DRIVER_IR_RAW) {
> +		/* Load raw decoders, if they aren't already */
> +		if (dev->raw_init) {
> +			IR_dprintk(1, "Loading raw decoders\n");
> +			ir_raw_init();
> +			dev->raw_init = true;
> +		}
>  		rc = ir_raw_event_register(dev);
>  		if (rc < 0)
>  			goto out_input;
> @@ -1176,8 +1182,6 @@ static int __init rc_core_init(void)
>  		return rc;
>  	}
>  
> -	/* Initialize/load the decoders/keymap code that will be used */
> -	ir_raw_init();
>  	rc_map_register(&empty_map);
>  
>  	return 0;

Uhm. How does this ever actually work? The only place I see raw_init set
to true is inside a conditional that requires it already be true. That's
not going to fly.

-- 
Jarod Wilson
jarod@redhat.com

Re: [PATCH] media: rc: Pospone ir raw decoders loading until really needed

[Ezequiel García]

>>       if (dev->driver_type == RC_DRIVER_IR_RAW) {
>> +             /* Load raw decoders, if they aren't already */
>> +             if (dev->raw_init) {
>
> The logic here seems to be inverted. it should be, instead !dev->raw_init.

Duh! Sorry about that. I'm *way* too accustomed to test things first.
Will pay more attention in the future.

Thanks for the review. I'm resending it now.
Ezequiel.