[PATCH] [media] pt3: remove bogus module_is_live() check

[PATCH] [media] pt3: remove bogus module_is_live() check

[Arnd Bergmann]

The new pt3 driver checks the module reference for presence
before dropping it, which fails to compile when modules
are disabled:

media/pci/pt3/pt3.c: In function 'pt3_attach_fe':
media/pci/pt3/pt3.c:433:6: error: implicit declaration of function 'module_is_live' [-Werror=implicit-function-declaration]
      module_is_live(pt3->adaps[i]->i2c_tuner->dev.driver->owner))

As far as I can tell however, this check is not needed at all, because
the module will not go away as long as pt3 is holding a reference on
it. Also the previous check for NULL pointer is not needed at all,
because module_put has the same check.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>

diff --git a/drivers/media/pci/pt3/pt3.c b/drivers/media/pci/pt3/pt3.c
index 90f86ce7a001..39305f07dc2e 100644
--- a/drivers/media/pci/pt3/pt3.c
+++ b/drivers/media/pci/pt3/pt3.c
@@ -429,14 +429,10 @@ static int pt3_attach_fe(struct pt3_board *pt3, int i)
 
 err_tuner:
 	i2c_unregister_device(pt3->adaps[i]->i2c_tuner);
-	if (pt3->adaps[i]->i2c_tuner->dev.driver->owner &&
-	    module_is_live(pt3->adaps[i]->i2c_tuner->dev.driver->owner))
-		module_put(pt3->adaps[i]->i2c_tuner->dev.driver->owner);
+	module_put(pt3->adaps[i]->i2c_tuner->dev.driver->owner);
 err_demod:
 	i2c_unregister_device(pt3->adaps[i]->i2c_demod);
-	if (pt3->adaps[i]->i2c_demod->dev.driver->owner &&
-	    module_is_live(pt3->adaps[i]->i2c_demod->dev.driver->owner))
-		module_put(pt3->adaps[i]->i2c_demod->dev.driver->owner);
+	module_put(pt3->adaps[i]->i2c_demod->dev.driver->owner);
 	return ret;
 }
 

Re: [PATCH] [media] pt3: remove bogus module_is_live() check

[Mauro Carvalho Chehab]

Hi Arnd,

Em Mon, 29 Sep 2014 16:28:55 +0200
Arnd Bergmann <arnd@arndb.de> escreveu:

> The new pt3 driver checks the module reference for presence
> before dropping it, which fails to compile when modules
> are disabled:
> 
> media/pci/pt3/pt3.c: In function 'pt3_attach_fe':
> media/pci/pt3/pt3.c:433:6: error: implicit declaration of function 'module_is_live' [-Werror=implicit-function-declaration]
>       module_is_live(pt3->adaps[i]->i2c_tuner->dev.driver->owner))
> 
> As far as I can tell however, this check is not needed at all, because
> the module will not go away as long as pt3 is holding a reference on
> it. Also the previous check for NULL pointer is not needed at all,
> because module_put has the same check.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Thanks for this patch. Antti was quicker than you ;)

I merged the patch on my tree yesterday. It should be popping up
on the tomorrow's -next (I think that sfr didn't release one today).

> 
> diff --git a/drivers/media/pci/pt3/pt3.c b/drivers/media/pci/pt3/pt3.c
> index 90f86ce7a001..39305f07dc2e 100644
> --- a/drivers/media/pci/pt3/pt3.c
> +++ b/drivers/media/pci/pt3/pt3.c
> @@ -429,14 +429,10 @@ static int pt3_attach_fe(struct pt3_board *pt3, int i)
>  
>  err_tuner:
>  	i2c_unregister_device(pt3->adaps[i]->i2c_tuner);
> -	if (pt3->adaps[i]->i2c_tuner->dev.driver->owner &&
> -	    module_is_live(pt3->adaps[i]->i2c_tuner->dev.driver->owner))
> -		module_put(pt3->adaps[i]->i2c_tuner->dev.driver->owner);
> +	module_put(pt3->adaps[i]->i2c_tuner->dev.driver->owner);
>  err_demod:
>  	i2c_unregister_device(pt3->adaps[i]->i2c_demod);
> -	if (pt3->adaps[i]->i2c_demod->dev.driver->owner &&
> -	    module_is_live(pt3->adaps[i]->i2c_demod->dev.driver->owner))
> -		module_put(pt3->adaps[i]->i2c_demod->dev.driver->owner);
> +	module_put(pt3->adaps[i]->i2c_demod->dev.driver->owner);
>  	return ret;
>  }
>  
>