On Wed, 30 Apr 2008, Andrew Morton wrote:
quoted text
> Does this make it feel better?
>
> --- a/drivers/media/video/tuner-core.c~a
> +++ a/drivers/media/video/tuner-core.c
> @@ -40,11 +40,11 @@
>  	typeof(&FUNCTION) __a = symbol_request(FUNCTION); \
>  	if (__a) { \
>  		__r = (int) __a(ARGS); \
> +		symbol_put(FUNCTION); \
>  	} else { \
>  		printk(KERN_ERR "TUNER: Unable to find " \
>  				"symbol "#FUNCTION"()\n"); \
>  	} \
> -	symbol_put(FUNCTION); \
>  	__r; \
>  })

Should the symbol_put be done at all?  When I wrote the code this is based
on, it would check if FUNCTION failed or not.  If it failed, the symbol was
put.  But if it worked, then one was returned a handle into FUNCTION's
module, and so the symbol was not put.  If it was, the module's refcount
would be zero but the caller would have a handle into the module.

So if FUNCTION does anything that creates references to the module, and it
doesn't inc it's own refcount, then the symbol_put shouldn't be done.
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/