On Tue, Jan 18, 2011 at 09:37:06AM +0200, Tomas Winkler wrote:
> easycap_main.c:34:5: warning: symbol 'easycap_debug' was not declared. Should 
> it be static?
> easycap_main.c:36:5: warning: symbol 'easycap_gain' was not declared. Should 
> it be static?
> 

First of all there are some code problems remaining so this will need
to be resent.

Second of all, No.  The patch description *DOES* need to be fixed.

Just copying the error messages doesn't tell me what the patch does
does or why.  For example, this patch adds a ->gain member to the
easycap struct.  That should have been mentioned in the commit message.
I shouldn't have to read the code to find out what you are trying to do.

Here is what a real commit message would look like:

Sparse gives the following warnings:

easycap_main.c:34:5: warning: symbol 'easycap_debug' was not declared.
        Should it be static?
easycap_main.c:36:5: warning: symbol 'easycap_gain' was not declared.
        Should it be static?

These two variables actually were declared in several places but
the code is a mess.  The variables are used in several files.  I've
fixed "easycap_debug" so it gets declared in one place only and included
properly.  For "easycap_gain" made it static and I created added a
->gain member to the easycap struct.  This seems cleaner than using a
global variable and later on we may make this controlable via sysfs.

> --- a/drivers/staging/easycap/easycap_ioctl.h
> +++ b/drivers/staging/easycap/easycap_ioctl.h
> @@ -27,7 +27,6 @@
>  #if !defined(EASYCAP_IOCTL_H)
>  #define EASYCAP_IOCTL_H
>  
> -extern int easycap_debug;
>  extern int easycap_gain;

You can remove the easycap_gain declaration here.

> --- a/drivers/staging/easycap/easycap_low.h
> +++ b/drivers/staging/easycap/easycap_low.h
> @@ -27,7 +27,6 @@
>  #if !defined(EASYCAP_LOW_H)
>  #define EASYCAP_LOW_H
>  
> -extern int easycap_debug;
>  extern int easycap_gain;

And again here.

regards,
dan carpenter
_______________________________________________
devel mailing list
[email protected]
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel

Reply via email to