[alsa-devel] [PATCH -mm] sb16: Shut up uninitialized var build warning

[Rene Herman]

On 09/02/2007 10:15 PM, Satyam Sharma wrote:

> sound/isa/sb16/sb16.c: In function ‘snd_sb16_isa_probe’:

Blah. Your message has:

	Content-Type: TEXT/PLAIN; charset=iso-2022-jp

This apparently is caused by a combination of GCC using groovy UTF tickmarks 
in its error messages when in a UTF locale and alpine believing it to be a 
great idea to automatically try for the "simplest" character set it can 
encode the content in. No idea why that means that iso-2022-jp is picked, 
but it is.

While I could actually read the message this time you should see what 
iso-2022-jp does to my font. It's scary. Best solution as far as I'm 
concerned is slap a few GCC developers (not that it wil help, but it'll 
certainly feel good) and then teach alpine to go for UTF-8 directly if 
US-ASCII won't do.

As to the content of this patch -- I'd almost say it's better to live with 
the warning than with that unitialized_var() thing. That ARRAY_SIZE is very 
much a compile time constant, so exactly how dumb must GCC get before we get 
to say to here and no further?

> --- linux-2.6.23-rc4-mm1/sound/isa/sb/sb16.c~fix	2007-09-02 21:41:51.000000000 +0530
> +++ linux-2.6.23-rc4-mm1/sound/isa/sb/sb16.c	2007-09-02 21:42:56.000000000 +0530
> @@ -556,7 +556,6 @@ static int __devinit snd_sb16_isa_match(
>  
>  static int __devinit snd_sb16_isa_probe(struct device *pdev, unsigned int dev)
>  {
> -	int err;
>  	static int possible_irqs[] = {5, 9, 10, 7, -1};
>  	static int possible_dmas8[] = {1, 3, 0, -1};
>  	static int possible_dmas16[] = {5, 6, 7, -1};
> @@ -585,6 +584,8 @@ static int __devinit snd_sb16_isa_probe(
>  	else {
>  		static int possible_ports[] = {0x220, 0x240, 0x260, 0x280};
>  		int i;
> +		int uninitialized_var(err);
> +
>  		for (i = 0; i < ARRAY_SIZE(possible_ports); i++) {
>  			port[dev] = possible_ports[i];
>  			err = snd_sb16_isa_probe1(dev, pdev);

Rene.