Re: [PATCH 2/2] staging: replace simple_strtoul with kstroul

[Andrew Morton <akpm@linux-foundation.org>]

On Fri, 20 Feb 2015 00:13:57 +0200 Ameen Ali <ameenali023@gmail.com> wrote:

> Subject: [PATCH 2/2] staging: replace simple_strtoul with kstroul

This isn't a staging patch.

s/kstroul/kstrtoul/

> Date: Fri, 20 Feb 2015 00:13:57 +0200
> X-Mailer: git-send-email 2.1.0
> 
> Replacing obsolete simple_stroul with kstroul

s/kstroul/kstrtoul/

> --- a/kernel/printk/printk.c
> +++ b/kernel/printk/printk.c
> @@ -518,7 +518,7 @@ struct devkmsg_user {
>  static ssize_t devkmsg_write(struct kiocb *iocb, struct iov_iter *from)
>  {
>  	char *buf, *line;
> -	int i;
> +	unsigned long int i;

It would be better to move this definition so it is local to the `if'
loop where it is used.  And to give it a meaningful name - using `i' is lame.

>  	int level = default_message_loglevel;
>  	int facility = 1;	/* LOG_USER */
>  	size_t len = iocb->ki_nbytes;
> @@ -549,7 +549,7 @@ static ssize_t devkmsg_write(struct kiocb *iocb, struct iov_iter *from)
>  	if (line[0] == '<') {
>  		char *endp = NULL;
>  
> -		i = simple_strtoul(line+1, &endp, 10);
> +		i = kstrtoul(line+1,10, &endp);
>  		if (endp && endp[0] == '>') {
>  			level = i & 7;
>  			if (i >> 3)

The interfaces are different - kstrtoul() can return a negative errno. 
afaict we'll only have issues if someone tried to print "<>..", but
this string is coming from userspace so we must be very paranoid.