Re: [PATCH 1/1 linux-next] exofs: convert simple_str to kstr

[Alexey Dobriyan]

On Thu, Apr 30, 2015 at 2:06 PM, Alexey Dobriyan <adobriyan@xxxxxxxxx> wrote:
>>   case Opt_pid:
>> + {
>> + int rc;
>> +
>>   if (0 == match_strlcpy(str, &args[0], sizeof(str)))
>>   return -EINVAL;
>> - opts->pid = simple_strtoull(str, NULL, 0);
>> + rc = kstrtoull(str, 0, &opts->pid);
>> + if (rc)
>> + return rc;
>
> kstrto*() functions were designed for proc/sysfs/sysctl/debugfs/...
> files accepting 1 value from shell:
>
>   echo 42 >/proc/foo
>
> To keep this simplest usecase simple kstrto*() functions accept
> one optional newline character before mandatory NUL.
> But in mount option parsing newlines aren't supposed to be.
>
> I'll resend parse_integer() shortly which is better suited for such code
> http://marc.info/?l=linux-kernel&m=142876674416451&w=4

It should be better:
* strsep/match_token/match_*() doesn't support u64 type,
* even if it did current code does useless copy:

char str[32];
token = match_token(p, tokens, args);
switch (token) {
case Opt_pid:
    if (0 == match_strlcpy(str, &args[0], sizeof(str)))
        return -EINVAL;
    opts->pid = simple_strtoull(str, NULL, 0);


With parse_integer() code would look like:

case Opt_pid:
    rv = parse_integer(args[0].from, 0, &opts->pid);
    if (rv < 0)
        return rv;
    if (args[0].from[rv] != '\0')
        return -EINVAL;

And if and when match_token interface will start verifying
that integer occupies whole options wihout junk at the end,
that second NUL check can be removed as well.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/