Re: A simple question of sys_

[Rofail Qu <rofail@gmail.com>]

2010/9/18 Randy Dunlap <rdunlap@xenotime.net>:
> On Fri, 17 Sep 2010 14:58:30 +0800 Rofail Qu wrote:
>
>> How to use macro IS_ERR() ?
>>
>> It defines as,
>> ...
>> #define IS_ERR_VALUE(x) unlikely((x) >= (unsigned long)-MAX_ERRNO)
>> static inline long __must_check IS_ERR(const void *ptr)
>> {
>>       return IS_ERR_VALUE((unsigned long)ptr);
>> }
>> ...
>> so when pass x as a pointer and x>=-MAX_ERRNO (including NULL or any
>> valid address),
>> IS_ERR() will return true!
>
> Since your conclusion is false, some part of your premise must have a problem.
> Can you find it?
Got it.
thanks any way.

>
>
>> IS_ERR(x) seems to use on judge if "x" is a valid error number, right?
>
> Yes, that's what it is for.
So in kernel, a bad pointer must have saved an valid error number, right?

>
>> So in sys_execve(),
>> ...
>>         long error;
>>       char* filename;
>>
>>       filename = getname(name);
>>       error = PTR_ERR(filename);
>>       if (IS_ERR(filename))     // <== should be IS_ERR((void *)error) or other?
>>               return error;
>>       error = do_execve(filename, argv, envp, regs);
>> ...
>>
>> Where i am wrong?
>
>
> ---
> ~Randy
> *** Remember to use Documentation/SubmitChecklist when testing your code ***
>