Re: [RFC PATCH 02/43] ptrace: cleanup ptrace_request()

[Arnd Bergmann <arnd@arndb.de>]

On Friday 27 August 2010, Namhyung Kim wrote:
> @@ -579,19 +580,19 @@ int ptrace_request(struct task_struct *child, long request,
>                 ret = ptrace_setoptions(child, data);
>                 break;
>         case PTRACE_GETEVENTMSG:
> -               ret = put_user(child->ptrace_message, (unsigned long __user *) data);
> +               ret = copy_to_user(datap, &child->ptrace_message,
> +                                  sizeof (child->ptrace_message)) ?
> +                       -EFAULT : 0;
>                 break;
>  
>         case PTRACE_GETSIGINFO:
>                 ret = ptrace_getsiginfo(child, &siginfo);
>                 if (!ret)
> -                       ret = copy_siginfo_to_user((siginfo_t __user *) data,
> -                                                  &siginfo);
> +                       ret = copy_siginfo_to_user(datap, &siginfo);
>                 break;

If you introduce two pointers, 'siginfo_t __user *usiginfo' and
'unsigned long __user *datap' instead of just 'datap', you can use
get_user/put_user everywhere, which would make the code better readable
and faster.

Otherwise this change looks good.

	Arnd