Re: [PATCH] Consolidate compat_sys_waitid

[Stephen Rothwell <sfr@canb.auug.org.au>]

On Tue, 15 Feb 2005 18:43:07 +0000 Matthew Wilcox <matthew@wil.cx> wrote:
>
> On Tue, Feb 15, 2005 at 02:01:49PM +1100, Stephen Rothwell wrote:
> 
> > +asmlinkage long compat_sys_waitid(u32 which, u32 pid,
> > +		struct compat_siginfo __user *uinfo, u32 options,
> > +		struct compat_rusage __user *uru)
> 
> Some subtle differences which I feel incompetent to diagnose ... ours
> looks like:
> 
> asmlinkage int compat_sys_waitid(int which, pid_t pid,
>                                  compat_siginfo_t __user *infop, int options,
>                                  struct compat_rusage __user *ru)

They are u32 in my version because we decided that the kernel zero extends
the compat syscall parameters.

> Other than variable names, we're identical to this point:
> 
>         /* Tell copy_siginfo_to_user that it was __SI_CHLD */
>         ksiginfo.si_code |= __SI_CHLD;
>         
>         if (compat_copy_siginfo_to_user(infop, &ksiginfo) != 0)

Every other relevant architecture has copy_siginfo_to_user32 which returns
0/-EFAULT.  And it is declared in linux/compat.h.  I like your name
better, but that is a different patch.

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/