[PATCH 0/4 RESEND] exec: unify compat/non-compat code

[Oleg Nesterov <oleg@redhat.com>]

On 12/01, Oleg Nesterov wrote:
>
> On 12/01, Milton Miller wrote:
> >
> > > +#ifdef CONFIG_COMPAT
> > > +int compat_do_execve(char * filename,
> > > +	compat_uptr_t __user *argv,
> > > +	compat_uptr_t __user *envp,
> > > +	struct pt_regs * regs)
> > > +{
> > > +	return do_execve_common(filename,
> > > +				(void __user*)argv, (void __user*)envp,
> >
> > Shouldn't these be compat_ptr(argv)?  (makes a difference on s390)
>
> I'll recheck, but I don't think so. Please note that compat_ptr()
> accepts "compat_uptr_t", not "compat_uptr_t *".
>
> argv should be correct as a pointer to user-space, otherwise the
> current code is buggy. For example, compat_do_execve() passes
> argv to compat_count() which does get_user(argv) without any
> conversion.

So, once again, this should not (and can not) be compat_ptr(argv) afaics.

I don't understand the s390 asm, but compat_wrapper.S:sys32_execve_wrapper
looks correct. If not, the current code is already buggy and s390 should
be fixed. argv/envp are not compat ptrs, they just point to compat_ data,
we should not do any conversion.

I am resending this series unchanged, plus the trivial 5/5 to document
acct_arg_size().

----------------------------------------------------------------------

execve code in fs/compat.c must die. It is very hard to maintain this
copy-and-paste horror. And the only reason for this duplication is that
argv/envp point to char* or compat_uptr_t depending on compat. We can
add the trivial helper which hides the difference and unify the code.

Oleg.