Re: [REGRESSION] x86/entry: Tracer no longer has opportunity to change the syscall number at entry via orig_ax

[Kees Cook <keescook@chromium.org>]

On Wed, Aug 19, 2020 at 09:44:39PM +0200, Thomas Gleixner wrote:
> On Wed, Aug 19 2020 at 10:14, Kyle Huey wrote:
> > tl;dr: after 27d6b4d14f5c3ab21c4aef87dd04055a2d7adf14 ptracer
> > modifications to orig_ax in a syscall entry trace stop are not honored
> > and this breaks our code.
> 
> My fault and I have no idead why none of the silly test cases
> noticed. Fix below.

Hmm, which were you trying? Looking just now, I see that the seccomp
selftests were failing for all their syscall-changing tests.

Regardless, I can confirm both the failure and the fix.

Reported-by: Kyle Huey <me@kylehuey.com>
Tested-by: Kees Cook <keescook@chromium.org>
Acked-by: Kees Cook <keescook@chromium.org>


kernelci.org is *so* close to having the kernel selftests actually
running with their builds. :)

https://github.com/kernelci/kernelci-core/issues/331

-Kees

> 
> Thanks,
> 
>         tglx
> ---
> diff --git a/kernel/entry/common.c b/kernel/entry/common.c
> index 9852e0d62d95..fcae019158ca 100644
> --- a/kernel/entry/common.c
> +++ b/kernel/entry/common.c
> @@ -65,7 +65,8 @@ static long syscall_trace_enter(struct pt_regs *regs, long syscall,
>  
>  	syscall_enter_audit(regs, syscall);
>  
> -	return ret ? : syscall;
> +	/* The above might have changed the syscall number */
> +	return ret ? : syscall_get_nr(current, regs);
>  }
>  
>  noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)

-- 
Kees Cook