Re: [PATCH v6 0/3] Harden spectrev2 userspace-userspace protection

[Peter Zijlstra <peterz@infradead.org>]

On Sat, Sep 22, 2018 at 03:30:07PM +0200, Thomas Gleixner wrote:
> On Sat, 22 Sep 2018, Thomas Gleixner wrote:
> > On Sat, 22 Sep 2018, Peter Zijlstra wrote:
> > > This has some unfortunate duplication.
> > > 
> > > Lets go with it for now, but I'll see if I can do something about that
> > > later.
> > 
> > Yes, I know. I tried to make the duplication smaller, but all attempts
> > ended up being a convoluted mess. I'll try again after applying more
> > coffee.
> 
> Lunch and coffee indeed made brain work better. The simple solution was way
> too obvious.
> --- a/kernel/ptrace.c
> +++ b/kernel/ptrace.c
> @@ -261,6 +261,9 @@ static int ptrace_check_attach(struct ta
>  
>  static int ptrace_has_cap(struct user_namespace *ns, unsigned int mode)
>  {
> +	if (mode & PTRACE_MODE_SCHED)
> +		return false;
> +
>  	if (mode & PTRACE_MODE_NOAUDIT)
>  		return has_ns_capability_noaudit(current, ns, CAP_SYS_PTRACE);
>  	else
> @@ -328,9 +331,16 @@ static int __ptrace_may_access(struct ta
>  	     !ptrace_has_cap(mm->user_ns, mode)))
>  	    return -EPERM;
>  
> +	if (mode & PTRACE_MODE_SCHED)
> +		return 0;
>  	return security_ptrace_access_check(task, mode);
>  }
>  
> +bool ptrace_may_access_sched(struct task_struct *task, unsigned int mode)
> +{
> +	return __ptrace_may_access(task, mode | PTRACE_MODE_SCHED);
> +}

Ha!, much nicer. Thanks!