Re: [RFC PATCH 1/3] sched: fix exit_mm vs membarrier (v3)

[Peter Zijlstra]

On Wed, Oct 07, 2020 at 10:57:52AM -0400, Mathieu Desnoyers wrote:
> ----- On Oct 7, 2020, at 10:29 AM, Peter Zijlstra peterz@xxxxxxxxxxxxx wrote:
> 
> > On Thu, Sep 24, 2020 at 01:25:06PM -0400, Mathieu Desnoyers wrote:
> >> diff --git a/kernel/exit.c b/kernel/exit.c
> >> index 733e80f334e7..0767a2dbf245 100644
> >> --- a/kernel/exit.c
> >> +++ b/kernel/exit.c
> >> @@ -475,7 +475,19 @@ static void exit_mm(void)
> >>  	BUG_ON(mm != current->active_mm);
> >>  	/* more a memory barrier than a real lock */
> >>  	task_lock(current);
> >> +	/*
> >> +	 * When a thread stops operating on an address space, the loop
> >> +	 * in membarrier_private_expedited() may not observe that
> >> +	 * tsk->mm, and the loop in membarrier_global_expedited() may
> >> +	 * not observe a MEMBARRIER_STATE_GLOBAL_EXPEDITED
> >> +	 * rq->membarrier_state, so those would not issue an IPI.
> >> +	 * Membarrier requires a memory barrier after accessing
> >> +	 * user-space memory, before clearing tsk->mm or the
> >> +	 * rq->membarrier_state.
> >> +	 */
> >> +	smp_mb__after_spinlock();
> >>  	current->mm = NULL;
> >> +	membarrier_update_current_mm(NULL);
> >>  	mmap_read_unlock(mm);
> >>  	enter_lazy_tlb(mm, current);
> >>  	task_unlock(current);
> > 
> > This site seems to be lacking in IRQ disabling. As proposed it will
> > explode on RT.
> 
> Right, so irq off is needed for accessing this_rq()'s fields safely,
> correct ?

Yes, but also we're having IRQs disabled on ever other site that mucks
with ->mm these days.