Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 18 Jul 2018 18:00:07 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Rik van Riel <riel@surriel.com>
Cc:     songliubraving@fb.com, linux-kernel@vger.kernel.org,
        dave.hansen@intel.com, hpa@zytor.com, tglx@linutronix.de,
        mingo@kernel.org, torvalds@linux-foundation.org,
        linux-tip-commits@vger.kernel.org
Subject: Re: [tip:x86/mm] x86/mm/tlb: Make lazy TLB mode lazier
Message-ID: <20180718160007.GS2494@hirez.programming.kicks-ass.net>
References: <20180716190337.26133-5-riel@surriel.com>
 <tip-ac0315896970d8589291e9d8a1569fc65967b7f1@git.kernel.org>
 <20180717113330.GU2476@hirez.programming.kicks-ass.net>
 <08AC2AF2-17DE-4416-BBBD-B6B950D20906@surriel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <08AC2AF2-17DE-4416-BBBD-B6B950D20906@surriel.com>
User-Agent: Mutt/1.10.0 (2018-05-17)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Jul 18, 2018 at 11:33:02AM -0400, Rik van Riel wrote:
> The tlb flush code first increments mm->context.tlb_gen, and then sends

> shootdown IPIs to CPUs that have this mm loaded and are not in lazy
> TLB mode.
> 
> At context switch time, we have to ensure that we check the tlb_gen after
> we load the old is_lazy state.
> 
> Maybe something like this?
>  
>                 /*
>                  * Read the tlb_gen to check whether a flush is needed.
>                  * If the TLB is up to date, just use it.
>                  * The TLB shootdown code first increments tlb_gen, and then
>                  * sends IPIs to CPUs that have this CPU loaded and are not
>                  * in lazy TLB mode. The barrier ensures we handle
>                  * cpu_tlbstate.is_lazy before tlb_gen, keeping this code
>                  * synchronized with the TLB flush code.
>                  */

Let me try and draw a diagram; that always works better for me that
text. So the relevant ordering is something like:

  CPU0 - switch_mm()			CPU1 - flush_tlb_mm_range()

  [W]	cpu_tlbstate.is_lazy = false;	[RmW]	next->tlb_gen++
	smp_mb()				MB (implied)
  [R]	tlb_gen = next->tlb_gen			native_flush_tlb_others()
					[R]	  cpu_tlbstate.is_lazy


Such that CPU1 either observes !lazy and flushes and/or CPU0 observes
the generation increment and forces a flush itself.

Either way, CPU0 gets flushed.


Also, I don't suppose you've looked at the paravirt instances of
flush_tlb_other() ? They don't elide the flushes because of lazy.