Re: [PATCH] Linux VM workaround for Knights Landing A/D leak

From: Dave Hansen <dave.hansen@linux.intel.com>
To: Nadav Amit <nadav.amit@gmail.com>,
	Lukasz Anaczkowski <lukasz.anaczkowski@intel.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
	linux-mm@kvack.org, Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>,
	ak@linux.intel.com, kirill.shutemov@linux.intel.com,
	mhocko@suse.com, Andrew Morton <akpm@linux-foundation.org>,
	"H. Peter Anvin" <hpa@zytor.com>,
	harish.srinivasappa@intel.com, lukasz.odzioba@intel.com
Subject: Re: [PATCH] Linux VM workaround for Knights Landing A/D leak
Date: Tue, 14 Jun 2016 10:18:25 -0700	[thread overview]
Message-ID: <57603C61.5000408@linux.intel.com> (raw)
In-Reply-To: <7FB15233-B347-4A87-9506-A9E10D331292@gmail.com>

On 06/14/2016 09:47 AM, Nadav Amit wrote:
> Lukasz Anaczkowski <lukasz.anaczkowski@intel.com> wrote:
> 
>> > From: Andi Kleen <ak@linux.intel.com>
>> > +void fix_pte_leak(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
>> > +{
> Here there should be a call to smp_mb__after_atomic() to synchronize with
> switch_mm. I submitted a similar patch, which is still pending (hint).
> 
>> > +	if (cpumask_any_but(mm_cpumask(mm), smp_processor_id()) < nr_cpu_ids) {
>> > +		trace_tlb_flush(TLB_LOCAL_SHOOTDOWN, TLB_FLUSH_ALL);
>> > +		flush_tlb_others(mm_cpumask(mm), mm, addr,
>> > +				 addr + PAGE_SIZE);
>> > +		mb();
>> > +		set_pte(ptep, __pte(0));
>> > +	}
>> > +}

Shouldn't that barrier be incorporated in the TLB flush code itself and
not every single caller (like this code is)?

It is insane to require individual TLB flushers to be concerned with the
barriers.