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:   Tue, 14 May 2019 13:49:19 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Nadav Amit <namit@vmware.com>
Cc:     Jan Stancek <jstancek@redhat.com>,
        Yang Shi <yang.shi@linux.alibaba.com>,
        Will Deacon <will.deacon@arm.com>,
        "minchan@kernel.org" <minchan@kernel.org>,
        "mgorman@suse.de" <mgorman@suse.de>,
        "stable@vger.kernel.org" <stable@vger.kernel.org>,
        "linux-mm@kvack.org" <linux-mm@kvack.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [v2 PATCH] mm: mmu_gather: remove __tlb_reset_range() for force
 flush
Message-ID: <20190514114919.GO2589@hirez.programming.kicks-ass.net>
References: <45c6096e-c3e0-4058-8669-75fbba415e07@email.android.com>
 <914836977.22577826.1557818139522.JavaMail.zimbra@redhat.com>
 <9E536319-815D-4425-B4B6-8786D415442C@vmware.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <9E536319-815D-4425-B4B6-8786D415442C@vmware.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, May 14, 2019 at 07:21:33AM +0000, Nadav Amit wrote:
> > On May 14, 2019, at 12:15 AM, Jan Stancek <jstancek@redhat.com> wrote:

> > Replacing fullmm with need_flush_all, brings the problem back / reproducer hangs.
> 
> Maybe setting need_flush_all does not have the right effect, but setting
> fullmm and then calling __tlb_reset_range() when the PTEs were already
> zapped seems strange.
> 
> fullmm is described as:
> 
>         /*
>          * we are in the middle of an operation to clear
>          * a full mm and can make some optimizations
>          */
> 
> And this not the case.

Correct; starting with fullmm would be wrong. For instance
tlb_start_vma() would do the wrong thing because it assumes the whole mm
is going away. But we're at tlb_finish_mmu() time and there the
difference doesn't matter anymore.

But yes, that's a wee abuse.