Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <d0edd1f3-4bad-f55a-e4f5-e202bc9d09f5@redhat.com>
Date:   Tue, 1 Aug 2023 19:09:55 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.13.0
Subject: Re: [PATCH v2 1/8] mm/gup: reintroduce FOLL_NUMA as
 FOLL_HONOR_NUMA_FAULT
Content-Language: en-US
To:     Peter Xu <peterx@redhat.com>
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        linux-fsdevel@vger.kernel.org, kvm@vger.kernel.org,
        linux-kselftest@vger.kernel.org,
        Andrew Morton <akpm@linux-foundation.org>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        liubo <liubo254@huawei.com>,
        Matthew Wilcox <willy@infradead.org>,
        Hugh Dickins <hughd@google.com>,
        Jason Gunthorpe <jgg@ziepe.ca>,
        John Hubbard <jhubbard@nvidia.com>,
        Mel Gorman <mgorman@suse.de>, Shuah Khan <shuah@kernel.org>,
        Paolo Bonzini <pbonzini@redhat.com>, stable@vger.kernel.org
References: <20230801124844.278698-1-david@redhat.com>
 <20230801124844.278698-2-david@redhat.com> <ZMkpM95vdc9wgs9T@x1n>
 <30d86a2d-4af2-d840-91be-2e68c73a07bd@redhat.com> <ZMk7MqApTDZXzwKX@x1n>
From:   David Hildenbrand <david@redhat.com>
Organization: Red Hat
In-Reply-To: <ZMk7MqApTDZXzwKX@x1n>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 01.08.23 19:04, Peter Xu wrote:
> On Tue, Aug 01, 2023 at 06:15:48PM +0200, David Hildenbrand wrote:
>> On 01.08.23 17:48, Peter Xu wrote:
>>> On Tue, Aug 01, 2023 at 02:48:37PM +0200, David Hildenbrand wrote:
>>>> @@ -2240,6 +2244,12 @@ static bool is_valid_gup_args(struct page **pages, int *locked,
>>>>    		gup_flags |= FOLL_UNLOCKABLE;
>>>>    	}
>>>> +	/*
>>>> +	 * For now, always trigger NUMA hinting faults. Some GUP users like
>>>> +	 * KVM really require it to benefit from autonuma.
>>>> +	 */
>>>> +	gup_flags |= FOLL_HONOR_NUMA_FAULT;
>>>
>>> Since at it, do we want to not set it for FOLL_REMOTE, which still sounds
>>> like a good thing to have?
>>
>> I thought about that, but decided against making that patch here more
>> complicated to eventually rip it again all out in #4.
> 
> I thought that was the whole point of having patch 4 separate, because we
> should assume patch 4 may not exist in (at least) some trees, so I ignored
> patch 4 when commenting here, and we should not assume it's destined to be
> removed here.

For me, the goal of this patch is to bring it *as close as possible* to 
the previous state as before, so we can backport it to stable without 
too many surprises (effectively, only a handful of FOLL_FORCE/ptrace 
user will get a different behavior).

I could add a separate patch that does the FOLL_REMOTE thing, but then, 
maybe we should do that if patch #4 runs into real trouble :)

But no strong opinion if this is what everybody wants in this patch.

> 
>>
>> I fully agree that FOLL_REMOTE does not make too much sense, but let's
>> rather keep it simple for this patch.
> 
> It's fine - I suppose this patch fixes whatever we're aware of that's
> broken with FOLL_NUMA's removal, so it can even be anything on top when
> needed.  I assume I'm happy to ack this with/without that change, then:
> 
> Acked-by: Peter Xu <peterx@redhat.com>

Thanks!

> 
> PS: I still hope that the other oneliner can be squashed here directly; it
> literally changes exact the same line above so reading this patch alone can
> be affected.  You said there you didn't want the commit message to be too
> long here, but this is definitely not long at all!  I bet you have similar
> understanding to me on defining "long commit message". :) I'd never worry
> that.  Your call.

No strong opinion, it just felt cleaner to not start adding what I have 
in that separate patch commit message in here.

-- 
Cheers,

David / dhildenb