Received-SPF: pass (google.com: domain of linux-kernel+bounces-72408-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Message-ID: <6da5c282-a522-4732-ad3c-73f6790c073e@linux.alibaba.com>
Date: Tue, 20 Feb 2024 14:28:10 +0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v6 0/4] Enable >0 order folio memory compaction
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Zi Yan <ziy@nvidia.com>, linux-mm@kvack.org,
 linux-kernel@vger.kernel.org, "Huang, Ying" <ying.huang@intel.com>,
 Ryan Roberts <ryan.roberts@arm.com>,
 "Matthew Wilcox (Oracle)" <willy@infradead.org>,
 David Hildenbrand <david@redhat.com>, "Yin, Fengwei"
 <fengwei.yin@intel.com>, Yu Zhao <yuzhao@google.com>,
 Vlastimil Babka <vbabka@suse.cz>,
 "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>,
 Johannes Weiner <hannes@cmpxchg.org>, Kemeng Shi
 <shikemeng@huaweicloud.com>, Mel Gorman <mgorman@techsingularity.net>,
 Rohan Puri <rohan.puri15@gmail.com>, Mcgrof Chamberlain <mcgrof@kernel.org>,
 Adam Manzanares <a.manzanares@samsung.com>,
 "Vishal Moola (Oracle)" <vishal.moola@gmail.com>
References: <20240216170432.1268753-1-zi.yan@sent.com>
 <20240219180617.89a0daaba45ce124ef01be50@linux-foundation.org>
 <9E3E6CD6-45DD-4E7F-B6DC-0897F38D5E81@nvidia.com>
 <4f0b3586-25e2-4d21-bbb6-41f644d3d77f@linux.alibaba.com>
 <20240219193048.cbd010579aa3d2c790f6e2b9@linux-foundation.org>
From: Baolin Wang <baolin.wang@linux.alibaba.com>
In-Reply-To: <20240219193048.cbd010579aa3d2c790f6e2b9@linux-foundation.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit


On 2024/2/20 11:30, Andrew Morton wrote:
> On Tue, 20 Feb 2024 11:00:39 +0800 Baolin Wang <baolin.wang@linux.alibaba.com> wrote:
> 
>>> The patch was based on top of my early version of this patchset, thus
>>> uses "cc->nr_migratepages -= 1 << order;" and
>>> "cc->nr_migratepages += 1 << order;", but now it is applied before
>>> mine. The change should be "cc->nr_migratepages--;" and
>>> "cc->nr_migratepages++;", respectively.
>>>
>>>>
>>>>> on nr_migratepages was based on this one, a better fixup
>>>>> for it might be below. Since before my patchset, compaction only deals with
>>>>> order-0 pages.
>>>>
>>>> I don't understand what this means.  The patchset you sent applies OK
>>>> to mm-unstable so what else is there to do?
>>>
>>> Your above fixup to Baolin's patch needs to be changed to the patch below
>>> and my "mm/compaction: add support for >0 order folio memory compaction" will
>>> need to be adjusted accordingly to be applied on top.
>>>
>>> Let me know if anything is unclear.
>>
>> Hi Andrew,
>>
>> To avoid conflicts, you can drop these two patches, and I will send a
>> new version with fixing the issue pointed by Vlastimilb on top of
>> "mm/compaction: add support for >0 order folio memory compaction".
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/commit/?h=mm-everything-2024-02-16-01-35&id=97f749c7c82f677f89bbf4f10de7816ce9b071fe
>>
>> https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/commit/?h=mm-everything-2024-02-16-01-35&id=ea87b0558293a5ad597bea606fe261f7b2650cda
> 
> Well I thought I'd fixed everything up 10 minutes ago.  Please take a
> look at next mm-unstable.

Sure. And I found a minor rebase error in the compaction_alloc() 
function while Zi Yan's original patch is correct.

+	cc->nr_freepages -= 1 << order;
  	cc->nr_migratepages--;
-
-	return dst;
+	return page_rmappable_folio(&dst->page);
  }

should change to be:
cc->nr_migratepages -= 1 << order;