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;
Subject: Re: [PATCH 6/9] mm/compaction: rename is_via_compact_memory to
 compaction_with_allocation_order
To:     Baolin Wang <baolin.wang@linux.alibaba.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, akpm@linux-foundation.org,
        mgorman@techsingularity.net, david@redhat.com
References: <20230805110711.2975149-1-shikemeng@huaweicloud.com>
 <20230805110711.2975149-7-shikemeng@huaweicloud.com>
 <111c3142-e20b-3e3c-f107-cbb64a16c2b0@linux.alibaba.com>
From:   Kemeng Shi <shikemeng@huaweicloud.com>
Message-ID: <d28fc70d-ea1e-4c27-a206-6c276e6e020e@huaweicloud.com>
Date:   Tue, 15 Aug 2023 20:04:36 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.5.0
MIME-Version: 1.0
In-Reply-To: <111c3142-e20b-3e3c-f107-cbb64a16c2b0@linux.alibaba.com>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Precedence: bulk


on 8/15/2023 4:58 PM, Baolin Wang wrote:
> 
> 
> On 8/5/2023 7:07 PM, Kemeng Shi wrote:
>> We have order = -1 via proactive compaction, the is_via_compact_memory is
>> not proper name anymore.
>> As cc->order informs the compaction to satisfy a allocation with that
>> order, so rename it to compaction_with_allocation_order.
>>
>> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
>> ---
>>   mm/compaction.c | 11 +++++------
>>   1 file changed, 5 insertions(+), 6 deletions(-)
>>
>> diff --git a/mm/compaction.c b/mm/compaction.c
>> index d8416d3dd445..b5a699ed526b 100644
>> --- a/mm/compaction.c
>> +++ b/mm/compaction.c
>> @@ -2055,12 +2055,11 @@ static isolate_migrate_t isolate_migratepages(struct compact_control *cc)
>>   }
>>     /*
>> - * order == -1 is expected when compacting via
>> - * /proc/sys/vm/compact_memory
>> + * compact to satisfy allocation with target order
>>    */
>> -static inline bool is_via_compact_memory(int order)
>> +static inline bool compaction_with_allocation_order(int order)
> 
> I know naming is hard, but this name is not good enough that can show the compaction mode. But the original one could.
> 
Yes, I agree with this, but name and comment of is_via_compact_memory may
mislead reader that order == -1 is equivalent to compaction from
/proc/sys/vm/compact_memory.
Actually, we have several approaches to trigger compaction with order == -1:
1. via /proc/sys/vm/compact_memory
2. via /sys/devices/system/node/nodex/compact
3. via proactive compact

Instead of indicate compaction is tirggerred by compact_memocy or anything,
order == -1 implies if compaction is triggerrred to meet allocation with high
order and we will stop compaction if allocation with target order will success.

>>   {
>> -    return order == -1;
>> +    return order != -1;
>>   }
>>     /*
>> @@ -2200,7 +2199,7 @@ static enum compact_result __compact_finished(struct compact_control *cc)
>>           goto out;
>>       }
>>   -    if (is_via_compact_memory(cc->order))
>> +    if (!compaction_with_allocation_order(cc->order))
>>           return COMPACT_CONTINUE;
>>         /*
>> @@ -2390,7 +2389,7 @@ compact_zone(struct compact_control *cc, struct capture_control *capc)
>>         cc->migratetype = gfp_migratetype(cc->gfp_mask);
>>   -    if (!is_via_compact_memory(cc->order)) {
>> +    if (compaction_with_allocation_order(cc->order)) {
>>           unsigned long watermark;
>>             /* Allocation can already succeed, nothing to do */
>