[PATCH] mm/compaction: fix set skip in fast_find_migrateblock

[PATCH] mm/compaction: fix set skip in fast_find_migrateblock

[Zhou Chuyi]

From: zhouchuyi <zhouchuyi@bytedance.com>

When we successfully find a pageblock in fast_find_migrateblock(), the block will be set skip-flag through set_pageblock_skip(). However, when entering isolate_migratepages_block(), the whole pageblock will be skipped due to the branch 'if (!valid_page && IS_ALIGNED(low_pfn, pageblock_nr_pages))'. Eventually we will goto isolate_abort and isolate nothing.
Signed-off-by: zhouchuyi <zhouchuyi@bytedance.com>
---
 mm/compaction.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/mm/compaction.c b/mm/compaction.c
index 1f89b969c..a1a2b50c8 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -1852,7 +1852,6 @@ static unsigned long fast_find_migrateblock(struct compact_control *cc)
 					pfn = cc->zone->zone_start_pfn;
 				cc->fast_search_fail = 0;
 				found_block = true;
-				set_pageblock_skip(freepage);
 				break;
 			}
 		}
-- 
2.20.1

Re: [PATCH] mm/compaction: fix set skip in fast_find_migrateblock

[Miaohe Lin]

Add Cc Andrew and linux-kernel email list.

On 2022/7/11 20:32, Zhou Chuyi wrote:
> From: zhouchuyi <zhouchuyi@bytedance.com>
> 
> When we successfully find a pageblock in fast_find_migrateblock(), the block will be set skip-flag through set_pageblock_skip(). However, when entering isolate_migratepages_block(), the whole pageblock will be skipped due to the branch 'if (!valid_page && IS_ALIGNED(low_pfn, pageblock_nr_pages))'. Eventually we will goto isolate_abort and isolate nothing.
> Signed-off-by: zhouchuyi <zhouchuyi@bytedance.com>

It seems we should tweak the commit log to satisfy the checkpatch.pl first.

> ---
>  mm/compaction.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/mm/compaction.c b/mm/compaction.c
> index 1f89b969c..a1a2b50c8 100644
> --- a/mm/compaction.c
> +++ b/mm/compaction.c
> @@ -1852,7 +1852,6 @@ static unsigned long fast_find_migrateblock(struct compact_control *cc)
>  					pfn = cc->zone->zone_start_pfn;
>  				cc->fast_search_fail = 0;
>  				found_block = true;
> -				set_pageblock_skip(freepage);

This looks like a real problem. Should we add a Fixes tag here? What's the runtime effect of it?

Thanks for your patch!

>  				break;
>  			}
>  		}
> 

Re: [External] Re: [PATCH] mm/compaction: fix set skip in fast_find_migrateblock

[周俶易]

在 2022/7/12 上午10:26, Miaohe Lin 写道:
> Add Cc Andrew and linux-kernel email list.
> 
> On 2022/7/11 20:32, Zhou Chuyi wrote:
>> From: zhouchuyi <zhouchuyi@bytedance.com>
>>
>> When we successfully find a pageblock in fast_find_migrateblock(), the block will be set skip-flag through set_pageblock_skip(). However, when entering isolate_migratepages_block(), the whole pageblock will be skipped due to the branch 'if (!valid_page && IS_ALIGNED(low_pfn, pageblock_nr_pages))'. Eventually we will goto isolate_abort and isolate nothing.
>> Signed-off-by: zhouchuyi <zhouchuyi@bytedance.com>
> 
> It seems we should tweak the commit log to satisfy the checkpatch.pl first.
> 
>> ---
>>   mm/compaction.c | 1 -
>>   1 file changed, 1 deletion(-)
>>
>> diff --git a/mm/compaction.c b/mm/compaction.c
>> index 1f89b969c..a1a2b50c8 100644
>> --- a/mm/compaction.c
>> +++ b/mm/compaction.c
>> @@ -1852,7 +1852,6 @@ static unsigned long fast_find_migrateblock(struct compact_control *cc)
>>   					pfn = cc->zone->zone_start_pfn;
>>   				cc->fast_search_fail = 0;
>>   				found_block = true;
>> -				set_pageblock_skip(freepage);
> 
> This looks like a real problem. Should we add a Fixes tag here? What's the runtime effect of it?
> 
> Thanks for your patch!
sorry for this mistakes, I'm a beginner of Linux kernel. I will fix this 
problems in v2
> 
>>   				break;
>>   			}
>>   		}
>>
>