[RFC PATCH] mm: memblock: fix wrong memmove size in memblock_merge_regions()

[RFC PATCH] mm: memblock: fix wrong memmove size in memblock_merge_regions()

[Lin Feng]

The memmove span covers from (next+1) to the end of the array, and the index
of next is (i+1), so the index of (next+1) is (i+2). So the size of remaining
array elements is (type->cnt - (i + 2)).

PS. It seems that memblock_merge_regions() could be made some improvement:
we need't memmove the remaining array elements until we find a none-mergable
element, but now we memmove everytime we find a neighboring compatible region.
I'm not sure if the trial is worth though.

Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Lin Feng <linfeng@cn.fujitsu.com>
---
 mm/memblock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/memblock.c b/mm/memblock.c
index 6259055..85ce056 100644
--- a/mm/memblock.c
+++ b/mm/memblock.c
@@ -314,7 +314,7 @@ static void __init_memblock memblock_merge_regions(struct memblock_type *type)
 		}
 
 		this->size += next->size;
-		memmove(next, next + 1, (type->cnt - (i + 1)) * sizeof(*next));
+		memmove(next, next + 1, (type->cnt - (i + 2)) * sizeof(*next));
 		type->cnt--;
 	}
 }
-- 
1.7.11.7

Re: [RFC PATCH] mm: memblock: fix wrong memmove size in memblock_merge_regions()

[Tejun Heo]

On Fri, Jan 04, 2013 at 05:10:50PM +0800, Lin Feng wrote:
> The memmove span covers from (next+1) to the end of the array, and the index
> of next is (i+1), so the index of (next+1) is (i+2). So the size of remaining
> array elements is (type->cnt - (i + 2)).
> 
> PS. It seems that memblock_merge_regions() could be made some improvement:
> we need't memmove the remaining array elements until we find a none-mergable
> element, but now we memmove everytime we find a neighboring compatible region.
> I'm not sure if the trial is worth though.
> 
> Cc: Tejun Heo <tj@kernel.org>
> Signed-off-by: Lin Feng <linfeng@cn.fujitsu.com>
> ---
>  mm/memblock.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/mm/memblock.c b/mm/memblock.c
> index 6259055..85ce056 100644
> --- a/mm/memblock.c
> +++ b/mm/memblock.c
> @@ -314,7 +314,7 @@ static void __init_memblock memblock_merge_regions(struct memblock_type *type)
>  		}
>  
>  		this->size += next->size;
> -		memmove(next, next + 1, (type->cnt - (i + 1)) * sizeof(*next));
> +		memmove(next, next + 1, (type->cnt - (i + 2)) * sizeof(*next));

Heh, that's confusing.  Nice catch.  Can you please also add a comment
explaning the index so that it's less confusing for the future readers?

Thanks.

-- 
tejun

Re: [RFC PATCH] mm: memblock: fix wrong memmove size in memblock_merge_regions()

[Lin Feng]

On 01/04/2013 10:56 PM, Tejun Heo wrote:
> On Fri, Jan 04, 2013 at 05:10:50PM +0800, Lin Feng wrote:
>> The memmove span covers from (next+1) to the end of the array, and the index
>> of next is (i+1), so the index of (next+1) is (i+2). So the size of remaining
>> array elements is (type->cnt - (i + 2)).
>>
>> PS. It seems that memblock_merge_regions() could be made some improvement:
>> we need't memmove the remaining array elements until we find a none-mergable
>> element, but now we memmove everytime we find a neighboring compatible region.
>> I'm not sure if the trial is worth though.
>>
>> Cc: Tejun Heo <tj@kernel.org>
>> Signed-off-by: Lin Feng <linfeng@cn.fujitsu.com>
>> ---
>>  mm/memblock.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/mm/memblock.c b/mm/memblock.c
>> index 6259055..85ce056 100644
>> --- a/mm/memblock.c
>> +++ b/mm/memblock.c
>> @@ -314,7 +314,7 @@ static void __init_memblock memblock_merge_regions(struct memblock_type *type)
>>  		}
>>  
>>  		this->size += next->size;
>> -		memmove(next, next + 1, (type->cnt - (i + 1)) * sizeof(*next));
>> +		memmove(next, next + 1, (type->cnt - (i + 2)) * sizeof(*next));
> 
> Heh, that's confusing.  Nice catch.  Can you please also add a comment
> explaning the index so that it's less confusing for the future readers?
Hi Tejun,

Glad to. I will resend it soon.

thanks,
linfeng
> 
> Thanks.
> 

Re: [RFC PATCH] mm: memblock: fix wrong memmove size in memblock_merge_regions()

[Lin Feng]

On 01/05/2013 09:04 AM, Wanpeng Li wrote:
> On Fri, Jan 04, 2013 at 05:10:50PM +0800, Lin Feng wrote:
>> The memmove span covers from (next+1) to the end of the array, and the index
>> of next is (i+1), so the index of (next+1) is (i+2). So the size of remaining
>> array elements is (type->cnt - (i + 2)).
>>
> 
> Make sense.
Hi Wanpeng,

Thanks for your review. I will add it in next version.

thanks,
linfeng
> 
> Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
> 
>> PS. It seems that memblock_merge_regions() could be made some improvement:
>> we need't memmove the remaining array elements until we find a none-mergable
>> element, but now we memmove everytime we find a neighboring compatible region.
>> I'm not sure if the trial is worth though.
>>
>> Cc: Tejun Heo <tj@kernel.org>
>> Signed-off-by: Lin Feng <linfeng@cn.fujitsu.com>
>> ---
>> mm/memblock.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/mm/memblock.c b/mm/memblock.c
>> index 6259055..85ce056 100644
>> --- a/mm/memblock.c
>> +++ b/mm/memblock.c
>> @@ -314,7 +314,7 @@ static void __init_memblock memblock_merge_regions(struct memblock_type *type)
>> 		}
>>
>> 		this->size += next->size;
>> -		memmove(next, next + 1, (type->cnt - (i + 1)) * sizeof(*next));
>> +		memmove(next, next + 1, (type->cnt - (i + 2)) * sizeof(*next));
>> 		type->cnt--;
>> 	}
>> }
>> -- 
>> 1.7.11.7
> 
>