Re: [PATCH] mm/mm_init.c: remove obsolete macro HASH_SMALL

[Miaohe Lin <linmiaohe@huawei.com>]

On 2023/6/17 18:51, Mike Rapoport wrote:
> On Sat, Jun 17, 2023 at 04:39:44PM +0800, Miaohe Lin wrote:
>> On 2023/6/17 15:56, Mike Rapoport wrote:
>>> On Sat, Jun 17, 2023 at 03:09:55PM +0800, Miaohe Lin wrote:
>>>> HASH_SMALL only works when parameter numentries is 0. But the sole caller
>>>> futex_init() never calls alloc_large_system_hash() with numentries set to
>>>> 0. 
>>>
>>
>> Thanks for your quick review.
>>
>>> Doesn't it? 
>>> What happens when CONFIG_BASE_SMALL is set?
>>
>> When CONFIG_BASE_SMALL is set, futex_hashsize is set to 16 and alloc_large_system_hash() is called with
>> numentries == 16 && flags == HASH_SMALL. But in the alloc_large_system_hash(), we have the below logic:
>>
>> alloc_large_system_hash()
>> {
>>   if (!numentries) { /* numentries == 16 here, so this code block is skipped. */
>>     ...
>>     if (unlikely(flags & HASH_SMALL)) { /* So as here. */
>>       ...
>>   }
>>   ...
>> }
>>
>> So HASH_SMALL is just unused. Or am I miss something?
> 
> You are right, I've missed that. 
> 
>> Thanks.
>>
>>>
>>>> So HASH_SMALL is obsolete and remove it.
>>>
>>>> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
>>>> ---
>>>>  include/linux/memblock.h |  2 --
>>>>  kernel/futex/core.c      |  3 +--
>>>>  mm/mm_init.c             | 10 +---------
>>>>  3 files changed, 2 insertions(+), 13 deletions(-)
>>>>
>>>> diff --git a/include/linux/memblock.h b/include/linux/memblock.h
>>>> index f71ff9f0ec81..346d80809517 100644
>>>> --- a/include/linux/memblock.h
>>>> +++ b/include/linux/memblock.h
>>>> @@ -581,8 +581,6 @@ extern void *alloc_large_system_hash(const char *tablename,
>>>>  				     unsigned long high_limit);
>>>>  
>>>>  #define HASH_EARLY	0x00000001	/* Allocating during early boot? */
>>>> -#define HASH_SMALL	0x00000002	/* sub-page allocation allowed, min
>>>> -					 * shift passed via *_hash_shift */
>>>>  #define HASH_ZERO	0x00000004	/* Zero allocated hash table */
> 
> Can you update HASH_ZERO to 0x2?

Will do. Thanks.