Re: [PATCH v2 1/6] mm/huge_memory.c: rework the function vma_adjust_trans_huge()

[Miaohe Lin <linmiaohe@huawei.com>]

Hi:
On 2021/3/17 4:40, Peter Xu wrote:
> On Tue, Mar 16, 2021 at 08:40:02AM -0400, Miaohe Lin wrote:
>> +static inline void split_huge_pmd_if_needed(struct vm_area_struct *vma, unsigned long address)
>> +{
>> +	/*
>> +	 * If the new address isn't hpage aligned and it could previously
>> +	 * contain an hugepage: check if we need to split an huge pmd.
>> +	 */
>> +	if (address & ~HPAGE_PMD_MASK &&
>> +	    range_in_vma(vma, address & HPAGE_PMD_MASK,
>> +			 (address & HPAGE_PMD_MASK) + HPAGE_PMD_SIZE))
> 
> Since you're at it, maybe use ALIGN/ALIGN_DOWN too against HPAGE_PMD_SIZE?
> 

Many thanks for reply. Sounds good. :) Do you mean this?

diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index bff92dea5ab3..ae16a82da823 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2301,44 +2301,38 @@ void split_huge_pmd_address(struct vm_area_struct *vma, unsigned long address,
        __split_huge_pmd(vma, pmd, address, freeze, page);
 }

+static inline void split_huge_pmd_if_needed(struct vm_area_struct *vma, unsigned long address)
+{
+       /*
+        * If the new address isn't hpage aligned and it could previously
+        * contain an hugepage: check if we need to split an huge pmd.
+        */
+       if (!IS_ALIGNED(address, HPAGE_PMD_SIZE) &&
+           range_in_vma(vma, ALIGN_DOWN(address, HPAGE_PMD_SIZE),
+                        ALIGN(address, HPAGE_PMD_SIZE)))
+               split_huge_pmd_address(vma, address, false, NULL);
+}
+

>> +		split_huge_pmd_address(vma, address, false, NULL);
>> +}
>