Re: [RFC 2/3] mm: export zap_page_range()

[David Hildenbrand <david@redhat.com>]

On 22.03.22 14:02, Xiaoguang Wang wrote:
> hi,
> 
>> On 18.03.22 10:55, Xiaoguang Wang wrote:
>>> Module target_core_user will use it to implement zero copy feature.
>>>
>>> Signed-off-by: Xiaoguang Wang <xiaoguang.wang@linux.alibaba.com>
>>> ---
>>>   mm/memory.c | 1 +
>>>   1 file changed, 1 insertion(+)
>>>
>>> diff --git a/mm/memory.c b/mm/memory.c
>>> index 1f745e4d11c2..9974d0406dad 100644
>>> --- a/mm/memory.c
>>> +++ b/mm/memory.c
>>> @@ -1664,6 +1664,7 @@ void zap_page_range(struct vm_area_struct *vma, unsigned long start,
>>>   	mmu_notifier_invalidate_range_end(&range);
>>>   	tlb_finish_mmu(&tlb);
>>>   }
>>> +EXPORT_SYMBOL_GPL(zap_page_range);
>>>   
>>>   /**
>>>    * zap_page_range_single - remove user pages in a given range
>> To which VMAs will you be applying zap_page_range? I assume only to some
>> special ones where you previously vm_insert_page(s)_mkspecial'ed pages,
>> not to some otherwise random VMAs, correct?
> Yes, you're right :)

I'd suggest exposing a dedicated function that performs sanity checks on
the vma (VM_PFNMAP ?) and only zaps within a single VMA.

Essentially zap_page_range_single(), excluding "struct zap_details
*details" and including sanity checks.

Reason is that we don't want anybody to blindly zap_page_range() within
random VMAs from a kernel module.

-- 
Thanks,

David / dhildenb