On 1/2/19 5:56 PM, Jerome Glisse wrote:
> On Wed, Jan 02, 2019 at 04:21:26PM -0800, Matthew Wilcox wrote:
>>
>> One of the paths in follow_pte_pmd() initialised the mmu_notifier_range
>> incorrectly.
>>
>> Signed-off-by: Matthew Wilcox <wi...@infradead.org>
>> Fixes: ac46d4f3c432 ("mm/mmu_notifier: use structure for
>> invalidate_range_start/end calls v2")
>> Tested-by: Dave Chinner <dchin...@redhat.com>
>
> Reviewed-by: Jérôme Glisse <jgli...@redhat.com>
>
>>
>> diff --git a/mm/memory.c b/mm/memory.c
>> index 2dd2f9ab57f4..21a650368be0 100644
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -4078,8 +4078,8 @@ static int __follow_pte_pmd(struct mm_struct *mm,
>> unsigned long address,
>> goto out;
>>
>> if (range) {
>> - range->start = address & PAGE_MASK;
>> - range->end = range->start + PAGE_SIZE;
>> + mmu_notifier_range_init(range, mm, address & PAGE_MASK,
>> + (address & PAGE_MASK) + PAGE_SIZE);
>> mmu_notifier_invalidate_range_start(range);
>> }
>> ptep = pte_offset_map_lock(mm, pmd, address, ptlp);
>
Looks correct to me, as well.
Having the range struct declared in separate places from the
mmu_notifier_range_init()
calls is not great. But I'm not sure I see a way to make it significantly
cleaner, given
that __follow_pte_pmd uses the range pointer as a way to decide to issue the
mmn calls.
thanks,
--
John Hubbard
NVIDIA
