diff mbox series

mm/rmap: Fix potential pte_unmap on an not mapped pte

Message ID 20210127093349.39081-1-linmiaohe@huawei.com (mailing list archive)
State New
Headers show
Series mm/rmap: Fix potential pte_unmap on an not mapped pte | expand

Commit Message

Miaohe Lin Jan. 27, 2021, 9:33 a.m. UTC
For PMD-mapped page (usually THP), pvmw->pte is NULL. For PTE-mapped THP,
pvmw->pte is mapped. But for HugeTLB pages, pvmw->pte is not mapped and set
to the relevant page table entry. So in page_vma_mapped_walk_done(), we may
do pte_unmap() for HugeTLB pte which is not mapped. Fix this by checking
pvmw->page against PageHuge before trying to do pte_unmap().

Fixes: commit ("ace71a19cec5 mm: introduce page_vma_mapped_walk()")
Signed-off-by: Hongxiang Lou <louhongxiang@huawei.com>
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Cc: <stable@vger.kernel.org>
---
 include/linux/rmap.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Andrew Morton Jan. 28, 2021, 12:09 a.m. UTC | #1
On Wed, 27 Jan 2021 04:33:49 -0500 Miaohe Lin <linmiaohe@huawei.com> wrote:

> For PMD-mapped page (usually THP), pvmw->pte is NULL. For PTE-mapped THP,
> pvmw->pte is mapped. But for HugeTLB pages, pvmw->pte is not mapped and set
> to the relevant page table entry. So in page_vma_mapped_walk_done(), we may
> do pte_unmap() for HugeTLB pte which is not mapped. Fix this by checking
> pvmw->page against PageHuge before trying to do pte_unmap().
> 

What are the runtime consequences of this?  Is there a workload which
is known to trigger it?

IOW, how do we justify a -stable backport of this fix?

>
> --- a/include/linux/rmap.h
> +++ b/include/linux/rmap.h
> @@ -213,7 +213,8 @@ struct page_vma_mapped_walk {
>  
>  static inline void page_vma_mapped_walk_done(struct page_vma_mapped_walk *pvmw)
>  {
> -	if (pvmw->pte)
> +	/* HugeTLB pte is set to the relevant page table entry without pte_mapped. */
> +	if (pvmw->pte && !PageHuge(pvmw->page))
>  		pte_unmap(pvmw->pte);
>  	if (pvmw->ptl)
>  		spin_unlock(pvmw->ptl);
> -- 
> 2.19.1
Miaohe Lin Jan. 28, 2021, 1:54 a.m. UTC | #2
Hi:
On 2021/1/28 8:09, Andrew Morton wrote:
> On Wed, 27 Jan 2021 04:33:49 -0500 Miaohe Lin <linmiaohe@huawei.com> wrote:
> 
>> For PMD-mapped page (usually THP), pvmw->pte is NULL. For PTE-mapped THP,
>> pvmw->pte is mapped. But for HugeTLB pages, pvmw->pte is not mapped and set
>> to the relevant page table entry. So in page_vma_mapped_walk_done(), we may
>> do pte_unmap() for HugeTLB pte which is not mapped. Fix this by checking
>> pvmw->page against PageHuge before trying to do pte_unmap().
>>
> 
> What are the runtime consequences of this?  Is there a workload which
> is known to trigger it?
> 

Not yet. This should not be backported. My bad. Sorry about it.

> IOW, how do we justify a -stable backport of this fix?
> >>
>> --- a/include/linux/rmap.h
>> +++ b/include/linux/rmap.h
>> @@ -213,7 +213,8 @@ struct page_vma_mapped_walk {
>>  
>>  static inline void page_vma_mapped_walk_done(struct page_vma_mapped_walk *pvmw)
>>  {
>> -	if (pvmw->pte)
>> +	/* HugeTLB pte is set to the relevant page table entry without pte_mapped. */
>> +	if (pvmw->pte && !PageHuge(pvmw->page))
>>  		pte_unmap(pvmw->pte);
>>  	if (pvmw->ptl)
>>  		spin_unlock(pvmw->ptl);
>> -- 
>> 2.19.1
> .
>
diff mbox series

Patch

diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index 70085ca1a3fc..def5c62c93b3 100644
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -213,7 +213,8 @@  struct page_vma_mapped_walk {
 
 static inline void page_vma_mapped_walk_done(struct page_vma_mapped_walk *pvmw)
 {
-	if (pvmw->pte)
+	/* HugeTLB pte is set to the relevant page table entry without pte_mapped. */
+	if (pvmw->pte && !PageHuge(pvmw->page))
 		pte_unmap(pvmw->pte);
 	if (pvmw->ptl)
 		spin_unlock(pvmw->ptl);