diff mbox series

ARC: mm: fix leakage of memory allocated for PTE

Message ID 20221017161127.24351-1-kozlov@synopsys.com
State New
Headers show
Series ARC: mm: fix leakage of memory allocated for PTE | expand

Commit Message

Pavel Kozlov Oct. 17, 2022, 4:11 p.m. UTC
From: Pavel Kozlov <pavel.kozlov@synopsys.com>

Since commit d9820ff ("ARC: mm: switch pgtable_t back to struct page *")
a memory leakage problem occurs. Memory allocated for page table entries
not released during process termination. This issue can be reproduced by
a small program that allocates a large amount of memory. After several
runs, you'll see that the amount of free memory has reduced and will
continue to reduce after each run. All ARC CPUs are effected by this
issue. The issue was introduced since the kernel stable release v5.15-rc1.

As described in commit d9820ff after switch pgtable_t back to struct
page *, a pointer to "struct page" and appropriate functions are used to
allocate and free a memory page for PTEs, but the pmd_pgtable macro hasn't
changed and returns the direct virtual address from the PMD (PGD) entry.
Than this address used as a parameter in the __pte_free() and as a result
this function couldn't release memory page allocated for PTEs.

Fix this issue by changing the pmd_pgtable macro and returning pointer to
struct page.

Fixes: d9820ff76f95 ("ARC: mm: switch pgtable_t back to struct page *")
Signed-off-by: Pavel Kozlov <pavel.kozlov@synopsys.com>
Cc: Vineet Gupta <vgupta@kernel.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: <stable@vger.kernel.org> # 4.15.x
---
 arch/arc/include/asm/pgtable-levels.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Vineet Gupta Oct. 19, 2022, 12:24 a.m. UTC | #1
On 10/17/22 09:11, Pavel.Kozlov@synopsys.com wrote:
> From: Pavel Kozlov <pavel.kozlov@synopsys.com>
>
> Since commit d9820ff ("ARC: mm: switch pgtable_t back to struct page *")
> a memory leakage problem occurs. Memory allocated for page table entries
> not released during process termination. This issue can be reproduced by
> a small program that allocates a large amount of memory. After several
> runs, you'll see that the amount of free memory has reduced and will
> continue to reduce after each run. All ARC CPUs are effected by this
> issue. The issue was introduced since the kernel stable release v5.15-rc1.
>
> As described in commit d9820ff after switch pgtable_t back to struct
> page *, a pointer to "struct page" and appropriate functions are used to
> allocate and free a memory page for PTEs, but the pmd_pgtable macro hasn't
> changed and returns the direct virtual address from the PMD (PGD) entry.
> Than this address used as a parameter in the __pte_free() and as a result
> this function couldn't release memory page allocated for PTEs.
>
> Fix this issue by changing the pmd_pgtable macro and returning pointer to
> struct page.

Good catch. Curious how did you find it. KMEMCHECK or some such or just oom.

> Fixes: d9820ff76f95 ("ARC: mm: switch pgtable_t back to struct page *")
> Signed-off-by: Pavel Kozlov <pavel.kozlov@synopsys.com>
> Cc: Vineet Gupta <vgupta@kernel.org>
> Cc: Mike Rapoport <rppt@kernel.org>
> Cc: <stable@vger.kernel.org> # 4.15.x

You meant 5.15.x

Added to for-curr.

Thx,
-Vineet

> ---
>   arch/arc/include/asm/pgtable-levels.h | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/arc/include/asm/pgtable-levels.h b/arch/arc/include/asm/pgtable-levels.h
> index 64ca25d199be..ef68758b69f7 100644
> --- a/arch/arc/include/asm/pgtable-levels.h
> +++ b/arch/arc/include/asm/pgtable-levels.h
> @@ -161,7 +161,7 @@
>   #define pmd_pfn(pmd)		((pmd_val(pmd) & PAGE_MASK) >> PAGE_SHIFT)
>   #define pmd_page(pmd)		virt_to_page(pmd_page_vaddr(pmd))
>   #define set_pmd(pmdp, pmd)	(*(pmdp) = pmd)
> -#define pmd_pgtable(pmd)	((pgtable_t) pmd_page_vaddr(pmd))
> +#define pmd_pgtable(pmd)	((pgtable_t) pmd_page(pmd))
>   
>   /*
>    * 4th level paging: pte
Pavel Kozlov Oct. 19, 2022, 1:35 p.m. UTC | #2
> Good catch. Curious how did you find it. KMEMCHECK or some such or just oom.

I've run glibc tests with 5.16 kernel and got many oom-killer messages and even kernel 
panic because lack of memory. These symptoms pointed to an issue. kmemleak didn't 
show anything. Didn't try kmemcheck. I prepared a small test and used git bisect to find
to the "blame" commit.

>> Cc: <stable@vger.kernel.org> # 4.15.x

> You meant 5.15.x

Yes, you right, that's my typo.

> Added to for-curr.

Thanks!

Regards,
Pavel
diff mbox series

Patch

diff --git a/arch/arc/include/asm/pgtable-levels.h b/arch/arc/include/asm/pgtable-levels.h
index 64ca25d199be..ef68758b69f7 100644
--- a/arch/arc/include/asm/pgtable-levels.h
+++ b/arch/arc/include/asm/pgtable-levels.h
@@ -161,7 +161,7 @@ 
 #define pmd_pfn(pmd)		((pmd_val(pmd) & PAGE_MASK) >> PAGE_SHIFT)
 #define pmd_page(pmd)		virt_to_page(pmd_page_vaddr(pmd))
 #define set_pmd(pmdp, pmd)	(*(pmdp) = pmd)
-#define pmd_pgtable(pmd)	((pgtable_t) pmd_page_vaddr(pmd))
+#define pmd_pgtable(pmd)	((pgtable_t) pmd_page(pmd))
 
 /*
  * 4th level paging: pte