[PATCH] s390/mm: fix pud table accounting

[PATCH] s390/mm: fix pud table accounting

[Heiko Carstens]

With "mm: account pud page tables" and "mm: consolidate page table
accounting" pud page table accounting was introduced which now results
in tons of warnings like this one on s390:

BUG: non-zero pgtables_bytes on freeing mm: -16384

Reason for this are our run-time folded page tables: by default new
processes start with three page table levels where the allocated pgd
is the same as the first pud. In this case there won't ever be a pud
allocated and therefore mm_inc_nr_puds() will also never be called.

However when freeing the address space free_pud_range() will call
exactly once mm_dec_nr_puds() which leads to misaccounting.

Therefore call mm_inc_nr_puds() within init_new_context() to fix
this. This is the same like we have it already for processes that run
with two page table levels (aka compat processes).

While at it also adjust the comment, since there is no "mm->nr_pmds"
anymore.

Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
---
 arch/s390/include/asm/mmu_context.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 3c9abedc323c..4f943d58cbac 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -43,6 +43,8 @@ static inline int init_new_context(struct task_struct *tsk,
 		mm->context.asce_limit = STACK_TOP_MAX;
 		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
 				   _ASCE_USER_BITS | _ASCE_TYPE_REGION3;
+		/* pgd_alloc() did not account this pud */
+		mm_inc_nr_puds(mm);
 		break;
 	case -PAGE_SIZE:
 		/* forked 5-level task, set new asce with new_mm->pgd */
@@ -58,7 +60,7 @@ static inline int init_new_context(struct task_struct *tsk,
 		/* forked 2-level compat task, set new asce with new mm->pgd */
 		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
 				   _ASCE_USER_BITS | _ASCE_TYPE_SEGMENT;
-		/* pgd_alloc() did not increase mm->nr_pmds */
+		/* pgd_alloc() did not account this pmd */
 		mm_inc_nr_pmds(mm);
 	}
 	crst_table_init((unsigned long *) mm->pgd, pgd_entry_type(mm));
-- 
2.13.5

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] s390/mm: fix pud table accounting

[Michal Hocko]

On Fri 03-11-17 10:05:51, Heiko Carstens wrote:
> With "mm: account pud page tables" and "mm: consolidate page table
> accounting" pud page table accounting was introduced which now results
> in tons of warnings like this one on s390:
> 
> BUG: non-zero pgtables_bytes on freeing mm: -16384
> 
> Reason for this are our run-time folded page tables: by default new
> processes start with three page table levels where the allocated pgd
> is the same as the first pud. In this case there won't ever be a pud
> allocated and therefore mm_inc_nr_puds() will also never be called.
> 
> However when freeing the address space free_pud_range() will call
> exactly once mm_dec_nr_puds() which leads to misaccounting.
> 
> Therefore call mm_inc_nr_puds() within init_new_context() to fix
> this. This is the same like we have it already for processes that run
> with two page table levels (aka compat processes).
> 
> While at it also adjust the comment, since there is no "mm->nr_pmds"
> anymore.

Subtle...

Thanks for the fix, I didn't have any idea about this when reviewing.

> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> Cc: Michal Hocko <mhocko@suse.com>
> Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
> ---
>  arch/s390/include/asm/mmu_context.h | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
> index 3c9abedc323c..4f943d58cbac 100644
> --- a/arch/s390/include/asm/mmu_context.h
> +++ b/arch/s390/include/asm/mmu_context.h
> @@ -43,6 +43,8 @@ static inline int init_new_context(struct task_struct *tsk,
>  		mm->context.asce_limit = STACK_TOP_MAX;
>  		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
>  				   _ASCE_USER_BITS | _ASCE_TYPE_REGION3;
> +		/* pgd_alloc() did not account this pud */
> +		mm_inc_nr_puds(mm);
>  		break;
>  	case -PAGE_SIZE:
>  		/* forked 5-level task, set new asce with new_mm->pgd */
> @@ -58,7 +60,7 @@ static inline int init_new_context(struct task_struct *tsk,
>  		/* forked 2-level compat task, set new asce with new mm->pgd */
>  		mm->context.asce = __pa(mm->pgd) | _ASCE_TABLE_LENGTH |
>  				   _ASCE_USER_BITS | _ASCE_TYPE_SEGMENT;
> -		/* pgd_alloc() did not increase mm->nr_pmds */
> +		/* pgd_alloc() did not account this pmd */
>  		mm_inc_nr_pmds(mm);
>  	}
>  	crst_table_init((unsigned long *) mm->pgd, pgd_entry_type(mm));
> -- 
> 2.13.5
> 

-- 
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] s390/mm: fix pud table accounting

[Kirill A. Shutemov]

On Fri, Nov 03, 2017 at 09:05:51AM +0000, Heiko Carstens wrote:
> With "mm: account pud page tables" and "mm: consolidate page table
> accounting" pud page table accounting was introduced which now results
> in tons of warnings like this one on s390:
> 
> BUG: non-zero pgtables_bytes on freeing mm: -16384
> 
> Reason for this are our run-time folded page tables: by default new
> processes start with three page table levels where the allocated pgd
> is the same as the first pud. In this case there won't ever be a pud
> allocated and therefore mm_inc_nr_puds() will also never be called.
> 
> However when freeing the address space free_pud_range() will call
> exactly once mm_dec_nr_puds() which leads to misaccounting.
> 
> Therefore call mm_inc_nr_puds() within init_new_context() to fix
> this. This is the same like we have it already for processes that run
> with two page table levels (aka compat processes).
> 
> While at it also adjust the comment, since there is no "mm->nr_pmds"
> anymore.

Thanks for tracking it down.

Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>

-- 
 Kirill A. Shutemov

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>