2019-05-02 15:31:42

by Mike Rapoport

Subject: [PATCH 08/15] mips: switch to generic version of pte allocation

MIPS allocates kernel PTE pages with

__get_free_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)

and user PTE pages with

alloc_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)

The PTE_ORDER is hardwired to zero, which makes MIPS implementation almost
identical to the generic one.

Switch MIPS to the generic version that does exactly the same thing for the
kernel page tables and adds __GFP_ACCOUNT for the user PTEs.

The pte_free_kernel() and pte_free() versions on mips are identical to the
generic ones and can be simply dropped.

Signed-off-by: Mike Rapoport <[email protected]>
---
arch/mips/include/asm/pgalloc.h | 33 ++-------------------------------
1 file changed, 2 insertions(+), 31 deletions(-)

diff --git a/arch/mips/include/asm/pgalloc.h b/arch/mips/include/asm/pgalloc.h
index 27808d9..aa16b85 100644
--- a/arch/mips/include/asm/pgalloc.h
+++ b/arch/mips/include/asm/pgalloc.h
@@ -13,6 +13,8 @@
#include <linux/mm.h>
#include <linux/sched.h>

+#include <asm-generic/pgalloc.h> /* for pte_{alloc,free}_one */
+
static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd,
pte_t *pte)
{
@@ -50,37 +52,6 @@ static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
free_pages((unsigned long)pgd, PGD_ORDER);
}

-static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm)
-{
- return (pte_t *)__get_free_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER);
-}
-
-static inline struct page *pte_alloc_one(struct mm_struct *mm)
-{
- struct page *pte;
-
- pte = alloc_pages(GFP_KERNEL, PTE_ORDER);
- if (!pte)
- return NULL;
- clear_highpage(pte);
- if (!pgtable_page_ctor(pte)) {
- __free_page(pte);
- return NULL;
- }
- return pte;
-}
-
-static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
-{
- free_pages((unsigned long)pte, PTE_ORDER);
-}
-
-static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
-{
- pgtable_page_dtor(pte);
- __free_pages(pte, PTE_ORDER);
-}
-
#define __pte_free_tlb(tlb,pte,address) \
do { \
pgtable_page_dtor(pte); \
--
2.7.4

2019-05-02 19:11:01

by Paul Burton

Subject: Re: [PATCH 08/15] mips: switch to generic version of pte allocation

Hi Mike,

On Thu, May 02, 2019 at 06:28:35PM +0300, Mike Rapoport wrote:
> MIPS allocates kernel PTE pages with
>
> __get_free_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)
>
> and user PTE pages with
>
> alloc_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)

That bit isn't quite true - we don't use __GFP_ZERO in pte_alloc_one() &
instead call clear_highpage() on the allocated page. Not that I have a
problem with using __GFP_ZERO - it seems like the more optimal choice.
It just might be worth mentioning the change & expected equivalent
behavior.

Otherwise:

Acked-by: Paul Burton <[email protected]>

Thanks,
Paul

2019-05-05 06:48:42

by Mike Rapoport

Subject: Re: [PATCH 08/15] mips: switch to generic version of pte allocation

On Thu, May 02, 2019 at 07:09:47PM +0000, Paul Burton wrote:
> Hi Mike,
>
> On Thu, May 02, 2019 at 06:28:35PM +0300, Mike Rapoport wrote:
> > MIPS allocates kernel PTE pages with
> >
> > __get_free_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)
> >
> > and user PTE pages with
> >
> > alloc_pages(GFP_KERNEL | __GFP_ZERO, PTE_ORDER)
>
> That bit isn't quite true - we don't use __GFP_ZERO in pte_alloc_one() &
> instead call clear_highpage() on the allocated page. Not that I have a
> problem with using __GFP_ZERO - it seems like the more optimal choice.
> It just might be worth mentioning the change & expected equivalent
> behavior.

You are right, I'll fix the changelog.

> Otherwise:
>
> Acked-by: Paul Burton <[email protected]>

Thanks.

> Thanks,
> Paul
>

--
Sincerely yours,
Mike.