2019-04-24 08:13:48

by sh liu

[permalink] [raw]
Subject: [PATCH] RISC-V: redefine PTRS_PER_PGD/PTRS_PER_PMD/PTRS_PER_PTE

Use the number of addresses to define the relevant macros.

Signed-off-by: damon <[email protected]>
---
arch/riscv/include/asm/pgtable-32.h | 1 +
arch/riscv/include/asm/pgtable-64.h | 3 ++-
arch/riscv/include/asm/pgtable.h | 8 ++++++--
3 files changed, 9 insertions(+), 3 deletions(-)

diff --git a/arch/riscv/include/asm/pgtable-32.h b/arch/riscv/include/asm/pgtable-32.h
index d61974b7..c9c8ca8 100644
--- a/arch/riscv/include/asm/pgtable-32.h
+++ b/arch/riscv/include/asm/pgtable-32.h
@@ -17,6 +17,7 @@
#include <asm-generic/pgtable-nopmd.h>
#include <linux/const.h>

+#define MAX_USER_VA_BITS 32
/* Size of region mapped by a page global directory */
#define PGDIR_SHIFT 22
#define PGDIR_SIZE (_AC(1, UL) << PGDIR_SHIFT)
diff --git a/arch/riscv/include/asm/pgtable-64.h b/arch/riscv/include/asm/pgtable-64.h
index 7aa0ea9..a56d4d0 100644
--- a/arch/riscv/include/asm/pgtable-64.h
+++ b/arch/riscv/include/asm/pgtable-64.h
@@ -16,6 +16,7 @@

#include <linux/const.h>

+#define MAX_USER_VA_BITS 39
#define PGDIR_SHIFT 30
/* Size of region mapped by a page global directory */
#define PGDIR_SIZE (_AC(1, UL) << PGDIR_SHIFT)
@@ -34,7 +35,7 @@
#define pmd_val(x) ((x).pmd)
#define __pmd(x) ((pmd_t) { (x) })

-#define PTRS_PER_PMD (PAGE_SIZE / sizeof(pmd_t))
+#define PTRS_PER_PMD (1 << (PGDIR_SHIFT - PMD_SHIFT))

static inline int pud_present(pud_t pud)
{
diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
index 1141364..9148043 100644
--- a/arch/riscv/include/asm/pgtable.h
+++ b/arch/riscv/include/asm/pgtable.h
@@ -33,9 +33,13 @@
#endif /* CONFIG_64BIT */

/* Number of entries in the page global directory */
-#define PTRS_PER_PGD (PAGE_SIZE / sizeof(pgd_t))
+#define PTRS_PER_PGD (1 << (MAX_USER_VA_BITS - PGDIR_SHIFT))
/* Number of entries in the page table */
-#define PTRS_PER_PTE (PAGE_SIZE / sizeof(pte_t))
+#ifdef __PAGETABLE_PMD_FOLDED
+#define PTRS_PER_PTE (1 << (PGDIR_SHIFT - PAGE_SHIFT))
+#else
+#define PTRS_PER_PTE (1 << (PMD_SHIFT - PAGE_SHIFT))
+#endif

/* Number of PGD entries that a user-mode program can use */
#define USER_PTRS_PER_PGD (TASK_SIZE / PGDIR_SIZE)
--
1.9.1


2019-04-24 08:16:13

by Anup Patel

[permalink] [raw]
Subject: Re: [PATCH] RISC-V: redefine PTRS_PER_PGD/PTRS_PER_PMD/PTRS_PER_PTE

On Wed, Apr 24, 2019 at 1:42 PM damon <[email protected]> wrote:
>
> Use the number of addresses to define the relevant macros.
>
> Signed-off-by: damon <[email protected]>
> ---
> arch/riscv/include/asm/pgtable-32.h | 1 +
> arch/riscv/include/asm/pgtable-64.h | 3 ++-
> arch/riscv/include/asm/pgtable.h | 8 ++++++--
> 3 files changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/arch/riscv/include/asm/pgtable-32.h b/arch/riscv/include/asm/pgtable-32.h
> index d61974b7..c9c8ca8 100644
> --- a/arch/riscv/include/asm/pgtable-32.h
> +++ b/arch/riscv/include/asm/pgtable-32.h
> @@ -17,6 +17,7 @@
> #include <asm-generic/pgtable-nopmd.h>
> #include <linux/const.h>
>
> +#define MAX_USER_VA_BITS 32
> /* Size of region mapped by a page global directory */
> #define PGDIR_SHIFT 22
> #define PGDIR_SIZE (_AC(1, UL) << PGDIR_SHIFT)
> diff --git a/arch/riscv/include/asm/pgtable-64.h b/arch/riscv/include/asm/pgtable-64.h
> index 7aa0ea9..a56d4d0 100644
> --- a/arch/riscv/include/asm/pgtable-64.h
> +++ b/arch/riscv/include/asm/pgtable-64.h
> @@ -16,6 +16,7 @@
>
> #include <linux/const.h>
>
> +#define MAX_USER_VA_BITS 39
> #define PGDIR_SHIFT 30
> /* Size of region mapped by a page global directory */
> #define PGDIR_SIZE (_AC(1, UL) << PGDIR_SHIFT)
> @@ -34,7 +35,7 @@
> #define pmd_val(x) ((x).pmd)
> #define __pmd(x) ((pmd_t) { (x) })
>
> -#define PTRS_PER_PMD (PAGE_SIZE / sizeof(pmd_t))
> +#define PTRS_PER_PMD (1 << (PGDIR_SHIFT - PMD_SHIFT))
>
> static inline int pud_present(pud_t pud)
> {
> diff --git a/arch/riscv/include/asm/pgtable.h b/arch/riscv/include/asm/pgtable.h
> index 1141364..9148043 100644
> --- a/arch/riscv/include/asm/pgtable.h
> +++ b/arch/riscv/include/asm/pgtable.h
> @@ -33,9 +33,13 @@
> #endif /* CONFIG_64BIT */
>
> /* Number of entries in the page global directory */
> -#define PTRS_PER_PGD (PAGE_SIZE / sizeof(pgd_t))
> +#define PTRS_PER_PGD (1 << (MAX_USER_VA_BITS - PGDIR_SHIFT))
> /* Number of entries in the page table */
> -#define PTRS_PER_PTE (PAGE_SIZE / sizeof(pte_t))
> +#ifdef __PAGETABLE_PMD_FOLDED
> +#define PTRS_PER_PTE (1 << (PGDIR_SHIFT - PAGE_SHIFT))
> +#else
> +#define PTRS_PER_PTE (1 << (PMD_SHIFT - PAGE_SHIFT))
> +#endif
>
> /* Number of PGD entries that a user-mode program can use */
> #define USER_PTRS_PER_PGD (TASK_SIZE / PGDIR_SIZE)
> --
> 1.9.1
>

LGTM.

Reviewed-by: Anup Patel <[email protected]>

Regards,
Anup

2019-06-28 18:53:26

by Paul Walmsley

[permalink] [raw]
Subject: Re: [PATCH] RISC-V: redefine PTRS_PER_PGD/PTRS_PER_PMD/PTRS_PER_PTE

On Wed, 24 Apr 2019, damon wrote:

> Use the number of addresses to define the relevant macros.
>
> Signed-off-by: damon <[email protected]>

This patch looks reasonable to me. But what's missing from the
description is the motivation. Is this a prerequisite for another patch
that you're planning to post? Or because you think this is clearer than
the original? Or something else? etc.


- Paul

2019-07-24 15:50:46

by sh liu

[permalink] [raw]
Subject: Re: [PATCH] RISC-V: redefine PTRS_PER_PGD/PTRS_PER_PMD/PTRS_PER_PTE

I think the previous description is unclear, and it is difficult for
readers to understand the meaning of these macros, because I never
understand. So I submitted this patch with reference to the definition
of arm. I think this way can make the reader easier to understand, and
I also think that this definition is more reasonable.

2019-06-29 2:52 GMT+08:00, Paul Walmsley <[email protected]>:
> On Wed, 24 Apr 2019, damon wrote:
>
>> Use the number of addresses to define the relevant macros.
>>
>> Signed-off-by: damon <[email protected]>
>
> This patch looks reasonable to me. But what's missing from the
> description is the motivation. Is this a prerequisite for another patch
> that you're planning to post? Or because you think this is clearer than
> the original? Or something else? etc.
>
>
> - Paul
>