[PATCH 14/17] 4level support for sh64

From: Andreas Kleen
Date: Mon Oct 25 2004 - 02:59:47 EST

Next message: Andreas Kleen: "[PATCH 8/17] 4level support for m32r"
Previous message: Andreas Kleen: "[PATCH 10/17] 4level support for ppc"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

4level support for sh64

sh64 converted

Signed-off-by: Andi Kleen <ak@xxxxxxx>

diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/cache.c linux-2.6.10rc1-4level/arch/sh64/mm/cache.c
--- linux-2.6.10rc1/arch/sh64/mm/cache.c 2004-08-15 19:45:12.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/cache.c 2004-10-25 04:48:09.000000000 +0200
@@ -581,7 +581,7 @@ static void sh64_dcache_purge_virt_page(
pte_t *pte;
pte_t entry;

- pgd = pgd_offset(mm, eaddr);
+ pgd = pml4_pgd_offset(pml4_offset(mm, eaddr),eaddr);
pmd = pmd_offset(pgd, eaddr);

if (pmd_none(*pmd) || pmd_bad(*pmd))
@@ -610,7 +610,7 @@ static void sh64_dcache_purge_user_page(
following page table traversal is safe even on SMP/pre-emptible. */

if (!mm) return; /* No way to find physical address of page */
- pgd = pgd_offset(mm, eaddr);
+ pgd = pml4_pgd_offset(pml4_offset(mm, eaddr), eaddr);
if (pgd_bad(*pgd)) return;

pmd = pmd_offset(pgd, eaddr);
diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/fault.c linux-2.6.10rc1-4level/arch/sh64/mm/fault.c
--- linux-2.6.10rc1/arch/sh64/mm/fault.c 2004-10-19 01:55:06.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/fault.c 2004-10-25 04:48:09.000000000 +0200
@@ -92,7 +92,7 @@ static pte_t *lookup_pte(struct mm_struc
pte_t *pte;
pte_t entry;

- dir = pgd_offset(mm, address);
+ dir = pml4_pgd_offset(pml4_offset(mm, address), address);
if (pgd_none(*dir)) {
return NULL;
}
diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/hugetlbpage.c linux-2.6.10rc1-4level/arch/sh64/mm/hugetlbpage.c
--- linux-2.6.10rc1/arch/sh64/mm/hugetlbpage.c 2004-08-15 19:45:12.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/hugetlbpage.c 2004-10-25 04:48:09.000000000 +0200
@@ -30,7 +30,7 @@ static pte_t *huge_pte_alloc(struct mm_s
pmd_t *pmd;
pte_t *pte = NULL;

- pgd = pgd_offset(mm, addr);
+ pgd = pml4_pgd_offset(pml4_offset(mm, addr), addr);
if (pgd) {
pmd = pmd_alloc(mm, pgd, addr);
if (pmd)
@@ -45,7 +45,7 @@ static pte_t *huge_pte_offset(struct mm_
pmd_t *pmd;
pte_t *pte = NULL;

- pgd = pgd_offset(mm, addr);
+ pgd = pml4_pgd_offset(pml4_offset(mm, addr), addr);
if (pgd) {
pmd = pmd_offset(pgd, addr);
if (pmd)
diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/init.c linux-2.6.10rc1-4level/arch/sh64/mm/init.c
--- linux-2.6.10rc1/arch/sh64/mm/init.c 2004-10-19 01:55:06.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/init.c 2004-10-25 04:48:09.000000000 +0200
@@ -32,7 +32,7 @@ DEFINE_PER_CPU(struct mmu_gather, mmu_ga
* Cache of MMU context last used.
*/
unsigned long mmu_context_cache;
-pgd_t * mmu_pdtp_cache;
+pml4_t * mmu_pdtp_cache;
int after_bootmem = 0;

/*
@@ -197,3 +197,19 @@ void free_initrd_mem(unsigned long start
}
#endif

+pgd_t *__pgd_alloc(struct mm_struct *mm, pml4_t *pml4, unsigned long addr)
+{
+ unsigned long *ret;
+
+ if ((ret = pgd_quicklist) != NULL) {
+ pgd_quicklist = (unsigned long *)(*ret);
+ ret[0] = 0;
+ pgtable_cache_size--;
+ } else
+ ret = (unsigned long *)get_pgd_slow();
+
+ if (ret) {
+ memset(ret, 0, USER_PTRS_PER_PGD * sizeof(pgd_t));
+ }
+ return (pgd_t *)ret;
+}
diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/ioremap.c linux-2.6.10rc1-4level/arch/sh64/mm/ioremap.c
--- linux-2.6.10rc1/arch/sh64/mm/ioremap.c 2004-08-15 19:45:12.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/ioremap.c 2004-10-25 04:48:09.000000000 +0200
@@ -97,7 +97,7 @@ static int remap_area_pages(unsigned lon
unsigned long end = address + size;

phys_addr -= address;
- dir = pgd_offset_k(address);
+ dir = pml4_pgd_offset_k(pml4_offset_k(address), address);
flush_cache_all();
if (address >= end)
BUG();
@@ -355,7 +355,7 @@ static void shmedia_mapioaddr(unsigned l

pr_debug("shmedia_mapiopage pa %08lx va %08lx\n", pa, va);

- pgdp = pgd_offset_k(va);
+ pgdp = pml4_pgd_offset_k(pml4_offset_k(va), va);
if (pgd_none(*pgdp) || !pgd_present(*pgdp)) {
pmdp = (pmd_t *)sh64_get_page();
set_pgd(pgdp, __pgd((unsigned long)pmdp | _KERNPG_TABLE));
@@ -388,7 +388,7 @@ static void shmedia_unmapioaddr(unsigned
pmd_t *pmdp;
pte_t *ptep;

- pgdp = pgd_offset_k(vaddr);
+ pgdp = pml4_pgd_offset_k(pml4_offset_k(vaddr), vaddr);
pmdp = pmd_offset(pgdp, vaddr);

if (pmd_none(*pmdp) || pmd_bad(*pmdp))
diff -urpN -X ../KDIFX linux-2.6.10rc1/arch/sh64/mm/tlbmiss.c linux-2.6.10rc1-4level/arch/sh64/mm/tlbmiss.c
--- linux-2.6.10rc1/arch/sh64/mm/tlbmiss.c 2004-10-19 01:55:06.000000000 +0200
+++ linux-2.6.10rc1-4level/arch/sh64/mm/tlbmiss.c 2004-10-25 04:48:09.000000000 +0200
@@ -99,7 +99,7 @@ static int handle_vmalloc_fault(struct m
static pte_t *pte;
pte_t entry;

- dir = pgd_offset_k(address);
+ dir = pml4_pgd_offset_k(pml4_offset_k(address), address);
pmd = pmd_offset(dir, address);

if (pmd_none(*pmd)) {
@@ -150,7 +150,7 @@ static int handle_tlbmiss(struct mm_stru
/* upper half - never has page table entries. */
return 0;
}
- dir = pgd_offset(mm, address);
+ dir = pml4_pgd_offset(pml4_offset(mm, address), address);
if (pgd_none(*dir)) {
return 0;
}
diff -urpN -X ../KDIFX linux-2.6.10rc1/include/asm-sh64/mmu_context.h linux-2.6.10rc1-4level/include/asm-sh64/mmu_context.h
--- linux-2.6.10rc1/include/asm-sh64/mmu_context.h 2004-08-15 19:45:49.000000000 +0200
+++ linux-2.6.10rc1-4level/include/asm-sh64/mmu_context.h 2004-10-25 04:48:10.000000000 +0200
@@ -31,7 +31,7 @@ extern unsigned long mmu_context_cache;

/* Current mm's pgd */
-extern pgd_t *mmu_pdtp_cache;
+extern pml4_t *mmu_pdtp_cache;

#define SR_ASID_MASK 0xffffffffff00ffffULL
#define SR_ASID_SHIFT 16
@@ -189,7 +189,7 @@ static __inline__ void switch_mm(struct
struct task_struct *tsk)
{
if (prev != next) {
- mmu_pdtp_cache = next->pgd;
+ mmu_pdtp_cache = next->pml4;
activate_context(next);
}
}
diff -urpN -X ../KDIFX linux-2.6.10rc1/include/asm-sh64/page.h linux-2.6.10rc1-4level/include/asm-sh64/page.h
--- linux-2.6.10rc1/include/asm-sh64/page.h 2004-08-15 19:45:49.000000000 +0200
+++ linux-2.6.10rc1-4level/include/asm-sh64/page.h 2004-10-25 04:48:10.000000000 +0200
@@ -130,6 +130,8 @@ extern __inline__ int get_order(unsigned
return order;
}

+#include <asm-generic/nopml4-page.h>
+
#endif

#endif /* __KERNEL__ */
diff -urpN -X ../KDIFX linux-2.6.10rc1/include/asm-sh64/pgalloc.h linux-2.6.10rc1-4level/include/asm-sh64/pgalloc.h
--- linux-2.6.10rc1/include/asm-sh64/pgalloc.h 2004-08-15 19:45:49.000000000 +0200
+++ linux-2.6.10rc1-4level/include/asm-sh64/pgalloc.h 2004-10-25 04:48:10.000000000 +0200
@@ -46,23 +46,6 @@ extern __inline__ pgd_t *get_pgd_slow(vo
return ret;
}

-extern __inline__ pgd_t *get_pgd_fast(void)
-{
- unsigned long *ret;
-
- if ((ret = pgd_quicklist) != NULL) {
- pgd_quicklist = (unsigned long *)(*ret);
- ret[0] = 0;
- pgtable_cache_size--;
- } else
- ret = (unsigned long *)get_pgd_slow();
-
- if (ret) {
- memset(ret, 0, USER_PTRS_PER_PGD * sizeof(pgd_t));
- }
- return (pgd_t *)ret;
-}
-
extern __inline__ void free_pgd_fast(pgd_t *pgd)
{
*(unsigned long *)pgd = (unsigned long) pgd_quicklist;
@@ -170,7 +153,6 @@ static __inline__ void pmd_free(pmd_t *p

#define check_pgt_cache() do { } while (0)
#define pgd_free(pgd) free_pgd_slow(pgd)
-#define pgd_alloc(mm) get_pgd_fast()

extern int do_check_pgt_cache(int, int);

@@ -199,4 +181,6 @@ static inline void pmd_populate(struct m
set_pmd(pmd, __pmd(_PAGE_TABLE + (unsigned long) page_address (pte)));
}

+#include <asm-generic/nopml4-pgalloc.h>
+
#endif /* __ASM_SH64_PGALLOC_H */
diff -urpN -X ../KDIFX linux-2.6.10rc1/include/asm-sh64/pgtable.h linux-2.6.10rc1-4level/include/asm-sh64/pgtable.h
--- linux-2.6.10rc1/include/asm-sh64/pgtable.h 2004-10-25 04:47:37.000000000 +0200
+++ linux-2.6.10rc1-4level/include/asm-sh64/pgtable.h 2004-10-25 04:48:10.000000000 +0200
@@ -146,11 +146,8 @@ static __inline__ void pmd_set(pmd_t *pm

/* To find an entry in a generic PGD. */
#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
+#define pgd_index_k(address) pgd_index(address)
#define __pgd_offset(address) pgd_index(address)
-#define pgd_offset(mm, address) ((mm)->pgd+pgd_index(address))
-
-/* To find an entry in a kernel PGD. */
-#define pgd_offset_k(address) pgd_offset(&init_mm, address)

/*
* PGD level access routines.
@@ -234,7 +231,7 @@ static inline pmd_t * pmd_offset(pgd_t *
#define pte_unmap_nested(pte) do { } while (0)

/* Round it up ! */
-#define USER_PTRS_PER_PGD ((TASK_SIZE+PGDIR_SIZE-1)/PGDIR_SIZE)
+#define USER_PPGDS_IN_LAST_PML4 ((TASK_SIZE+PGDIR_SIZE-1)/PGDIR_SIZE)
#define FIRST_USER_PGD_NR 0

#ifndef __ASSEMBLY__
@@ -495,5 +492,6 @@ extern void update_mmu_cache(struct vm_a
extern pgd_t swapper_pg_dir[PTRS_PER_PGD];

#include <asm-generic/pgtable.h>
+#include <asm-generic/nopml4-pgtable.h>

#endif /* __ASM_SH64_PGTABLE_H */
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Andreas Kleen: "[PATCH 8/17] 4level support for m32r"
Previous message: Andreas Kleen: "[PATCH 10/17] 4level support for ppc"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]