aboutsummaryrefslogtreecommitdiffstats
path: root/include/asm-powerpc/pgalloc-64.h
diff options
context:
space:
mode:
authorMartin Schwidefsky <schwidefsky@de.ibm.com>2008-02-08 07:22:04 -0500
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2008-02-08 12:22:42 -0500
commit2f569afd9ced9ebec9a6eb3dbf6f83429be0a7b4 (patch)
tree23a31763887d9505e62e9d7cc8ec2fa4b86bd380 /include/asm-powerpc/pgalloc-64.h
parent13214adf738abc92b0a00c0763fd3be79eebaa7c (diff)
CONFIG_HIGHPTE vs. sub-page page tables.
Background: I've implemented 1K/2K page tables for s390. These sub-page page tables are required to properly support the s390 virtualization instruction with KVM. The SIE instruction requires that the page tables have 256 page table entries (pte) followed by 256 page status table entries (pgste). The pgstes are only required if the process is using the SIE instruction. The pgstes are updated by the hardware and by the hypervisor for a number of reasons, one of them is dirty and reference bit tracking. To avoid wasting memory the standard pte table allocation should return 1K/2K (31/64 bit) and 2K/4K if the process is using SIE. Problem: Page size on s390 is 4K, page table size is 1K or 2K. That means the s390 version for pte_alloc_one cannot return a pointer to a struct page. Trouble is that with the CONFIG_HIGHPTE feature on x86 pte_alloc_one cannot return a pointer to a pte either, since that would require more than 32 bit for the return value of pte_alloc_one (and the pte * would not be accessible since its not kmapped). Solution: The only solution I found to this dilemma is a new typedef: a pgtable_t. For s390 pgtable_t will be a (pte *) - to be introduced with a later patch. For everybody else it will be a (struct page *). The additional problem with the initialization of the ptl lock and the NR_PAGETABLE accounting is solved with a constructor pgtable_page_ctor and a destructor pgtable_page_dtor. The page table allocation and free functions need to call these two whenever a page table page is allocated or freed. pmd_populate will get a pgtable_t instead of a struct page pointer. To get the pgtable_t back from a pmd entry that has been installed with pmd_populate a new function pmd_pgtable is added. It replaces the pmd_page call in free_pte_range and apply_to_pte_range. Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: <linux-arch@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'include/asm-powerpc/pgalloc-64.h')
-rw-r--r--include/asm-powerpc/pgalloc-64.h27
1 files changed, 20 insertions, 7 deletions
diff --git a/include/asm-powerpc/pgalloc-64.h b/include/asm-powerpc/pgalloc-64.h
index 5afae8593931..68980990f62a 100644
--- a/include/asm-powerpc/pgalloc-64.h
+++ b/include/asm-powerpc/pgalloc-64.h
@@ -58,6 +58,7 @@ static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd)
58#define pmd_populate(mm, pmd, pte_page) \ 58#define pmd_populate(mm, pmd, pte_page) \
59 pmd_populate_kernel(mm, pmd, page_address(pte_page)) 59 pmd_populate_kernel(mm, pmd, page_address(pte_page))
60#define pmd_populate_kernel(mm, pmd, pte) pmd_set(pmd, (unsigned long)(pte)) 60#define pmd_populate_kernel(mm, pmd, pte) pmd_set(pmd, (unsigned long)(pte))
61#define pmd_pgtable(pmd) pmd_page(pmd)
61 62
62 63
63#else /* CONFIG_PPC_64K_PAGES */ 64#else /* CONFIG_PPC_64K_PAGES */
@@ -72,6 +73,7 @@ static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd,
72 73
73#define pmd_populate(mm, pmd, pte_page) \ 74#define pmd_populate(mm, pmd, pte_page) \
74 pmd_populate_kernel(mm, pmd, page_address(pte_page)) 75 pmd_populate_kernel(mm, pmd, page_address(pte_page))
76#define pmd_pgtable(pmd) pmd_page(pmd)
75 77
76#endif /* CONFIG_PPC_64K_PAGES */ 78#endif /* CONFIG_PPC_64K_PAGES */
77 79
@@ -92,11 +94,18 @@ static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
92 return (pte_t *)__get_free_page(GFP_KERNEL | __GFP_REPEAT | __GFP_ZERO); 94 return (pte_t *)__get_free_page(GFP_KERNEL | __GFP_REPEAT | __GFP_ZERO);
93} 95}
94 96
95static inline struct page *pte_alloc_one(struct mm_struct *mm, 97static inline pgtable_t pte_alloc_one(struct mm_struct *mm,
96 unsigned long address) 98 unsigned long address)
97{ 99{
98 pte_t *pte = pte_alloc_one_kernel(mm, address); 100 struct page *page;
99 return pte ? virt_to_page(pte) : NULL; 101 pte_t *pte;
102
103 pte = pte_alloc_one_kernel(mm, address);
104 if (!pte)
105 return NULL;
106 page = virt_to_page(pte);
107 pgtable_page_ctor(page);
108 return page;
100} 109}
101 110
102static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte) 111static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
@@ -104,8 +113,9 @@ static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
104 free_page((unsigned long)pte); 113 free_page((unsigned long)pte);
105} 114}
106 115
107static inline void pte_free(struct mm_struct *mm, struct page *ptepage) 116static inline void pte_free(struct mm_struct *mm, pgtable_t ptepage)
108{ 117{
118 pgtable_page_dtor(ptepage);
109 __free_page(ptepage); 119 __free_page(ptepage);
110} 120}
111 121
@@ -136,9 +146,12 @@ static inline void pgtable_free(pgtable_free_t pgf)
136 146
137extern void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf); 147extern void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf);
138 148
139#define __pte_free_tlb(tlb, ptepage) \ 149#define __pte_free_tlb(tlb,ptepage) \
150do { \
151 pgtable_page_dtor(ptepage); \
140 pgtable_free_tlb(tlb, pgtable_free_cache(page_address(ptepage), \ 152 pgtable_free_tlb(tlb, pgtable_free_cache(page_address(ptepage), \
141 PTE_NONCACHE_NUM, PTE_TABLE_SIZE-1)) 153 PTE_NONCACHE_NUM, PTE_TABLE_SIZE-1)); \
154} while (0)
142#define __pmd_free_tlb(tlb, pmd) \ 155#define __pmd_free_tlb(tlb, pmd) \
143 pgtable_free_tlb(tlb, pgtable_free_cache(pmd, \ 156 pgtable_free_tlb(tlb, pgtable_free_cache(pmd, \
144 PMD_CACHE_NUM, PMD_TABLE_SIZE-1)) 157 PMD_CACHE_NUM, PMD_TABLE_SIZE-1))