sh: pmd rework.

Remove extra bits from the pmd structure and store a kernel logical
address rather than a physical address. This allows it to be directly
dereferenced. Another piece of wierdness inherited from x86.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

arch/sh/mm/fault.c

@@ -46,6 +46,45 @@ asmlinkage void __kprobes do_page_fault(struct pt_regs *regs,
 	mm = tsk->mm;
 	si_code = SEGV_MAPERR;
 
+	if (unlikely(address >= TASK_SIZE)) {
+		/*
+		 * Synchronize this task's top level page-table
+		 * with the 'reference' page table.
+		 *
+		 * Do _not_ use "tsk" here. We might be inside
+		 * an interrupt in the middle of a task switch..
+		 */
+		int offset = pgd_index(address);
+		pgd_t *pgd, *pgd_k;
+		pud_t *pud, *pud_k;
+		pmd_t *pmd, *pmd_k;
+
+		pgd = get_TTB() + offset;
+		pgd_k = swapper_pg_dir + offset;
+
+		/* This will never happen with the folded page table. */
+		if (!pgd_present(*pgd)) {
+			if (!pgd_present(*pgd_k))
+				goto bad_area_nosemaphore;
+			set_pgd(pgd, *pgd_k);
+			return;
+		}
+
+		pud = pud_offset(pgd, address);
+		pud_k = pud_offset(pgd_k, address);
+		if (pud_present(*pud) || !pud_present(*pud_k))
+			goto bad_area_nosemaphore;
+		set_pud(pud, *pud_k);
+
+		pmd = pmd_offset(pud, address);
+		pmd_k = pmd_offset(pud_k, address);
+		if (pmd_present(*pmd) || !pmd_present(*pmd_k))
+			goto bad_area_nosemaphore;
+		set_pmd(pmd, *pmd_k);
+
+		return;
+	}
+
 	/*
 	 * If we're in an interrupt or have no user
 	 * context, we must not take the fault..
@@ -109,6 +148,7 @@ survive:
 bad_area:
 	up_read(&mm->mmap_sem);
 
+bad_area_nosemaphore:
 	if (user_mode(regs)) {
 		info.si_signo = SIGSEGV;
 		info.si_errno = 0;

arch/sh/mm/init.c

@@ -84,30 +84,22 @@ static void set_pte_phys(unsigned long addr, unsigned long phys, pgprot_t prot)
 	pmd_t *pmd;
 	pte_t *pte;
 
-	pgd = swapper_pg_dir + pgd_index(addr);
+	pgd = pgd_offset_k(addr);
 	if (pgd_none(*pgd)) {
 		pgd_ERROR(*pgd);
 		return;
 	}
 
-	pud = pud_offset(pgd, addr);
-	if (pud_none(*pud)) {
-		pmd = (pmd_t *)get_zeroed_page(GFP_ATOMIC);
-		set_pud(pud, __pud(__pa(pmd) | _PAGE_TABLE));
-		if (pmd != pmd_offset(pud, 0)) {
-			pud_ERROR(*pud);
-			return;
-		}
+	pud = pud_alloc(NULL, pgd, addr);
+	if (unlikely(!pud)) {
+		pud_ERROR(*pud);
+		return;
 	}
 
-	pmd = pmd_offset(pud, addr);
-	if (pmd_none(*pmd)) {
-		pte = (pte_t *)get_zeroed_page(GFP_ATOMIC);
-		set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE));
-		if (pte != pte_offset_kernel(pmd, 0)) {
-			pmd_ERROR(*pmd);
-			return;
-		}
+	pmd = pmd_alloc(NULL, pud, addr);
+	if (unlikely(!pmd)) {
+		pmd_ERROR(*pmd);
+		return;
 	}
 
 	pte = pte_offset_kernel(pmd, addr);

include/asm-sh/pgalloc.h

@@ -1,13 +1,16 @@
 #ifndef __ASM_SH_PGALLOC_H
 #define __ASM_SH_PGALLOC_H
 
-#define pmd_populate_kernel(mm, pmd, pte) \
-		set_pmd(pmd, __pmd(_PAGE_TABLE + __pa(pte)))
+static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd,
+				       pte_t *pte)
+{
+	set_pmd(pmd, __pmd((unsigned long)pte));
+}
 
 static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd,
 				struct page *pte)
 {
-	set_pmd(pmd, __pmd(_PAGE_TABLE + page_to_phys(pte)));
+	set_pmd(pmd, __pmd((unsigned long)page_address(pte)));
 }
 
 /*
@@ -15,7 +18,16 @@ static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd,
  */
 static inline pgd_t *pgd_alloc(struct mm_struct *mm)
 {
-	return (pgd_t *)__get_free_page(GFP_KERNEL | __GFP_REPEAT | __GFP_ZERO);
+	pgd_t *pgd = (pgd_t *)__get_free_page(GFP_KERNEL | __GFP_REPEAT);
+
+	if (pgd) {
+		memset(pgd, 0, USER_PTRS_PER_PGD * sizeof(pgd_t));
+		memcpy(pgd + USER_PTRS_PER_PGD,
+		       swapper_pg_dir + USER_PTRS_PER_PGD,
+		       (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
+	}
+
+	return pgd;
 }
 
 static inline void pgd_free(pgd_t *pgd)

include/asm-sh/pgtable.h

@@ -203,26 +203,18 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #ifndef __ASSEMBLY__
 
 #if defined(CONFIG_X2TLB) /* SH-X2 TLB */
-#define _PAGE_TABLE \
-	(_PAGE_PRESENT | _PAGE_ACCESSED | _PAGE_DIRTY | \
-	 _PAGE_EXT(_PAGE_EXT_USER_READ | _PAGE_EXT_USER_WRITE))
-
-#define _KERNPG_TABLE \
-	(_PAGE_PRESENT | _PAGE_ACCESSED | _PAGE_DIRTY | \
-	 _PAGE_EXT(_PAGE_EXT_KERN_READ | _PAGE_EXT_KERN_WRITE))
-
 #define PAGE_NONE	__pgprot(_PAGE_PROTNONE | _PAGE_CACHABLE | \
 				 _PAGE_ACCESSED | _PAGE_FLAGS_HARD)
 
 #define PAGE_SHARED	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_USER_READ | \
-				  	   _PAGE_EXT_USER_WRITE))
+					   _PAGE_EXT_USER_WRITE))
 
 #define PAGE_EXECREAD	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_USER_EXEC | \
-				 	   _PAGE_EXT_USER_READ))
+					   _PAGE_EXT_USER_READ))
 
 #define PAGE_COPY	PAGE_EXECREAD
 
@@ -237,14 +229,14 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #define PAGE_RWX	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_USER_WRITE | \
-				 	   _PAGE_EXT_USER_READ  | \
+					   _PAGE_EXT_USER_READ  | \
 					   _PAGE_EXT_USER_EXEC))
 
 #define PAGE_KERNEL	__pgprot(_PAGE_PRESENT | _PAGE_CACHABLE | \
 				 _PAGE_DIRTY | _PAGE_ACCESSED | \
 				 _PAGE_HW_SHARED | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_KERN_READ | \
-				 	   _PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_KERN_WRITE | \
 					   _PAGE_EXT_KERN_EXEC))
 
 #define PAGE_KERNEL_NOCACHE \
@@ -252,30 +244,25 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 				 _PAGE_ACCESSED | _PAGE_HW_SHARED | \
 				 _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_KERN_READ | \
-				 	   _PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_KERN_WRITE | \
 					   _PAGE_EXT_KERN_EXEC))
 
 #define PAGE_KERNEL_RO	__pgprot(_PAGE_PRESENT | _PAGE_CACHABLE | \
 				 _PAGE_DIRTY | _PAGE_ACCESSED | \
 				 _PAGE_HW_SHARED | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_KERN_READ | \
-				 	   _PAGE_EXT_KERN_EXEC))
+					   _PAGE_EXT_KERN_EXEC))
 
 #define PAGE_KERNEL_PCC(slot, type) \
 			__pgprot(_PAGE_PRESENT | _PAGE_DIRTY | \
 				 _PAGE_ACCESSED | _PAGE_FLAGS_HARD | \
 				 _PAGE_EXT(_PAGE_EXT_KERN_READ | \
-				 	   _PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_KERN_WRITE | \
 					   _PAGE_EXT_KERN_EXEC) \
 				 (slot ? _PAGE_PCC_AREA5 : _PAGE_PCC_AREA6) | \
 				 (type))
 
 #elif defined(CONFIG_MMU) /* SH-X TLB */
-#define _PAGE_TABLE \
-	(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED | _PAGE_DIRTY)
-#define _KERNPG_TABLE \
-	(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
-
 #define PAGE_NONE	__pgprot(_PAGE_PROTNONE | _PAGE_CACHABLE | \
 				 _PAGE_ACCESSED | _PAGE_FLAGS_HARD)
 
@@ -390,9 +377,9 @@ static inline void set_pte(pte_t *ptep, pte_t pte)
 #define pte_clear(mm,addr,xp) do { set_pte_at(mm, addr, xp, __pte(0)); } while (0)
 
 #define pmd_none(x)	(!pmd_val(x))
-#define pmd_present(x)	(pmd_val(x) & _PAGE_PRESENT)
+#define pmd_present(x)	(pmd_val(x))
 #define pmd_clear(xp)	do { set_pmd(xp, __pmd(0)); } while (0)
-#define	pmd_bad(x)	((pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE)
+#define	pmd_bad(x)	(pmd_val(x) & ~PAGE_MASK)
 
 #define pages_to_mb(x)	((x) >> (20-PAGE_SHIFT))
 #define pte_page(x)	phys_to_page(pte_val(x)&PTE_PHYS_MASK)
@@ -477,11 +464,8 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
 	return pte;
 }
 
-#define pmd_page_vaddr(pmd) \
-((unsigned long) __va(pmd_val(pmd) & PAGE_MASK))
-
-#define pmd_page(pmd) \
-	(phys_to_page(pmd_val(pmd)))
+#define pmd_page_vaddr(pmd)	pmd_val(pmd)
+#define pmd_page(pmd)		(virt_to_page(pmd_val(pmd)))
 
 /* to find an entry in a page-table-directory. */
 #define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))