sh: Fix up extended mode TLB for SH-X2+ cores.

The extended mode TLB requires both 64-bit PTEs and a 64-bit pgprot,
correspondingly, the PGD also has to be 64-bits, so fix that up.

The kernel and user permission bits really are decoupled in early
cuts of the silicon, which means that we also have to set corresponding
kernel permissions on user pages or we end up with user pages that the
kernel simply can't touch (!).

Finally, with those things corrected, really enable MMUCR.ME and
correct the PTEA value (this simply needs to be the upper 32-bits
of the PTE, with the size and protection bit encoding).

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

4 files changed, 88 insertions, 72 deletions

diff --git a/arch/sh/mm/tlb-sh4.c b/arch/sh/mm/tlb-sh4.c
index 13fde8cc7179..2d1dd6044307 100644
--- a/arch/sh/mm/tlb-sh4.c
+++ b/arch/sh/mm/tlb-sh4.c
@@ -4,27 +4,14 @@
  * SH-4 specific TLB operations
  *
  * Copyright (C) 1999  Niibe Yutaka
- * Copyright (C) 2002  Paul Mundt
+ * Copyright (C) 2002 - 2007 Paul Mundt
  *
  * Released under the terms of the GNU GPL v2.0.
  */
-#include <linux/signal.h>
-#include <linux/sched.h>
 #include <linux/kernel.h>
-#include <linux/errno.h>
-#include <linux/string.h>
-#include <linux/types.h>
-#include <linux/ptrace.h>
-#include <linux/mman.h>
 #include <linux/mm.h>
-#include <linux/smp.h>
-#include <linux/smp_lock.h>
-#include <linux/interrupt.h>
-
+#include <linux/io.h>
 #include <asm/system.h>
-#include <asm/io.h>
-#include <asm/uaccess.h>
-#include <asm/pgalloc.h>
 #include <asm/mmu_context.h>
 #include <asm/cacheflush.h>
 
@@ -62,12 +49,22 @@ void update_mmu_cache(struct vm_area_struct * vma,
 	vpn = (address & MMU_VPN_MASK) | get_asid();
 	ctrl_outl(vpn, MMU_PTEH);
 
-	pteval = pte_val(pte);
+	pteval = pte.pte_low;
 
 	/* Set PTEA register */
+#ifdef CONFIG_X2TLB
+	/*
+	 * For the extended mode TLB this is trivial, only the ESZ and
+	 * EPR bits need to be written out to PTEA, with the remainder of
+	 * the protection bits (with the exception of the compat-mode SZ
+	 * and PR bits, which are cleared) being written out in PTEL.
+	 */
+	ctrl_outl(pte.pte_high, MMU_PTEA);
+#else
 	if (cpu_data->flags & CPU_HAS_PTEA)
 		/* TODO: make this look less hacky */
 		ctrl_outl(((pteval >> 28) & 0xe) | (pteval & 0x1), MMU_PTEA);
+#endif
 
 	/* Set PTEL register */
 	pteval &= _PAGE_FLAGS_HARDWARE_MASK; /* drop software flags */
@@ -98,4 +95,3 @@ void local_flush_tlb_one(unsigned long asid, unsigned long page)
 	ctrl_outl(data, addr);
 	back_to_P1();
 }
-
diff --git a/include/asm-sh/cpu-sh4/mmu_context.h b/include/asm-sh/cpu-sh4/mmu_context.h
index ff4c5fbbfaf0..979acddc0f8e 100644
--- a/include/asm-sh/cpu-sh4/mmu_context.h
+++ b/include/asm-sh/cpu-sh4/mmu_context.h
@@ -22,13 +22,21 @@
 #define MMU_UTLB_ADDRESS_ARRAY	0xF6000000
 #define MMU_PAGE_ASSOC_BIT	0x80
 
-#define MMU_NTLB_ENTRIES	64	/* for 7750 */
+#ifdef CONFIG_X2TLB
+#define MMUCR_ME		(1 << 7)
+#else
+#define MMUCR_ME		(0)
+#endif
+
 #ifdef CONFIG_SH_STORE_QUEUES
-#define MMU_CONTROL_INIT	0x05	/* SQMD=0, SV=0, TI=1, AT=1 */
+#define MMUCR_SQMD		(1 << 9)
 #else
-#define MMU_CONTROL_INIT	0x205	/* SQMD=1, SV=0, TI=1, AT=1 */
+#define MMUCR_SQMD		(0)
 #endif
 
+#define MMU_NTLB_ENTRIES	64
+#define MMU_CONTROL_INIT	(0x05|MMUCR_SQMD|MMUCR_ME)
+
 #define MMU_ITLB_DATA_ARRAY	0xF3000000
 #define MMU_UTLB_DATA_ARRAY	0xF7000000
 
diff --git a/include/asm-sh/page.h b/include/asm-sh/page.h
index 48b718e7455c..cb3d46c59eab 100644
--- a/include/asm-sh/page.h
+++ b/include/asm-sh/page.h
@@ -88,6 +88,7 @@ extern void __copy_user_page(void *to, void *from, void *orig_to);
 #ifdef CONFIG_X2TLB
 typedef struct { unsigned long pte_low, pte_high; } pte_t;
 typedef struct { unsigned long long pgprot; } pgprot_t;
+typedef struct { unsigned long long pgd; } pgd_t;
 #define pte_val(x) \
 	((x).pte_low | ((unsigned long long)(x).pte_high << 32))
 #define __pte(x) \
@@ -95,12 +96,11 @@ typedef struct { unsigned long long pgprot; } pgprot_t;
 #else
 typedef struct { unsigned long pte_low; } pte_t;
 typedef struct { unsigned long pgprot; } pgprot_t;
+typedef struct { unsigned long pgd; } pgd_t;
 #define pte_val(x)	((x).pte_low)
 #define __pte(x) ((pte_t) { (x) } )
 #endif
 
-typedef struct { unsigned long pgd; } pgd_t;
-
 #define pgd_val(x)	((x).pgd)
 #define pgprot_val(x)	((x).pgprot)
 
diff --git a/include/asm-sh/pgtable.h b/include/asm-sh/pgtable.h
index 54ad5037fe40..cf0dd2b648c2 100644
--- a/include/asm-sh/pgtable.h
+++ b/include/asm-sh/pgtable.h
@@ -42,13 +42,12 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 
 /* PGD bits */
 #define PGDIR_SHIFT	(PTE_SHIFT + PTE_BITS)
-#define PGDIR_BITS	(32 - PGDIR_SHIFT)
 #define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
 #define PGDIR_MASK	(~(PGDIR_SIZE-1))
 
 /* Entries per level */
 #define PTRS_PER_PTE	(PAGE_SIZE / (1 << PTE_MAGNITUDE))
-#define PTRS_PER_PGD	(PAGE_SIZE / 4)
+#define PTRS_PER_PGD	(PAGE_SIZE / sizeof(pgd_t))
 
 #define USER_PTRS_PER_PGD	(TASK_SIZE/PGDIR_SIZE)
 #define FIRST_USER_ADDRESS	0
@@ -100,17 +99,18 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #define _PAGE_HW_SHARED	0x002		/* SH-bit  : shared among processes */
 #define _PAGE_DIRTY	0x004		/* D-bit   : page changed */
 #define _PAGE_CACHABLE	0x008		/* C-bit   : cachable */
-#ifndef CONFIG_X2TLB
-# define _PAGE_SZ0	0x010		/* SZ0-bit : Size of page */
-# define _PAGE_RW	0x020		/* PR0-bit : write access allowed */
-# define _PAGE_USER	0x040		/* PR1-bit : user space access allowed*/
-# define _PAGE_SZ1	0x080		/* SZ1-bit : Size of page (on SH-4) */
-#endif
+#define _PAGE_SZ0	0x010		/* SZ0-bit : Size of page */
+#define _PAGE_RW	0x020		/* PR0-bit : write access allowed */
+#define _PAGE_USER	0x040		/* PR1-bit : user space access allowed*/
+#define _PAGE_SZ1	0x080		/* SZ1-bit : Size of page (on SH-4) */
 #define _PAGE_PRESENT	0x100		/* V-bit   : page is valid */
 #define _PAGE_PROTNONE	0x200		/* software: if not present  */
 #define _PAGE_ACCESSED	0x400		/* software: page referenced */
 #define _PAGE_FILE	_PAGE_WT	/* software: pagecache or swap? */
 
+#define _PAGE_SZ_MASK	(_PAGE_SZ0 | _PAGE_SZ1)
+#define _PAGE_PR_MASK	(_PAGE_RW | _PAGE_USER)
+
 /* Extended mode bits */
 #define _PAGE_EXT_ESZ0		0x0010	/* ESZ0-bit: Size of page */
 #define _PAGE_EXT_ESZ1		0x0020	/* ESZ1-bit: Size of page */
@@ -126,11 +126,7 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #define _PAGE_EXT_KERN_READ	0x2000	/* EPR5-bit: Kernel space readable */
 
 /* Wrapper for extended mode pgprot twiddling */
-#ifdef CONFIG_X2TLB
-# define _PAGE_EXT(x)		((unsigned long long)(x) << 32)
-#else
-# define _PAGE_EXT(x)		(0)
-#endif
+#define _PAGE_EXT(x)		((unsigned long long)(x) << 32)
 
 /* software: moves to PTEA.TC (Timing Control) */
 #define _PAGE_PCC_AREA5	0x00000000	/* use BSC registers for area5 */
@@ -146,10 +142,14 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 #define _PAGE_PCC_ATR16	0x60000001	/* Attribute Memory space, 6 bit bus */
 
 /* Mask which drops unused bits from the PTEL value */
-#ifdef CONFIG_CPU_SH3
+#if defined(CONFIG_CPU_SH3)
 #define _PAGE_CLEAR_FLAGS	(_PAGE_PROTNONE | _PAGE_ACCESSED| \
 				 _PAGE_FILE	| _PAGE_SZ1	| \
 				 _PAGE_HW_SHARED)
+#elif defined(CONFIG_X2TLB)
+/* Get rid of the legacy PR/SZ bits when using extended mode */
+#define _PAGE_CLEAR_FLAGS	(_PAGE_PROTNONE | _PAGE_ACCESSED | \
+				 _PAGE_FILE | _PAGE_PR_MASK | _PAGE_SZ_MASK)
 #else
 #define _PAGE_CLEAR_FLAGS	(_PAGE_PROTNONE | _PAGE_ACCESSED | _PAGE_FILE)
 #endif
@@ -212,27 +212,36 @@ extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
 
 #define PAGE_SHARED	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
-				 _PAGE_EXT(_PAGE_EXT_USER_READ | \
+				 _PAGE_EXT(_PAGE_EXT_KERN_READ  | \
+					   _PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_USER_READ  | \
 					   _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(_PAGE_EXT_KERN_EXEC | \
+					   _PAGE_EXT_KERN_READ | \
+					   _PAGE_EXT_USER_EXEC | \
 					   _PAGE_EXT_USER_READ))
 
 #define PAGE_COPY	PAGE_EXECREAD
 
 #define PAGE_READONLY	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
-				 _PAGE_EXT(_PAGE_EXT_USER_READ))
+				 _PAGE_EXT(_PAGE_EXT_KERN_READ | \
+					   _PAGE_EXT_USER_READ))
 
 #define PAGE_WRITEONLY	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
-				 _PAGE_EXT(_PAGE_EXT_USER_WRITE))
+				 _PAGE_EXT(_PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_USER_WRITE))
 
 #define PAGE_RWX	__pgprot(_PAGE_PRESENT | _PAGE_ACCESSED | \
 				 _PAGE_CACHABLE | _PAGE_FLAGS_HARD | \
-				 _PAGE_EXT(_PAGE_EXT_USER_WRITE | \
+				 _PAGE_EXT(_PAGE_EXT_KERN_WRITE | \
+					   _PAGE_EXT_KERN_READ  | \
+					   _PAGE_EXT_KERN_EXEC  | \
+					   _PAGE_EXT_USER_WRITE | \
 					   _PAGE_EXT_USER_READ  | \
 					   _PAGE_EXT_USER_EXEC))
 
@@ -373,11 +382,15 @@ static inline void set_pte(pte_t *ptep, pte_t pte)
 #define set_pmd(pmdptr, pmdval) (*(pmdptr) = pmdval)
 
 #define pte_pfn(x)		((unsigned long)(((x).pte_low >> PAGE_SHIFT)))
-#define pfn_pte(pfn, prot)	__pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
-#define pfn_pmd(pfn, prot)	__pmd(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
 
-#define pte_none(x)	(!pte_val(x))
-#define pte_present(x)	(pte_val(x) & (_PAGE_PRESENT | _PAGE_PROTNONE))
+#define pfn_pte(pfn, prot) \
+	__pte(((unsigned long long)(pfn) << PAGE_SHIFT) | pgprot_val(prot))
+#define pfn_pmd(pfn, prot) \
+	__pmd(((unsigned long long)(pfn) << PAGE_SHIFT) | pgprot_val(prot))
+
+#define pte_none(x)		(!pte_val(x))
+#define pte_present(x)		((x).pte_low & (_PAGE_PRESENT | _PAGE_PROTNONE))
+
 #define pte_clear(mm,addr,xp) do { set_pte_at(mm, addr, xp, __pte(0)); } while (0)
 
 #define pmd_none(x)	(!pmd_val(x))
@@ -392,15 +405,15 @@ static inline void set_pte(pte_t *ptep, pte_t pte)
  * The following only work if pte_present() is true.
  * Undefined behaviour if not..
  */
-#define pte_not_present(pte)	(!(pte_val(pte) & _PAGE_PRESENT))
-#define pte_dirty(pte)		(pte_val(pte) & _PAGE_DIRTY)
-#define pte_young(pte)		(pte_val(pte) & _PAGE_ACCESSED)
-#define pte_file(pte)		(pte_val(pte) & _PAGE_FILE)
+#define pte_not_present(pte)	(!((pte).pte_low & _PAGE_PRESENT))
+#define pte_dirty(pte)		((pte).pte_low & _PAGE_DIRTY)
+#define pte_young(pte)		((pte).pte_low & _PAGE_ACCESSED)
+#define pte_file(pte)		((pte).pte_low & _PAGE_FILE)
 
 #ifdef CONFIG_X2TLB
 #define pte_write(pte)		((pte).pte_high & _PAGE_EXT_USER_WRITE)
 #else
-#define pte_write(pte)		(pte_val(pte) & _PAGE_RW)
+#define pte_write(pte)		((pte).pte_low & _PAGE_RW)
 #endif
 
 #define PTE_BIT_FUNC(h,fn,op) \
@@ -429,17 +442,10 @@ PTE_BIT_FUNC(low, mkyoung, |= _PAGE_ACCESSED);
 /*
  * Macro and implementation to make a page protection as uncachable.
  */
-#define pgprot_noncached pgprot_noncached
+#define pgprot_writecombine(prot) \
+	__pgprot(pgprot_val(prot) & ~_PAGE_CACHABLE)
 
-static inline pgprot_t pgprot_noncached(pgprot_t _prot)
-{
-	unsigned long prot = pgprot_val(_prot);
-
-	prot &= ~_PAGE_CACHABLE;
-	return __pgprot(prot);
-}
-
-#define pgprot_writecombine(prot) __pgprot(pgprot_val(prot) & ~_PAGE_CACHABLE)
+#define pgprot_noncached	 pgprot_writecombine
 
 /*
  * Conversion functions: convert a page and protection to a page entry,
@@ -451,28 +457,33 @@ static inline pgprot_t pgprot_noncached(pgprot_t _prot)
 
 static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
 {
-	set_pte(&pte, __pte((pte_val(pte) & _PAGE_CHG_MASK) |
-			    pgprot_val(newprot)));
+	pte.pte_low &= _PAGE_CHG_MASK;
+	pte.pte_low |= pgprot_val(newprot);
+
+#ifdef CONFIG_X2TLB
+	pte.pte_high |= pgprot_val(newprot) >> 32;
+#endif
+
 	return pte;
 }
 
-#define pmd_page_vaddr(pmd)	pmd_val(pmd)
+#define pmd_page_vaddr(pmd)	((unsigned long)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))
-#define pgd_offset(mm, address) ((mm)->pgd+pgd_index(address))
+#define pgd_index(address)	(((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
+#define pgd_offset(mm, address)	((mm)->pgd+pgd_index(address))
 
 /* to find an entry in a kernel page-table-directory */
-#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+#define pgd_offset_k(address)	pgd_offset(&init_mm, address)
 
 /* Find an entry in the third-level page table.. */
-#define pte_index(address) \
-		((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
+#define pte_index(address)	((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
 #define pte_offset_kernel(dir, address) \
 	((pte_t *) pmd_page_vaddr(*(dir)) + pte_index(address))
-#define pte_offset_map(dir, address) pte_offset_kernel(dir, address)
-#define pte_offset_map_nested(dir, address) pte_offset_kernel(dir, address)
+#define pte_offset_map(dir, address)		pte_offset_kernel(dir, address)
+#define pte_offset_map_nested(dir, address)	pte_offset_kernel(dir, address)
+
 #define pte_unmap(pte)		do { } while (0)
 #define pte_unmap_nested(pte)	do { } while (0)
 
@@ -480,13 +491,14 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
 #define pte_ERROR(e) \
 	printk("%s:%d: bad pte %p(%08lx%08lx).\n", __FILE__, __LINE__, \
 	       &(e), (e).pte_high, (e).pte_low)
+#define pgd_ERROR(e) \
+	printk("%s:%d: bad pgd %016llx.\n", __FILE__, __LINE__, pgd_val(e))
 #else
 #define pte_ERROR(e) \
 	printk("%s:%d: bad pte %08lx.\n", __FILE__, __LINE__, pte_val(e))
-#endif
-
 #define pgd_ERROR(e) \
 	printk("%s:%d: bad pgd %08lx.\n", __FILE__, __LINE__, pgd_val(e))
+#endif
 
 struct vm_area_struct;
 extern void update_mmu_cache(struct vm_area_struct * vma,