MIPS: Add support for XPA.

Add support for extended physical addressing (XPA) so that
32-bit platforms can access equal to or greater than 40 bits
of physical addresses.

NOTE:
      1) XPA and EVA are not the same and cannot be used
         simultaneously.
      2) If you configure your kernel for XPA, the PTEs
         and all address sizes become 64-bit.
      3) Your platform MUST have working HIGHMEM support.

Signed-off-by: Steven J. Hill <Steven.Hill@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/9355/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

arch/mips/Kconfig

@@ -377,6 +377,7 @@ config MIPS_MALTA
 	select SYS_HAS_CPU_MIPS32_R1
 	select SYS_HAS_CPU_MIPS32_R2
 	select SYS_HAS_CPU_MIPS32_R3_5
+	select SYS_HAS_CPU_MIPS32_R5
 	select SYS_HAS_CPU_MIPS32_R6
 	select SYS_HAS_CPU_MIPS64_R1
 	select SYS_HAS_CPU_MIPS64_R2
@@ -386,6 +387,7 @@ config MIPS_MALTA
 	select SYS_SUPPORTS_32BIT_KERNEL
 	select SYS_SUPPORTS_64BIT_KERNEL
 	select SYS_SUPPORTS_BIG_ENDIAN
+	select SYS_SUPPORTS_HIGHMEM
 	select SYS_SUPPORTS_LITTLE_ENDIAN
 	select SYS_SUPPORTS_MICROMIPS
 	select SYS_SUPPORTS_MIPS_CMP
@@ -1596,6 +1598,33 @@ config CPU_MIPS32_3_5_EVA
 	  One of its primary benefits is an increase in the maximum size
 	  of lowmem (up to 3GB). If unsure, say 'N' here.
 
+config CPU_MIPS32_R5_FEATURES
+	bool "MIPS32 Release 5 Features"
+	depends on SYS_HAS_CPU_MIPS32_R5
+	depends on CPU_MIPS32_R2
+	help
+	  Choose this option to build a kernel for release 2 or later of the
+	  MIPS32 architecture including features from release 5 such as
+	  support for Extended Physical Addressing (XPA).
+
+config CPU_MIPS32_R5_XPA
+	bool "Extended Physical Addressing (XPA)"
+	depends on CPU_MIPS32_R5_FEATURES
+	depends on !EVA
+	depends on !PAGE_SIZE_4KB
+	depends on SYS_SUPPORTS_HIGHMEM
+	select XPA
+	select HIGHMEM
+	select ARCH_PHYS_ADDR_T_64BIT
+	default n
+	help
+	  Choose this option if you want to enable the Extended Physical
+	  Addressing (XPA) on your MIPS32 core (such as P5600 series). The
+	  benefit is to increase physical addressing equal to or greater
+	  than 40 bits. Note that this has the side effect of turning on
+	  64-bit addressing which in turn makes the PTEs 64-bit in size.
+	  If unsure, say 'N' here.
+
 if CPU_LOONGSON2F
 config CPU_NOP_WORKAROUNDS
 	bool
@@ -1699,6 +1728,9 @@ config SYS_HAS_CPU_MIPS32_R2
 config SYS_HAS_CPU_MIPS32_R3_5
 	bool
 
+config SYS_HAS_CPU_MIPS32_R5
+	bool
+
 config SYS_HAS_CPU_MIPS32_R6
 	bool
 
@@ -1836,6 +1868,9 @@ config CPU_MIPSR6
 config EVA
 	bool
 
+config XPA
+	bool
+
 config SYS_SUPPORTS_32BIT_KERNEL
 	bool
 config SYS_SUPPORTS_64BIT_KERNEL

arch/mips/include/asm/cpu-features.h

@@ -139,6 +139,9 @@
 # endif
 #endif
 
+#ifndef cpu_has_xpa
+#define cpu_has_xpa		(cpu_data[0].options & MIPS_CPU_XPA)
+#endif
 #ifndef cpu_has_vtag_icache
 #define cpu_has_vtag_icache	(cpu_data[0].icache.flags & MIPS_CACHE_VTAG)
 #endif

arch/mips/include/asm/cpu.h

@@ -377,6 +377,7 @@ enum cpu_type_enum {
 #define MIPS_CPU_MAAR		0x400000000ull /* MAAR(I) registers are present */
 #define MIPS_CPU_FRE		0x800000000ull /* FRE & UFE bits implemented */
 #define MIPS_CPU_RW_LLB		0x1000000000ull /* LLADDR/LLB writes are allowed */
+#define MIPS_CPU_XPA		0x2000000000ull /* CPU supports Extended Physical Addressing */
 
 /*
  * CPU ASE encodings

arch/mips/include/asm/pgtable-32.h

@@ -105,13 +105,16 @@ static inline void pmd_clear(pmd_t *pmdp)
 
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
 #define pte_page(x)		pfn_to_page(pte_pfn(x))
-#define pte_pfn(x)		((unsigned long)((x).pte_high >> 6))
+#define pte_pfn(x)		(((unsigned long)((x).pte_high >> _PFN_SHIFT)) | (unsigned long)((x).pte_low << _PAGE_PRESENT_SHIFT))
 static inline pte_t
 pfn_pte(unsigned long pfn, pgprot_t prot)
 {
 	pte_t pte;
-	pte.pte_high = (pfn << 6) | (pgprot_val(prot) & 0x3f);
-	pte.pte_low = pgprot_val(prot);
+
+	pte.pte_low = (pfn >> _PAGE_PRESENT_SHIFT) |
+				(pgprot_val(prot) & ~_PFNX_MASK);
+	pte.pte_high = (pfn << _PFN_SHIFT) |
+				(pgprot_val(prot) & ~_PFN_MASK);
 	return pte;
 }
 
@@ -166,9 +169,9 @@ pfn_pte(unsigned long pfn, pgprot_t prot)
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
 
 /* Swap entries must have VALID and GLOBAL bits cleared. */
-#define __swp_type(x)			(((x).val >> 2) & 0x1f)
-#define __swp_offset(x)			 ((x).val >> 7)
-#define __swp_entry(type,offset)	((swp_entry_t)	{ ((type) << 2) | ((offset) << 7) })
+#define __swp_type(x)			(((x).val >> 4) & 0x1f)
+#define __swp_offset(x)			 ((x).val >> 9)
+#define __swp_entry(type,offset)	((swp_entry_t)  { ((type) << 4) | ((offset) << 9) })
 #define __pte_to_swp_entry(pte)		((swp_entry_t) { (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t) { 0, (x).val })
 

arch/mips/include/asm/pgtable-bits.h

@@ -37,7 +37,11 @@
 /*
  * The following bits are implemented by the TLB hardware
  */
-#define _PAGE_GLOBAL_SHIFT	0
+#define _PAGE_NO_EXEC_SHIFT	0
+#define _PAGE_NO_EXEC		(1 << _PAGE_NO_EXEC_SHIFT)
+#define _PAGE_NO_READ_SHIFT	(_PAGE_NO_EXEC_SHIFT + 1)
+#define _PAGE_NO_READ		(1 << _PAGE_NO_READ_SHIFT)
+#define _PAGE_GLOBAL_SHIFT	(_PAGE_NO_READ_SHIFT + 1)
 #define _PAGE_GLOBAL		(1 << _PAGE_GLOBAL_SHIFT)
 #define _PAGE_VALID_SHIFT	(_PAGE_GLOBAL_SHIFT + 1)
 #define _PAGE_VALID		(1 << _PAGE_VALID_SHIFT)
@@ -49,7 +53,7 @@
 /*
  * The following bits are implemented in software
  */
-#define _PAGE_PRESENT_SHIFT	(_CACHE_SHIFT + 3)
+#define _PAGE_PRESENT_SHIFT	(24)
 #define _PAGE_PRESENT		(1 << _PAGE_PRESENT_SHIFT)
 #define _PAGE_READ_SHIFT	(_PAGE_PRESENT_SHIFT + 1)
 #define _PAGE_READ		(1 << _PAGE_READ_SHIFT)
@@ -62,6 +66,11 @@
 
 #define _PFN_SHIFT		(PAGE_SHIFT - 12 + _CACHE_SHIFT + 3)
 
+/*
+ * Bits for extended EntryLo0/EntryLo1 registers
+ */
+#define _PFNX_MASK		0xffffff
+
 #elif defined(CONFIG_CPU_R3000) || defined(CONFIG_CPU_TX39XX)
 
 /*

arch/mips/include/asm/pgtable.h

@@ -133,7 +133,7 @@ extern void set_pte_at(struct mm_struct *mm, unsigned long addr, pte_t *ptep,
 
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
 
-#define pte_none(pte)		(!(((pte).pte_low | (pte).pte_high) & ~_PAGE_GLOBAL))
+#define pte_none(pte)		(!(((pte).pte_high) & ~_PAGE_GLOBAL))
 #define pte_present(pte)	((pte).pte_low & _PAGE_PRESENT)
 
 static inline void set_pte(pte_t *ptep, pte_t pte)
@@ -142,16 +142,14 @@ static inline void set_pte(pte_t *ptep, pte_t pte)
 	smp_wmb();
 	ptep->pte_low = pte.pte_low;
 
-	if (pte.pte_low & _PAGE_GLOBAL) {
+	if (pte.pte_high & _PAGE_GLOBAL) {
 		pte_t *buddy = ptep_buddy(ptep);
 		/*
 		 * Make sure the buddy is global too (if it's !none,
 		 * it better already be global)
 		 */
-		if (pte_none(*buddy)) {
-			buddy->pte_low	|= _PAGE_GLOBAL;
+		if (pte_none(*buddy))
 			buddy->pte_high |= _PAGE_GLOBAL;
-		}
 	}
 }
 
@@ -161,8 +159,8 @@ static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *pt
 
 	htw_stop();
 	/* Preserve global status for the pair */
-	if (ptep_buddy(ptep)->pte_low & _PAGE_GLOBAL)
-		null.pte_low = null.pte_high = _PAGE_GLOBAL;
+	if (ptep_buddy(ptep)->pte_high & _PAGE_GLOBAL)
+		null.pte_high = _PAGE_GLOBAL;
 
 	set_pte_at(mm, addr, ptep, null);
 	htw_start();
@@ -242,52 +240,46 @@ static inline int pte_young(pte_t pte)	{ return pte.pte_low & _PAGE_ACCESSED; }
 
 static inline pte_t pte_wrprotect(pte_t pte)
 {
-	pte.pte_low  &= ~(_PAGE_WRITE | _PAGE_SILENT_WRITE);
+	pte.pte_low  &= ~_PAGE_WRITE;
 	pte.pte_high &= ~_PAGE_SILENT_WRITE;
 	return pte;
 }
 
 static inline pte_t pte_mkclean(pte_t pte)
 {
-	pte.pte_low  &= ~(_PAGE_MODIFIED | _PAGE_SILENT_WRITE);
+	pte.pte_low  &= ~_PAGE_MODIFIED;
 	pte.pte_high &= ~_PAGE_SILENT_WRITE;
 	return pte;
 }
 
 static inline pte_t pte_mkold(pte_t pte)
 {
-	pte.pte_low  &= ~(_PAGE_ACCESSED | _PAGE_SILENT_READ);
+	pte.pte_low  &= ~_PAGE_ACCESSED;
 	pte.pte_high &= ~_PAGE_SILENT_READ;
 	return pte;
 }
 
 static inline pte_t pte_mkwrite(pte_t pte)
 {
 	pte.pte_low |= _PAGE_WRITE;
-	if (pte.pte_low & _PAGE_MODIFIED) {
-		pte.pte_low  |= _PAGE_SILENT_WRITE;
+	if (pte.pte_low & _PAGE_MODIFIED)
 		pte.pte_high |= _PAGE_SILENT_WRITE;
-	}
 	return pte;
 }
 
 static inline pte_t pte_mkdirty(pte_t pte)
 {
 	pte.pte_low |= _PAGE_MODIFIED;
-	if (pte.pte_low & _PAGE_WRITE) {
-		pte.pte_low  |= _PAGE_SILENT_WRITE;
+	if (pte.pte_low & _PAGE_WRITE)
 		pte.pte_high |= _PAGE_SILENT_WRITE;
-	}
 	return pte;
 }
 
 static inline pte_t pte_mkyoung(pte_t pte)
 {
 	pte.pte_low |= _PAGE_ACCESSED;
-	if (pte.pte_low & _PAGE_READ) {
-		pte.pte_low  |= _PAGE_SILENT_READ;
+	if (pte.pte_low & _PAGE_READ)
 		pte.pte_high |= _PAGE_SILENT_READ;
-	}
 	return pte;
 }
 #else
@@ -391,10 +383,10 @@ static inline pgprot_t pgprot_writecombine(pgprot_t _prot)
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
 static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
 {
-	pte.pte_low  &= _PAGE_CHG_MASK;
+	pte.pte_low  &= (_PAGE_MODIFIED | _PAGE_ACCESSED | _PFNX_MASK);
 	pte.pte_high &= (_PFN_MASK | _CACHE_MASK);
-	pte.pte_low  |= pgprot_val(newprot);
-	pte.pte_high |= pgprot_val(newprot) & ~(_PFN_MASK | _CACHE_MASK);
+	pte.pte_low  |= pgprot_val(newprot) & ~_PFNX_MASK;
+	pte.pte_high |= pgprot_val(newprot) & ~_PFN_MASK;
 	return pte;
 }
 #else

arch/mips/kernel/cpu-probe.c

@@ -516,6 +516,10 @@ static inline unsigned int decode_config5(struct cpuinfo_mips *c)
 		c->options |= MIPS_CPU_MAAR;
 	if (config5 & MIPS_CONF5_LLB)
 		c->options |= MIPS_CPU_RW_LLB;
+#ifdef CONFIG_XPA
+	if (config5 & MIPS_CONF5_MVH)
+		c->options |= MIPS_CPU_XPA;
+#endif
 
 	return config5 & MIPS_CONF_M;
 }

arch/mips/kernel/proc.c

@@ -120,6 +120,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
 	if (cpu_has_msa)	seq_printf(m, "%s", " msa");
 	if (cpu_has_eva)	seq_printf(m, "%s", " eva");
 	if (cpu_has_htw)	seq_printf(m, "%s", " htw");
+	if (cpu_has_xpa)	seq_printf(m, "%s", " xpa");
 	seq_printf(m, "\n");
 
 	if (cpu_has_mmips) {

arch/mips/mm/init.c

@@ -96,7 +96,7 @@ static void *__kmap_pgprot(struct page *page, unsigned long addr, pgprot_t prot)
 	vaddr = __fix_to_virt(FIX_CMAP_END - idx);
 	pte = mk_pte(page, prot);
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
-	entrylo = pte.pte_high;
+	entrylo = pte_to_entrylo(pte.pte_high);
 #else
 	entrylo = pte_to_entrylo(pte_val(pte));
 #endif
@@ -106,6 +106,11 @@ static void *__kmap_pgprot(struct page *page, unsigned long addr, pgprot_t prot)
 	write_c0_entryhi(vaddr & (PAGE_MASK << 1));
 	write_c0_entrylo0(entrylo);
 	write_c0_entrylo1(entrylo);
+#ifdef CONFIG_XPA
+	entrylo = (pte.pte_low & _PFNX_MASK);
+	writex_c0_entrylo0(entrylo);
+	writex_c0_entrylo1(entrylo);
+#endif
 	tlbidx = read_c0_wired();
 	write_c0_wired(tlbidx + 1);
 	write_c0_index(tlbidx);

arch/mips/mm/tlb-r4k.c

@@ -333,9 +333,17 @@ void __update_tlb(struct vm_area_struct * vma, unsigned long address, pte_t pte)
 		ptep = pte_offset_map(pmdp, address);
 
 #if defined(CONFIG_PHYS_ADDR_T_64BIT) && defined(CONFIG_CPU_MIPS32)
+#ifdef CONFIG_XPA
+		write_c0_entrylo0(pte_to_entrylo(ptep->pte_high));
+		writex_c0_entrylo0(ptep->pte_low & _PFNX_MASK);
+		ptep++;
+		write_c0_entrylo1(pte_to_entrylo(ptep->pte_high));
+		writex_c0_entrylo1(ptep->pte_low & _PFNX_MASK);
+#else
 		write_c0_entrylo0(ptep->pte_high);
 		ptep++;
 		write_c0_entrylo1(ptep->pte_high);
+#endif
 #else
 		write_c0_entrylo0(pte_to_entrylo(pte_val(*ptep++)));
 		write_c0_entrylo1(pte_to_entrylo(pte_val(*ptep)));
@@ -355,6 +363,9 @@ void __update_tlb(struct vm_area_struct * vma, unsigned long address, pte_t pte)
 void add_wired_entry(unsigned long entrylo0, unsigned long entrylo1,
 		     unsigned long entryhi, unsigned long pagemask)
 {
+#ifdef CONFIG_XPA
+	panic("Broken for XPA kernels");
+#else
 	unsigned long flags;
 	unsigned long wired;
 	unsigned long old_pagemask;
@@ -383,6 +394,7 @@ void add_wired_entry(unsigned long entrylo0, unsigned long entrylo1,
 	write_c0_pagemask(old_pagemask);
 	local_flush_tlb_all();
 	local_irq_restore(flags);
+#endif
 }
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE