Merge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus

Pull MIPS updates from Ralf Baechle:
 "This is the main pull request for MIPS for Linux 4.1.  Most
  noteworthy:

   - Add more Octeon-optimized crypto functions
   - Octeon crypto preemption and locking fixes
   - Little endian support for Octeon
   - Use correct CSR to soft reset Octeons
   - Support LEDs on the Octeon-based DSR-1000N
   - Fix PCI interrupt mapping for the Octeon-based DSR-1000N
   - Mark prom_free_prom_memory() as __init for a number of systems
   - Support for Imagination's Pistachio SOC.  This includes arch and
     CLK bits.  I'd like to merge pinctrl bits later
   - Improve parallelism of csum_partial for certain pipelines
   - Organize DTB files in subdirs like other architectures
   - Implement read_sched_clock for all MIPS platforms other than
     Octeon
   - Massive series of 38 fixes and cleanups for the FPU emulator /
     kernel
   - Further FPU remulator work to support new features.  This sits on a
     separate branch which also has been pulled into the 4.1 KVM branch
   - Clean up and fixes for the SEAD3 eval board; remove unused file
   - Various updates for Netlogic platforms
   - A number of small updates for Loongson 3 platforms
   - Increase the memory limit for ATH79 platforms to 256MB
   - A fair number of fixes and updates for BCM47xx platforms
   - Finish the implementation of XPA support
   - MIPS FDC support.  No, not floppy controller but Fast Debug Channel :)
   - Detect the R16000 used in SGI legacy platforms
   - Fix Kconfig dependencies for the SSB bus support"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (265 commits)
  MIPS: Makefile: Fix MIPS ASE detection code
  MIPS: asm: elf: Set O32 default FPU flags
  MIPS: BCM47XX: Fix detecting Microsoft MN-700 & Asus WL500G
  MIPS: Kconfig: Disable SMP/CPS for 64-bit
  MIPS: Hibernate: flush TLB entries earlier
  MIPS: smp-cps: cpu_set FPU mask if FPU present
  MIPS: lose_fpu(): Disable FPU when MSA enabled
  MIPS: ralink: add missing symbol for RALINK_ILL_ACC
  MIPS: ralink: Fix bad config symbol in PCI makefile.
  SSB: fix Kconfig dependencies
  MIPS: Malta: Detect and fix bad memsize values
  Revert "MIPS: Avoid pipeline stalls on some MIPS32R2 cores."
  MIPS: Octeon: Delete override of cpu_has_mips_r2_exec_hazard.
  MIPS: Fix cpu_has_mips_r2_exec_hazard.
  MIPS: kernel: entry.S: Set correct ISA level for mips_ihb
  MIPS: asm: spinlock: Fix addiu instruction for R10000_LLSC_WAR case
  MIPS: r4kcache: Use correct base register for MIPS R6 cache flushes
  MIPS: Kconfig: Fix typo for the r2-to-r6 emulator kernel parameter
  MIPS: unaligned: Fix regular load/store instruction emulation for EVA
  MIPS: unaligned: Surround load/store macros in do {} while statements
  ...

7 files changed, 148 insertions, 67 deletions

diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c
index 3f8059602765..0dbb65a51ce5 100644
--- a/arch/mips/mm/c-r4k.c
+++ b/arch/mips/mm/c-r4k.c
@@ -430,6 +430,7 @@ static inline void local_r4k___flush_cache_all(void * args)
 	case CPU_R10000:
 	case CPU_R12000:
 	case CPU_R14000:
+	case CPU_R16000:
 		/*
 		 * These caches are inclusive caches, that is, if something
 		 * is not cached in the S-cache, we know it also won't be
@@ -506,7 +507,7 @@ static inline void local_r4k_flush_cache_mm(void * args)
 
 	/*
 	 * Kludge alert.  For obscure reasons R4000SC and R4400SC go nuts if we
-	 * only flush the primary caches but R10000 and R12000 behave sane ...
+	 * only flush the primary caches but R1x000 behave sane ...
 	 * R4000SC and R4400SC indexed S-cache ops also invalidate primary
 	 * caches, so we can bail out early.
 	 */
@@ -888,33 +889,39 @@ static inline void rm7k_erratum31(void)
 	}
 }
 
-static inline void alias_74k_erratum(struct cpuinfo_mips *c)
+static inline int alias_74k_erratum(struct cpuinfo_mips *c)
 {
 	unsigned int imp = c->processor_id & PRID_IMP_MASK;
 	unsigned int rev = c->processor_id & PRID_REV_MASK;
+	int present = 0;
 
 	/*
 	 * Early versions of the 74K do not update the cache tags on a
 	 * vtag miss/ptag hit which can occur in the case of KSEG0/KUSEG
-	 * aliases. In this case it is better to treat the cache as always
-	 * having aliases.
+	 * aliases.  In this case it is better to treat the cache as always
+	 * having aliases.  Also disable the synonym tag update feature
+	 * where available.  In this case no opportunistic tag update will
+	 * happen where a load causes a virtual address miss but a physical
+	 * address hit during a D-cache look-up.
 	 */
 	switch (imp) {
 	case PRID_IMP_74K:
 		if (rev <= PRID_REV_ENCODE_332(2, 4, 0))
-			c->dcache.flags |= MIPS_CACHE_VTAG;
+			present = 1;
 		if (rev == PRID_REV_ENCODE_332(2, 4, 0))
 			write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
 		break;
 	case PRID_IMP_1074K:
 		if (rev <= PRID_REV_ENCODE_332(1, 1, 0)) {
-			c->dcache.flags |= MIPS_CACHE_VTAG;
+			present = 1;
 			write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
 		}
 		break;
 	default:
 		BUG();
 	}
+
+	return present;
 }
 
 static void b5k_instruction_hazard(void)
@@ -938,6 +945,7 @@ static void probe_pcache(void)
 	struct cpuinfo_mips *c = &current_cpu_data;
 	unsigned int config = read_c0_config();
 	unsigned int prid = read_c0_prid();
+	int has_74k_erratum = 0;
 	unsigned long config1;
 	unsigned int lsize;
 
@@ -1012,6 +1020,7 @@ static void probe_pcache(void)
 	case CPU_R10000:
 	case CPU_R12000:
 	case CPU_R14000:
+	case CPU_R16000:
 		icache_size = 1 << (12 + ((config & R10K_CONF_IC) >> 29));
 		c->icache.linesz = 64;
 		c->icache.ways = 2;
@@ -1223,8 +1232,8 @@ static void probe_pcache(void)
 		dcache_size / (c->dcache.linesz * c->dcache.ways) : 0;
 
 	/*
-	 * R10000 and R12000 P-caches are odd in a positive way.  They're 32kB
-	 * 2-way virtually indexed so normally would suffer from aliases.  So
+	 * R1x000 P-caches are odd in a positive way.  They're 32kB 2-way
+	 * virtually indexed so normally would suffer from aliases.  So
 	 * normally they'd suffer from aliases but magic in the hardware deals
 	 * with that for us so we don't need to take care ourselves.
 	 */
@@ -1240,11 +1249,12 @@ static void probe_pcache(void)
 	case CPU_R10000:
 	case CPU_R12000:
 	case CPU_R14000:
+	case CPU_R16000:
 		break;
 
 	case CPU_74K:
 	case CPU_1074K:
-		alias_74k_erratum(c);
+		has_74k_erratum = alias_74k_erratum(c);
 		/* Fall through. */
 	case CPU_M14KC:
 	case CPU_M14KEC:
@@ -1259,7 +1269,7 @@ static void probe_pcache(void)
 		if (!(read_c0_config7() & MIPS_CONF7_IAR) &&
 		    (c->icache.waysize > PAGE_SIZE))
 			c->icache.flags |= MIPS_CACHE_ALIASES;
-		if (read_c0_config7() & MIPS_CONF7_AR) {
+		if (!has_74k_erratum && (read_c0_config7() & MIPS_CONF7_AR)) {
 			/*
 			 * Effectively physically indexed dcache,
 			 * thus no virtual aliases.
@@ -1268,7 +1278,7 @@ static void probe_pcache(void)
 			break;
 		}
 	default:
-		if (c->dcache.waysize > PAGE_SIZE)
+		if (has_74k_erratum || c->dcache.waysize > PAGE_SIZE)
 			c->dcache.flags |= MIPS_CACHE_ALIASES;
 	}
 
@@ -1438,6 +1448,7 @@ static void setup_scache(void)
 	case CPU_R10000:
 	case CPU_R12000:
 	case CPU_R14000:
+	case CPU_R16000:
 		scache_size = 0x80000 << ((config & R10K_CONF_SS) >> 16);
 		c->scache.linesz = 64 << ((config >> 13) & 1);
 		c->scache.ways = 2;
diff --git a/arch/mips/mm/cache.c b/arch/mips/mm/cache.c
index 7e3ea7766822..77d96db8253c 100644
--- a/arch/mips/mm/cache.c
+++ b/arch/mips/mm/cache.c
@@ -119,36 +119,37 @@ void __flush_anon_page(struct page *page, unsigned long vmaddr)
 
 EXPORT_SYMBOL(__flush_anon_page);
 
-static void mips_flush_dcache_from_pte(pte_t pteval, unsigned long address)
+void __flush_icache_page(struct vm_area_struct *vma, struct page *page)
+{
+	unsigned long addr;
+
+	if (PageHighMem(page))
+		return;
+
+	addr = (unsigned long) page_address(page);
+	flush_data_cache_page(addr);
+}
+EXPORT_SYMBOL_GPL(__flush_icache_page);
+
+void __update_cache(struct vm_area_struct *vma, unsigned long address,
+	pte_t pte)
 {
 	struct page *page;
-	unsigned long pfn = pte_pfn(pteval);
+	unsigned long pfn, addr;
+	int exec = (vma->vm_flags & VM_EXEC) && !cpu_has_ic_fills_f_dc;
 
+	pfn = pte_pfn(pte);
 	if (unlikely(!pfn_valid(pfn)))
 		return;
-
 	page = pfn_to_page(pfn);
 	if (page_mapping(page) && Page_dcache_dirty(page)) {
-		unsigned long page_addr = (unsigned long) page_address(page);
-
-		if (!cpu_has_ic_fills_f_dc ||
-		    pages_do_alias(page_addr, address & PAGE_MASK))
-			flush_data_cache_page(page_addr);
+		addr = (unsigned long) page_address(page);
+		if (exec || pages_do_alias(addr, address & PAGE_MASK))
+			flush_data_cache_page(addr);
 		ClearPageDcacheDirty(page);
 	}
 }
 
-void set_pte_at(struct mm_struct *mm, unsigned long addr,
-        pte_t *ptep, pte_t pteval)
-{
-        if (cpu_has_dc_aliases || !cpu_has_ic_fills_f_dc) {
-                if (pte_present(pteval))
-                        mips_flush_dcache_from_pte(pteval, addr);
-        }
-
-        set_pte(ptep, pteval);
-}
-
 unsigned long _page_cachable_default;
 EXPORT_SYMBOL(_page_cachable_default);
 
diff --git a/arch/mips/mm/dma-default.c b/arch/mips/mm/dma-default.c
index af5f046e627e..609d1241b0c4 100644
--- a/arch/mips/mm/dma-default.c
+++ b/arch/mips/mm/dma-default.c
@@ -258,7 +258,7 @@ static void mips_dma_unmap_page(struct device *dev, dma_addr_t dma_addr,
 	if (cpu_needs_post_dma_flush(dev))
 		__dma_sync(dma_addr_to_page(dev, dma_addr),
 			   dma_addr & ~PAGE_MASK, size, direction);
-
+	plat_post_dma_flush(dev);
 	plat_unmap_dma_mem(dev, dma_addr, size, direction);
 }
 
@@ -312,6 +312,7 @@ static void mips_dma_sync_single_for_cpu(struct device *dev,
 	if (cpu_needs_post_dma_flush(dev))
 		__dma_sync(dma_addr_to_page(dev, dma_handle),
 			   dma_handle & ~PAGE_MASK, size, direction);
+	plat_post_dma_flush(dev);
 }
 
 static void mips_dma_sync_single_for_device(struct device *dev,
@@ -331,6 +332,7 @@ static void mips_dma_sync_sg_for_cpu(struct device *dev,
 		for (i = 0; i < nelems; i++, sg++)
 			__dma_sync(sg_page(sg), sg->offset, sg->length,
 				   direction);
+	plat_post_dma_flush(dev);
 }
 
 static void mips_dma_sync_sg_for_device(struct device *dev,
diff --git a/arch/mips/mm/init.c b/arch/mips/mm/init.c
index 448cde372af0..faa5c9822ecc 100644
--- a/arch/mips/mm/init.c
+++ b/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);
diff --git a/arch/mips/mm/page.c b/arch/mips/mm/page.c
index 3f85f921801b..885d73ffd6fb 100644
--- a/arch/mips/mm/page.c
+++ b/arch/mips/mm/page.c
@@ -157,6 +157,7 @@ static void set_prefetch_parameters(void)
 		case CPU_R10000:
 		case CPU_R12000:
 		case CPU_R14000:
+		case CPU_R16000:
 			/*
 			 * Those values have been experimentally tuned for an
 			 * Origin 200.
diff --git a/arch/mips/mm/tlb-r4k.c b/arch/mips/mm/tlb-r4k.c
index b2afa49beab0..a27a088e6f9f 100644
--- a/arch/mips/mm/tlb-r4k.c
+++ b/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
@@ -477,7 +489,8 @@ static void r4k_tlb_configure(void)
 	write_c0_wired(0);
 	if (current_cpu_type() == CPU_R10000 ||
 	    current_cpu_type() == CPU_R12000 ||
-	    current_cpu_type() == CPU_R14000)
+	    current_cpu_type() == CPU_R14000 ||
+	    current_cpu_type() == CPU_R16000)
 		write_c0_framemask(0);
 
 	if (cpu_has_rixi) {
diff --git a/arch/mips/mm/tlbex.c b/arch/mips/mm/tlbex.c
index d75ff73a2012..97c87027c17f 100644
--- a/arch/mips/mm/tlbex.c
+++ b/arch/mips/mm/tlbex.c
@@ -35,6 +35,17 @@
 #include <asm/uasm.h>
 #include <asm/setup.h>
 
+static int __cpuinitdata mips_xpa_disabled;
+
+static int __init xpa_disable(char *s)
+{
+	mips_xpa_disabled = 1;
+
+	return 1;
+}
+
+__setup("noxpa", xpa_disable);
+
 /*
  * TLB load/store/modify handlers.
  *
@@ -231,14 +242,14 @@ static void output_pgtable_bits_defines(void)
 	pr_define("_PAGE_HUGE_SHIFT %d\n", _PAGE_HUGE_SHIFT);
 	pr_define("_PAGE_SPLITTING_SHIFT %d\n", _PAGE_SPLITTING_SHIFT);
 #endif
+#ifdef CONFIG_CPU_MIPSR2
 	if (cpu_has_rixi) {
 #ifdef _PAGE_NO_EXEC_SHIFT
 		pr_define("_PAGE_NO_EXEC_SHIFT %d\n", _PAGE_NO_EXEC_SHIFT);
-#endif
-#ifdef _PAGE_NO_READ_SHIFT
 		pr_define("_PAGE_NO_READ_SHIFT %d\n", _PAGE_NO_READ_SHIFT);
 #endif
 	}
+#endif
 	pr_define("_PAGE_GLOBAL_SHIFT %d\n", _PAGE_GLOBAL_SHIFT);
 	pr_define("_PAGE_VALID_SHIFT %d\n", _PAGE_VALID_SHIFT);
 	pr_define("_PAGE_DIRTY_SHIFT %d\n", _PAGE_DIRTY_SHIFT);
@@ -501,26 +512,9 @@ static void build_tlb_write_entry(u32 **p, struct uasm_label **l,
 	case tlb_indexed: tlbw = uasm_i_tlbwi; break;
 	}
 
-	if (cpu_has_mips_r2_exec_hazard) {
-		/*
-		 * The architecture spec says an ehb is required here,
-		 * but a number of cores do not have the hazard and
-		 * using an ehb causes an expensive pipeline stall.
-		 */
-		switch (current_cpu_type()) {
-		case CPU_M14KC:
-		case CPU_74K:
-		case CPU_1074K:
-		case CPU_PROAPTIV:
-		case CPU_P5600:
-		case CPU_M5150:
-		case CPU_QEMU_GENERIC:
-			break;
-
-		default:
+	if (cpu_has_mips_r2_r6) {
+		if (cpu_has_mips_r2_exec_hazard)
 			uasm_i_ehb(p);
-			break;
-		}
 		tlbw(p);
 		return;
 	}
@@ -569,6 +563,7 @@ static void build_tlb_write_entry(u32 **p, struct uasm_label **l,
 	case CPU_R10000:
 	case CPU_R12000:
 	case CPU_R14000:
+	case CPU_R16000:
 	case CPU_4KC:
 	case CPU_4KEC:
 	case CPU_M14KC:
@@ -1027,12 +1022,27 @@ static void build_update_entries(u32 **p, unsigned int tmp, unsigned int ptep)
 	} else {
 		int pte_off_even = sizeof(pte_t) / 2;
 		int pte_off_odd = pte_off_even + sizeof(pte_t);
+#ifdef CONFIG_XPA
+		const int scratch = 1; /* Our extra working register */
 
-		/* The pte entries are pre-shifted */
-		uasm_i_lw(p, tmp, pte_off_even, ptep); /* get even pte */
-		UASM_i_MTC0(p, tmp, C0_ENTRYLO0); /* load it */
-		uasm_i_lw(p, ptep, pte_off_odd, ptep); /* get odd pte */
-		UASM_i_MTC0(p, ptep, C0_ENTRYLO1); /* load it */
+		uasm_i_addu(p, scratch, 0, ptep);
+#endif
+		uasm_i_lw(p, tmp, pte_off_even, ptep); /* even pte */
+		uasm_i_lw(p, ptep, pte_off_odd, ptep); /* odd pte */
+		UASM_i_ROTR(p, tmp, tmp, ilog2(_PAGE_GLOBAL));
+		UASM_i_ROTR(p, ptep, ptep, ilog2(_PAGE_GLOBAL));
+		UASM_i_MTC0(p, tmp, C0_ENTRYLO0);
+		UASM_i_MTC0(p, ptep, C0_ENTRYLO1);
+#ifdef CONFIG_XPA
+		uasm_i_lw(p, tmp, 0, scratch);
+		uasm_i_lw(p, ptep, sizeof(pte_t), scratch);
+		uasm_i_lui(p, scratch, 0xff);
+		uasm_i_ori(p, scratch, scratch, 0xffff);
+		uasm_i_and(p, tmp, scratch, tmp);
+		uasm_i_and(p, ptep, scratch, ptep);
+		uasm_i_mthc0(p, tmp, C0_ENTRYLO0);
+		uasm_i_mthc0(p, ptep, C0_ENTRYLO1);
+#endif
 	}
 #else
 	UASM_i_LW(p, tmp, 0, ptep); /* get even pte */
@@ -1533,8 +1543,14 @@ iPTE_SW(u32 **p, struct uasm_reloc **r, unsigned int pte, unsigned int ptr,
 {
 #ifdef CONFIG_PHYS_ADDR_T_64BIT
 	unsigned int hwmode = mode & (_PAGE_VALID | _PAGE_DIRTY);
-#endif
 
+	if (!cpu_has_64bits) {
+		const int scratch = 1; /* Our extra working register */
+
+		uasm_i_lui(p, scratch, (mode >> 16));
+		uasm_i_or(p, pte, pte, scratch);
+	} else
+#endif
 	uasm_i_ori(p, pte, pte, mode);
 #ifdef CONFIG_SMP
 # ifdef CONFIG_PHYS_ADDR_T_64BIT
@@ -1598,15 +1614,17 @@ build_pte_present(u32 **p, struct uasm_reloc **r,
 			uasm_il_bbit0(p, r, pte, ilog2(_PAGE_PRESENT), lid);
 			uasm_i_nop(p);
 		} else {
-			uasm_i_andi(p, t, pte, _PAGE_PRESENT);
+			uasm_i_srl(p, t, pte, _PAGE_PRESENT_SHIFT);
+			uasm_i_andi(p, t, t, 1);
 			uasm_il_beqz(p, r, t, lid);
 			if (pte == t)
 				/* You lose the SMP race :-(*/
 				iPTE_LW(p, pte, ptr);
 		}
 	} else {
-		uasm_i_andi(p, t, pte, _PAGE_PRESENT | _PAGE_READ);
-		uasm_i_xori(p, t, t, _PAGE_PRESENT | _PAGE_READ);
+		uasm_i_srl(p, t, pte, _PAGE_PRESENT_SHIFT);
+		uasm_i_andi(p, t, t, 3);
+		uasm_i_xori(p, t, t, 3);
 		uasm_il_bnez(p, r, t, lid);
 		if (pte == t)
 			/* You lose the SMP race :-(*/
@@ -1635,8 +1653,9 @@ build_pte_writable(u32 **p, struct uasm_reloc **r,
 {
 	int t = scratch >= 0 ? scratch : pte;
 
-	uasm_i_andi(p, t, pte, _PAGE_PRESENT | _PAGE_WRITE);
-	uasm_i_xori(p, t, t, _PAGE_PRESENT | _PAGE_WRITE);
+	uasm_i_srl(p, t, pte, _PAGE_PRESENT_SHIFT);
+	uasm_i_andi(p, t, t, 5);
+	uasm_i_xori(p, t, t, 5);
 	uasm_il_bnez(p, r, t, lid);
 	if (pte == t)
 		/* You lose the SMP race :-(*/
@@ -1672,7 +1691,8 @@ build_pte_modifiable(u32 **p, struct uasm_reloc **r,
 		uasm_i_nop(p);
 	} else {
 		int t = scratch >= 0 ? scratch : pte;
-		uasm_i_andi(p, t, pte, _PAGE_WRITE);
+		uasm_i_srl(p, t, pte, _PAGE_WRITE_SHIFT);
+		uasm_i_andi(p, t, t, 1);
 		uasm_il_beqz(p, r, t, lid);
 		if (pte == t)
 			/* You lose the SMP race :-(*/
@@ -2285,6 +2305,11 @@ static void config_htw_params(void)
 
 	pwsize = ilog2(PTRS_PER_PGD) << MIPS_PWSIZE_GDW_SHIFT;
 	pwsize |= ilog2(PTRS_PER_PTE) << MIPS_PWSIZE_PTW_SHIFT;
+
+	/* If XPA has been enabled, PTEs are 64-bit in size. */
+	if (read_c0_pagegrain() & PG_ELPA)
+		pwsize |= 1;
+
 	write_c0_pwsize(pwsize);
 
 	/* Make sure everything is set before we enable the HTW */
@@ -2298,6 +2323,28 @@ static void config_htw_params(void)
 	print_htw_config();
 }
 
+static void config_xpa_params(void)
+{
+#ifdef CONFIG_XPA
+	unsigned int pagegrain;
+
+	if (mips_xpa_disabled) {
+		pr_info("Extended Physical Addressing (XPA) disabled\n");
+		return;
+	}
+
+	pagegrain = read_c0_pagegrain();
+	write_c0_pagegrain(pagegrain | PG_ELPA);
+	back_to_back_c0_hazard();
+	pagegrain = read_c0_pagegrain();
+
+	if (pagegrain & PG_ELPA)
+		pr_info("Extended Physical Addressing (XPA) enabled\n");
+	else
+		panic("Extended Physical Addressing (XPA) disabled");
+#endif
+}
+
 void build_tlb_refill_handler(void)
 {
 	/*
@@ -2362,8 +2409,9 @@ void build_tlb_refill_handler(void)
 		}
 		if (cpu_has_local_ebase)
 			build_r4000_tlb_refill_handler();
+		if (cpu_has_xpa)
+			config_xpa_params();
 		if (cpu_has_htw)
 			config_htw_params();
-
 	}
 }