From 964cf28f9d10f4e5229e4365258c292bc5c856b2 Mon Sep 17 00:00:00 2001
From: Vineet Gupta <vgupta@synopsys.com>
Date: Fri, 2 Oct 2015 19:20:27 +0530
Subject: ARC: boot log: move helper macros to header for reuse

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 arch/arc/mm/cache.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

(limited to 'arch/arc/mm/cache.c')

diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index 0d1a6e96839f..ae3b772ecc4d 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -37,7 +37,6 @@ char *arc_cache_mumbojumbo(int c, char *buf, int len)
 	int n = 0;
 	struct cpuinfo_arc_cache *p;
 
-#define IS_USED_RUN(v)		((v) ? "" : "(disabled) ")
 #define PR_CACHE(p, cfg, str)						\
 	if (!(p)->ver)							\
 		n += scnprintf(buf + n, len - n, str"\t\t: N/A\n");	\
@@ -47,7 +46,7 @@ char *arc_cache_mumbojumbo(int c, char *buf, int len)
 			(p)->sz_k, (p)->assoc, (p)->line_len,		\
 			(p)->vipt ? "VIPT" : "PIPT",			\
 			(p)->alias ? " aliasing" : "",			\
-			IS_ENABLED(cfg) ? "" : " (not used)");
+			IS_USED_CFG(cfg));
 
 	PR_CACHE(&cpuinfo_arc700[c].icache, CONFIG_ARC_HAS_ICACHE, "I-Cache");
 	PR_CACHE(&cpuinfo_arc700[c].dcache, CONFIG_ARC_HAS_DCACHE, "D-Cache");
@@ -63,7 +62,7 @@ char *arc_cache_mumbojumbo(int c, char *buf, int len)
 
 	if (ioc_exists)
 		n += scnprintf(buf + n, len - n, "IOC\t\t:%s\n",
-				IS_USED_RUN(ioc_enable));
+				IS_DISABLED_RUN(ioc_enable));
 
 	return buf;
 }
-- 
cgit v1.2.3


From 336e2136e1353db8e9e731c27381ee0735656a8a Mon Sep 17 00:00:00 2001
From: Vineet Gupta <vgupta@synopsys.com>
Date: Thu, 5 Mar 2015 17:06:31 +0530
Subject: ARC: mm: preps ahead of HIGHMEM support

Before we plug in highmem support, some of code needs to be ready for it
 - copy_user_highpage() needs to be using the kmap_atomic API
 - mk_pte() can't assume page_address()
 - do_page_fault() can't assume VMALLOC_END is end of kernel vaddr space

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 arch/arc/include/asm/pgtable.h |  9 +--------
 arch/arc/mm/cache.c            | 16 +++++++++++-----
 arch/arc/mm/fault.c            | 13 ++++++++++---
 3 files changed, 22 insertions(+), 16 deletions(-)

(limited to 'arch/arc/mm/cache.c')

diff --git a/arch/arc/include/asm/pgtable.h b/arch/arc/include/asm/pgtable.h
index f7c7273cd537..bd771351a1d1 100644
--- a/arch/arc/include/asm/pgtable.h
+++ b/arch/arc/include/asm/pgtable.h
@@ -270,13 +270,7 @@ static inline void pmd_set(pmd_t *pmdp, pte_t *ptep)
 		(unsigned long)(((pte_val(x) - CONFIG_LINUX_LINK_BASE) >> \
 				PAGE_SHIFT)))
 
-#define mk_pte(page, pgprot)						\
-({									\
-	pte_t pte;							\
-	pte_val(pte) = __pa(page_address(page)) + pgprot_val(pgprot);	\
-	pte;								\
-})
-
+#define mk_pte(page, prot)	pfn_pte(page_to_pfn(page), prot)
 #define pte_pfn(pte)		(pte_val(pte) >> PAGE_SHIFT)
 #define pfn_pte(pfn, prot)	(__pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot)))
 #define __pte_index(addr)	(((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
@@ -360,7 +354,6 @@ static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
 #define pgd_offset_fast(mm, addr)	pgd_offset(mm, addr)
 #endif
 
-extern void paging_init(void);
 extern pgd_t swapper_pg_dir[] __aligned(PAGE_SIZE);
 void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
 		      pte_t *ptep);
diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index ae3b772ecc4d..521fb2bf90bd 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -806,8 +806,8 @@ void flush_anon_page(struct vm_area_struct *vma, struct page *page,
 void copy_user_highpage(struct page *to, struct page *from,
 	unsigned long u_vaddr, struct vm_area_struct *vma)
 {
-	unsigned long kfrom = (unsigned long)page_address(from);
-	unsigned long kto = (unsigned long)page_address(to);
+	void *kfrom = kmap_atomic(from);
+	void *kto = kmap_atomic(to);
 	int clean_src_k_mappings = 0;
 
 	/*
@@ -817,13 +817,16 @@ void copy_user_highpage(struct page *to, struct page *from,
 	 *
 	 * Note that while @u_vaddr refers to DST page's userspace vaddr, it is
 	 * equally valid for SRC page as well
+	 *
+	 * For !VIPT cache, all of this gets compiled out as
+	 * addr_not_cache_congruent() is 0
 	 */
 	if (page_mapped(from) && addr_not_cache_congruent(kfrom, u_vaddr)) {
-		__flush_dcache_page(kfrom, u_vaddr);
+		__flush_dcache_page((unsigned long)kfrom, u_vaddr);
 		clean_src_k_mappings = 1;
 	}
 
-	copy_page((void *)kto, (void *)kfrom);
+	copy_page(kto, kfrom);
 
 	/*
 	 * Mark DST page K-mapping as dirty for a later finalization by
@@ -840,11 +843,14 @@ void copy_user_highpage(struct page *to, struct page *from,
 	 * sync the kernel mapping back to physical page
 	 */
 	if (clean_src_k_mappings) {
-		__flush_dcache_page(kfrom, kfrom);
+		__flush_dcache_page((unsigned long)kfrom, (unsigned long)kfrom);
 		set_bit(PG_dc_clean, &from->flags);
 	} else {
 		clear_bit(PG_dc_clean, &from->flags);
 	}
+
+	kunmap_atomic(kto);
+	kunmap_atomic(kfrom);
 }
 
 void clear_user_page(void *to, unsigned long u_vaddr, struct page *page)
diff --git a/arch/arc/mm/fault.c b/arch/arc/mm/fault.c
index d948e4e9d89c..af63f4a13e60 100644
--- a/arch/arc/mm/fault.c
+++ b/arch/arc/mm/fault.c
@@ -18,7 +18,14 @@
 #include <asm/pgalloc.h>
 #include <asm/mmu.h>
 
-static int handle_vmalloc_fault(unsigned long address)
+/*
+ * kernel virtual address is required to implement vmalloc/pkmap/fixmap
+ * Refer to asm/processor.h for System Memory Map
+ *
+ * It simply copies the PMD entry (pointer to 2nd level page table or hugepage)
+ * from swapper pgdir to task pgdir. The 2nd level table/page is thus shared
+ */
+noinline static int handle_kernel_vaddr_fault(unsigned long address)
 {
 	/*
 	 * Synchronize this task's top level page-table
@@ -72,8 +79,8 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
 	 * only copy the information from the master page table,
 	 * nothing more.
 	 */
-	if (address >= VMALLOC_START && address <= VMALLOC_END) {
-		ret = handle_vmalloc_fault(address);
+	if (address >= VMALLOC_START) {
+		ret = handle_kernel_vaddr_fault(address);
 		if (unlikely(ret))
 			goto bad_area_nosemaphore;
 		else
-- 
cgit v1.2.3


From 28b4af729fc4f7ee748c4bccb50ba5a6066418eb Mon Sep 17 00:00:00 2001
From: Vineet Gupta <vgupta@synopsys.com>
Date: Mon, 14 Sep 2015 18:43:42 -0700
Subject: ARC: mm: PAE40: switch to using phys_addr_t for physical addresses

That way a single flip of phys_addr_t to 64 bit ensures all places
dealing with physical addresses get correct data

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 arch/arc/include/asm/cacheflush.h |  8 ++++----
 arch/arc/mm/cache.c               | 26 +++++++++++++-------------
 arch/arc/mm/tlb.c                 | 10 +++++-----
 3 files changed, 22 insertions(+), 22 deletions(-)

(limited to 'arch/arc/mm/cache.c')

diff --git a/arch/arc/include/asm/cacheflush.h b/arch/arc/include/asm/cacheflush.h
index 0992d3dbcc65..fbe3587c4f36 100644
--- a/arch/arc/include/asm/cacheflush.h
+++ b/arch/arc/include/asm/cacheflush.h
@@ -31,10 +31,10 @@
 
 void flush_cache_all(void);
 
-void flush_icache_range(unsigned long start, unsigned long end);
-void __sync_icache_dcache(unsigned long paddr, unsigned long vaddr, int len);
-void __inv_icache_page(unsigned long paddr, unsigned long vaddr);
-void __flush_dcache_page(unsigned long paddr, unsigned long vaddr);
+void flush_icache_range(unsigned long kstart, unsigned long kend);
+void __sync_icache_dcache(phys_addr_t paddr, unsigned long vaddr, int len);
+void __inv_icache_page(phys_addr_t paddr, unsigned long vaddr);
+void __flush_dcache_page(phys_addr_t paddr, unsigned long vaddr);
 
 #define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1
 
diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index 521fb2bf90bd..91e28d791d56 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -25,7 +25,7 @@ static int l2_line_sz;
 int ioc_exists;
 volatile int slc_enable = 1, ioc_enable = 1;
 
-void (*_cache_line_loop_ic_fn)(unsigned long paddr, unsigned long vaddr,
+void (*_cache_line_loop_ic_fn)(phys_addr_t paddr, unsigned long vaddr,
 			       unsigned long sz, const int cacheop);
 
 void (*__dma_cache_wback_inv)(unsigned long start, unsigned long sz);
@@ -216,7 +216,7 @@ slc_chk:
  */
 
 static inline
-void __cache_line_loop_v2(unsigned long paddr, unsigned long vaddr,
+void __cache_line_loop_v2(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz, const int op)
 {
 	unsigned int aux_cmd;
@@ -254,7 +254,7 @@ void __cache_line_loop_v2(unsigned long paddr, unsigned long vaddr,
 }
 
 static inline
-void __cache_line_loop_v3(unsigned long paddr, unsigned long vaddr,
+void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz, const int op)
 {
 	unsigned int aux_cmd, aux_tag;
@@ -308,7 +308,7 @@ void __cache_line_loop_v3(unsigned long paddr, unsigned long vaddr,
  * specified in PTAG (similar to MMU v3)
  */
 static inline
-void __cache_line_loop_v4(unsigned long paddr, unsigned long vaddr,
+void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz, const int cacheop)
 {
 	unsigned int aux_cmd;
@@ -412,7 +412,7 @@ static inline void __dc_entire_op(const int op)
 /*
  * D-Cache Line ops: Per Line INV (discard or wback+discard) or FLUSH (wback)
  */
-static inline void __dc_line_op(unsigned long paddr, unsigned long vaddr,
+static inline void __dc_line_op(phys_addr_t paddr, unsigned long vaddr,
 				unsigned long sz, const int op)
 {
 	unsigned long flags;
@@ -445,7 +445,7 @@ static inline void __ic_entire_inv(void)
 }
 
 static inline void
-__ic_line_inv_vaddr_local(unsigned long paddr, unsigned long vaddr,
+__ic_line_inv_vaddr_local(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz)
 {
 	unsigned long flags;
@@ -462,7 +462,7 @@ __ic_line_inv_vaddr_local(unsigned long paddr, unsigned long vaddr,
 #else
 
 struct ic_inv_args {
-	unsigned long paddr, vaddr;
+	phys_addr_t paddr, vaddr;
 	int sz;
 };
 
@@ -473,7 +473,7 @@ static void __ic_line_inv_vaddr_helper(void *info)
         __ic_line_inv_vaddr_local(ic_inv->paddr, ic_inv->vaddr, ic_inv->sz);
 }
 
-static void __ic_line_inv_vaddr(unsigned long paddr, unsigned long vaddr,
+static void __ic_line_inv_vaddr(phys_addr_t paddr, unsigned long vaddr,
 				unsigned long sz)
 {
 	struct ic_inv_args ic_inv = {
@@ -494,7 +494,7 @@ static void __ic_line_inv_vaddr(unsigned long paddr, unsigned long vaddr,
 
 #endif /* CONFIG_ARC_HAS_ICACHE */
 
-noinline void slc_op(unsigned long paddr, unsigned long sz, const int op)
+noinline void slc_op(phys_addr_t paddr, unsigned long sz, const int op)
 {
 #ifdef CONFIG_ISA_ARCV2
 	/*
@@ -584,7 +584,7 @@ void flush_dcache_page(struct page *page)
 	} else if (page_mapped(page)) {
 
 		/* kernel reading from page with U-mapping */
-		unsigned long paddr = (unsigned long)page_address(page);
+		phys_addr_t paddr = (unsigned long)page_address(page);
 		unsigned long vaddr = page->index << PAGE_CACHE_SHIFT;
 
 		if (addr_not_cache_congruent(paddr, vaddr))
@@ -732,14 +732,14 @@ EXPORT_SYMBOL(flush_icache_range);
  *    builtin kernel page will not have any virtual mappings.
  *    kprobe on loadable module will be kernel vaddr.
  */
-void __sync_icache_dcache(unsigned long paddr, unsigned long vaddr, int len)
+void __sync_icache_dcache(phys_addr_t paddr, unsigned long vaddr, int len)
 {
 	__dc_line_op(paddr, vaddr, len, OP_FLUSH_N_INV);
 	__ic_line_inv_vaddr(paddr, vaddr, len);
 }
 
 /* wrapper to compile time eliminate alignment checks in flush loop */
-void __inv_icache_page(unsigned long paddr, unsigned long vaddr)
+void __inv_icache_page(phys_addr_t paddr, unsigned long vaddr)
 {
 	__ic_line_inv_vaddr(paddr, vaddr, PAGE_SIZE);
 }
@@ -748,7 +748,7 @@ void __inv_icache_page(unsigned long paddr, unsigned long vaddr)
  * wrapper to clearout kernel or userspace mappings of a page
  * For kernel mappings @vaddr == @paddr
  */
-void __flush_dcache_page(unsigned long paddr, unsigned long vaddr)
+void __flush_dcache_page(phys_addr_t paddr, unsigned long vaddr)
 {
 	__dc_line_op(paddr, vaddr & PAGE_MASK, PAGE_SIZE, OP_FLUSH_N_INV);
 }
diff --git a/arch/arc/mm/tlb.c b/arch/arc/mm/tlb.c
index 5dcae21dd8dc..b5f28dc0f924 100644
--- a/arch/arc/mm/tlb.c
+++ b/arch/arc/mm/tlb.c
@@ -499,7 +499,7 @@ void flush_tlb_kernel_range(unsigned long start, unsigned long end)
 /*
  * Routine to create a TLB entry
  */
-void create_tlb(struct vm_area_struct *vma, unsigned long address, pte_t *ptep)
+void create_tlb(struct vm_area_struct *vma, unsigned long vaddr, pte_t *ptep)
 {
 	unsigned long flags;
 	unsigned int asid_or_sasid, rwx;
@@ -535,9 +535,9 @@ void create_tlb(struct vm_area_struct *vma, unsigned long address, pte_t *ptep)
 
 	local_irq_save(flags);
 
-	tlb_paranoid_check(asid_mm(vma->vm_mm, smp_processor_id()), address);
+	tlb_paranoid_check(asid_mm(vma->vm_mm, smp_processor_id()), vaddr);
 
-	address &= PAGE_MASK;
+	vaddr &= PAGE_MASK;
 
 	/* update this PTE credentials */
 	pte_val(*ptep) |= (_PAGE_PRESENT | _PAGE_ACCESSED);
@@ -547,7 +547,7 @@ void create_tlb(struct vm_area_struct *vma, unsigned long address, pte_t *ptep)
 	/* ASID for this task */
 	asid_or_sasid = read_aux_reg(ARC_REG_PID) & 0xff;
 
-	pd0 = address | asid_or_sasid | (pte_val(*ptep) & PTE_BITS_IN_PD0);
+	pd0 = vaddr | asid_or_sasid | (pte_val(*ptep) & PTE_BITS_IN_PD0);
 
 	/*
 	 * ARC MMU provides fully orthogonal access bits for K/U mode,
@@ -583,7 +583,7 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long vaddr_unaligned,
 		      pte_t *ptep)
 {
 	unsigned long vaddr = vaddr_unaligned & PAGE_MASK;
-	unsigned long paddr = pte_val(*ptep) & PAGE_MASK;
+	phys_addr_t paddr = pte_val(*ptep) & PAGE_MASK;
 	struct page *page = pfn_to_page(pte_pfn(*ptep));
 
 	create_tlb(vma, vaddr, ptep);
-- 
cgit v1.2.3


From 5a364c2a1762e8a78721fafc93144509c0b6cb84 Mon Sep 17 00:00:00 2001
From: Vineet Gupta <vgupta@synopsys.com>
Date: Fri, 6 Feb 2015 18:44:57 +0300
Subject: ARC: mm: PAE40 support

This is the first working implementation of 40-bit physical address
extension on ARCv2.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
---
 arch/arc/Kconfig               | 15 ++++++++++++++
 arch/arc/include/asm/cache.h   |  2 ++
 arch/arc/include/asm/mmu.h     |  7 +++++++
 arch/arc/include/asm/page.h    |  4 ++++
 arch/arc/include/asm/pgalloc.h |  6 +++---
 arch/arc/include/asm/pgtable.h |  8 +++++++-
 arch/arc/mm/cache.c            | 44 ++++++++++++++++++++++++++++++++++++++----
 arch/arc/mm/tlb.c              | 27 +++++++++++++++++++++-----
 arch/arc/mm/tlbex.S            | 11 ++++++++++-
 9 files changed, 110 insertions(+), 14 deletions(-)

(limited to 'arch/arc/mm/cache.c')

diff --git a/arch/arc/Kconfig b/arch/arc/Kconfig
index fd9632f4ddc8..2c2ac3f3ff80 100644
--- a/arch/arc/Kconfig
+++ b/arch/arc/Kconfig
@@ -453,6 +453,21 @@ config HIGHMEM
 	  kernel. Enable this to potentially allow access to rest of 2G and PAE
 	  in future
 
+config ARC_HAS_PAE40
+	bool "Support for the 40-bit Physical Address Extension"
+	default n
+	depends on ISA_ARCV2
+	select HIGHMEM
+	help
+	  Enable access to physical memory beyond 4G, only supported on
+	  ARC cores with 40 bit Physical Addressing support
+
+config ARCH_PHYS_ADDR_T_64BIT
+	def_bool ARC_HAS_PAE40
+
+config ARCH_DMA_ADDR_T_64BIT
+	bool
+
 config ARC_CURR_IN_REG
 	bool "Dedicate Register r25 for current_task pointer"
 	default y
diff --git a/arch/arc/include/asm/cache.h b/arch/arc/include/asm/cache.h
index e23ea6e7633a..abf06e81c929 100644
--- a/arch/arc/include/asm/cache.h
+++ b/arch/arc/include/asm/cache.h
@@ -65,6 +65,7 @@ extern int ioc_exists;
 #if defined(CONFIG_ARC_MMU_V3) || defined(CONFIG_ARC_MMU_V4)
 #define ARC_REG_IC_PTAG		0x1E
 #endif
+#define ARC_REG_IC_PTAG_HI	0x1F
 
 /* Bit val in IC_CTRL */
 #define IC_CTRL_CACHE_DISABLE   0x1
@@ -77,6 +78,7 @@ extern int ioc_exists;
 #define ARC_REG_DC_FLSH		0x4B
 #define ARC_REG_DC_FLDL		0x4C
 #define ARC_REG_DC_PTAG		0x5C
+#define ARC_REG_DC_PTAG_HI	0x5F
 
 /* Bit val in DC_CTRL */
 #define DC_CTRL_INV_MODE_FLUSH  0x40
diff --git a/arch/arc/include/asm/mmu.h b/arch/arc/include/asm/mmu.h
index 0f9c3eb5327e..b144d7ca7d20 100644
--- a/arch/arc/include/asm/mmu.h
+++ b/arch/arc/include/asm/mmu.h
@@ -24,6 +24,7 @@
 #if (CONFIG_ARC_MMU_VER < 4)
 #define ARC_REG_TLBPD0		0x405
 #define ARC_REG_TLBPD1		0x406
+#define ARC_REG_TLBPD1HI	0	/* Dummy: allows code sharing with ARC700 */
 #define ARC_REG_TLBINDEX	0x407
 #define ARC_REG_TLBCOMMAND	0x408
 #define ARC_REG_PID		0x409
@@ -31,6 +32,7 @@
 #else
 #define ARC_REG_TLBPD0		0x460
 #define ARC_REG_TLBPD1		0x461
+#define ARC_REG_TLBPD1HI	0x463
 #define ARC_REG_TLBINDEX	0x464
 #define ARC_REG_TLBCOMMAND	0x465
 #define ARC_REG_PID		0x468
@@ -83,6 +85,11 @@ void arc_mmu_init(void);
 extern char *arc_mmu_mumbojumbo(int cpu_id, char *buf, int len);
 void read_decode_mmu_bcr(void);
 
+static inline int is_pae40_enabled(void)
+{
+	return IS_ENABLED(CONFIG_ARC_HAS_PAE40);
+}
+
 #endif	/* !__ASSEMBLY__ */
 
 #endif
diff --git a/arch/arc/include/asm/page.h b/arch/arc/include/asm/page.h
index 37706837ef75..429957f1c236 100644
--- a/arch/arc/include/asm/page.h
+++ b/arch/arc/include/asm/page.h
@@ -56,7 +56,11 @@ typedef struct {
 
 #else /* !STRICT_MM_TYPECHECKS */
 
+#ifdef CONFIG_ARC_HAS_PAE40
+typedef unsigned long long pte_t;
+#else
 typedef unsigned long pte_t;
+#endif
 typedef unsigned long pgd_t;
 typedef unsigned long pgprot_t;
 
diff --git a/arch/arc/include/asm/pgalloc.h b/arch/arc/include/asm/pgalloc.h
index 9149b5ca26d7..86ed671286df 100644
--- a/arch/arc/include/asm/pgalloc.h
+++ b/arch/arc/include/asm/pgalloc.h
@@ -49,7 +49,7 @@ pmd_populate(struct mm_struct *mm, pmd_t *pmd, pgtable_t ptep)
 
 static inline int __get_order_pgd(void)
 {
-	return get_order(PTRS_PER_PGD * 4);
+	return get_order(PTRS_PER_PGD * sizeof(pgd_t));
 }
 
 static inline pgd_t *pgd_alloc(struct mm_struct *mm)
@@ -87,7 +87,7 @@ static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
 
 static inline int __get_order_pte(void)
 {
-	return get_order(PTRS_PER_PTE * 4);
+	return get_order(PTRS_PER_PTE * sizeof(pte_t));
 }
 
 static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
@@ -110,7 +110,7 @@ pte_alloc_one(struct mm_struct *mm, unsigned long address)
 	pte_pg = (pgtable_t)__get_free_pages(GFP_KERNEL | __GFP_REPEAT, __get_order_pte());
 	if (!pte_pg)
 		return 0;
-	memzero((void *)pte_pg, PTRS_PER_PTE * 4);
+	memzero((void *)pte_pg, PTRS_PER_PTE * sizeof(pte_t));
 	page = virt_to_page(pte_pg);
 	if (!pgtable_page_ctor(page)) {
 		__free_page(page);
diff --git a/arch/arc/include/asm/pgtable.h b/arch/arc/include/asm/pgtable.h
index bd771351a1d1..57af2f05ae84 100644
--- a/arch/arc/include/asm/pgtable.h
+++ b/arch/arc/include/asm/pgtable.h
@@ -134,7 +134,12 @@
 /* Masks for actual TLB "PD"s */
 #define PTE_BITS_IN_PD0		(_PAGE_GLOBAL | _PAGE_PRESENT | _PAGE_HW_SZ)
 #define PTE_BITS_RWX		(_PAGE_EXECUTE | _PAGE_WRITE | _PAGE_READ)
+
+#ifdef CONFIG_ARC_HAS_PAE40
+#define PTE_BITS_NON_RWX_IN_PD1	(0xff00000000 | PAGE_MASK | _PAGE_CACHEABLE)
+#else
 #define PTE_BITS_NON_RWX_IN_PD1	(PAGE_MASK | _PAGE_CACHEABLE)
+#endif
 
 /**************************************************************************
  * Mapping of vm_flags (Generic VM) to PTE flags (arch specific)
@@ -272,7 +277,8 @@ static inline void pmd_set(pmd_t *pmdp, pte_t *ptep)
 
 #define mk_pte(page, prot)	pfn_pte(page_to_pfn(page), prot)
 #define pte_pfn(pte)		(pte_val(pte) >> PAGE_SHIFT)
-#define pfn_pte(pfn, prot)	(__pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot)))
+#define pfn_pte(pfn, prot)	(__pte(((pte_t)(pfn) << PAGE_SHIFT) | \
+				 pgprot_val(prot)))
 #define __pte_index(addr)	(((addr) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
 
 /*
diff --git a/arch/arc/mm/cache.c b/arch/arc/mm/cache.c
index 91e28d791d56..ff7ff6cbb811 100644
--- a/arch/arc/mm/cache.c
+++ b/arch/arc/mm/cache.c
@@ -253,6 +253,10 @@ void __cache_line_loop_v2(phys_addr_t paddr, unsigned long vaddr,
 	}
 }
 
+/*
+ * For ARC700 MMUv3 I-cache and D-cache flushes
+ * Also reused for HS38 aliasing I-cache configuration
+ */
 static inline
 void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
 			  unsigned long sz, const int op)
@@ -289,6 +293,16 @@ void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
 	if (full_page)
 		write_aux_reg(aux_tag, paddr);
 
+	/*
+	 * This is technically for MMU v4, using the MMU v3 programming model
+	 * Special work for HS38 aliasing I-cache configuratino with PAE40
+	 *   - upper 8 bits of paddr need to be written into PTAG_HI
+	 *   - (and needs to be written before the lower 32 bits)
+	 * Note that PTAG_HI is hoisted outside the line loop
+	 */
+	if (is_pae40_enabled() && op == OP_INV_IC)
+		write_aux_reg(ARC_REG_IC_PTAG_HI, (u64)paddr >> 32);
+
 	while (num_lines-- > 0) {
 		if (!full_page) {
 			write_aux_reg(aux_tag, paddr);
@@ -301,11 +315,17 @@ void __cache_line_loop_v3(phys_addr_t paddr, unsigned long vaddr,
 }
 
 /*
- * In HS38x (MMU v4), although icache is VIPT, only paddr is needed for cache
- * maintenance ops (in IVIL reg), as long as icache doesn't alias.
+ * In HS38x (MMU v4), I-cache is VIPT (can alias), D-cache is PIPT
+ * Here's how cache ops are implemented
+ *
+ *  - D-cache: only paddr needed (in DC_IVDL/DC_FLDL)
+ *  - I-cache Non Aliasing: Despite VIPT, only paddr needed (in IC_IVIL)
+ *  - I-cache Aliasing: Both vaddr and paddr needed (in IC_IVIL, IC_PTAG
+ *    respectively, similar to MMU v3 programming model, hence
+ *    __cache_line_loop_v3() is used)
  *
- * For Aliasing icache, vaddr is also needed (in IVIL), while paddr is
- * specified in PTAG (similar to MMU v3)
+ * If PAE40 is enabled, independent of aliasing considerations, the higher bits
+ * needs to be written into PTAG_HI
  */
 static inline
 void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
@@ -335,6 +355,22 @@ void __cache_line_loop_v4(phys_addr_t paddr, unsigned long vaddr,
 
 	num_lines = DIV_ROUND_UP(sz, L1_CACHE_BYTES);
 
+	/*
+	 * For HS38 PAE40 configuration
+	 *   - upper 8 bits of paddr need to be written into PTAG_HI
+	 *   - (and needs to be written before the lower 32 bits)
+	 */
+	if (is_pae40_enabled()) {
+		if (cacheop == OP_INV_IC)
+			/*
+			 * Non aliasing I-cache in HS38,
+			 * aliasing I-cache handled in __cache_line_loop_v3()
+			 */
+			write_aux_reg(ARC_REG_IC_PTAG_HI, (u64)paddr >> 32);
+		else
+			write_aux_reg(ARC_REG_DC_PTAG_HI, (u64)paddr >> 32);
+	}
+
 	while (num_lines-- > 0) {
 		write_aux_reg(aux_cmd, paddr);
 		paddr += L1_CACHE_BYTES;
diff --git a/arch/arc/mm/tlb.c b/arch/arc/mm/tlb.c
index b5f28dc0f924..0ee739846847 100644
--- a/arch/arc/mm/tlb.c
+++ b/arch/arc/mm/tlb.c
@@ -109,6 +109,10 @@ DEFINE_PER_CPU(unsigned int, asid_cache) = MM_CTXT_FIRST_CYCLE;
 static inline void __tlb_entry_erase(void)
 {
 	write_aux_reg(ARC_REG_TLBPD1, 0);
+
+	if (is_pae40_enabled())
+		write_aux_reg(ARC_REG_TLBPD1HI, 0);
+
 	write_aux_reg(ARC_REG_TLBPD0, 0);
 	write_aux_reg(ARC_REG_TLBCOMMAND, TLBWrite);
 }
@@ -182,7 +186,7 @@ static void utlb_invalidate(void)
 
 }
 
-static void tlb_entry_insert(unsigned int pd0, unsigned int pd1)
+static void tlb_entry_insert(unsigned int pd0, pte_t pd1)
 {
 	unsigned int idx;
 
@@ -225,10 +229,14 @@ static void tlb_entry_erase(unsigned int vaddr_n_asid)
 	write_aux_reg(ARC_REG_TLBCOMMAND, TLBDeleteEntry);
 }
 
-static void tlb_entry_insert(unsigned int pd0, unsigned int pd1)
+static void tlb_entry_insert(unsigned int pd0, pte_t pd1)
 {
 	write_aux_reg(ARC_REG_TLBPD0, pd0);
 	write_aux_reg(ARC_REG_TLBPD1, pd1);
+
+	if (is_pae40_enabled())
+		write_aux_reg(ARC_REG_TLBPD1HI, (u64)pd1 >> 32);
+
 	write_aux_reg(ARC_REG_TLBCOMMAND, TLBInsertEntry);
 }
 
@@ -249,6 +257,10 @@ noinline void local_flush_tlb_all(void)
 
 	/* Load PD0 and PD1 with template for a Blank Entry */
 	write_aux_reg(ARC_REG_TLBPD1, 0);
+
+	if (is_pae40_enabled())
+		write_aux_reg(ARC_REG_TLBPD1HI, 0);
+
 	write_aux_reg(ARC_REG_TLBPD0, 0);
 
 	for (entry = 0; entry < num_tlb; entry++) {
@@ -503,7 +515,8 @@ void create_tlb(struct vm_area_struct *vma, unsigned long vaddr, pte_t *ptep)
 {
 	unsigned long flags;
 	unsigned int asid_or_sasid, rwx;
-	unsigned long pd0, pd1;
+	unsigned long pd0;
+	pte_t pd1;
 
 	/*
 	 * create_tlb() assumes that current->mm == vma->mm, since
@@ -785,10 +798,11 @@ char *arc_mmu_mumbojumbo(int cpu_id, char *buf, int len)
 			  IS_USED_CFG(CONFIG_TRANSPARENT_HUGEPAGE));
 
 	n += scnprintf(buf + n, len - n,
-		      "MMU [v%x]\t: %dK PAGE, %sJTLB %d (%dx%d), uDTLB %d, uITLB %d\n",
+		      "MMU [v%x]\t: %dk PAGE, %sJTLB %d (%dx%d), uDTLB %d, uITLB %d %s%s\n",
 		       p_mmu->ver, p_mmu->pg_sz_k, super_pg,
 		       p_mmu->sets * p_mmu->ways, p_mmu->sets, p_mmu->ways,
-		       p_mmu->u_dtlb, p_mmu->u_itlb);
+		       p_mmu->u_dtlb, p_mmu->u_itlb,
+		       IS_AVAIL2(p_mmu->pae, "PAE40 ", CONFIG_ARC_HAS_PAE40));
 
 	return buf;
 }
@@ -821,6 +835,9 @@ void arc_mmu_init(void)
 		panic("MMU Super pg size != Linux HPAGE_PMD_SIZE (%luM)\n",
 		      (unsigned long)TO_MB(HPAGE_PMD_SIZE));
 
+	if (IS_ENABLED(CONFIG_ARC_HAS_PAE40) && !mmu->pae)
+		panic("Hardware doesn't support PAE40\n");
+
 	/* Enable the MMU */
 	write_aux_reg(ARC_REG_PID, MMU_ENABLE);
 
diff --git a/arch/arc/mm/tlbex.S b/arch/arc/mm/tlbex.S
index 8d1b81990bc1..63860adc4814 100644
--- a/arch/arc/mm/tlbex.S
+++ b/arch/arc/mm/tlbex.S
@@ -223,12 +223,16 @@ ex_saved_reg1:
 	; (2) y = x & (PTRS_PER_PTE - 1) -> to get index
 	; (3) z = (pgtbl + y * 4)
 
+#ifdef CONFIG_ARC_HAS_PAE40
+#define PTE_SIZE_LOG	3	/* 8 == 2 ^ 3 */
+#else
 #define PTE_SIZE_LOG	2	/* 4 == 2 ^ 2 */
+#endif
 
 	; multiply in step (3) above avoided by shifting lesser in step (1)
 	lsr     r0, r2, ( PAGE_SHIFT - PTE_SIZE_LOG )
 	and     r0, r0, ( (PTRS_PER_PTE - 1) << PTE_SIZE_LOG )
-	ld.aw   r0, [r1, r0]		; r0: PTE
+	ld.aw   r0, [r1, r0]            ; r0: PTE (lower word only for PAE40)
 					; r1: PTE ptr
 
 2:
@@ -247,6 +251,7 @@ ex_saved_reg1:
 ;-----------------------------------------------------------------
 ; Convert Linux PTE entry into TLB entry
 ; A one-word PTE entry is programmed as two-word TLB Entry [PD0:PD1] in mmu
+;    (for PAE40, two-words PTE, while three-word TLB Entry [PD0:PD1:PD1HI])
 ; IN: r0 = PTE, r1 = ptr to PTE
 
 .macro CONV_PTE_TO_TLB
@@ -259,6 +264,10 @@ ex_saved_reg1:
 	or  r3, r3, r2
 
 	sr  r3, [ARC_REG_TLBPD1]    	; paddr[31..13] | Kr Kw Kx Ur Uw Ux | C
+#ifdef	CONFIG_ARC_HAS_PAE40
+	ld	r3, [r1, 4]		; paddr[39..32]
+	sr	r3, [ARC_REG_TLBPD1HI]
+#endif
 
 	and r2, r0, PTE_BITS_IN_PD0 ; Extract other PTE flags: (V)alid, (G)lb
 
-- 
cgit v1.2.3