Add the rt linux 4.1.3-rt3 as base

Import the rt linux 4.1.3-rt3 as OPNFV kvm base.

It's from git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-4.1.y-rt and
the base is:

commit 0917f823c59692d751951bf5ea699a2d1e2f26a2
Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Date:   Sat Jul 25 12:13:34 2015 +0200

    Prepare v4.1.3-rt3

    Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>

We lose all the git history this way and it's not good. We
should apply another opnfv project repo in future.

Change-Id: I87543d81c9df70d99c5001fbdf646b202c19f423
Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com>

9 files changed, 1235 insertions, 0 deletions

diff --git a/kernel/arch/avr32/mm/Makefile b/kernel/arch/avr32/mm/Makefile
new file mode 100644
index 000000000..0066491f9
--- /dev/null
+++ b/kernel/arch/avr32/mm/Makefile
@@ -0,0 +1,6 @@
+#
+# Makefile for the Linux/AVR32 kernel.
+#
+
+obj-y				+= init.o clear_page.o copy_page.o dma-coherent.o
+obj-y				+= ioremap.o cache.o fault.o tlb.o
diff --git a/kernel/arch/avr32/mm/cache.c b/kernel/arch/avr32/mm/cache.c
new file mode 100644
index 000000000..85d635cd7
--- /dev/null
+++ b/kernel/arch/avr32/mm/cache.c
@@ -0,0 +1,163 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/highmem.h>
+#include <linux/unistd.h>
+
+#include <asm/cacheflush.h>
+#include <asm/cachectl.h>
+#include <asm/processor.h>
+#include <asm/uaccess.h>
+#include <asm/syscalls.h>
+
+/*
+ * If you attempt to flush anything more than this, you need superuser
+ * privileges.  The value is completely arbitrary.
+ */
+#define CACHEFLUSH_MAX_LEN	1024
+
+void invalidate_dcache_region(void *start, size_t size)
+{
+	unsigned long v, begin, end, linesz, mask;
+
+	linesz = boot_cpu_data.dcache.linesz;
+	mask = linesz - 1;
+
+	/* when first and/or last cachelines are shared, flush them
+	 * instead of invalidating ... never discard valid data!
+	 */
+	begin = (unsigned long)start;
+	end = begin + size;
+
+	if (begin & mask) {
+		flush_dcache_line(start);
+		begin += linesz;
+	}
+	if (end & mask) {
+		flush_dcache_line((void *)end);
+		end &= ~mask;
+	}
+
+	/* remaining cachelines only need invalidation */
+	for (v = begin; v < end; v += linesz)
+		invalidate_dcache_line((void *)v);
+	flush_write_buffer();
+}
+
+void clean_dcache_region(void *start, size_t size)
+{
+	unsigned long v, begin, end, linesz;
+
+	linesz = boot_cpu_data.dcache.linesz;
+	begin = (unsigned long)start & ~(linesz - 1);
+	end = ((unsigned long)start + size + linesz - 1) & ~(linesz - 1);
+
+	for (v = begin; v < end; v += linesz)
+		clean_dcache_line((void *)v);
+	flush_write_buffer();
+}
+
+void flush_dcache_region(void *start, size_t size)
+{
+	unsigned long v, begin, end, linesz;
+
+	linesz = boot_cpu_data.dcache.linesz;
+	begin = (unsigned long)start & ~(linesz - 1);
+	end = ((unsigned long)start + size + linesz - 1) & ~(linesz - 1);
+
+	for (v = begin; v < end; v += linesz)
+		flush_dcache_line((void *)v);
+	flush_write_buffer();
+}
+
+void invalidate_icache_region(void *start, size_t size)
+{
+	unsigned long v, begin, end, linesz;
+
+	linesz = boot_cpu_data.icache.linesz;
+	begin = (unsigned long)start & ~(linesz - 1);
+	end = ((unsigned long)start + size + linesz - 1) & ~(linesz - 1);
+
+	for (v = begin; v < end; v += linesz)
+		invalidate_icache_line((void *)v);
+}
+
+static inline void __flush_icache_range(unsigned long start, unsigned long end)
+{
+	unsigned long v, linesz;
+
+	linesz = boot_cpu_data.dcache.linesz;
+	for (v = start; v < end; v += linesz) {
+		clean_dcache_line((void *)v);
+		invalidate_icache_line((void *)v);
+	}
+
+	flush_write_buffer();
+}
+
+/*
+ * This one is called after a module has been loaded.
+ */
+void flush_icache_range(unsigned long start, unsigned long end)
+{
+	unsigned long linesz;
+
+	linesz = boot_cpu_data.dcache.linesz;
+	__flush_icache_range(start & ~(linesz - 1),
+			     (end + linesz - 1) & ~(linesz - 1));
+}
+EXPORT_SYMBOL(flush_icache_range);
+
+/*
+ * This one is called from __do_fault() and do_swap_page().
+ */
+void flush_icache_page(struct vm_area_struct *vma, struct page *page)
+{
+	if (vma->vm_flags & VM_EXEC) {
+		void *v = page_address(page);
+		__flush_icache_range((unsigned long)v, (unsigned long)v + PAGE_SIZE);
+	}
+}
+
+asmlinkage int sys_cacheflush(int operation, void __user *addr, size_t len)
+{
+	int ret;
+
+	if (len > CACHEFLUSH_MAX_LEN) {
+		ret = -EPERM;
+		if (!capable(CAP_SYS_ADMIN))
+			goto out;
+	}
+
+	ret = -EFAULT;
+	if (!access_ok(VERIFY_WRITE, addr, len))
+		goto out;
+
+	switch (operation) {
+	case CACHE_IFLUSH:
+		flush_icache_range((unsigned long)addr,
+				   (unsigned long)addr + len);
+		ret = 0;
+		break;
+	default:
+		ret = -EINVAL;
+	}
+
+out:
+	return ret;
+}
+
+void copy_to_user_page(struct vm_area_struct *vma, struct page *page,
+		unsigned long vaddr, void *dst, const void *src,
+		unsigned long len)
+{
+	memcpy(dst, src, len);
+	if (vma->vm_flags & VM_EXEC)
+		flush_icache_range((unsigned long)dst,
+				(unsigned long)dst + len);
+}
diff --git a/kernel/arch/avr32/mm/clear_page.S b/kernel/arch/avr32/mm/clear_page.S
new file mode 100644
index 000000000..5d70dca00
--- /dev/null
+++ b/kernel/arch/avr32/mm/clear_page.S
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/page.h>
+
+/*
+ * clear_page
+ * r12: P1 address (to)
+ */
+	.text
+	.global clear_page
+clear_page:
+	sub	r9, r12, -PAGE_SIZE
+	mov     r10, 0
+	mov	r11, 0
+0:      st.d    r12++, r10
+	cp      r12, r9
+	brne	0b
+	mov     pc, lr
diff --git a/kernel/arch/avr32/mm/copy_page.S b/kernel/arch/avr32/mm/copy_page.S
new file mode 100644
index 000000000..c2b375294
--- /dev/null
+++ b/kernel/arch/avr32/mm/copy_page.S
@@ -0,0 +1,28 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/linkage.h>
+#include <asm/page.h>
+
+/*
+ * copy_page
+ *
+ * r12		to (P1 address)
+ * r11		from (P1 address)
+ * r8-r10	scratch
+ */
+	.text
+	.global copy_page
+copy_page:
+	sub	r10, r11, -(1 << PAGE_SHIFT)
+	/* pref	r11[0] */
+1:	/* pref	r11[8] */
+	ld.d	r8, r11++
+	st.d	r12++, r8
+	cp	r11, r10
+	brlo	1b
+	mov	pc, lr
diff --git a/kernel/arch/avr32/mm/dma-coherent.c b/kernel/arch/avr32/mm/dma-coherent.c
new file mode 100644
index 000000000..50cdb5b10
--- /dev/null
+++ b/kernel/arch/avr32/mm/dma-coherent.c
@@ -0,0 +1,152 @@
+/*
+ *  Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/gfp.h>
+#include <linux/export.h>
+
+#include <asm/addrspace.h>
+#include <asm/cacheflush.h>
+
+void dma_cache_sync(struct device *dev, void *vaddr, size_t size, int direction)
+{
+	/*
+	 * No need to sync an uncached area
+	 */
+	if (PXSEG(vaddr) == P2SEG)
+		return;
+
+	switch (direction) {
+	case DMA_FROM_DEVICE:		/* invalidate only */
+		invalidate_dcache_region(vaddr, size);
+		break;
+	case DMA_TO_DEVICE:		/* writeback only */
+		clean_dcache_region(vaddr, size);
+		break;
+	case DMA_BIDIRECTIONAL:		/* writeback and invalidate */
+		flush_dcache_region(vaddr, size);
+		break;
+	default:
+		BUG();
+	}
+}
+EXPORT_SYMBOL(dma_cache_sync);
+
+static struct page *__dma_alloc(struct device *dev, size_t size,
+				dma_addr_t *handle, gfp_t gfp)
+{
+	struct page *page, *free, *end;
+	int order;
+
+	/* Following is a work-around (a.k.a. hack) to prevent pages
+	 * with __GFP_COMP being passed to split_page() which cannot
+	 * handle them.  The real problem is that this flag probably
+	 * should be 0 on AVR32 as it is not supported on this
+	 * platform--see CONFIG_HUGETLB_PAGE. */
+	gfp &= ~(__GFP_COMP);
+
+	size = PAGE_ALIGN(size);
+	order = get_order(size);
+
+	page = alloc_pages(gfp, order);
+	if (!page)
+		return NULL;
+	split_page(page, order);
+
+	/*
+	 * When accessing physical memory with valid cache data, we
+	 * get a cache hit even if the virtual memory region is marked
+	 * as uncached.
+	 *
+	 * Since the memory is newly allocated, there is no point in
+	 * doing a writeback. If the previous owner cares, he should
+	 * have flushed the cache before releasing the memory.
+	 */
+	invalidate_dcache_region(phys_to_virt(page_to_phys(page)), size);
+
+	*handle = page_to_bus(page);
+	free = page + (size >> PAGE_SHIFT);
+	end = page + (1 << order);
+
+	/*
+	 * Free any unused pages
+	 */
+	while (free < end) {
+		__free_page(free);
+		free++;
+	}
+
+	return page;
+}
+
+static void __dma_free(struct device *dev, size_t size,
+		       struct page *page, dma_addr_t handle)
+{
+	struct page *end = page + (PAGE_ALIGN(size) >> PAGE_SHIFT);
+
+	while (page < end)
+		__free_page(page++);
+}
+
+void *dma_alloc_coherent(struct device *dev, size_t size,
+			 dma_addr_t *handle, gfp_t gfp)
+{
+	struct page *page;
+	void *ret = NULL;
+
+	page = __dma_alloc(dev, size, handle, gfp);
+	if (page)
+		ret = phys_to_uncached(page_to_phys(page));
+
+	return ret;
+}
+EXPORT_SYMBOL(dma_alloc_coherent);
+
+void dma_free_coherent(struct device *dev, size_t size,
+		       void *cpu_addr, dma_addr_t handle)
+{
+	void *addr = phys_to_cached(uncached_to_phys(cpu_addr));
+	struct page *page;
+
+	pr_debug("dma_free_coherent addr %p (phys %08lx) size %u\n",
+		 cpu_addr, (unsigned long)handle, (unsigned)size);
+	BUG_ON(!virt_addr_valid(addr));
+	page = virt_to_page(addr);
+	__dma_free(dev, size, page, handle);
+}
+EXPORT_SYMBOL(dma_free_coherent);
+
+void *dma_alloc_writecombine(struct device *dev, size_t size,
+			     dma_addr_t *handle, gfp_t gfp)
+{
+	struct page *page;
+	dma_addr_t phys;
+
+	page = __dma_alloc(dev, size, handle, gfp);
+	if (!page)
+		return NULL;
+
+	phys = page_to_phys(page);
+	*handle = phys;
+
+	/* Now, map the page into P3 with write-combining turned on */
+	return __ioremap(phys, size, _PAGE_BUFFER);
+}
+EXPORT_SYMBOL(dma_alloc_writecombine);
+
+void dma_free_writecombine(struct device *dev, size_t size,
+			   void *cpu_addr, dma_addr_t handle)
+{
+	struct page *page;
+
+	iounmap(cpu_addr);
+
+	page = phys_to_page(handle);
+	__dma_free(dev, size, page, handle);
+}
+EXPORT_SYMBOL(dma_free_writecombine);
diff --git a/kernel/arch/avr32/mm/fault.c b/kernel/arch/avr32/mm/fault.c
new file mode 100644
index 000000000..c03533937
--- /dev/null
+++ b/kernel/arch/avr32/mm/fault.c
@@ -0,0 +1,268 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * Based on linux/arch/sh/mm/fault.c:
+ *   Copyright (C) 1999  Niibe Yutaka
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/kdebug.h>
+#include <linux/kprobes.h>
+#include <linux/uaccess.h>
+
+#include <asm/mmu_context.h>
+#include <asm/sysreg.h>
+#include <asm/tlb.h>
+
+#ifdef CONFIG_KPROBES
+static inline int notify_page_fault(struct pt_regs *regs, int trap)
+{
+	int ret = 0;
+
+	if (!user_mode(regs)) {
+		if (kprobe_running() && kprobe_fault_handler(regs, trap))
+			ret = 1;
+	}
+
+	return ret;
+}
+#else
+static inline int notify_page_fault(struct pt_regs *regs, int trap)
+{
+	return 0;
+}
+#endif
+
+int exception_trace = 1;
+
+/*
+ * This routine handles page faults. It determines the address and the
+ * problem, and then passes it off to one of the appropriate routines.
+ *
+ * ecr is the Exception Cause Register. Possible values are:
+ *   6:  Protection fault (instruction access)
+ *   15: Protection fault (read access)
+ *   16: Protection fault (write access)
+ *   20: Page not found (instruction access)
+ *   24: Page not found (read access)
+ *   28: Page not found (write access)
+ */
+asmlinkage void do_page_fault(unsigned long ecr, struct pt_regs *regs)
+{
+	struct task_struct *tsk;
+	struct mm_struct *mm;
+	struct vm_area_struct *vma;
+	const struct exception_table_entry *fixup;
+	unsigned long address;
+	unsigned long page;
+	long signr;
+	int code;
+	int fault;
+	unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
+
+	if (notify_page_fault(regs, ecr))
+		return;
+
+	address = sysreg_read(TLBEAR);
+
+	tsk = current;
+	mm = tsk->mm;
+
+	signr = SIGSEGV;
+	code = SEGV_MAPERR;
+
+	/*
+	 * If we're in an interrupt or have no user context, we must
+	 * not take the fault...
+	 */
+	if (faulthandler_disabled() || !mm || regs->sr & SYSREG_BIT(GM))
+		goto no_context;
+
+	local_irq_enable();
+
+	if (user_mode(regs))
+		flags |= FAULT_FLAG_USER;
+retry:
+	down_read(&mm->mmap_sem);
+
+	vma = find_vma(mm, address);
+	if (!vma)
+		goto bad_area;
+	if (vma->vm_start <= address)
+		goto good_area;
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+	if (expand_stack(vma, address))
+		goto bad_area;
+
+	/*
+	 * Ok, we have a good vm_area for this memory access, so we
+	 * can handle it...
+	 */
+good_area:
+	code = SEGV_ACCERR;
+
+	switch (ecr) {
+	case ECR_PROTECTION_X:
+	case ECR_TLB_MISS_X:
+		if (!(vma->vm_flags & VM_EXEC))
+			goto bad_area;
+		break;
+	case ECR_PROTECTION_R:
+	case ECR_TLB_MISS_R:
+		if (!(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)))
+			goto bad_area;
+		break;
+	case ECR_PROTECTION_W:
+	case ECR_TLB_MISS_W:
+		if (!(vma->vm_flags & VM_WRITE))
+			goto bad_area;
+		flags |= FAULT_FLAG_WRITE;
+		break;
+	default:
+		panic("Unhandled case %lu in do_page_fault!", ecr);
+	}
+
+	/*
+	 * If for any reason at all we couldn't handle the fault, make
+	 * sure we exit gracefully rather than endlessly redo the
+	 * fault.
+	 */
+	fault = handle_mm_fault(mm, vma, address, flags);
+
+	if ((fault & VM_FAULT_RETRY) && fatal_signal_pending(current))
+		return;
+
+	if (unlikely(fault & VM_FAULT_ERROR)) {
+		if (fault & VM_FAULT_OOM)
+			goto out_of_memory;
+		else if (fault & VM_FAULT_SIGSEGV)
+			goto bad_area;
+		else if (fault & VM_FAULT_SIGBUS)
+			goto do_sigbus;
+		BUG();
+	}
+
+	if (flags & FAULT_FLAG_ALLOW_RETRY) {
+		if (fault & VM_FAULT_MAJOR)
+			tsk->maj_flt++;
+		else
+			tsk->min_flt++;
+		if (fault & VM_FAULT_RETRY) {
+			flags &= ~FAULT_FLAG_ALLOW_RETRY;
+			flags |= FAULT_FLAG_TRIED;
+
+			/*
+			 * No need to up_read(&mm->mmap_sem) as we would have
+			 * already released it in __lock_page_or_retry() in
+			 * mm/filemap.c.
+			 */
+			goto retry;
+		}
+	}
+
+	up_read(&mm->mmap_sem);
+	return;
+
+	/*
+	 * Something tried to access memory that isn't in our memory
+	 * map. Fix it, but check if it's kernel or user first...
+	 */
+bad_area:
+	up_read(&mm->mmap_sem);
+
+	if (user_mode(regs)) {
+		if (exception_trace && printk_ratelimit())
+			printk("%s%s[%d]: segfault at %08lx pc %08lx "
+			       "sp %08lx ecr %lu\n",
+			       is_global_init(tsk) ? KERN_EMERG : KERN_INFO,
+			       tsk->comm, tsk->pid, address, regs->pc,
+			       regs->sp, ecr);
+		_exception(SIGSEGV, regs, code, address);
+		return;
+	}
+
+no_context:
+	/* Are we prepared to handle this kernel fault? */
+	fixup = search_exception_tables(regs->pc);
+	if (fixup) {
+		regs->pc = fixup->fixup;
+		return;
+	}
+
+	/*
+	 * Oops. The kernel tried to access some bad page. We'll have
+	 * to terminate things with extreme prejudice.
+	 */
+	if (address < PAGE_SIZE)
+		printk(KERN_ALERT
+		       "Unable to handle kernel NULL pointer dereference");
+	else
+		printk(KERN_ALERT
+		       "Unable to handle kernel paging request");
+	printk(" at virtual address %08lx\n", address);
+
+	page = sysreg_read(PTBR);
+	printk(KERN_ALERT "ptbr = %08lx", page);
+	if (address >= TASK_SIZE)
+		page = (unsigned long)swapper_pg_dir;
+	if (page) {
+		page = ((unsigned long *)page)[address >> 22];
+		printk(" pgd = %08lx", page);
+		if (page & _PAGE_PRESENT) {
+			page &= PAGE_MASK;
+			address &= 0x003ff000;
+			page = ((unsigned long *)__va(page))[address >> PAGE_SHIFT];
+			printk(" pte = %08lx", page);
+		}
+	}
+	printk("\n");
+	die("Kernel access of bad area", regs, signr);
+	return;
+
+	/*
+	 * We ran out of memory, or some other thing happened to us
+	 * that made us unable to handle the page fault gracefully.
+	 */
+out_of_memory:
+	up_read(&mm->mmap_sem);
+	if (!user_mode(regs))
+		goto no_context;
+	pagefault_out_of_memory();
+	return;
+
+do_sigbus:
+	up_read(&mm->mmap_sem);
+
+	/* Kernel mode? Handle exceptions or die */
+	signr = SIGBUS;
+	code = BUS_ADRERR;
+	if (!user_mode(regs))
+		goto no_context;
+
+	if (exception_trace)
+		printk("%s%s[%d]: bus error at %08lx pc %08lx "
+		       "sp %08lx ecr %lu\n",
+		       is_global_init(tsk) ? KERN_EMERG : KERN_INFO,
+		       tsk->comm, tsk->pid, address, regs->pc,
+		       regs->sp, ecr);
+
+	_exception(SIGBUS, regs, BUS_ADRERR, address);
+}
+
+asmlinkage void do_bus_error(unsigned long addr, int write_access,
+			     struct pt_regs *regs)
+{
+	printk(KERN_ALERT
+	       "Bus error at physical address 0x%08lx (%s access)\n",
+	       addr, write_access ? "write" : "read");
+	printk(KERN_INFO "DTLB dump:\n");
+	dump_dtlb();
+	die("Bus Error", regs, SIGKILL);
+}
diff --git a/kernel/arch/avr32/mm/init.c b/kernel/arch/avr32/mm/init.c
new file mode 100644
index 000000000..def5391d9
--- /dev/null
+++ b/kernel/arch/avr32/mm/init.c
@@ -0,0 +1,125 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/gfp.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/init.h>
+#include <linux/mmzone.h>
+#include <linux/module.h>
+#include <linux/bootmem.h>
+#include <linux/pagemap.h>
+#include <linux/nodemask.h>
+
+#include <asm/page.h>
+#include <asm/mmu_context.h>
+#include <asm/tlb.h>
+#include <asm/io.h>
+#include <asm/dma.h>
+#include <asm/setup.h>
+#include <asm/sections.h>
+
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __page_aligned_data;
+
+struct page *empty_zero_page;
+EXPORT_SYMBOL(empty_zero_page);
+
+/*
+ * Cache of MMU context last used.
+ */
+unsigned long mmu_context_cache = NO_CONTEXT;
+
+/*
+ * paging_init() sets up the page tables
+ *
+ * This routine also unmaps the page at virtual kernel address 0, so
+ * that we can trap those pesky NULL-reference errors in the kernel.
+ */
+void __init paging_init(void)
+{
+	extern unsigned long _evba;
+	void *zero_page;
+	int nid;
+
+	/*
+	 * Make sure we can handle exceptions before enabling
+	 * paging. Not that we should ever _get_ any exceptions this
+	 * early, but you never know...
+	 */
+	printk("Exception vectors start at %p\n", &_evba);
+	sysreg_write(EVBA, (unsigned long)&_evba);
+
+	/*
+	 * Since we are ready to handle exceptions now, we should let
+	 * the CPU generate them...
+	 */
+	__asm__ __volatile__ ("csrf %0" : : "i"(SR_EM_BIT));
+
+	/*
+	 * Allocate the zero page. The allocator will panic if it
+	 * can't satisfy the request, so no need to check.
+	 */
+	zero_page = alloc_bootmem_low_pages_node(NODE_DATA(0),
+						 PAGE_SIZE);
+
+	sysreg_write(PTBR, (unsigned long)swapper_pg_dir);
+	enable_mmu();
+	printk ("CPU: Paging enabled\n");
+
+	for_each_online_node(nid) {
+		pg_data_t *pgdat = NODE_DATA(nid);
+		unsigned long zones_size[MAX_NR_ZONES];
+		unsigned long low, start_pfn;
+
+		start_pfn = pgdat->bdata->node_min_pfn;
+		low = pgdat->bdata->node_low_pfn;
+
+		memset(zones_size, 0, sizeof(zones_size));
+		zones_size[ZONE_NORMAL] = low - start_pfn;
+
+		printk("Node %u: start_pfn = 0x%lx, low = 0x%lx\n",
+		       nid, start_pfn, low);
+
+		free_area_init_node(nid, zones_size, start_pfn, NULL);
+
+		printk("Node %u: mem_map starts at %p\n",
+		       pgdat->node_id, pgdat->node_mem_map);
+	}
+
+	mem_map = NODE_DATA(0)->node_mem_map;
+
+	empty_zero_page = virt_to_page(zero_page);
+	flush_dcache_page(empty_zero_page);
+}
+
+void __init mem_init(void)
+{
+	pg_data_t *pgdat;
+
+	high_memory = NULL;
+	for_each_online_pgdat(pgdat)
+		high_memory = max_t(void *, high_memory,
+				    __va(pgdat_end_pfn(pgdat) << PAGE_SHIFT));
+
+	set_max_mapnr(MAP_NR(high_memory));
+	free_all_bootmem();
+	mem_init_print_info(NULL);
+}
+
+void free_initmem(void)
+{
+	free_initmem_default(-1);
+}
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void free_initrd_mem(unsigned long start, unsigned long end)
+{
+	free_reserved_area((void *)start, (void *)end, -1, "initrd");
+}
+#endif
diff --git a/kernel/arch/avr32/mm/ioremap.c b/kernel/arch/avr32/mm/ioremap.c
new file mode 100644
index 000000000..7def0d84c
--- /dev/null
+++ b/kernel/arch/avr32/mm/ioremap.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+
+#include <asm/pgtable.h>
+#include <asm/addrspace.h>
+
+/*
+ * Re-map an arbitrary physical address space into the kernel virtual
+ * address space. Needed when the kernel wants to access physical
+ * memory directly.
+ */
+void __iomem *__ioremap(unsigned long phys_addr, size_t size,
+			unsigned long flags)
+{
+	unsigned long addr;
+	struct vm_struct *area;
+	unsigned long offset, last_addr;
+	pgprot_t prot;
+
+	/*
+	 * Check if we can simply use the P4 segment. This area is
+	 * uncacheable, so if caching/buffering is requested, we can't
+	 * use it.
+	 */
+	if ((phys_addr >= P4SEG) && (flags == 0))
+		return (void __iomem *)phys_addr;
+
+	/* Don't allow wraparound or zero size */
+	last_addr = phys_addr + size - 1;
+	if (!size || last_addr < phys_addr)
+		return NULL;
+
+	/*
+	 * XXX: When mapping regular RAM, we'd better make damn sure
+	 * it's never used for anything else.  But this is really the
+	 * caller's responsibility...
+	 */
+	if (PHYSADDR(P2SEGADDR(phys_addr)) == phys_addr)
+		return (void __iomem *)P2SEGADDR(phys_addr);
+
+	/* Mappings have to be page-aligned */
+	offset = phys_addr & ~PAGE_MASK;
+	phys_addr &= PAGE_MASK;
+	size = PAGE_ALIGN(last_addr + 1) - phys_addr;
+
+	prot = __pgprot(_PAGE_PRESENT | _PAGE_GLOBAL | _PAGE_RW | _PAGE_DIRTY
+			| _PAGE_ACCESSED | _PAGE_TYPE_SMALL | flags);
+
+	/*
+	 * Ok, go for it..
+	 */
+	area = get_vm_area(size, VM_IOREMAP);
+	if (!area)
+		return NULL;
+	area->phys_addr = phys_addr;
+	addr = (unsigned long )area->addr;
+	if (ioremap_page_range(addr, addr + size, phys_addr, prot)) {
+		vunmap((void *)addr);
+		return NULL;
+	}
+
+	return (void __iomem *)(offset + (char *)addr);
+}
+EXPORT_SYMBOL(__ioremap);
+
+void __iounmap(void __iomem *addr)
+{
+	struct vm_struct *p;
+
+	if ((unsigned long)addr >= P4SEG)
+		return;
+	if (PXSEG(addr) == P2SEG)
+		return;
+
+	p = remove_vm_area((void *)(PAGE_MASK & (unsigned long __force)addr));
+	if (unlikely(!p)) {
+		printk (KERN_ERR "iounmap: bad address %p\n", addr);
+		return;
+	}
+
+	kfree (p);
+}
+EXPORT_SYMBOL(__iounmap);
diff --git a/kernel/arch/avr32/mm/tlb.c b/kernel/arch/avr32/mm/tlb.c
new file mode 100644
index 000000000..0da23109f
--- /dev/null
+++ b/kernel/arch/avr32/mm/tlb.c
@@ -0,0 +1,375 @@
+/*
+ * AVR32 TLB operations
+ *
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/mm.h>
+
+#include <asm/mmu_context.h>
+
+/* TODO: Get the correct number from the CONFIG1 system register */
+#define NR_TLB_ENTRIES 32
+
+static void show_dtlb_entry(unsigned int index)
+{
+	u32 tlbehi, tlbehi_save, tlbelo, mmucr, mmucr_save;
+	unsigned long flags;
+
+	local_irq_save(flags);
+	mmucr_save = sysreg_read(MMUCR);
+	tlbehi_save = sysreg_read(TLBEHI);
+	mmucr = SYSREG_BFINS(DRP, index, mmucr_save);
+	sysreg_write(MMUCR, mmucr);
+
+	__builtin_tlbr();
+	cpu_sync_pipeline();
+
+	tlbehi = sysreg_read(TLBEHI);
+	tlbelo = sysreg_read(TLBELO);
+
+	printk("%2u: %c %c %02x   %05x %05x %o  %o  %c %c %c %c\n",
+	       index,
+	       SYSREG_BFEXT(TLBEHI_V, tlbehi) ? '1' : '0',
+	       SYSREG_BFEXT(G, tlbelo) ? '1' : '0',
+	       SYSREG_BFEXT(ASID, tlbehi),
+	       SYSREG_BFEXT(VPN, tlbehi) >> 2,
+	       SYSREG_BFEXT(PFN, tlbelo) >> 2,
+	       SYSREG_BFEXT(AP, tlbelo),
+	       SYSREG_BFEXT(SZ, tlbelo),
+	       SYSREG_BFEXT(TLBELO_C, tlbelo) ? 'C' : ' ',
+	       SYSREG_BFEXT(B, tlbelo) ? 'B' : ' ',
+	       SYSREG_BFEXT(W, tlbelo) ? 'W' : ' ',
+	       SYSREG_BFEXT(TLBELO_D, tlbelo) ? 'D' : ' ');
+
+	sysreg_write(MMUCR, mmucr_save);
+	sysreg_write(TLBEHI, tlbehi_save);
+	cpu_sync_pipeline();
+	local_irq_restore(flags);
+}
+
+void dump_dtlb(void)
+{
+	unsigned int i;
+
+	printk("ID  V G ASID VPN   PFN   AP SZ C B W D\n");
+	for (i = 0; i < NR_TLB_ENTRIES; i++)
+		show_dtlb_entry(i);
+}
+
+static void update_dtlb(unsigned long address, pte_t pte)
+{
+	u32 tlbehi;
+	u32 mmucr;
+
+	/*
+	 * We're not changing the ASID here, so no need to flush the
+	 * pipeline.
+	 */
+	tlbehi = sysreg_read(TLBEHI);
+	tlbehi = SYSREG_BF(ASID, SYSREG_BFEXT(ASID, tlbehi));
+	tlbehi |= address & MMU_VPN_MASK;
+	tlbehi |= SYSREG_BIT(TLBEHI_V);
+	sysreg_write(TLBEHI, tlbehi);
+
+	/* Does this mapping already exist? */
+	__builtin_tlbs();
+	mmucr = sysreg_read(MMUCR);
+
+	if (mmucr & SYSREG_BIT(MMUCR_N)) {
+		/* Not found -- pick a not-recently-accessed entry */
+		unsigned int rp;
+		u32 tlbar = sysreg_read(TLBARLO);
+
+		rp = 32 - fls(tlbar);
+		if (rp == 32) {
+			rp = 0;
+			sysreg_write(TLBARLO, -1L);
+		}
+
+		mmucr = SYSREG_BFINS(DRP, rp, mmucr);
+		sysreg_write(MMUCR, mmucr);
+	}
+
+	sysreg_write(TLBELO, pte_val(pte) & _PAGE_FLAGS_HARDWARE_MASK);
+
+	/* Let's go */
+	__builtin_tlbw();
+}
+
+void update_mmu_cache(struct vm_area_struct *vma,
+		      unsigned long address, pte_t *ptep)
+{
+	unsigned long flags;
+
+	/* ptrace may call this routine */
+	if (vma && current->active_mm != vma->vm_mm)
+		return;
+
+	local_irq_save(flags);
+	update_dtlb(address, *ptep);
+	local_irq_restore(flags);
+}
+
+static void __flush_tlb_page(unsigned long asid, unsigned long page)
+{
+	u32 mmucr, tlbehi;
+
+	/*
+	 * Caller is responsible for masking out non-PFN bits in page
+	 * and changing the current ASID if necessary. This means that
+	 * we don't need to flush the pipeline after writing TLBEHI.
+	 */
+	tlbehi = page | asid;
+	sysreg_write(TLBEHI, tlbehi);
+
+	__builtin_tlbs();
+	mmucr = sysreg_read(MMUCR);
+
+	if (!(mmucr & SYSREG_BIT(MMUCR_N))) {
+		unsigned int entry;
+		u32 tlbarlo;
+
+		/* Clear the "valid" bit */
+		sysreg_write(TLBEHI, tlbehi);
+
+		/* mark the entry as "not accessed" */
+		entry = SYSREG_BFEXT(DRP, mmucr);
+		tlbarlo = sysreg_read(TLBARLO);
+		tlbarlo |= (0x80000000UL >> entry);
+		sysreg_write(TLBARLO, tlbarlo);
+
+		/* update the entry with valid bit clear */
+		__builtin_tlbw();
+	}
+}
+
+void flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
+{
+	if (vma->vm_mm && vma->vm_mm->context != NO_CONTEXT) {
+		unsigned long flags, asid;
+		unsigned long saved_asid = MMU_NO_ASID;
+
+		asid = vma->vm_mm->context & MMU_CONTEXT_ASID_MASK;
+		page &= PAGE_MASK;
+
+		local_irq_save(flags);
+		if (vma->vm_mm != current->mm) {
+			saved_asid = get_asid();
+			set_asid(asid);
+		}
+
+		__flush_tlb_page(asid, page);
+
+		if (saved_asid != MMU_NO_ASID)
+			set_asid(saved_asid);
+		local_irq_restore(flags);
+	}
+}
+
+void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
+		     unsigned long end)
+{
+	struct mm_struct *mm = vma->vm_mm;
+
+	if (mm->context != NO_CONTEXT) {
+		unsigned long flags;
+		int size;
+
+		local_irq_save(flags);
+		size = (end - start + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+
+		if (size > (MMU_DTLB_ENTRIES / 4)) { /* Too many entries to flush */
+			mm->context = NO_CONTEXT;
+			if (mm == current->mm)
+				activate_context(mm);
+		} else {
+			unsigned long asid;
+			unsigned long saved_asid;
+
+			asid = mm->context & MMU_CONTEXT_ASID_MASK;
+			saved_asid = MMU_NO_ASID;
+
+			start &= PAGE_MASK;
+			end += (PAGE_SIZE - 1);
+			end &= PAGE_MASK;
+
+			if (mm != current->mm) {
+				saved_asid = get_asid();
+				set_asid(asid);
+			}
+
+			while (start < end) {
+				__flush_tlb_page(asid, start);
+				start += PAGE_SIZE;
+			}
+			if (saved_asid != MMU_NO_ASID)
+				set_asid(saved_asid);
+		}
+		local_irq_restore(flags);
+	}
+}
+
+/*
+ * This function depends on the pages to be flushed having the G
+ * (global) bit set in their pte. This is true for all
+ * PAGE_KERNEL(_RO) pages.
+ */
+void flush_tlb_kernel_range(unsigned long start, unsigned long end)
+{
+	unsigned long flags;
+	int size;
+
+	size = (end - start + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+	if (size > (MMU_DTLB_ENTRIES / 4)) { /* Too many entries to flush */
+		flush_tlb_all();
+	} else {
+		unsigned long asid;
+
+		local_irq_save(flags);
+		asid = get_asid();
+
+		start &= PAGE_MASK;
+		end += (PAGE_SIZE - 1);
+		end &= PAGE_MASK;
+
+		while (start < end) {
+			__flush_tlb_page(asid, start);
+			start += PAGE_SIZE;
+		}
+		local_irq_restore(flags);
+	}
+}
+
+void flush_tlb_mm(struct mm_struct *mm)
+{
+	/* Invalidate all TLB entries of this process by getting a new ASID */
+	if (mm->context != NO_CONTEXT) {
+		unsigned long flags;
+
+		local_irq_save(flags);
+		mm->context = NO_CONTEXT;
+		if (mm == current->mm)
+			activate_context(mm);
+		local_irq_restore(flags);
+	}
+}
+
+void flush_tlb_all(void)
+{
+	unsigned long flags;
+
+	local_irq_save(flags);
+	sysreg_write(MMUCR, sysreg_read(MMUCR) | SYSREG_BIT(MMUCR_I));
+	local_irq_restore(flags);
+}
+
+#ifdef CONFIG_PROC_FS
+
+#include <linux/seq_file.h>
+#include <linux/proc_fs.h>
+#include <linux/init.h>
+
+static void *tlb_start(struct seq_file *tlb, loff_t *pos)
+{
+	static unsigned long tlb_index;
+
+	if (*pos >= NR_TLB_ENTRIES)
+		return NULL;
+
+	tlb_index = 0;
+	return &tlb_index;
+}
+
+static void *tlb_next(struct seq_file *tlb, void *v, loff_t *pos)
+{
+	unsigned long *index = v;
+
+	if (*index >= NR_TLB_ENTRIES - 1)
+		return NULL;
+
+	++*pos;
+	++*index;
+	return index;
+}
+
+static void tlb_stop(struct seq_file *tlb, void *v)
+{
+
+}
+
+static int tlb_show(struct seq_file *tlb, void *v)
+{
+	unsigned int tlbehi, tlbehi_save, tlbelo, mmucr, mmucr_save;
+	unsigned long flags;
+	unsigned long *index = v;
+
+	if (*index == 0)
+		seq_puts(tlb, "ID  V G ASID VPN   PFN   AP SZ C B W D\n");
+
+	BUG_ON(*index >= NR_TLB_ENTRIES);
+
+	local_irq_save(flags);
+	mmucr_save = sysreg_read(MMUCR);
+	tlbehi_save = sysreg_read(TLBEHI);
+	mmucr = SYSREG_BFINS(DRP, *index, mmucr_save);
+	sysreg_write(MMUCR, mmucr);
+
+	/* TLBR might change the ASID */
+	__builtin_tlbr();
+	cpu_sync_pipeline();
+
+	tlbehi = sysreg_read(TLBEHI);
+	tlbelo = sysreg_read(TLBELO);
+
+	sysreg_write(MMUCR, mmucr_save);
+	sysreg_write(TLBEHI, tlbehi_save);
+	cpu_sync_pipeline();
+	local_irq_restore(flags);
+
+	seq_printf(tlb, "%2lu: %c %c %02x   %05x %05x %o  %o  %c %c %c %c\n",
+		   *index,
+		   SYSREG_BFEXT(TLBEHI_V, tlbehi) ? '1' : '0',
+		   SYSREG_BFEXT(G, tlbelo) ? '1' : '0',
+		   SYSREG_BFEXT(ASID, tlbehi),
+		   SYSREG_BFEXT(VPN, tlbehi) >> 2,
+		   SYSREG_BFEXT(PFN, tlbelo) >> 2,
+		   SYSREG_BFEXT(AP, tlbelo),
+		   SYSREG_BFEXT(SZ, tlbelo),
+		   SYSREG_BFEXT(TLBELO_C, tlbelo) ? '1' : '0',
+		   SYSREG_BFEXT(B, tlbelo) ? '1' : '0',
+		   SYSREG_BFEXT(W, tlbelo) ? '1' : '0',
+		   SYSREG_BFEXT(TLBELO_D, tlbelo) ? '1' : '0');
+
+	return 0;
+}
+
+static const struct seq_operations tlb_ops = {
+	.start		= tlb_start,
+	.next		= tlb_next,
+	.stop		= tlb_stop,
+	.show		= tlb_show,
+};
+
+static int tlb_open(struct inode *inode, struct file *file)
+{
+	return seq_open(file, &tlb_ops);
+}
+
+static const struct file_operations proc_tlb_operations = {
+	.open		= tlb_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+
+static int __init proctlb_init(void)
+{
+	proc_create("tlb", 0, NULL, &proc_tlb_operations);
+	return 0;
+}
+late_initcall(proctlb_init);
+#endif /* CONFIG_PROC_FS */