[deliverable/linux.git] / arch / arm64 / mm / mmu.c

/*
 * Based on arch/arm/mm/mmu.c
 *
 * Copyright (C) 1995-2005 Russell King
 * Copyright (C) 2012 ARM Ltd.
 *
 * 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.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */

#include <linux/export.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/mman.h>
#include <linux/nodemask.h>
#include <linux/memblock.h>
#include <linux/fs.h>
#include <linux/io.h>

#include <asm/cputype.h>
#include <asm/fixmap.h>
#include <asm/sections.h>
#include <asm/setup.h>
#include <asm/sizes.h>
#include <asm/tlb.h>
#include <asm/memblock.h>
#include <asm/mmu_context.h>

#include "mm.h"

/*
 * Empty_zero_page is a special page that is used for zero-initialized data
 * and COW.
 */
struct page *empty_zero_page;
EXPORT_SYMBOL(empty_zero_page);

pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn,
			      unsigned long size, pgprot_t vma_prot)
{
	if (!pfn_valid(pfn))
		return pgprot_noncached(vma_prot);
	else if (file->f_flags & O_SYNC)
		return pgprot_writecombine(vma_prot);
	return vma_prot;
}
EXPORT_SYMBOL(phys_mem_access_prot);

static void __init *early_alloc(unsigned long sz)
{
	void *ptr = __va(memblock_alloc(sz, sz));
	memset(ptr, 0, sz);
	return ptr;
}

static void __init alloc_init_pte(pmd_t *pmd, unsigned long addr,
				  unsigned long end, unsigned long pfn,
				  pgprot_t prot)
{
	pte_t *pte;

	if (pmd_none(*pmd)) {
		pte = early_alloc(PTRS_PER_PTE * sizeof(pte_t));
		__pmd_populate(pmd, __pa(pte), PMD_TYPE_TABLE);
	}
	BUG_ON(pmd_bad(*pmd));

	pte = pte_offset_kernel(pmd, addr);
	do {
		set_pte(pte, pfn_pte(pfn, prot));
		pfn++;
	} while (pte++, addr += PAGE_SIZE, addr != end);
}

static void __init alloc_init_pmd(struct mm_struct *mm, pud_t *pud,
				  unsigned long addr, unsigned long end,
				  phys_addr_t phys, pgprot_t prot)
{
	pmd_t *pmd;
	unsigned long next;

	/*
	 * Check for initial section mappings in the pgd/pud and remove them.
	 */
	if (pud_none(*pud) || pud_bad(*pud)) {
		pmd = early_alloc(PTRS_PER_PMD * sizeof(pmd_t));
		pud_populate(mm, pud, pmd);
	}

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		/* try section mapping first */
		if (((addr | next | phys) & ~SECTION_MASK) == 0) {
			pmd_t old_pmd =*pmd;
			set_pmd(pmd, __pmd(phys |
					   pgprot_val(mk_sect_prot(prot))));
			/*
			 * Check for previous table entries created during
			 * boot (__create_page_tables) and flush them.
			 */
			if (!pmd_none(old_pmd))
				flush_tlb_all();
		} else {
			alloc_init_pte(pmd, addr, next, __phys_to_pfn(phys),
				       prot);
		}
		phys += next - addr;
	} while (pmd++, addr = next, addr != end);
}

static void __init alloc_init_pud(struct mm_struct *mm, pgd_t *pgd,
				  unsigned long addr, unsigned long end,
				  phys_addr_t phys, pgprot_t prot)
{
	pud_t *pud;
	unsigned long next;

	if (pgd_none(*pgd)) {
		pud = early_alloc(PTRS_PER_PUD * sizeof(pud_t));
		pgd_populate(mm, pgd, pud);
	}
	BUG_ON(pgd_bad(*pgd));

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);

		/*
		 * For 4K granule only, attempt to put down a 1GB block
		 */
		if ((PAGE_SHIFT == 12) &&
		    ((addr | next | phys) & ~PUD_MASK) == 0) {
			pud_t old_pud = *pud;
			set_pud(pud, __pud(phys |
					   pgprot_val(mk_sect_prot(prot))));

			/*
			 * If we have an old value for a pud, it will
			 * be pointing to a pmd table that we no longer
			 * need (from swapper_pg_dir).
			 *
			 * Look up the old pmd table and free it.
			 */
			if (!pud_none(old_pud)) {
				phys_addr_t table = __pa(pmd_offset(&old_pud, 0));
				memblock_free(table, PAGE_SIZE);
				flush_tlb_all();
			}
		} else {
			alloc_init_pmd(mm, pud, addr, next, phys, prot);
		}
		phys += next - addr;
	} while (pud++, addr = next, addr != end);
}

/*
 * Create the page directory entries and any necessary page tables for the
 * mapping specified by 'md'.
 */
static void __init __create_mapping(struct mm_struct *mm, pgd_t *pgd,
				    phys_addr_t phys, unsigned long virt,
				    phys_addr_t size, pgprot_t prot)
{
	unsigned long addr, length, end, next;

	addr = virt & PAGE_MASK;
	length = PAGE_ALIGN(size + (virt & ~PAGE_MASK));

	end = addr + length;
	do {
		next = pgd_addr_end(addr, end);
		alloc_init_pud(mm, pgd, addr, next, phys, prot);
		phys += next - addr;
	} while (pgd++, addr = next, addr != end);
}

static void __init create_mapping(phys_addr_t phys, unsigned long virt,
				  phys_addr_t size)
{
	if (virt < VMALLOC_START) {
		pr_warn("BUG: not creating mapping for %pa at 0x%016lx - outside kernel range\n",
			&phys, virt);
		return;
	}
	__create_mapping(&init_mm, pgd_offset_k(virt & PAGE_MASK), phys, virt,
			 size, PAGE_KERNEL_EXEC);
}

void __init create_pgd_mapping(struct mm_struct *mm, phys_addr_t phys,
			       unsigned long virt, phys_addr_t size,
			       pgprot_t prot)
{
	__create_mapping(mm, pgd_offset(mm, virt), phys, virt, size, prot);
}

static void __init map_mem(void)
{
	struct memblock_region *reg;
	phys_addr_t limit;

	/*
	 * Temporarily limit the memblock range. We need to do this as
	 * create_mapping requires puds, pmds and ptes to be allocated from
	 * memory addressable from the initial direct kernel mapping.
	 *
	 * The initial direct kernel mapping, located at swapper_pg_dir, gives
	 * us PUD_SIZE (4K pages) or PMD_SIZE (64K pages) memory starting from
	 * PHYS_OFFSET (which must be aligned to 2MB as per
	 * Documentation/arm64/booting.txt).
	 */
	if (IS_ENABLED(CONFIG_ARM64_64K_PAGES))
		limit = PHYS_OFFSET + PMD_SIZE;
	else
		limit = PHYS_OFFSET + PUD_SIZE;
	memblock_set_current_limit(limit);

	/* map all the memory banks */
	for_each_memblock(memory, reg) {
		phys_addr_t start = reg->base;
		phys_addr_t end = start + reg->size;

		if (start >= end)
			break;

#ifndef CONFIG_ARM64_64K_PAGES
		/*
		 * For the first memory bank align the start address and
		 * current memblock limit to prevent create_mapping() from
		 * allocating pte page tables from unmapped memory.
		 * When 64K pages are enabled, the pte page table for the
		 * first PGDIR_SIZE is already present in swapper_pg_dir.
		 */
		if (start < limit)
			start = ALIGN(start, PMD_SIZE);
		if (end < limit) {
			limit = end & PMD_MASK;
			memblock_set_current_limit(limit);
		}
#endif

		create_mapping(start, __phys_to_virt(start), end - start);
	}

	/* Limit no longer required. */
	memblock_set_current_limit(MEMBLOCK_ALLOC_ANYWHERE);
}

/*
 * paging_init() sets up the page tables, initialises the zone memory
 * maps and sets up the zero page.
 */
void __init paging_init(void)
{
	void *zero_page;

	map_mem();

	/*
	 * Finally flush the caches and tlb to ensure that we're in a
	 * consistent state.
	 */
	flush_cache_all();
	flush_tlb_all();

	/* allocate the zero page. */
	zero_page = early_alloc(PAGE_SIZE);

	bootmem_init();

	empty_zero_page = virt_to_page(zero_page);

	/*
	 * TTBR0 is only used for the identity mapping at this stage. Make it
	 * point to zero page to avoid speculatively fetching new entries.
	 */
	cpu_set_reserved_ttbr0();
	flush_tlb_all();
}

/*
 * Enable the identity mapping to allow the MMU disabling.
 */
void setup_mm_for_reboot(void)
{
	cpu_switch_mm(idmap_pg_dir, &init_mm);
	flush_tlb_all();
}

/*
 * Check whether a kernel address is valid (derived from arch/x86/).
 */
int kern_addr_valid(unsigned long addr)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

	if ((((long)addr) >> VA_BITS) != -1UL)
		return 0;

	pgd = pgd_offset_k(addr);
	if (pgd_none(*pgd))
		return 0;

	pud = pud_offset(pgd, addr);
	if (pud_none(*pud))
		return 0;

	if (pud_sect(*pud))
		return pfn_valid(pud_pfn(*pud));

	pmd = pmd_offset(pud, addr);
	if (pmd_none(*pmd))
		return 0;

	if (pmd_sect(*pmd))
		return pfn_valid(pmd_pfn(*pmd));

	pte = pte_offset_kernel(pmd, addr);
	if (pte_none(*pte))
		return 0;

	return pfn_valid(pte_pfn(*pte));
}
#ifdef CONFIG_SPARSEMEM_VMEMMAP
#ifdef CONFIG_ARM64_64K_PAGES
int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
{
	return vmemmap_populate_basepages(start, end, node);
}
#else	/* !CONFIG_ARM64_64K_PAGES */
int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
{
	unsigned long addr = start;
	unsigned long next;
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;

	do {
		next = pmd_addr_end(addr, end);

		pgd = vmemmap_pgd_populate(addr, node);
		if (!pgd)
			return -ENOMEM;

		pud = vmemmap_pud_populate(pgd, addr, node);
		if (!pud)
			return -ENOMEM;

		pmd = pmd_offset(pud, addr);
		if (pmd_none(*pmd)) {
			void *p = NULL;

			p = vmemmap_alloc_block_buf(PMD_SIZE, node);
			if (!p)
				return -ENOMEM;

			set_pmd(pmd, __pmd(__pa(p) | PROT_SECT_NORMAL));
		} else
			vmemmap_verify((pte_t *)pmd, node, addr, next);
	} while (addr = next, addr != end);

	return 0;
}
#endif	/* CONFIG_ARM64_64K_PAGES */
void vmemmap_free(unsigned long start, unsigned long end)
{
}
#endif	/* CONFIG_SPARSEMEM_VMEMMAP */

static pte_t bm_pte[PTRS_PER_PTE] __page_aligned_bss;
#if CONFIG_ARM64_PGTABLE_LEVELS > 2
static pmd_t bm_pmd[PTRS_PER_PMD] __page_aligned_bss;
#endif
#if CONFIG_ARM64_PGTABLE_LEVELS > 3
static pud_t bm_pud[PTRS_PER_PUD] __page_aligned_bss;
#endif

static inline pud_t * fixmap_pud(unsigned long addr)
{
	pgd_t *pgd = pgd_offset_k(addr);

	BUG_ON(pgd_none(*pgd) || pgd_bad(*pgd));

	return pud_offset(pgd, addr);
}

static inline pmd_t * fixmap_pmd(unsigned long addr)
{
	pud_t *pud = fixmap_pud(addr);

	BUG_ON(pud_none(*pud) || pud_bad(*pud));

	return pmd_offset(pud, addr);
}

static inline pte_t * fixmap_pte(unsigned long addr)
{
	pmd_t *pmd = fixmap_pmd(addr);

	BUG_ON(pmd_none(*pmd) || pmd_bad(*pmd));

	return pte_offset_kernel(pmd, addr);
}

void __init early_fixmap_init(void)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	unsigned long addr = FIXADDR_START;

	pgd = pgd_offset_k(addr);
	pgd_populate(&init_mm, pgd, bm_pud);
	pud = pud_offset(pgd, addr);
	pud_populate(&init_mm, pud, bm_pmd);
	pmd = pmd_offset(pud, addr);
	pmd_populate_kernel(&init_mm, pmd, bm_pte);

	/*
	 * The boot-ioremap range spans multiple pmds, for which
	 * we are not preparted:
	 */
	BUILD_BUG_ON((__fix_to_virt(FIX_BTMAP_BEGIN) >> PMD_SHIFT)
		     != (__fix_to_virt(FIX_BTMAP_END) >> PMD_SHIFT));

	if ((pmd != fixmap_pmd(fix_to_virt(FIX_BTMAP_BEGIN)))
	     || pmd != fixmap_pmd(fix_to_virt(FIX_BTMAP_END))) {
		WARN_ON(1);
		pr_warn("pmd %p != %p, %p\n",
			pmd, fixmap_pmd(fix_to_virt(FIX_BTMAP_BEGIN)),
			fixmap_pmd(fix_to_virt(FIX_BTMAP_END)));
		pr_warn("fix_to_virt(FIX_BTMAP_BEGIN): %08lx\n",
			fix_to_virt(FIX_BTMAP_BEGIN));
		pr_warn("fix_to_virt(FIX_BTMAP_END):   %08lx\n",
			fix_to_virt(FIX_BTMAP_END));

		pr_warn("FIX_BTMAP_END:       %d\n", FIX_BTMAP_END);
		pr_warn("FIX_BTMAP_BEGIN:     %d\n", FIX_BTMAP_BEGIN);
	}
}

void __set_fixmap(enum fixed_addresses idx,
			       phys_addr_t phys, pgprot_t flags)
{
	unsigned long addr = __fix_to_virt(idx);
	pte_t *pte;

	if (idx >= __end_of_fixed_addresses) {
		BUG();
		return;
	}

	pte = fixmap_pte(addr);

	if (pgprot_val(flags)) {
		set_pte(pte, pfn_pte(phys >> PAGE_SHIFT, flags));
	} else {
		pte_clear(&init_mm, addr, pte);
		flush_tlb_kernel_range(addr, addr+PAGE_SIZE);
	}
}
Commit	Line	Data
c1cc1552 CM	1	/*
	2	* Based on arch/arm/mm/mmu.c
	3	*
	4	* Copyright (C) 1995-2005 Russell King
	5	* Copyright (C) 2012 ARM Ltd.
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License version 2 as
	9	* published by the Free Software Foundation.
	10	*
	11	* This program is distributed in the hope that it will be useful,
	12	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	13	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	14	* GNU General Public License for more details.
	15	*
	16	* You should have received a copy of the GNU General Public License
	17	* along with this program. If not, see <http://www.gnu.org/licenses/>.
	18	*/
	19
	20	#include <linux/export.h>
	21	#include <linux/kernel.h>
	22	#include <linux/errno.h>
	23	#include <linux/init.h>
	24	#include <linux/mman.h>
	25	#include <linux/nodemask.h>
	26	#include <linux/memblock.h>
	27	#include <linux/fs.h>
2475ff9d	28	#include <linux/io.h>
c1cc1552 CM	29
c1cc1552 CM	30	#include <asm/cputype.h>
af86e597	31	#include <asm/fixmap.h>
c1cc1552 CM	32	#include <asm/sections.h>
	33	#include <asm/setup.h>
	34	#include <asm/sizes.h>
	35	#include <asm/tlb.h>
c79b954b	36	#include <asm/memblock.h>
c1cc1552 CM	37	#include <asm/mmu_context.h>
	38
	39	#include "mm.h"
	40
	41	/*
	42	* Empty_zero_page is a special page that is used for zero-initialized data
	43	* and COW.
	44	*/
	45	struct page *empty_zero_page;
	46	EXPORT_SYMBOL(empty_zero_page);
	47
c1cc1552 CM	48	pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn,
	49	unsigned long size, pgprot_t vma_prot)
	50	{
	51	if (!pfn_valid(pfn))
	52	return pgprot_noncached(vma_prot);
	53	else if (file->f_flags & O_SYNC)
	54	return pgprot_writecombine(vma_prot);
	55	return vma_prot;
	56	}
	57	EXPORT_SYMBOL(phys_mem_access_prot);
	58
	59	static void __init *early_alloc(unsigned long sz)
	60	{
	61	void *ptr = __va(memblock_alloc(sz, sz));
	62	memset(ptr, 0, sz);
	63	return ptr;
	64	}
	65
	66	static void __init alloc_init_pte(pmd_t *pmd, unsigned long addr,
d7ecbddf MS	67	unsigned long end, unsigned long pfn,
d7ecbddf MS	68	pgprot_t prot)
c1cc1552 CM	69	{
	70	pte_t *pte;
	71
	72	if (pmd_none(*pmd)) {
	73	pte = early_alloc(PTRS_PER_PTE * sizeof(pte_t));
	74	__pmd_populate(pmd, __pa(pte), PMD_TYPE_TABLE);
	75	}
	76	BUG_ON(pmd_bad(*pmd));
	77
	78	pte = pte_offset_kernel(pmd, addr);
	79	do {
d7ecbddf	80	set_pte(pte, pfn_pte(pfn, prot));
c1cc1552 CM	81	pfn++;
	82	} while (pte++, addr += PAGE_SIZE, addr != end);
	83	}
	84
e1e1fdda AB	85	static void __init alloc_init_pmd(struct mm_struct mm, pud_t pud,
e1e1fdda AB	86	unsigned long addr, unsigned long end,
8ce837ce	87	phys_addr_t phys, pgprot_t prot)
c1cc1552 CM	88	{
	89	pmd_t *pmd;
	90	unsigned long next;
	91
	92	/*
	93	* Check for initial section mappings in the pgd/pud and remove them.
	94	*/
	95	if (pud_none(pud) \|\| pud_bad(pud)) {
	96	pmd = early_alloc(PTRS_PER_PMD * sizeof(pmd_t));
e1e1fdda	97	pud_populate(mm, pud, pmd);
c1cc1552 CM	98	}
	99
	100	pmd = pmd_offset(pud, addr);
	101	do {
	102	next = pmd_addr_end(addr, end);
	103	/* try section mapping first */
a55f9929 CM	104	if (((addr \| next \| phys) & ~SECTION_MASK) == 0) {
a55f9929 CM	105	pmd_t old_pmd =*pmd;
8ce837ce AB	106	set_pmd(pmd, __pmd(phys \|
8ce837ce AB	107	pgprot_val(mk_sect_prot(prot))));
a55f9929 CM	108	/*
	109	* Check for previous table entries created during
	110	* boot (__create_page_tables) and flush them.
	111	*/
	112	if (!pmd_none(old_pmd))
	113	flush_tlb_all();
	114	} else {
d7ecbddf	115	alloc_init_pte(pmd, addr, next, __phys_to_pfn(phys),
8ce837ce	116	prot);
a55f9929	117	}
c1cc1552 CM	118	phys += next - addr;
	119	} while (pmd++, addr = next, addr != end);
	120	}
	121
e1e1fdda AB	122	static void __init alloc_init_pud(struct mm_struct mm, pgd_t pgd,
e1e1fdda AB	123	unsigned long addr, unsigned long end,
8ce837ce	124	phys_addr_t phys, pgprot_t prot)
c1cc1552	125	{
c79b954b	126	pud_t *pud;
c1cc1552 CM	127	unsigned long next;
c1cc1552 CM	128
c79b954b JL	129	if (pgd_none(*pgd)) {
c79b954b JL	130	pud = early_alloc(PTRS_PER_PUD * sizeof(pud_t));
e1e1fdda	131	pgd_populate(mm, pgd, pud);
c79b954b JL	132	}
	133	BUG_ON(pgd_bad(*pgd));
	134
	135	pud = pud_offset(pgd, addr);
c1cc1552 CM	136	do {
c1cc1552 CM	137	next = pud_addr_end(addr, end);
206a2a73 SC	138
	139	/*
	140	* For 4K granule only, attempt to put down a 1GB block
	141	*/
8ce837ce	142	if ((PAGE_SHIFT == 12) &&
206a2a73 SC	143	((addr \| next \| phys) & ~PUD_MASK) == 0) {
206a2a73 SC	144	pud_t old_pud = *pud;
8ce837ce AB	145	set_pud(pud, __pud(phys \|
8ce837ce AB	146	pgprot_val(mk_sect_prot(prot))));
206a2a73 SC	147
	148	/*
	149	* If we have an old value for a pud, it will
	150	* be pointing to a pmd table that we no longer
	151	* need (from swapper_pg_dir).
	152	*
	153	* Look up the old pmd table and free it.
	154	*/
	155	if (!pud_none(old_pud)) {
	156	phys_addr_t table = __pa(pmd_offset(&old_pud, 0));
	157	memblock_free(table, PAGE_SIZE);
	158	flush_tlb_all();
	159	}
	160	} else {
8ce837ce	161	alloc_init_pmd(mm, pud, addr, next, phys, prot);
206a2a73	162	}
c1cc1552 CM	163	phys += next - addr;
	164	} while (pud++, addr = next, addr != end);
	165	}
	166
	167	/*
	168	* Create the page directory entries and any necessary page tables for the
	169	* mapping specified by 'md'.
	170	*/
e1e1fdda AB	171	static void __init __create_mapping(struct mm_struct mm, pgd_t pgd,
e1e1fdda AB	172	phys_addr_t phys, unsigned long virt,
8ce837ce	173	phys_addr_t size, pgprot_t prot)
c1cc1552 CM	174	{
c1cc1552 CM	175	unsigned long addr, length, end, next;
c1cc1552 CM	176
	177	addr = virt & PAGE_MASK;
	178	length = PAGE_ALIGN(size + (virt & ~PAGE_MASK));
	179
c1cc1552 CM	180	end = addr + length;
	181	do {
	182	next = pgd_addr_end(addr, end);
8ce837ce	183	alloc_init_pud(mm, pgd, addr, next, phys, prot);
c1cc1552 CM	184	phys += next - addr;
	185	} while (pgd++, addr = next, addr != end);
	186	}
	187
d7ecbddf MS	188	static void __init create_mapping(phys_addr_t phys, unsigned long virt,
	189	phys_addr_t size)
	190	{
	191	if (virt < VMALLOC_START) {
	192	pr_warn("BUG: not creating mapping for %pa at 0x%016lx - outside kernel range\n",
	193	&phys, virt);
	194	return;
	195	}
e1e1fdda	196	__create_mapping(&init_mm, pgd_offset_k(virt & PAGE_MASK), phys, virt,
8ce837ce	197	size, PAGE_KERNEL_EXEC);
d7ecbddf MS	198	}
d7ecbddf MS	199
8ce837ce AB	200	void __init create_pgd_mapping(struct mm_struct *mm, phys_addr_t phys,
	201	unsigned long virt, phys_addr_t size,
	202	pgprot_t prot)
	203	{
	204	__create_mapping(mm, pgd_offset(mm, virt), phys, virt, size, prot);
d7ecbddf MS	205	}
d7ecbddf MS	206
c1cc1552 CM	207	static void __init map_mem(void)
	208	{
	209	struct memblock_region *reg;
e25208f7	210	phys_addr_t limit;
c1cc1552	211
f6bc87c3 SC	212	/*
	213	* Temporarily limit the memblock range. We need to do this as
	214	* create_mapping requires puds, pmds and ptes to be allocated from
	215	* memory addressable from the initial direct kernel mapping.
	216	*
3dec0fe4 CM	217	* The initial direct kernel mapping, located at swapper_pg_dir, gives
	218	* us PUD_SIZE (4K pages) or PMD_SIZE (64K pages) memory starting from
	219	* PHYS_OFFSET (which must be aligned to 2MB as per
	220	* Documentation/arm64/booting.txt).
f6bc87c3	221	*/
3dec0fe4 CM	222	if (IS_ENABLED(CONFIG_ARM64_64K_PAGES))
	223	limit = PHYS_OFFSET + PMD_SIZE;
	224	else
	225	limit = PHYS_OFFSET + PUD_SIZE;
e25208f7	226	memblock_set_current_limit(limit);
f6bc87c3	227
c1cc1552 CM	228	/* map all the memory banks */
	229	for_each_memblock(memory, reg) {
	230	phys_addr_t start = reg->base;
	231	phys_addr_t end = start + reg->size;
	232
	233	if (start >= end)
	234	break;
	235
e25208f7 CM	236	#ifndef CONFIG_ARM64_64K_PAGES
	237	/*
	238	* For the first memory bank align the start address and
	239	* current memblock limit to prevent create_mapping() from
	240	* allocating pte page tables from unmapped memory.
	241	* When 64K pages are enabled, the pte page table for the
	242	* first PGDIR_SIZE is already present in swapper_pg_dir.
	243	*/
	244	if (start < limit)
	245	start = ALIGN(start, PMD_SIZE);
	246	if (end < limit) {
	247	limit = end & PMD_MASK;
	248	memblock_set_current_limit(limit);
	249	}
	250	#endif
	251
c1cc1552 CM	252	create_mapping(start, __phys_to_virt(start), end - start);
c1cc1552 CM	253	}
f6bc87c3 SC	254
	255	/* Limit no longer required. */
	256	memblock_set_current_limit(MEMBLOCK_ALLOC_ANYWHERE);
c1cc1552 CM	257	}
	258
	259	/*
	260	* paging_init() sets up the page tables, initialises the zone memory
	261	* maps and sets up the zero page.
	262	*/
	263	void __init paging_init(void)
	264	{
	265	void *zero_page;
	266
c1cc1552 CM	267	map_mem();
	268
	269	/*
	270	* Finally flush the caches and tlb to ensure that we're in a
	271	* consistent state.
	272	*/
	273	flush_cache_all();
	274	flush_tlb_all();
	275
	276	/* allocate the zero page. */
	277	zero_page = early_alloc(PAGE_SIZE);
	278
	279	bootmem_init();
	280
	281	empty_zero_page = virt_to_page(zero_page);
c1cc1552 CM	282
	283	/*
	284	* TTBR0 is only used for the identity mapping at this stage. Make it
	285	* point to zero page to avoid speculatively fetching new entries.
	286	*/
	287	cpu_set_reserved_ttbr0();
	288	flush_tlb_all();
	289	}
	290
	291	/*
	292	* Enable the identity mapping to allow the MMU disabling.
	293	*/
	294	void setup_mm_for_reboot(void)
	295	{
	296	cpu_switch_mm(idmap_pg_dir, &init_mm);
	297	flush_tlb_all();
	298	}
	299
	300	/*
	301	* Check whether a kernel address is valid (derived from arch/x86/).
	302	*/
	303	int kern_addr_valid(unsigned long addr)
	304	{
	305	pgd_t *pgd;
	306	pud_t *pud;
	307	pmd_t *pmd;
	308	pte_t *pte;
	309
	310	if ((((long)addr) >> VA_BITS) != -1UL)
	311	return 0;
	312
	313	pgd = pgd_offset_k(addr);
	314	if (pgd_none(*pgd))
	315	return 0;
	316
	317	pud = pud_offset(pgd, addr);
	318	if (pud_none(*pud))
	319	return 0;
	320
206a2a73 SC	321	if (pud_sect(*pud))
	322	return pfn_valid(pud_pfn(*pud));
	323
c1cc1552 CM	324	pmd = pmd_offset(pud, addr);
	325	if (pmd_none(*pmd))
	326	return 0;
	327
da6e4cb6 DA	328	if (pmd_sect(*pmd))
	329	return pfn_valid(pmd_pfn(*pmd));
	330
c1cc1552 CM	331	pte = pte_offset_kernel(pmd, addr);
	332	if (pte_none(*pte))
	333	return 0;
	334
	335	return pfn_valid(pte_pfn(*pte));
	336	}
	337	#ifdef CONFIG_SPARSEMEM_VMEMMAP
	338	#ifdef CONFIG_ARM64_64K_PAGES
0aad818b	339	int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
c1cc1552	340	{
0aad818b	341	return vmemmap_populate_basepages(start, end, node);
c1cc1552 CM	342	}
c1cc1552 CM	343	#else /* !CONFIG_ARM64_64K_PAGES */
0aad818b	344	int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
c1cc1552	345	{
0aad818b	346	unsigned long addr = start;
c1cc1552 CM	347	unsigned long next;
	348	pgd_t *pgd;
	349	pud_t *pud;
	350	pmd_t *pmd;
	351
	352	do {
	353	next = pmd_addr_end(addr, end);
	354
	355	pgd = vmemmap_pgd_populate(addr, node);
	356	if (!pgd)
	357	return -ENOMEM;
	358
	359	pud = vmemmap_pud_populate(pgd, addr, node);
	360	if (!pud)
	361	return -ENOMEM;
	362
	363	pmd = pmd_offset(pud, addr);
	364	if (pmd_none(*pmd)) {
	365	void *p = NULL;
	366
	367	p = vmemmap_alloc_block_buf(PMD_SIZE, node);
	368	if (!p)
	369	return -ENOMEM;
	370
a501e324	371	set_pmd(pmd, __pmd(__pa(p) \| PROT_SECT_NORMAL));
c1cc1552 CM	372	} else
	373	vmemmap_verify((pte_t *)pmd, node, addr, next);
	374	} while (addr = next, addr != end);
	375
	376	return 0;
	377	}
	378	#endif /* CONFIG_ARM64_64K_PAGES */
0aad818b	379	void vmemmap_free(unsigned long start, unsigned long end)
0197518c TC	380	{
0197518c TC	381	}
c1cc1552	382	#endif /* CONFIG_SPARSEMEM_VMEMMAP */
af86e597 LA	383
	384	static pte_t bm_pte[PTRS_PER_PTE] __page_aligned_bss;
	385	#if CONFIG_ARM64_PGTABLE_LEVELS > 2
	386	static pmd_t bm_pmd[PTRS_PER_PMD] __page_aligned_bss;
	387	#endif
	388	#if CONFIG_ARM64_PGTABLE_LEVELS > 3
	389	static pud_t bm_pud[PTRS_PER_PUD] __page_aligned_bss;
	390	#endif
	391
	392	static inline pud_t * fixmap_pud(unsigned long addr)
	393	{
	394	pgd_t *pgd = pgd_offset_k(addr);
	395
	396	BUG_ON(pgd_none(pgd) \|\| pgd_bad(pgd));
	397
	398	return pud_offset(pgd, addr);
	399	}
	400
	401	static inline pmd_t * fixmap_pmd(unsigned long addr)
	402	{
	403	pud_t *pud = fixmap_pud(addr);
	404
	405	BUG_ON(pud_none(pud) \|\| pud_bad(pud));
	406
	407	return pmd_offset(pud, addr);
	408	}
	409
	410	static inline pte_t * fixmap_pte(unsigned long addr)
	411	{
	412	pmd_t *pmd = fixmap_pmd(addr);
	413
	414	BUG_ON(pmd_none(pmd) \|\| pmd_bad(pmd));
	415
	416	return pte_offset_kernel(pmd, addr);
	417	}
	418
	419	void __init early_fixmap_init(void)
	420	{
	421	pgd_t *pgd;
	422	pud_t *pud;
	423	pmd_t *pmd;
	424	unsigned long addr = FIXADDR_START;
	425
	426	pgd = pgd_offset_k(addr);
	427	pgd_populate(&init_mm, pgd, bm_pud);
	428	pud = pud_offset(pgd, addr);
	429	pud_populate(&init_mm, pud, bm_pmd);
	430	pmd = pmd_offset(pud, addr);
	431	pmd_populate_kernel(&init_mm, pmd, bm_pte);
	432
	433	/*
	434	* The boot-ioremap range spans multiple pmds, for which
	435	* we are not preparted:
	436	*/
	437	BUILD_BUG_ON((__fix_to_virt(FIX_BTMAP_BEGIN) >> PMD_SHIFT)
	438	!= (__fix_to_virt(FIX_BTMAP_END) >> PMD_SHIFT));
	439
	440	if ((pmd != fixmap_pmd(fix_to_virt(FIX_BTMAP_BEGIN)))
	441	\|\| pmd != fixmap_pmd(fix_to_virt(FIX_BTMAP_END))) {
	442	WARN_ON(1);
	443	pr_warn("pmd %p != %p, %p\n",
	444	pmd, fixmap_pmd(fix_to_virt(FIX_BTMAP_BEGIN)),
	445	fixmap_pmd(fix_to_virt(FIX_BTMAP_END)));
	446	pr_warn("fix_to_virt(FIX_BTMAP_BEGIN): %08lx\n",
447	fix_to_virt(FIX_BTMAP_BEGIN));
448	pr_warn("fix_to_virt(FIX_BTMAP_END): %08lx\n",
449	fix_to_virt(FIX_BTMAP_END));
450
451	pr_warn("FIX_BTMAP_END: %d\n", FIX_BTMAP_END);
452	pr_warn("FIX_BTMAP_BEGIN: %d\n", FIX_BTMAP_BEGIN);
453	}
454	}
455
456	void __set_fixmap(enum fixed_addresses idx,
457	phys_addr_t phys, pgprot_t flags)
458	{
459	unsigned long addr = __fix_to_virt(idx);
460	pte_t *pte;
461
462	if (idx >= __end_of_fixed_addresses) {
463	BUG();
464	return;
465	}
466
467	pte = fixmap_pte(addr);
468
469	if (pgprot_val(flags)) {
470	set_pte(pte, pfn_pte(phys >> PAGE_SHIFT, flags));
471	} else {
472	pte_clear(&init_mm, addr, pte);
473	flush_tlb_kernel_range(addr, addr+PAGE_SIZE);
474	}
475	}