2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* linux/arch/x86_64/mm/init.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 1995 Linus Torvalds
|
|
|
|
* Copyright (C) 2000 Pavel Machek <pavel@suse.cz>
|
|
|
|
* Copyright (C) 2002,2003 Andi Kleen <ak@suse.de>
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/signal.h>
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/ptrace.h>
|
|
|
|
#include <linux/mman.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/swap.h>
|
|
|
|
#include <linux/smp.h>
|
|
|
|
#include <linux/init.h>
|
2008-05-12 21:43:36 +08:00
|
|
|
#include <linux/initrd.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/bootmem.h>
|
|
|
|
#include <linux/proc_fs.h>
|
2005-11-06 00:25:53 +08:00
|
|
|
#include <linux/pci.h>
|
[PATCH] x86: tighten kernel image page access rights
On x86-64, kernel memory freed after init can be entirely unmapped instead
of just getting 'poisoned' by overwriting with a debug pattern.
On i386 and x86-64 (under CONFIG_DEBUG_RODATA), kernel text and bug table
can also be write-protected.
Compared to the first version, this one prevents re-creating deleted
mappings in the kernel image range on x86-64, if those got removed
previously. This, together with the original changes, prevents temporarily
having inconsistent mappings when cacheability attributes are being
changed on such pages (e.g. from AGP code). While on i386 such duplicate
mappings don't exist, the same change is done there, too, both for
consistency and because checking pte_present() before using various other
pte_XXX functions is a requirement anyway. At once, i386 code gets
adjusted to use pte_huge() instead of open coding this.
AK: split out cpa() changes
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2007-05-03 01:27:10 +08:00
|
|
|
#include <linux/pfn.h>
|
2006-06-27 17:53:52 +08:00
|
|
|
#include <linux/poison.h>
|
2006-01-12 05:44:42 +08:00
|
|
|
#include <linux/dma-mapping.h>
|
2006-01-17 14:03:41 +08:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/memory_hotplug.h>
|
2007-05-03 01:27:11 +08:00
|
|
|
#include <linux/nmi.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include <asm/processor.h>
|
2008-10-12 21:06:29 +08:00
|
|
|
#include <asm/bios_ebda.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <asm/system.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
#include <asm/pgtable.h>
|
|
|
|
#include <asm/pgalloc.h>
|
|
|
|
#include <asm/dma.h>
|
|
|
|
#include <asm/fixmap.h>
|
|
|
|
#include <asm/e820.h>
|
|
|
|
#include <asm/apic.h>
|
|
|
|
#include <asm/tlb.h>
|
|
|
|
#include <asm/mmu_context.h>
|
|
|
|
#include <asm/proto.h>
|
|
|
|
#include <asm/smp.h>
|
2005-11-06 00:25:53 +08:00
|
|
|
#include <asm/sections.h>
|
2008-01-30 20:30:17 +08:00
|
|
|
#include <asm/kdebug.h>
|
2008-01-30 20:30:17 +08:00
|
|
|
#include <asm/numa.h>
|
2008-02-13 04:12:01 +08:00
|
|
|
#include <asm/cacheflush.h>
|
2009-03-05 20:55:08 +08:00
|
|
|
#include <asm/init.h>
|
2010-02-03 05:44:16 +08:00
|
|
|
#include <linux/bootmem.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-11-06 00:25:53 +08:00
|
|
|
static unsigned long dma_reserve __initdata;
|
|
|
|
|
2008-04-17 23:40:45 +08:00
|
|
|
static int __init parse_direct_gbpages_off(char *arg)
|
|
|
|
{
|
|
|
|
direct_gbpages = 0;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
early_param("nogbpages", parse_direct_gbpages_off);
|
|
|
|
|
|
|
|
static int __init parse_direct_gbpages_on(char *arg)
|
|
|
|
{
|
|
|
|
direct_gbpages = 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
early_param("gbpages", parse_direct_gbpages_on);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* NOTE: pagetable_init alloc all the fixmap pagetables contiguous on the
|
|
|
|
* physical space so we can cache the place of the first one and move
|
|
|
|
* around without checking the pgd every time.
|
|
|
|
*/
|
|
|
|
|
2008-09-08 06:21:13 +08:00
|
|
|
pteval_t __supported_pte_mask __read_mostly = ~_PAGE_IOMAP;
|
2008-09-05 15:58:28 +08:00
|
|
|
EXPORT_SYMBOL_GPL(__supported_pte_mask);
|
|
|
|
|
|
|
|
int force_personality32;
|
|
|
|
|
2008-09-05 16:23:26 +08:00
|
|
|
/*
|
|
|
|
* noexec32=on|off
|
|
|
|
* Control non executable heap for 32bit processes.
|
|
|
|
* To control the stack too use noexec=off
|
|
|
|
*
|
|
|
|
* on PROT_READ does not imply PROT_EXEC for 32-bit processes (default)
|
|
|
|
* off PROT_READ implies PROT_EXEC
|
|
|
|
*/
|
2008-09-05 15:58:28 +08:00
|
|
|
static int __init nonx32_setup(char *str)
|
|
|
|
{
|
|
|
|
if (!strcmp(str, "on"))
|
|
|
|
force_personality32 &= ~READ_IMPLIES_EXEC;
|
|
|
|
else if (!strcmp(str, "off"))
|
|
|
|
force_personality32 |= READ_IMPLIES_EXEC;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("noexec32=", nonx32_setup);
|
|
|
|
|
2008-08-16 00:32:24 +08:00
|
|
|
/*
|
|
|
|
* NOTE: This function is marked __ref because it calls __init function
|
|
|
|
* (alloc_bootmem_pages). It's safe to do it ONLY when after_bootmem == 0.
|
|
|
|
*/
|
|
|
|
static __ref void *spp_getpage(void)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
void *ptr;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (after_bootmem)
|
2009-02-22 18:28:25 +08:00
|
|
|
ptr = (void *) get_zeroed_page(GFP_ATOMIC | __GFP_NOTRACK);
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
|
|
|
ptr = alloc_bootmem_pages(PAGE_SIZE);
|
2008-01-30 20:34:10 +08:00
|
|
|
|
|
|
|
if (!ptr || ((unsigned long)ptr & ~PAGE_MASK)) {
|
|
|
|
panic("set_pte_phys: cannot allocate page data %s\n",
|
|
|
|
after_bootmem ? "after bootmem" : "");
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
pr_debug("spp_getpage %p\n", ptr);
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return ptr;
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-03-04 04:02:57 +08:00
|
|
|
static pud_t *fill_pud(pgd_t *pgd, unsigned long vaddr)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-02-24 10:57:21 +08:00
|
|
|
if (pgd_none(*pgd)) {
|
|
|
|
pud_t *pud = (pud_t *)spp_getpage();
|
|
|
|
pgd_populate(&init_mm, pgd, pud);
|
|
|
|
if (pud != pud_offset(pgd, 0))
|
|
|
|
printk(KERN_ERR "PAGETABLE BUG #00! %p <-> %p\n",
|
|
|
|
pud, pud_offset(pgd, 0));
|
|
|
|
}
|
|
|
|
return pud_offset(pgd, vaddr);
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-03-04 04:02:57 +08:00
|
|
|
static pmd_t *fill_pmd(pud_t *pud, unsigned long vaddr)
|
2009-02-24 10:57:21 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
if (pud_none(*pud)) {
|
2009-02-24 10:57:21 +08:00
|
|
|
pmd_t *pmd = (pmd_t *) spp_getpage();
|
2008-06-25 12:19:02 +08:00
|
|
|
pud_populate(&init_mm, pud, pmd);
|
2009-02-24 10:57:21 +08:00
|
|
|
if (pmd != pmd_offset(pud, 0))
|
2008-01-30 20:34:10 +08:00
|
|
|
printk(KERN_ERR "PAGETABLE BUG #01! %p <-> %p\n",
|
2009-02-24 10:57:21 +08:00
|
|
|
pmd, pmd_offset(pud, 0));
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2009-02-24 10:57:21 +08:00
|
|
|
return pmd_offset(pud, vaddr);
|
|
|
|
}
|
|
|
|
|
2009-03-04 04:02:57 +08:00
|
|
|
static pte_t *fill_pte(pmd_t *pmd, unsigned long vaddr)
|
2009-02-24 10:57:21 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
if (pmd_none(*pmd)) {
|
2009-02-24 10:57:21 +08:00
|
|
|
pte_t *pte = (pte_t *) spp_getpage();
|
2008-06-25 12:19:02 +08:00
|
|
|
pmd_populate_kernel(&init_mm, pmd, pte);
|
2009-02-24 10:57:21 +08:00
|
|
|
if (pte != pte_offset_kernel(pmd, 0))
|
2008-01-30 20:34:10 +08:00
|
|
|
printk(KERN_ERR "PAGETABLE BUG #02!\n");
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2009-02-24 10:57:21 +08:00
|
|
|
return pte_offset_kernel(pmd, vaddr);
|
|
|
|
}
|
|
|
|
|
|
|
|
void set_pte_vaddr_pud(pud_t *pud_page, unsigned long vaddr, pte_t new_pte)
|
|
|
|
{
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd;
|
|
|
|
pte_t *pte;
|
|
|
|
|
|
|
|
pud = pud_page + pud_index(vaddr);
|
|
|
|
pmd = fill_pmd(pud, vaddr);
|
|
|
|
pte = fill_pte(pmd, vaddr);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
set_pte(pte, new_pte);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* It's enough to flush this one mapping.
|
|
|
|
* (PGE mappings get flushed as well)
|
|
|
|
*/
|
|
|
|
__flush_tlb_one(vaddr);
|
|
|
|
}
|
|
|
|
|
2009-02-24 10:57:21 +08:00
|
|
|
void set_pte_vaddr(unsigned long vaddr, pte_t pteval)
|
2008-06-25 12:19:22 +08:00
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud_page;
|
|
|
|
|
|
|
|
pr_debug("set_pte_vaddr %lx to %lx\n", vaddr, native_pte_val(pteval));
|
|
|
|
|
|
|
|
pgd = pgd_offset_k(vaddr);
|
|
|
|
if (pgd_none(*pgd)) {
|
|
|
|
printk(KERN_ERR
|
|
|
|
"PGD FIXMAP MISSING, it should be setup in head.S!\n");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
pud_page = (pud_t*)pgd_page_vaddr(*pgd);
|
|
|
|
set_pte_vaddr_pud(pud_page, vaddr, pteval);
|
|
|
|
}
|
|
|
|
|
2009-02-24 10:57:21 +08:00
|
|
|
pmd_t * __init populate_extra_pmd(unsigned long vaddr)
|
2009-02-20 15:29:09 +08:00
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud;
|
|
|
|
|
|
|
|
pgd = pgd_offset_k(vaddr);
|
2009-02-24 10:57:21 +08:00
|
|
|
pud = fill_pud(pgd, vaddr);
|
|
|
|
return fill_pmd(pud, vaddr);
|
|
|
|
}
|
|
|
|
|
|
|
|
pte_t * __init populate_extra_pte(unsigned long vaddr)
|
|
|
|
{
|
|
|
|
pmd_t *pmd;
|
2009-02-20 15:29:09 +08:00
|
|
|
|
2009-02-24 10:57:21 +08:00
|
|
|
pmd = populate_extra_pmd(vaddr);
|
|
|
|
return fill_pte(pmd, vaddr);
|
2009-02-20 15:29:09 +08:00
|
|
|
}
|
|
|
|
|
2008-07-02 03:45:32 +08:00
|
|
|
/*
|
|
|
|
* Create large page table mappings for a range of physical addresses.
|
|
|
|
*/
|
|
|
|
static void __init __init_extra_mapping(unsigned long phys, unsigned long size,
|
|
|
|
pgprot_t prot)
|
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd;
|
|
|
|
|
|
|
|
BUG_ON((phys & ~PMD_MASK) || (size & ~PMD_MASK));
|
|
|
|
for (; size; phys += PMD_SIZE, size -= PMD_SIZE) {
|
|
|
|
pgd = pgd_offset_k((unsigned long)__va(phys));
|
|
|
|
if (pgd_none(*pgd)) {
|
|
|
|
pud = (pud_t *) spp_getpage();
|
|
|
|
set_pgd(pgd, __pgd(__pa(pud) | _KERNPG_TABLE |
|
|
|
|
_PAGE_USER));
|
|
|
|
}
|
|
|
|
pud = pud_offset(pgd, (unsigned long)__va(phys));
|
|
|
|
if (pud_none(*pud)) {
|
|
|
|
pmd = (pmd_t *) spp_getpage();
|
|
|
|
set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE |
|
|
|
|
_PAGE_USER));
|
|
|
|
}
|
|
|
|
pmd = pmd_offset(pud, phys);
|
|
|
|
BUG_ON(!pmd_none(*pmd));
|
|
|
|
set_pmd(pmd, __pmd(phys | pgprot_val(prot)));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
void __init init_extra_mapping_wb(unsigned long phys, unsigned long size)
|
|
|
|
{
|
|
|
|
__init_extra_mapping(phys, size, PAGE_KERNEL_LARGE);
|
|
|
|
}
|
|
|
|
|
|
|
|
void __init init_extra_mapping_uc(unsigned long phys, unsigned long size)
|
|
|
|
{
|
|
|
|
__init_extra_mapping(phys, size, PAGE_KERNEL_LARGE_NOCACHE);
|
|
|
|
}
|
|
|
|
|
2008-02-16 00:29:12 +08:00
|
|
|
/*
|
2008-02-21 18:04:11 +08:00
|
|
|
* The head.S code sets up the kernel high mapping:
|
|
|
|
*
|
|
|
|
* from __START_KERNEL_map to __START_KERNEL_map + size (== _end-_text)
|
2008-02-16 00:29:12 +08:00
|
|
|
*
|
|
|
|
* phys_addr holds the negative offset to the kernel, which is added
|
|
|
|
* to the compile time generated pmds. This results in invalid pmds up
|
|
|
|
* to the point where we hit the physaddr 0 mapping.
|
|
|
|
*
|
|
|
|
* We limit the mappings to the region from _text to _end. _end is
|
|
|
|
* rounded up to the 2MB boundary. This catches the invalid pmds as
|
|
|
|
* well, as they are located before _text:
|
|
|
|
*/
|
|
|
|
void __init cleanup_highmap(void)
|
|
|
|
{
|
|
|
|
unsigned long vaddr = __START_KERNEL_map;
|
2008-07-25 22:48:57 +08:00
|
|
|
unsigned long end = roundup((unsigned long)_end, PMD_SIZE) - 1;
|
2008-02-16 00:29:12 +08:00
|
|
|
pmd_t *pmd = level2_kernel_pgt;
|
|
|
|
pmd_t *last_pmd = pmd + PTRS_PER_PMD;
|
|
|
|
|
|
|
|
for (; pmd < last_pmd; pmd++, vaddr += PMD_SIZE) {
|
2008-05-29 02:36:07 +08:00
|
|
|
if (pmd_none(*pmd))
|
2008-02-16 00:29:12 +08:00
|
|
|
continue;
|
|
|
|
if (vaddr < (unsigned long) _text || vaddr > end)
|
|
|
|
set_pmd(pmd, __pmd(0));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-08-21 21:28:42 +08:00
|
|
|
static __ref void *alloc_low_page(unsigned long *phys)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2009-03-05 20:55:06 +08:00
|
|
|
unsigned long pfn = e820_table_end++;
|
2005-04-17 06:20:36 +08:00
|
|
|
void *adr;
|
|
|
|
|
2006-01-17 14:03:41 +08:00
|
|
|
if (after_bootmem) {
|
2009-02-22 18:28:25 +08:00
|
|
|
adr = (void *)get_zeroed_page(GFP_ATOMIC | __GFP_NOTRACK);
|
2006-01-17 14:03:41 +08:00
|
|
|
*phys = __pa(adr);
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2006-01-17 14:03:41 +08:00
|
|
|
return adr;
|
|
|
|
}
|
|
|
|
|
2009-03-05 20:55:06 +08:00
|
|
|
if (pfn >= e820_table_top)
|
2008-01-30 20:34:10 +08:00
|
|
|
panic("alloc_low_page: ran out of memory");
|
2007-05-03 01:27:06 +08:00
|
|
|
|
2008-09-08 06:21:15 +08:00
|
|
|
adr = early_memremap(pfn * PAGE_SIZE, PAGE_SIZE);
|
2006-01-17 14:03:41 +08:00
|
|
|
memset(adr, 0, PAGE_SIZE);
|
2007-05-03 01:27:06 +08:00
|
|
|
*phys = pfn * PAGE_SIZE;
|
|
|
|
return adr;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-08-21 21:28:42 +08:00
|
|
|
static __ref void unmap_low_page(void *adr)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2006-01-17 14:03:41 +08:00
|
|
|
if (after_bootmem)
|
|
|
|
return;
|
|
|
|
|
2007-05-03 01:27:06 +08:00
|
|
|
early_iounmap(adr, PAGE_SIZE);
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-07-10 11:15:02 +08:00
|
|
|
static unsigned long __meminit
|
2008-10-08 04:58:46 +08:00
|
|
|
phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
|
|
|
|
pgprot_t prot)
|
2008-06-25 12:19:19 +08:00
|
|
|
{
|
|
|
|
unsigned pages = 0;
|
2008-07-10 11:15:02 +08:00
|
|
|
unsigned long last_map_addr = end;
|
2008-06-25 12:19:19 +08:00
|
|
|
int i;
|
2008-07-10 11:15:02 +08:00
|
|
|
|
2008-06-25 12:19:19 +08:00
|
|
|
pte_t *pte = pte_page + pte_index(addr);
|
|
|
|
|
|
|
|
for(i = pte_index(addr); i < PTRS_PER_PTE; i++, addr += PAGE_SIZE, pte++) {
|
|
|
|
|
|
|
|
if (addr >= end) {
|
|
|
|
if (!after_bootmem) {
|
|
|
|
for(; i < PTRS_PER_PTE; i++, pte++)
|
|
|
|
set_pte(pte, __pte(0));
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2008-10-08 04:58:46 +08:00
|
|
|
/*
|
|
|
|
* We will re-use the existing mapping.
|
|
|
|
* Xen for example has some special requirements, like mapping
|
|
|
|
* pagetable pages as RO. So assume someone who pre-setup
|
|
|
|
* these mappings are more intelligent.
|
|
|
|
*/
|
2008-10-26 13:58:21 +08:00
|
|
|
if (pte_val(*pte)) {
|
|
|
|
pages++;
|
2008-06-25 12:19:19 +08:00
|
|
|
continue;
|
2008-10-26 13:58:21 +08:00
|
|
|
}
|
2008-06-25 12:19:19 +08:00
|
|
|
|
|
|
|
if (0)
|
|
|
|
printk(" pte=%p addr=%lx pte=%016lx\n",
|
|
|
|
pte, addr, pfn_pte(addr >> PAGE_SHIFT, PAGE_KERNEL).pte);
|
|
|
|
pages++;
|
2008-10-08 04:58:46 +08:00
|
|
|
set_pte(pte, pfn_pte(addr >> PAGE_SHIFT, prot));
|
2008-07-10 11:15:02 +08:00
|
|
|
last_map_addr = (addr & PAGE_MASK) + PAGE_SIZE;
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
2008-09-24 05:00:38 +08:00
|
|
|
|
2008-06-25 12:19:19 +08:00
|
|
|
update_page_count(PG_LEVEL_4K, pages);
|
2008-07-10 11:15:02 +08:00
|
|
|
|
|
|
|
return last_map_addr;
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
|
|
|
|
2008-07-10 11:15:02 +08:00
|
|
|
static unsigned long __meminit
|
2008-10-08 04:58:46 +08:00
|
|
|
phys_pte_update(pmd_t *pmd, unsigned long address, unsigned long end,
|
|
|
|
pgprot_t prot)
|
2008-06-25 12:19:19 +08:00
|
|
|
{
|
|
|
|
pte_t *pte = (pte_t *)pmd_page_vaddr(*pmd);
|
|
|
|
|
2008-10-08 04:58:46 +08:00
|
|
|
return phys_pte_init(pte, address, end, prot);
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
|
|
|
|
2008-03-12 10:53:28 +08:00
|
|
|
static unsigned long __meminit
|
2008-07-08 16:41:05 +08:00
|
|
|
phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
|
2008-10-08 04:58:46 +08:00
|
|
|
unsigned long page_size_mask, pgprot_t prot)
|
2006-01-17 14:03:41 +08:00
|
|
|
{
|
2008-05-02 17:46:49 +08:00
|
|
|
unsigned long pages = 0;
|
2008-07-10 11:15:02 +08:00
|
|
|
unsigned long last_map_addr = end;
|
2008-05-02 17:46:49 +08:00
|
|
|
|
2006-09-26 16:52:36 +08:00
|
|
|
int i = pmd_index(address);
|
2006-01-17 14:03:41 +08:00
|
|
|
|
2006-09-26 16:52:36 +08:00
|
|
|
for (; i < PTRS_PER_PMD; i++, address += PMD_SIZE) {
|
2008-06-25 12:19:19 +08:00
|
|
|
unsigned long pte_phys;
|
2006-09-26 16:52:36 +08:00
|
|
|
pmd_t *pmd = pmd_page + pmd_index(address);
|
2008-06-25 12:19:19 +08:00
|
|
|
pte_t *pte;
|
2008-10-08 04:58:46 +08:00
|
|
|
pgprot_t new_prot = prot;
|
2006-01-17 14:03:41 +08:00
|
|
|
|
2006-06-26 19:59:02 +08:00
|
|
|
if (address >= end) {
|
2008-01-30 20:34:10 +08:00
|
|
|
if (!after_bootmem) {
|
2006-06-26 19:59:02 +08:00
|
|
|
for (; i < PTRS_PER_PMD; i++, pmd++)
|
|
|
|
set_pmd(pmd, __pmd(0));
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2006-01-17 14:03:41 +08:00
|
|
|
break;
|
|
|
|
}
|
2006-09-26 16:52:36 +08:00
|
|
|
|
2008-06-25 12:19:19 +08:00
|
|
|
if (pmd_val(*pmd)) {
|
2008-08-21 21:27:22 +08:00
|
|
|
if (!pmd_large(*pmd)) {
|
|
|
|
spin_lock(&init_mm.page_table_lock);
|
2008-07-10 11:15:02 +08:00
|
|
|
last_map_addr = phys_pte_update(pmd, address,
|
2008-10-08 04:58:46 +08:00
|
|
|
end, prot);
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2008-09-24 05:00:38 +08:00
|
|
|
continue;
|
2008-08-21 21:27:22 +08:00
|
|
|
}
|
2008-10-08 04:58:46 +08:00
|
|
|
/*
|
|
|
|
* If we are ok with PG_LEVEL_2M mapping, then we will
|
|
|
|
* use the existing mapping,
|
|
|
|
*
|
|
|
|
* Otherwise, we will split the large page mapping but
|
|
|
|
* use the same existing protection bits except for
|
|
|
|
* large page, so that we don't violate Intel's TLB
|
|
|
|
* Application note (317080) which says, while changing
|
|
|
|
* the page sizes, new and old translations should
|
|
|
|
* not differ with respect to page frame and
|
|
|
|
* attributes.
|
|
|
|
*/
|
2008-10-26 13:58:21 +08:00
|
|
|
if (page_size_mask & (1 << PG_LEVEL_2M)) {
|
|
|
|
pages++;
|
2008-10-08 04:58:46 +08:00
|
|
|
continue;
|
2008-10-26 13:58:21 +08:00
|
|
|
}
|
2008-10-08 04:58:46 +08:00
|
|
|
new_prot = pte_pgprot(pte_clrhuge(*(pte_t *)pmd));
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
|
|
|
|
2008-07-08 16:41:05 +08:00
|
|
|
if (page_size_mask & (1<<PG_LEVEL_2M)) {
|
2008-06-25 12:19:19 +08:00
|
|
|
pages++;
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_lock(&init_mm.page_table_lock);
|
2008-06-25 12:19:19 +08:00
|
|
|
set_pte((pte_t *)pmd,
|
2008-10-08 04:58:46 +08:00
|
|
|
pfn_pte(address >> PAGE_SHIFT,
|
|
|
|
__pgprot(pgprot_val(prot) | _PAGE_PSE)));
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2008-07-10 11:15:02 +08:00
|
|
|
last_map_addr = (address & PMD_MASK) + PMD_SIZE;
|
2006-09-26 16:52:36 +08:00
|
|
|
continue;
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
2006-09-26 16:52:36 +08:00
|
|
|
|
2008-06-25 12:19:19 +08:00
|
|
|
pte = alloc_low_page(&pte_phys);
|
2008-10-08 04:58:46 +08:00
|
|
|
last_map_addr = phys_pte_init(pte, address, end, new_prot);
|
2008-06-25 12:19:19 +08:00
|
|
|
unmap_low_page(pte);
|
|
|
|
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_lock(&init_mm.page_table_lock);
|
2008-06-25 12:19:19 +08:00
|
|
|
pmd_populate_kernel(&init_mm, pmd, __va(pte_phys));
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2006-01-17 14:03:41 +08:00
|
|
|
}
|
2008-05-02 17:46:49 +08:00
|
|
|
update_page_count(PG_LEVEL_2M, pages);
|
2008-07-10 11:15:02 +08:00
|
|
|
return last_map_addr;
|
2006-01-17 14:03:41 +08:00
|
|
|
}
|
|
|
|
|
2008-03-12 10:53:28 +08:00
|
|
|
static unsigned long __meminit
|
2008-07-08 16:41:05 +08:00
|
|
|
phys_pmd_update(pud_t *pud, unsigned long address, unsigned long end,
|
2008-10-08 04:58:46 +08:00
|
|
|
unsigned long page_size_mask, pgprot_t prot)
|
2006-01-17 14:03:41 +08:00
|
|
|
{
|
2008-01-30 20:34:10 +08:00
|
|
|
pmd_t *pmd = pmd_offset(pud, 0);
|
2008-03-12 10:53:28 +08:00
|
|
|
unsigned long last_map_addr;
|
|
|
|
|
2008-10-08 04:58:46 +08:00
|
|
|
last_map_addr = phys_pmd_init(pmd, address, end, page_size_mask, prot);
|
2006-09-26 16:52:36 +08:00
|
|
|
__flush_tlb_all();
|
2008-03-12 10:53:28 +08:00
|
|
|
return last_map_addr;
|
2006-01-17 14:03:41 +08:00
|
|
|
}
|
|
|
|
|
2008-03-12 10:53:28 +08:00
|
|
|
static unsigned long __meminit
|
2008-07-08 16:41:05 +08:00
|
|
|
phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
|
|
|
|
unsigned long page_size_mask)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2008-05-02 17:46:49 +08:00
|
|
|
unsigned long pages = 0;
|
2008-03-12 10:53:28 +08:00
|
|
|
unsigned long last_map_addr = end;
|
2006-09-26 16:52:36 +08:00
|
|
|
int i = pud_index(addr);
|
2006-01-17 14:03:41 +08:00
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
for (; i < PTRS_PER_PUD; i++, addr = (addr & PUD_MASK) + PUD_SIZE) {
|
2006-09-26 16:52:36 +08:00
|
|
|
unsigned long pmd_phys;
|
|
|
|
pud_t *pud = pud_page + pud_index(addr);
|
2005-04-17 06:20:36 +08:00
|
|
|
pmd_t *pmd;
|
2008-10-08 04:58:46 +08:00
|
|
|
pgprot_t prot = PAGE_KERNEL;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-09-26 16:52:36 +08:00
|
|
|
if (addr >= end)
|
2005-04-17 06:20:36 +08:00
|
|
|
break;
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
if (!after_bootmem &&
|
|
|
|
!e820_any_mapped(addr, addr+PUD_SIZE, 0)) {
|
|
|
|
set_pud(pud, __pud(0));
|
2005-04-17 06:20:36 +08:00
|
|
|
continue;
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-09-26 16:52:36 +08:00
|
|
|
if (pud_val(*pud)) {
|
2008-09-24 05:00:38 +08:00
|
|
|
if (!pud_large(*pud)) {
|
2008-07-08 16:41:05 +08:00
|
|
|
last_map_addr = phys_pmd_update(pud, addr, end,
|
2008-10-08 04:58:46 +08:00
|
|
|
page_size_mask, prot);
|
2008-09-24 05:00:38 +08:00
|
|
|
continue;
|
|
|
|
}
|
2008-10-08 04:58:46 +08:00
|
|
|
/*
|
|
|
|
* If we are ok with PG_LEVEL_1G mapping, then we will
|
|
|
|
* use the existing mapping.
|
|
|
|
*
|
|
|
|
* Otherwise, we will split the gbpage mapping but use
|
|
|
|
* the same existing protection bits except for large
|
|
|
|
* page, so that we don't violate Intel's TLB
|
|
|
|
* Application note (317080) which says, while changing
|
|
|
|
* the page sizes, new and old translations should
|
|
|
|
* not differ with respect to page frame and
|
|
|
|
* attributes.
|
|
|
|
*/
|
2008-10-26 13:58:21 +08:00
|
|
|
if (page_size_mask & (1 << PG_LEVEL_1G)) {
|
|
|
|
pages++;
|
2008-10-08 04:58:46 +08:00
|
|
|
continue;
|
2008-10-26 13:58:21 +08:00
|
|
|
}
|
2008-10-08 04:58:46 +08:00
|
|
|
prot = pte_pgprot(pte_clrhuge(*(pte_t *)pud));
|
2008-04-17 23:40:45 +08:00
|
|
|
}
|
|
|
|
|
2008-07-08 16:41:05 +08:00
|
|
|
if (page_size_mask & (1<<PG_LEVEL_1G)) {
|
2008-05-02 17:46:49 +08:00
|
|
|
pages++;
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_lock(&init_mm.page_table_lock);
|
2008-04-17 23:40:45 +08:00
|
|
|
set_pte((pte_t *)pud,
|
|
|
|
pfn_pte(addr >> PAGE_SHIFT, PAGE_KERNEL_LARGE));
|
2008-08-21 21:27:22 +08:00
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2008-03-12 10:53:28 +08:00
|
|
|
last_map_addr = (addr & PUD_MASK) + PUD_SIZE;
|
2006-09-26 16:52:36 +08:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2007-05-03 01:27:06 +08:00
|
|
|
pmd = alloc_low_page(&pmd_phys);
|
2008-10-08 04:58:46 +08:00
|
|
|
last_map_addr = phys_pmd_init(pmd, addr, end, page_size_mask,
|
|
|
|
prot);
|
2008-06-25 12:19:19 +08:00
|
|
|
unmap_low_page(pmd);
|
2008-08-21 21:27:22 +08:00
|
|
|
|
|
|
|
spin_lock(&init_mm.page_table_lock);
|
2008-06-25 12:19:19 +08:00
|
|
|
pud_populate(&init_mm, pud, __va(pmd_phys));
|
2006-01-17 14:03:41 +08:00
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-01-30 20:33:54 +08:00
|
|
|
__flush_tlb_all();
|
2008-09-24 05:00:38 +08:00
|
|
|
|
2008-05-02 17:46:49 +08:00
|
|
|
update_page_count(PG_LEVEL_1G, pages);
|
2008-03-12 10:53:28 +08:00
|
|
|
|
2008-06-25 05:56:20 +08:00
|
|
|
return last_map_addr;
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-06-25 12:19:19 +08:00
|
|
|
static unsigned long __meminit
|
2008-07-08 16:41:05 +08:00
|
|
|
phys_pud_update(pgd_t *pgd, unsigned long addr, unsigned long end,
|
|
|
|
unsigned long page_size_mask)
|
2008-06-25 12:19:19 +08:00
|
|
|
{
|
|
|
|
pud_t *pud;
|
|
|
|
|
|
|
|
pud = (pud_t *)pgd_page_vaddr(*pgd);
|
|
|
|
|
2008-07-08 16:41:05 +08:00
|
|
|
return phys_pud_init(pud, addr, end, page_size_mask);
|
2008-06-25 12:19:19 +08:00
|
|
|
}
|
|
|
|
|
2009-06-12 12:57:52 +08:00
|
|
|
unsigned long __meminit
|
2009-03-05 20:55:05 +08:00
|
|
|
kernel_physical_mapping_init(unsigned long start,
|
|
|
|
unsigned long end,
|
|
|
|
unsigned long page_size_mask)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-07-08 16:41:05 +08:00
|
|
|
unsigned long next, last_map_addr = end;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
start = (unsigned long)__va(start);
|
|
|
|
end = (unsigned long)__va(end);
|
|
|
|
|
|
|
|
for (; start < end; start = next) {
|
2006-01-17 14:03:41 +08:00
|
|
|
pgd_t *pgd = pgd_offset_k(start);
|
2008-01-30 20:34:10 +08:00
|
|
|
unsigned long pud_phys;
|
2006-01-17 14:03:41 +08:00
|
|
|
pud_t *pud;
|
|
|
|
|
2008-07-17 00:11:59 +08:00
|
|
|
next = (start + PGDIR_SIZE) & PGDIR_MASK;
|
2008-06-25 12:19:19 +08:00
|
|
|
if (next > end)
|
|
|
|
next = end;
|
|
|
|
|
|
|
|
if (pgd_val(*pgd)) {
|
2008-07-08 16:41:05 +08:00
|
|
|
last_map_addr = phys_pud_update(pgd, __pa(start),
|
|
|
|
__pa(end), page_size_mask);
|
2008-06-25 12:19:19 +08:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2008-08-21 21:27:22 +08:00
|
|
|
pud = alloc_low_page(&pud_phys);
|
2008-07-08 16:41:05 +08:00
|
|
|
last_map_addr = phys_pud_init(pud, __pa(start), __pa(next),
|
|
|
|
page_size_mask);
|
2008-06-25 12:19:19 +08:00
|
|
|
unmap_low_page(pud);
|
2008-08-21 21:27:22 +08:00
|
|
|
|
|
|
|
spin_lock(&init_mm.page_table_lock);
|
|
|
|
pgd_populate(&init_mm, pgd, __va(pud_phys));
|
|
|
|
spin_unlock(&init_mm.page_table_lock);
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2008-09-24 05:00:38 +08:00
|
|
|
__flush_tlb_all();
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-07-08 16:41:05 +08:00
|
|
|
return last_map_addr;
|
|
|
|
}
|
2008-07-10 11:15:02 +08:00
|
|
|
|
2005-06-23 15:08:06 +08:00
|
|
|
#ifndef CONFIG_NUMA
|
2009-09-26 06:20:00 +08:00
|
|
|
void __init initmem_init(unsigned long start_pfn, unsigned long end_pfn,
|
|
|
|
int acpi, int k8)
|
2008-06-22 17:44:49 +08:00
|
|
|
{
|
2010-02-10 17:20:20 +08:00
|
|
|
#ifndef CONFIG_NO_BOOTMEM
|
2008-06-22 17:44:49 +08:00
|
|
|
unsigned long bootmap_size, bootmap;
|
|
|
|
|
|
|
|
bootmap_size = bootmem_bootmap_pages(end_pfn)<<PAGE_SHIFT;
|
|
|
|
bootmap = find_e820_area(0, end_pfn<<PAGE_SHIFT, bootmap_size,
|
|
|
|
PAGE_SIZE);
|
|
|
|
if (bootmap == -1L)
|
|
|
|
panic("Cannot find bootmem map of size %ld\n", bootmap_size);
|
2010-02-10 17:20:15 +08:00
|
|
|
reserve_early(bootmap, bootmap + bootmap_size, "BOOTMAP");
|
2008-06-23 18:06:14 +08:00
|
|
|
/* don't touch min_low_pfn */
|
|
|
|
bootmap_size = init_bootmem_node(NODE_DATA(0), bootmap >> PAGE_SHIFT,
|
|
|
|
0, end_pfn);
|
2008-06-22 17:44:49 +08:00
|
|
|
e820_register_active_regions(0, start_pfn, end_pfn);
|
|
|
|
free_bootmem_with_active_regions(0, end_pfn);
|
2010-02-10 17:20:20 +08:00
|
|
|
#else
|
|
|
|
e820_register_active_regions(0, start_pfn, end_pfn);
|
|
|
|
#endif
|
2008-06-22 17:44:49 +08:00
|
|
|
}
|
2009-05-07 20:35:41 +08:00
|
|
|
#endif
|
2008-06-22 17:44:49 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
void __init paging_init(void)
|
|
|
|
{
|
2006-10-11 16:20:39 +08:00
|
|
|
unsigned long max_zone_pfns[MAX_NR_ZONES];
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2006-10-11 16:20:39 +08:00
|
|
|
memset(max_zone_pfns, 0, sizeof(max_zone_pfns));
|
|
|
|
max_zone_pfns[ZONE_DMA] = MAX_DMA_PFN;
|
|
|
|
max_zone_pfns[ZONE_DMA32] = MAX_DMA32_PFN;
|
2008-06-25 13:14:09 +08:00
|
|
|
max_zone_pfns[ZONE_NORMAL] = max_pfn;
|
2006-10-11 16:20:39 +08:00
|
|
|
|
2009-05-07 20:35:41 +08:00
|
|
|
sparse_memory_present_with_active_regions(MAX_NUMNODES);
|
2006-01-17 14:03:41 +08:00
|
|
|
sparse_init();
|
2009-07-09 00:50:19 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* clear the default setting with node 0
|
|
|
|
* note: don't use nodes_clear here, that is really clearing when
|
|
|
|
* numa support is not compiled in, and later node_set_state
|
|
|
|
* will not set it back.
|
|
|
|
*/
|
|
|
|
node_clear_state(0, N_NORMAL_MEMORY);
|
|
|
|
|
2006-09-27 16:49:52 +08:00
|
|
|
free_area_init_nodes(max_zone_pfns);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2006-01-17 14:03:41 +08:00
|
|
|
/*
|
|
|
|
* Memory hotplug specific functions
|
|
|
|
*/
|
2006-06-27 17:53:30 +08:00
|
|
|
#ifdef CONFIG_MEMORY_HOTPLUG
|
2010-02-03 05:44:16 +08:00
|
|
|
/*
|
|
|
|
* After memory hotplug the variables max_pfn, max_low_pfn and high_memory need
|
|
|
|
* updating.
|
|
|
|
*/
|
|
|
|
static void update_end_of_memory_vars(u64 start, u64 size)
|
|
|
|
{
|
|
|
|
unsigned long end_pfn = PFN_UP(start + size);
|
|
|
|
|
|
|
|
if (end_pfn > max_pfn) {
|
|
|
|
max_pfn = end_pfn;
|
|
|
|
max_low_pfn = end_pfn;
|
|
|
|
high_memory = (void *)__va(max_pfn * PAGE_SIZE - 1) + 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2006-04-08 01:49:15 +08:00
|
|
|
/*
|
|
|
|
* Memory is added always to NORMAL zone. This means you will never get
|
|
|
|
* additional DMA/DMA32 memory.
|
|
|
|
*/
|
2006-06-27 17:53:30 +08:00
|
|
|
int arch_add_memory(int nid, u64 start, u64 size)
|
2006-01-17 14:03:41 +08:00
|
|
|
{
|
2006-06-27 17:53:30 +08:00
|
|
|
struct pglist_data *pgdat = NODE_DATA(nid);
|
2006-09-26 14:31:09 +08:00
|
|
|
struct zone *zone = pgdat->node_zones + ZONE_NORMAL;
|
2008-03-12 10:53:28 +08:00
|
|
|
unsigned long last_mapped_pfn, start_pfn = start >> PAGE_SHIFT;
|
2006-01-17 14:03:41 +08:00
|
|
|
unsigned long nr_pages = size >> PAGE_SHIFT;
|
|
|
|
int ret;
|
|
|
|
|
2008-10-28 04:03:18 +08:00
|
|
|
last_mapped_pfn = init_memory_mapping(start, start + size);
|
2008-03-12 10:53:28 +08:00
|
|
|
if (last_mapped_pfn > max_pfn_mapped)
|
|
|
|
max_pfn_mapped = last_mapped_pfn;
|
2006-10-01 14:27:09 +08:00
|
|
|
|
2009-01-07 06:39:14 +08:00
|
|
|
ret = __add_pages(nid, zone, start_pfn, nr_pages);
|
2008-10-29 07:43:14 +08:00
|
|
|
WARN_ON_ONCE(ret);
|
2006-01-17 14:03:41 +08:00
|
|
|
|
2010-02-03 05:44:16 +08:00
|
|
|
/* update max_pfn, max_low_pfn and high_memory */
|
|
|
|
update_end_of_memory_vars(start, size);
|
|
|
|
|
2006-01-17 14:03:41 +08:00
|
|
|
return ret;
|
|
|
|
}
|
2006-06-27 17:53:30 +08:00
|
|
|
EXPORT_SYMBOL_GPL(arch_add_memory);
|
2006-01-17 14:03:41 +08:00
|
|
|
|
2006-11-19 14:19:40 +08:00
|
|
|
#if !defined(CONFIG_ACPI_NUMA) && defined(CONFIG_NUMA)
|
2006-10-01 14:27:06 +08:00
|
|
|
int memory_add_physaddr_to_nid(u64 start)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
2006-10-01 14:27:07 +08:00
|
|
|
EXPORT_SYMBOL_GPL(memory_add_physaddr_to_nid);
|
2006-10-01 14:27:06 +08:00
|
|
|
#endif
|
|
|
|
|
2006-10-01 14:27:09 +08:00
|
|
|
#endif /* CONFIG_MEMORY_HOTPLUG */
|
|
|
|
|
2009-09-23 07:45:49 +08:00
|
|
|
static struct kcore_list kcore_vsyscall;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
void __init mem_init(void)
|
|
|
|
{
|
2005-09-13 00:49:24 +08:00
|
|
|
long codesize, reservedpages, datasize, initsize;
|
2008-10-15 09:59:18 +08:00
|
|
|
unsigned long absent_pages;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-06-26 19:58:11 +08:00
|
|
|
pci_iommu_alloc();
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-01-30 20:32:36 +08:00
|
|
|
/* clear_bss() already clear the empty_zero_page */
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
reservedpages = 0;
|
|
|
|
|
|
|
|
/* this will put all low memory onto the freelists */
|
2005-06-23 15:08:06 +08:00
|
|
|
#ifdef CONFIG_NUMA
|
2005-09-13 00:49:24 +08:00
|
|
|
totalram_pages = numa_free_all_bootmem();
|
2005-04-17 06:20:36 +08:00
|
|
|
#else
|
2005-09-13 00:49:24 +08:00
|
|
|
totalram_pages = free_all_bootmem();
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif
|
2008-10-15 09:59:18 +08:00
|
|
|
|
|
|
|
absent_pages = absent_pages_in_range(0, max_pfn);
|
|
|
|
reservedpages = max_pfn - totalram_pages - absent_pages;
|
2005-04-17 06:20:36 +08:00
|
|
|
after_bootmem = 1;
|
|
|
|
|
|
|
|
codesize = (unsigned long) &_etext - (unsigned long) &_text;
|
|
|
|
datasize = (unsigned long) &_edata - (unsigned long) &_etext;
|
|
|
|
initsize = (unsigned long) &__init_end - (unsigned long) &__init_begin;
|
|
|
|
|
|
|
|
/* Register memory areas for /proc/kcore */
|
2008-01-30 20:34:10 +08:00
|
|
|
kclist_add(&kcore_vsyscall, (void *)VSYSCALL_START,
|
2009-09-23 07:45:43 +08:00
|
|
|
VSYSCALL_END - VSYSCALL_START, KCORE_OTHER);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
printk(KERN_INFO "Memory: %luk/%luk available (%ldk kernel code, "
|
2008-10-15 09:59:18 +08:00
|
|
|
"%ldk absent, %ldk reserved, %ldk data, %ldk init)\n",
|
2009-09-22 08:02:36 +08:00
|
|
|
nr_free_pages() << (PAGE_SHIFT-10),
|
2008-06-25 13:14:09 +08:00
|
|
|
max_pfn << (PAGE_SHIFT-10),
|
2005-04-17 06:20:36 +08:00
|
|
|
codesize >> 10,
|
2008-10-15 09:59:18 +08:00
|
|
|
absent_pages << (PAGE_SHIFT-10),
|
2005-04-17 06:20:36 +08:00
|
|
|
reservedpages << (PAGE_SHIFT-10),
|
|
|
|
datasize >> 10,
|
|
|
|
initsize >> 10);
|
|
|
|
}
|
|
|
|
|
2006-01-06 16:12:04 +08:00
|
|
|
#ifdef CONFIG_DEBUG_RODATA
|
2008-01-30 20:34:08 +08:00
|
|
|
const int rodata_test_data = 0xC3;
|
|
|
|
EXPORT_SYMBOL_GPL(rodata_test_data);
|
2006-01-06 16:12:04 +08:00
|
|
|
|
2009-10-29 10:46:56 +08:00
|
|
|
int kernel_set_to_readonly;
|
2009-02-18 06:57:30 +08:00
|
|
|
|
|
|
|
void set_kernel_text_rw(void)
|
|
|
|
{
|
2009-10-15 05:46:55 +08:00
|
|
|
unsigned long start = PFN_ALIGN(_text);
|
2009-10-29 10:46:58 +08:00
|
|
|
unsigned long end = PFN_ALIGN(__stop___ex_table);
|
2009-02-18 06:57:30 +08:00
|
|
|
|
|
|
|
if (!kernel_set_to_readonly)
|
|
|
|
return;
|
|
|
|
|
|
|
|
pr_debug("Set kernel text: %lx - %lx for read write\n",
|
|
|
|
start, end);
|
|
|
|
|
2009-10-29 10:46:58 +08:00
|
|
|
/*
|
|
|
|
* Make the kernel identity mapping for text RW. Kernel text
|
|
|
|
* mapping will always be RO. Refer to the comment in
|
|
|
|
* static_protections() in pageattr.c
|
|
|
|
*/
|
2009-02-18 06:57:30 +08:00
|
|
|
set_memory_rw(start, (end - start) >> PAGE_SHIFT);
|
|
|
|
}
|
|
|
|
|
|
|
|
void set_kernel_text_ro(void)
|
|
|
|
{
|
2009-10-15 05:46:55 +08:00
|
|
|
unsigned long start = PFN_ALIGN(_text);
|
2009-10-29 10:46:58 +08:00
|
|
|
unsigned long end = PFN_ALIGN(__stop___ex_table);
|
2009-02-18 06:57:30 +08:00
|
|
|
|
|
|
|
if (!kernel_set_to_readonly)
|
|
|
|
return;
|
|
|
|
|
|
|
|
pr_debug("Set kernel text: %lx - %lx for read only\n",
|
|
|
|
start, end);
|
|
|
|
|
2009-10-29 10:46:58 +08:00
|
|
|
/*
|
|
|
|
* Set the kernel identity mapping for text RO.
|
|
|
|
*/
|
2009-02-18 06:57:30 +08:00
|
|
|
set_memory_ro(start, (end - start) >> PAGE_SHIFT);
|
|
|
|
}
|
|
|
|
|
2006-01-06 16:12:04 +08:00
|
|
|
void mark_rodata_ro(void)
|
|
|
|
{
|
2009-10-15 05:46:56 +08:00
|
|
|
unsigned long start = PFN_ALIGN(_text);
|
2008-05-13 03:20:56 +08:00
|
|
|
unsigned long rodata_start =
|
|
|
|
((unsigned long)__start_rodata + PAGE_SIZE - 1) & PAGE_MASK;
|
2009-10-15 05:46:56 +08:00
|
|
|
unsigned long end = (unsigned long) &__end_rodata_hpage_align;
|
|
|
|
unsigned long text_end = PAGE_ALIGN((unsigned long) &__stop___ex_table);
|
|
|
|
unsigned long rodata_end = PAGE_ALIGN((unsigned long) &__end_rodata);
|
|
|
|
unsigned long data_start = (unsigned long) &_sdata;
|
2008-05-13 03:20:56 +08:00
|
|
|
|
[PATCH] x86: tighten kernel image page access rights
On x86-64, kernel memory freed after init can be entirely unmapped instead
of just getting 'poisoned' by overwriting with a debug pattern.
On i386 and x86-64 (under CONFIG_DEBUG_RODATA), kernel text and bug table
can also be write-protected.
Compared to the first version, this one prevents re-creating deleted
mappings in the kernel image range on x86-64, if those got removed
previously. This, together with the original changes, prevents temporarily
having inconsistent mappings when cacheability attributes are being
changed on such pages (e.g. from AGP code). While on i386 such duplicate
mappings don't exist, the same change is done there, too, both for
consistency and because checking pte_present() before using various other
pte_XXX functions is a requirement anyway. At once, i386 code gets
adjusted to use pte_huge() instead of open coding this.
AK: split out cpa() changes
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
2007-05-03 01:27:10 +08:00
|
|
|
printk(KERN_INFO "Write protecting the kernel read-only data: %luk\n",
|
Revert "[PATCH] x86: __pa and __pa_symbol address space separation"
This was broken. It adds complexity, for no good reason. Rather than
separate __pa() and __pa_symbol(), we should deprecate __pa_symbol(),
and preferably __pa() too - and just use "virt_to_phys()" instead, which
is more readable and has nicer semantics.
However, right now, just undo the separation, and make __pa_symbol() be
the exact same as __pa(). That fixes the bugs this patch introduced,
and we can do the fairly obvious cleanups later.
Do the new __phys_addr() function (which is now the actual workhorse for
the unified __pa()/__pa_symbol()) as a real external function, that way
all the potential issues with compile/link-time optimizations of
constant symbol addresses go away, and we can also, if we choose to, add
more sanity-checking of the argument.
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-05-07 23:44:24 +08:00
|
|
|
(end - start) >> 10);
|
2008-02-07 05:39:45 +08:00
|
|
|
set_memory_ro(start, (end - start) >> PAGE_SHIFT);
|
|
|
|
|
2009-02-18 06:57:30 +08:00
|
|
|
kernel_set_to_readonly = 1;
|
|
|
|
|
2008-02-07 05:39:45 +08:00
|
|
|
/*
|
|
|
|
* The rodata section (but not the kernel text!) should also be
|
|
|
|
* not-executable.
|
|
|
|
*/
|
2008-05-13 03:21:01 +08:00
|
|
|
set_memory_nx(rodata_start, (end - rodata_start) >> PAGE_SHIFT);
|
2006-01-06 16:12:04 +08:00
|
|
|
|
2008-01-30 20:34:09 +08:00
|
|
|
rodata_test();
|
|
|
|
|
2008-01-30 20:33:42 +08:00
|
|
|
#ifdef CONFIG_CPA_DEBUG
|
2008-01-30 20:34:10 +08:00
|
|
|
printk(KERN_INFO "Testing CPA: undo %lx-%lx\n", start, end);
|
2008-01-30 20:34:06 +08:00
|
|
|
set_memory_rw(start, (end-start) >> PAGE_SHIFT);
|
2008-01-30 20:33:42 +08:00
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
printk(KERN_INFO "Testing CPA: again\n");
|
2008-01-30 20:34:06 +08:00
|
|
|
set_memory_ro(start, (end-start) >> PAGE_SHIFT);
|
2008-01-30 20:33:42 +08:00
|
|
|
#endif
|
2009-10-15 05:46:56 +08:00
|
|
|
|
|
|
|
free_init_pages("unused kernel memory",
|
|
|
|
(unsigned long) page_address(virt_to_page(text_end)),
|
|
|
|
(unsigned long)
|
|
|
|
page_address(virt_to_page(rodata_start)));
|
|
|
|
free_init_pages("unused kernel memory",
|
|
|
|
(unsigned long) page_address(virt_to_page(rodata_end)),
|
|
|
|
(unsigned long) page_address(virt_to_page(data_start)));
|
2006-01-06 16:12:04 +08:00
|
|
|
}
|
2008-02-03 04:42:20 +08:00
|
|
|
|
2006-01-06 16:12:04 +08:00
|
|
|
#endif
|
|
|
|
|
2008-06-13 17:00:56 +08:00
|
|
|
int __init reserve_bootmem_generic(unsigned long phys, unsigned long len,
|
|
|
|
int flags)
|
2008-01-30 20:34:10 +08:00
|
|
|
{
|
2005-06-23 15:08:06 +08:00
|
|
|
#ifdef CONFIG_NUMA
|
2008-03-19 03:50:21 +08:00
|
|
|
int nid, next_nid;
|
2008-06-24 05:02:36 +08:00
|
|
|
int ret;
|
2006-11-14 23:57:46 +08:00
|
|
|
#endif
|
|
|
|
unsigned long pfn = phys >> PAGE_SHIFT;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2008-06-25 13:14:09 +08:00
|
|
|
if (pfn >= max_pfn) {
|
2008-01-30 20:34:10 +08:00
|
|
|
/*
|
|
|
|
* This can happen with kdump kernels when accessing
|
|
|
|
* firmware tables:
|
|
|
|
*/
|
2008-03-22 04:27:10 +08:00
|
|
|
if (pfn < max_pfn_mapped)
|
2008-06-08 21:46:30 +08:00
|
|
|
return -EFAULT;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2008-06-24 05:02:36 +08:00
|
|
|
printk(KERN_ERR "reserve_bootmem: illegal reserve %lx %lu\n",
|
2006-11-14 23:57:46 +08:00
|
|
|
phys, len);
|
2008-06-08 21:46:30 +08:00
|
|
|
return -EFAULT;
|
2006-11-14 23:57:46 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Should check here against the e820 map to avoid double free */
|
|
|
|
#ifdef CONFIG_NUMA
|
2008-03-19 03:50:21 +08:00
|
|
|
nid = phys_to_nid(phys);
|
|
|
|
next_nid = phys_to_nid(phys + len - 1);
|
|
|
|
if (nid == next_nid)
|
2008-06-08 21:46:30 +08:00
|
|
|
ret = reserve_bootmem_node(NODE_DATA(nid), phys, len, flags);
|
2008-03-19 03:50:21 +08:00
|
|
|
else
|
2008-06-08 21:46:30 +08:00
|
|
|
ret = reserve_bootmem(phys, len, flags);
|
|
|
|
|
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
#else
|
2009-08-21 16:34:45 +08:00
|
|
|
reserve_bootmem(phys, len, flags);
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif
|
2008-03-19 03:50:21 +08:00
|
|
|
|
2006-09-27 16:49:56 +08:00
|
|
|
if (phys+len <= MAX_DMA_PFN*PAGE_SIZE) {
|
2005-11-06 00:25:53 +08:00
|
|
|
dma_reserve += len / PAGE_SIZE;
|
2006-09-27 16:49:56 +08:00
|
|
|
set_dma_reserve(dma_reserve);
|
|
|
|
}
|
2008-06-08 21:46:30 +08:00
|
|
|
|
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
int kern_addr_valid(unsigned long addr)
|
|
|
|
{
|
2005-04-17 06:20:36 +08:00
|
|
|
unsigned long above = ((long)addr) >> __VIRTUAL_MASK_SHIFT;
|
2008-01-30 20:34:10 +08:00
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd;
|
|
|
|
pte_t *pte;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
if (above != 0 && above != -1UL)
|
2008-01-30 20:34:10 +08:00
|
|
|
return 0;
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
pgd = pgd_offset_k(addr);
|
|
|
|
if (pgd_none(*pgd))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
pud = pud_offset(pgd, addr);
|
|
|
|
if (pud_none(*pud))
|
2008-01-30 20:34:10 +08:00
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
pmd = pmd_offset(pud, addr);
|
|
|
|
if (pmd_none(*pmd))
|
|
|
|
return 0;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (pmd_large(*pmd))
|
|
|
|
return pfn_valid(pmd_pfn(*pmd));
|
|
|
|
|
|
|
|
pte = pte_offset_kernel(pmd, addr);
|
|
|
|
if (pte_none(*pte))
|
|
|
|
return 0;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return pfn_valid(pte_pfn(*pte));
|
|
|
|
}
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
/*
|
|
|
|
* A pseudo VMA to allow ptrace access for the vsyscall page. This only
|
|
|
|
* covers the 64bit vsyscall page now. 32bit has a real VMA now and does
|
|
|
|
* not need special handling anymore:
|
|
|
|
*/
|
2005-04-17 06:20:36 +08:00
|
|
|
static struct vm_area_struct gate_vma = {
|
2008-01-30 20:34:10 +08:00
|
|
|
.vm_start = VSYSCALL_START,
|
|
|
|
.vm_end = VSYSCALL_START + (VSYSCALL_MAPPED_PAGES * PAGE_SIZE),
|
|
|
|
.vm_page_prot = PAGE_READONLY_EXEC,
|
|
|
|
.vm_flags = VM_READ | VM_EXEC
|
2005-04-17 06:20:36 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
struct vm_area_struct *get_gate_vma(struct task_struct *tsk)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_IA32_EMULATION
|
2005-04-17 06:24:55 +08:00
|
|
|
if (test_tsk_thread_flag(tsk, TIF_IA32))
|
|
|
|
return NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif
|
|
|
|
return &gate_vma;
|
|
|
|
}
|
|
|
|
|
|
|
|
int in_gate_area(struct task_struct *task, unsigned long addr)
|
|
|
|
{
|
|
|
|
struct vm_area_struct *vma = get_gate_vma(task);
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:24:55 +08:00
|
|
|
if (!vma)
|
|
|
|
return 0;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return (addr >= vma->vm_start) && (addr < vma->vm_end);
|
|
|
|
}
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
/*
|
|
|
|
* Use this when you have no reliable task/vma, typically from interrupt
|
|
|
|
* context. It is less reliable than using the task's vma and may give
|
|
|
|
* false positives:
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
int in_gate_area_no_task(unsigned long addr)
|
|
|
|
{
|
2005-04-17 06:24:55 +08:00
|
|
|
return (addr >= VSYSCALL_START) && (addr < VSYSCALL_END);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2007-06-01 15:46:28 +08:00
|
|
|
|
2007-07-21 23:10:01 +08:00
|
|
|
const char *arch_vma_name(struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
if (vma->vm_mm && vma->vm_start == (long)vma->vm_mm->context.vdso)
|
|
|
|
return "[vdso]";
|
|
|
|
if (vma == &gate_vma)
|
|
|
|
return "[vsyscall]";
|
|
|
|
return NULL;
|
|
|
|
}
|
2007-10-16 16:24:15 +08:00
|
|
|
|
|
|
|
#ifdef CONFIG_SPARSEMEM_VMEMMAP
|
|
|
|
/*
|
|
|
|
* Initialise the sparsemem vmemmap using huge-pages at the PMD level.
|
|
|
|
*/
|
2008-04-12 16:19:24 +08:00
|
|
|
static long __meminitdata addr_start, addr_end;
|
|
|
|
static void __meminitdata *p_start, *p_end;
|
|
|
|
static int __meminitdata node_start;
|
|
|
|
|
2008-01-30 20:34:10 +08:00
|
|
|
int __meminit
|
|
|
|
vmemmap_populate(struct page *start_page, unsigned long size, int node)
|
2007-10-16 16:24:15 +08:00
|
|
|
{
|
|
|
|
unsigned long addr = (unsigned long)start_page;
|
|
|
|
unsigned long end = (unsigned long)(start_page + size);
|
|
|
|
unsigned long next;
|
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd;
|
|
|
|
|
|
|
|
for (; addr < end; addr = next) {
|
2008-06-25 12:19:20 +08:00
|
|
|
void *p = NULL;
|
2007-10-16 16:24:15 +08:00
|
|
|
|
|
|
|
pgd = vmemmap_pgd_populate(addr, node);
|
|
|
|
if (!pgd)
|
|
|
|
return -ENOMEM;
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2007-10-16 16:24:15 +08:00
|
|
|
pud = vmemmap_pud_populate(pgd, addr, node);
|
|
|
|
if (!pud)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2008-06-25 12:19:20 +08:00
|
|
|
if (!cpu_has_pse) {
|
|
|
|
next = (addr + PAGE_SIZE) & PAGE_MASK;
|
|
|
|
pmd = vmemmap_pmd_populate(pud, addr, node);
|
|
|
|
|
|
|
|
if (!pmd)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
p = vmemmap_pte_populate(pmd, addr, node);
|
2008-01-30 20:34:10 +08:00
|
|
|
|
2007-10-16 16:24:15 +08:00
|
|
|
if (!p)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2008-06-25 12:19:20 +08:00
|
|
|
addr_end = addr + PAGE_SIZE;
|
|
|
|
p_end = p + PAGE_SIZE;
|
2008-01-30 20:34:10 +08:00
|
|
|
} else {
|
2008-06-25 12:19:20 +08:00
|
|
|
next = pmd_addr_end(addr, end);
|
|
|
|
|
|
|
|
pmd = pmd_offset(pud, addr);
|
|
|
|
if (pmd_none(*pmd)) {
|
|
|
|
pte_t entry;
|
|
|
|
|
2010-02-10 17:20:22 +08:00
|
|
|
p = vmemmap_alloc_block_buf(PMD_SIZE, node);
|
2008-06-25 12:19:20 +08:00
|
|
|
if (!p)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
entry = pfn_pte(__pa(p) >> PAGE_SHIFT,
|
|
|
|
PAGE_KERNEL_LARGE);
|
|
|
|
set_pmd(pmd, __pmd(pte_val(entry)));
|
|
|
|
|
|
|
|
/* check to see if we have contiguous blocks */
|
|
|
|
if (p_end != p || node_start != node) {
|
|
|
|
if (p_start)
|
|
|
|
printk(KERN_DEBUG " [%lx-%lx] PMD -> [%p-%p] on node %d\n",
|
|
|
|
addr_start, addr_end-1, p_start, p_end-1, node_start);
|
|
|
|
addr_start = addr;
|
|
|
|
node_start = node;
|
|
|
|
p_start = p;
|
|
|
|
}
|
2008-07-04 03:29:34 +08:00
|
|
|
|
|
|
|
addr_end = addr + PMD_SIZE;
|
|
|
|
p_end = p + PMD_SIZE;
|
2008-06-25 12:19:20 +08:00
|
|
|
} else
|
|
|
|
vmemmap_verify((pte_t *)pmd, node, addr, next);
|
2008-01-30 20:34:10 +08:00
|
|
|
}
|
2008-06-25 12:19:20 +08:00
|
|
|
|
2007-10-16 16:24:15 +08:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
2008-04-12 16:19:24 +08:00
|
|
|
|
|
|
|
void __meminit vmemmap_populate_print_last(void)
|
|
|
|
{
|
|
|
|
if (p_start) {
|
|
|
|
printk(KERN_DEBUG " [%lx-%lx] PMD -> [%p-%p] on node %d\n",
|
|
|
|
addr_start, addr_end-1, p_start, p_end-1, node_start);
|
|
|
|
p_start = NULL;
|
|
|
|
p_end = NULL;
|
|
|
|
node_start = 0;
|
|
|
|
}
|
|
|
|
}
|
2007-10-16 16:24:15 +08:00
|
|
|
#endif
|