2005-04-16 18:20:36 -04:00
|
|
|
/*
|
|
|
|
* linux/arch/i386/mm/init.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 1995 Linus Torvalds
|
|
|
|
*
|
|
|
|
* Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
#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/hugetlb.h>
|
|
|
|
#include <linux/swap.h>
|
|
|
|
#include <linux/smp.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/pagemap.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>
2007-05-02 13:27:10 -04:00
|
|
|
#include <linux/pfn.h>
|
2006-06-27 05:53:52 -04:00
|
|
|
#include <linux/poison.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <linux/bootmem.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/proc_fs.h>
|
2005-10-29 21:16:57 -04:00
|
|
|
#include <linux/memory_hotplug.h>
|
2005-11-13 19:06:51 -05:00
|
|
|
#include <linux/initrd.h>
|
2006-06-23 05:04:49 -04:00
|
|
|
#include <linux/cpumask.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
#include <asm/processor.h>
|
|
|
|
#include <asm/system.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
#include <asm/pgtable.h>
|
|
|
|
#include <asm/dma.h>
|
|
|
|
#include <asm/fixmap.h>
|
|
|
|
#include <asm/e820.h>
|
|
|
|
#include <asm/apic.h>
|
|
|
|
#include <asm/tlb.h>
|
|
|
|
#include <asm/tlbflush.h>
|
2008-01-30 07:33:39 -05:00
|
|
|
#include <asm/pgalloc.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <asm/sections.h>
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
#include <asm/paravirt.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
unsigned int __VMALLOC_RESERVE = 128 << 20;
|
|
|
|
|
|
|
|
DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
|
|
|
|
unsigned long highstart_pfn, highend_pfn;
|
|
|
|
|
|
|
|
static int noinline do_test_wp_bit(void);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Creates a middle page table and puts a pointer to it in the
|
|
|
|
* given global directory entry. This only returns the gd entry
|
|
|
|
* in non-PAE compilation mode, since the middle layer is folded.
|
|
|
|
*/
|
|
|
|
static pmd_t * __init one_md_table_init(pgd_t *pgd)
|
|
|
|
{
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd_table;
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_PAE
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
if (!(pgd_val(*pgd) & _PAGE_PRESENT)) {
|
|
|
|
pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
|
|
|
|
|
2008-01-30 07:33:39 -05:00
|
|
|
paravirt_alloc_pd(&init_mm, __pa(pmd_table) >> PAGE_SHIFT);
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
|
|
|
|
pud = pud_offset(pgd, 0);
|
|
|
|
if (pmd_table != pmd_offset(pud, 0))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
#endif
|
2005-04-16 18:20:36 -04:00
|
|
|
pud = pud_offset(pgd, 0);
|
|
|
|
pmd_table = pmd_offset(pud, 0);
|
|
|
|
return pmd_table;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Create a page table and place a pointer to it in a middle page
|
|
|
|
* directory entry.
|
|
|
|
*/
|
|
|
|
static pte_t * __init one_page_table_init(pmd_t *pmd)
|
|
|
|
{
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
if (!(pmd_val(*pmd) & _PAGE_PRESENT)) {
|
2007-10-17 12:04:34 -04:00
|
|
|
pte_t *page_table = NULL;
|
|
|
|
|
|
|
|
#ifdef CONFIG_DEBUG_PAGEALLOC
|
|
|
|
page_table = (pte_t *) alloc_bootmem_pages(PAGE_SIZE);
|
|
|
|
#endif
|
|
|
|
if (!page_table)
|
|
|
|
page_table =
|
|
|
|
(pte_t *)alloc_bootmem_low_pages(PAGE_SIZE);
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
|
2007-07-17 21:37:03 -04:00
|
|
|
paravirt_alloc_pt(&init_mm, __pa(page_table) >> PAGE_SHIFT);
|
2005-04-16 18:20:36 -04:00
|
|
|
set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE));
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
BUG_ON(page_table != pte_offset_kernel(pmd, 0));
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
2007-10-17 12:04:34 -04:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
return pte_offset_kernel(pmd, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This function initializes a certain range of kernel virtual memory
|
|
|
|
* with new bootmem page tables, everywhere page tables are missing in
|
|
|
|
* the given range.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* NOTE: The pagetables are allocated contiguous on the physical space
|
|
|
|
* so we can cache the place of the first one and move around without
|
|
|
|
* checking the pgd every time.
|
|
|
|
*/
|
|
|
|
static void __init page_table_range_init (unsigned long start, unsigned long end, pgd_t *pgd_base)
|
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
pmd_t *pmd;
|
|
|
|
int pgd_idx, pmd_idx;
|
|
|
|
unsigned long vaddr;
|
|
|
|
|
|
|
|
vaddr = start;
|
|
|
|
pgd_idx = pgd_index(vaddr);
|
|
|
|
pmd_idx = pmd_index(vaddr);
|
|
|
|
pgd = pgd_base + pgd_idx;
|
|
|
|
|
|
|
|
for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) {
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
pmd = one_md_table_init(pgd);
|
|
|
|
pmd = pmd + pmd_index(vaddr);
|
2005-04-16 18:20:36 -04:00
|
|
|
for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) {
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
one_page_table_init(pmd);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
vaddr += PMD_SIZE;
|
|
|
|
}
|
|
|
|
pmd_idx = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int is_kernel_text(unsigned long addr)
|
|
|
|
{
|
|
|
|
if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end)
|
|
|
|
return 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This maps the physical memory to kernel virtual address space, a total
|
|
|
|
* of max_low_pfn pages, by creating page tables starting from address
|
|
|
|
* PAGE_OFFSET.
|
|
|
|
*/
|
|
|
|
static void __init kernel_physical_mapping_init(pgd_t *pgd_base)
|
|
|
|
{
|
|
|
|
unsigned long pfn;
|
|
|
|
pgd_t *pgd;
|
|
|
|
pmd_t *pmd;
|
|
|
|
pte_t *pte;
|
|
|
|
int pgd_idx, pmd_idx, pte_ofs;
|
|
|
|
|
|
|
|
pgd_idx = pgd_index(PAGE_OFFSET);
|
|
|
|
pgd = pgd_base + pgd_idx;
|
|
|
|
pfn = 0;
|
|
|
|
|
|
|
|
for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
|
|
|
|
pmd = one_md_table_init(pgd);
|
|
|
|
if (pfn >= max_low_pfn)
|
|
|
|
continue;
|
2008-01-30 07:31:09 -05:00
|
|
|
for (pmd_idx = 0;
|
|
|
|
pmd_idx < PTRS_PER_PMD && pfn < max_low_pfn;
|
|
|
|
pmd++, pmd_idx++) {
|
2005-04-16 18:20:36 -04:00
|
|
|
unsigned int address = pfn * PAGE_SIZE + PAGE_OFFSET;
|
|
|
|
|
2008-01-30 07:31:09 -05:00
|
|
|
/* Map with big pages if possible, otherwise
|
|
|
|
create normal page tables. */
|
2005-04-16 18:20:36 -04:00
|
|
|
if (cpu_has_pse) {
|
2008-01-30 07:31:09 -05:00
|
|
|
unsigned int address2;
|
|
|
|
pgprot_t prot = PAGE_KERNEL_LARGE;
|
|
|
|
|
|
|
|
address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE +
|
|
|
|
PAGE_OFFSET + PAGE_SIZE-1;
|
|
|
|
|
|
|
|
if (is_kernel_text(address) ||
|
|
|
|
is_kernel_text(address2))
|
|
|
|
prot = PAGE_KERNEL_LARGE_EXEC;
|
|
|
|
|
|
|
|
set_pmd(pmd, pfn_pmd(pfn, prot));
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
pfn += PTRS_PER_PTE;
|
|
|
|
} else {
|
|
|
|
pte = one_page_table_init(pmd);
|
|
|
|
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
for (pte_ofs = 0;
|
|
|
|
pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
|
|
|
|
pte++, pfn++, pte_ofs++, address += PAGE_SIZE) {
|
2008-01-30 07:31:09 -05:00
|
|
|
pgprot_t prot = PAGE_KERNEL;
|
|
|
|
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
if (is_kernel_text(address))
|
2008-01-30 07:31:09 -05:00
|
|
|
prot = PAGE_KERNEL_EXEC;
|
|
|
|
|
|
|
|
set_pte(pte, pfn_pte(pfn, prot));
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline int page_kills_ppro(unsigned long pagenr)
|
|
|
|
{
|
|
|
|
if (pagenr >= 0x70000 && pagenr <= 0x7003F)
|
|
|
|
return 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
pte_t *kmap_pte;
|
|
|
|
pgprot_t kmap_prot;
|
|
|
|
|
|
|
|
#define kmap_get_fixmap_pte(vaddr) \
|
|
|
|
pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr), vaddr), (vaddr)), (vaddr))
|
|
|
|
|
|
|
|
static void __init kmap_init(void)
|
|
|
|
{
|
|
|
|
unsigned long kmap_vstart;
|
|
|
|
|
|
|
|
/* cache the first kmap pte */
|
|
|
|
kmap_vstart = __fix_to_virt(FIX_KMAP_BEGIN);
|
|
|
|
kmap_pte = kmap_get_fixmap_pte(kmap_vstart);
|
|
|
|
|
|
|
|
kmap_prot = PAGE_KERNEL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __init permanent_kmaps_init(pgd_t *pgd_base)
|
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
pud_t *pud;
|
|
|
|
pmd_t *pmd;
|
|
|
|
pte_t *pte;
|
|
|
|
unsigned long vaddr;
|
|
|
|
|
|
|
|
vaddr = PKMAP_BASE;
|
|
|
|
page_table_range_init(vaddr, vaddr + PAGE_SIZE*LAST_PKMAP, pgd_base);
|
|
|
|
|
|
|
|
pgd = swapper_pg_dir + pgd_index(vaddr);
|
|
|
|
pud = pud_offset(pgd, vaddr);
|
|
|
|
pmd = pmd_offset(pud, vaddr);
|
|
|
|
pte = pte_offset_kernel(pmd, vaddr);
|
|
|
|
pkmap_page_table = pte;
|
|
|
|
}
|
|
|
|
|
2006-01-17 01:03:44 -05:00
|
|
|
static void __meminit free_new_highpage(struct page *page)
|
2005-10-29 21:16:57 -04:00
|
|
|
{
|
2006-03-22 03:08:40 -05:00
|
|
|
init_page_count(page);
|
2005-10-29 21:16:57 -04:00
|
|
|
__free_page(page);
|
|
|
|
totalhigh_pages++;
|
|
|
|
}
|
|
|
|
|
|
|
|
void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
|
|
|
|
ClearPageReserved(page);
|
2005-10-29 21:16:57 -04:00
|
|
|
free_new_highpage(page);
|
2005-04-16 18:20:36 -04:00
|
|
|
} else
|
|
|
|
SetPageReserved(page);
|
|
|
|
}
|
|
|
|
|
2007-01-10 19:52:44 -05:00
|
|
|
static int __meminit add_one_highpage_hotplug(struct page *page, unsigned long pfn)
|
2005-10-29 21:16:57 -04:00
|
|
|
{
|
|
|
|
free_new_highpage(page);
|
|
|
|
totalram_pages++;
|
|
|
|
#ifdef CONFIG_FLATMEM
|
|
|
|
max_mapnr = max(pfn, max_mapnr);
|
|
|
|
#endif
|
|
|
|
num_physpages++;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Not currently handling the NUMA case.
|
|
|
|
* Assuming single node and all memory that
|
|
|
|
* has been added dynamically that would be
|
|
|
|
* onlined here is in HIGHMEM
|
|
|
|
*/
|
2007-01-10 19:52:44 -05:00
|
|
|
void __meminit online_page(struct page *page)
|
2005-10-29 21:16:57 -04:00
|
|
|
{
|
|
|
|
ClearPageReserved(page);
|
|
|
|
add_one_highpage_hotplug(page, page_to_pfn(page));
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-06-23 03:07:57 -04:00
|
|
|
#ifdef CONFIG_NUMA
|
|
|
|
extern void set_highmem_pages_init(int);
|
|
|
|
#else
|
2005-04-16 18:20:36 -04:00
|
|
|
static void __init set_highmem_pages_init(int bad_ppro)
|
|
|
|
{
|
|
|
|
int pfn;
|
2008-01-15 10:44:37 -05:00
|
|
|
for (pfn = highstart_pfn; pfn < highend_pfn; pfn++) {
|
|
|
|
/*
|
|
|
|
* Holes under sparsemem might not have no mem_map[]:
|
|
|
|
*/
|
|
|
|
if (pfn_valid(pfn))
|
|
|
|
add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
|
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
totalram_pages += totalhigh_pages;
|
|
|
|
}
|
2005-06-23 03:07:57 -04:00
|
|
|
#endif /* CONFIG_FLATMEM */
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
#else
|
|
|
|
#define kmap_init() do { } while (0)
|
|
|
|
#define permanent_kmaps_init(pgd_base) do { } while (0)
|
|
|
|
#define set_highmem_pages_init(bad_ppro) do { } while (0)
|
|
|
|
#endif /* CONFIG_HIGHMEM */
|
|
|
|
|
2008-01-30 07:33:50 -05:00
|
|
|
pteval_t __PAGE_KERNEL = _PAGE_KERNEL;
|
2005-06-23 03:08:33 -04:00
|
|
|
EXPORT_SYMBOL(__PAGE_KERNEL);
|
2008-01-30 07:33:50 -05:00
|
|
|
pteval_t __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2005-06-23 03:07:57 -04:00
|
|
|
#ifdef CONFIG_NUMA
|
2005-04-16 18:20:36 -04:00
|
|
|
extern void __init remap_numa_kva(void);
|
2005-06-23 03:07:57 -04:00
|
|
|
#else
|
|
|
|
#define remap_numa_kva() do {} while (0)
|
2005-04-16 18:20:36 -04:00
|
|
|
#endif
|
|
|
|
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
void __init native_pagetable_setup_start(pgd_t *base)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
int i;
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Init entries of the first-level page table to the
|
|
|
|
* zero page, if they haven't already been set up.
|
|
|
|
*
|
|
|
|
* In a normal native boot, we'll be running on a
|
|
|
|
* pagetable rooted in swapper_pg_dir, but not in PAE
|
|
|
|
* mode, so this will end up clobbering the mappings
|
|
|
|
* for the lower 24Mbytes of the address space,
|
|
|
|
* without affecting the kernel address space.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < USER_PTRS_PER_PGD; i++)
|
|
|
|
set_pgd(&base[i],
|
|
|
|
__pgd(__pa(empty_zero_page) | _PAGE_PRESENT));
|
|
|
|
|
|
|
|
/* Make sure kernel address space is empty so that a pagetable
|
|
|
|
will be allocated for it. */
|
|
|
|
memset(&base[USER_PTRS_PER_PGD], 0,
|
|
|
|
KERNEL_PGD_PTRS * sizeof(pgd_t));
|
2007-02-13 07:26:21 -05:00
|
|
|
#else
|
2008-01-30 07:33:39 -05:00
|
|
|
paravirt_alloc_pd(&init_mm, __pa(base) >> PAGE_SHIFT);
|
2005-04-16 18:20:36 -04:00
|
|
|
#endif
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
void __init native_pagetable_setup_done(pgd_t *base)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
/*
|
|
|
|
* Add low memory identity-mappings - SMP needs it when
|
|
|
|
* starting up on an AP from real-mode. In the non-PAE
|
|
|
|
* case we already have these mappings through head.S.
|
|
|
|
* All user-space mappings are explicitly cleared after
|
|
|
|
* SMP startup.
|
|
|
|
*/
|
|
|
|
set_pgd(&base[0], base[USER_PTRS_PER_PGD]);
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Build a proper pagetable for the kernel mappings. Up until this
|
|
|
|
* point, we've been running on some set of pagetables constructed by
|
|
|
|
* the boot process.
|
|
|
|
*
|
|
|
|
* If we're booting on native hardware, this will be a pagetable
|
|
|
|
* constructed in arch/i386/kernel/head.S, and not running in PAE mode
|
|
|
|
* (even if we'll end up running in PAE). The root of the pagetable
|
|
|
|
* will be swapper_pg_dir.
|
|
|
|
*
|
|
|
|
* If we're booting paravirtualized under a hypervisor, then there are
|
|
|
|
* more options: we may already be running PAE, and the pagetable may
|
|
|
|
* or may not be based in swapper_pg_dir. In any case,
|
|
|
|
* paravirt_pagetable_setup_start() will set up swapper_pg_dir
|
|
|
|
* appropriately for the rest of the initialization to work.
|
|
|
|
*
|
|
|
|
* In general, pagetable_init() assumes that the pagetable may already
|
|
|
|
* be partially populated, and so it avoids stomping on any existing
|
|
|
|
* mappings.
|
|
|
|
*/
|
|
|
|
static void __init pagetable_init (void)
|
|
|
|
{
|
|
|
|
unsigned long vaddr, end;
|
|
|
|
pgd_t *pgd_base = swapper_pg_dir;
|
|
|
|
|
|
|
|
paravirt_pagetable_setup_start(pgd_base);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
/* Enable PSE if available */
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
if (cpu_has_pse)
|
2005-04-16 18:20:36 -04:00
|
|
|
set_in_cr4(X86_CR4_PSE);
|
|
|
|
|
|
|
|
/* Enable PGE if available */
|
|
|
|
if (cpu_has_pge) {
|
|
|
|
set_in_cr4(X86_CR4_PGE);
|
|
|
|
__PAGE_KERNEL |= _PAGE_GLOBAL;
|
|
|
|
__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
kernel_physical_mapping_init(pgd_base);
|
|
|
|
remap_numa_kva();
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Fixed mappings, only the page table structure has to be
|
|
|
|
* created - mappings will be set by set_fixmap():
|
|
|
|
*/
|
2008-01-30 07:33:44 -05:00
|
|
|
early_ioremap_clear();
|
2005-04-16 18:20:36 -04:00
|
|
|
vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
end = (FIXADDR_TOP + PMD_SIZE - 1) & PMD_MASK;
|
|
|
|
page_table_range_init(vaddr, end, pgd_base);
|
2008-01-30 07:33:44 -05:00
|
|
|
early_ioremap_reset();
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
permanent_kmaps_init(pgd_base);
|
|
|
|
|
[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable
This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.
In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory. When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.
When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.
In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary. Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.
In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable. This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.
A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables. This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.
And:
+From: "Eric W. Biederman" <ebiederm@xmission.com>
If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S. So we need to redo the effort here,
so we pick up PSE, PGE and the like.
Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
2007-05-02 13:27:13 -04:00
|
|
|
paravirt_pagetable_setup_done(pgd_base);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2007-07-29 17:24:36 -04:00
|
|
|
#if defined(CONFIG_HIBERNATION) || defined(CONFIG_ACPI)
|
2005-04-16 18:20:36 -04:00
|
|
|
/*
|
|
|
|
* Swap suspend & friends need this for resume because things like the intel-agp
|
|
|
|
* driver might have split up a kernel 4MB mapping.
|
|
|
|
*/
|
|
|
|
char __nosavedata swsusp_pg_dir[PAGE_SIZE]
|
|
|
|
__attribute__ ((aligned (PAGE_SIZE)));
|
|
|
|
|
|
|
|
static inline void save_pg_dir(void)
|
|
|
|
{
|
|
|
|
memcpy(swsusp_pg_dir, swapper_pg_dir, PAGE_SIZE);
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
static inline void save_pg_dir(void)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
void zap_low_mappings (void)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
save_pg_dir();
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Zap initial low-memory mappings.
|
|
|
|
*
|
|
|
|
* Note that "pgd_clear()" doesn't do it for
|
|
|
|
* us, because pgd_clear() is a no-op on i386.
|
|
|
|
*/
|
|
|
|
for (i = 0; i < USER_PTRS_PER_PGD; i++)
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
|
|
|
|
#else
|
|
|
|
set_pgd(swapper_pg_dir+i, __pgd(0));
|
|
|
|
#endif
|
|
|
|
flush_tlb_all();
|
|
|
|
}
|
|
|
|
|
2007-07-21 11:10:26 -04:00
|
|
|
int nx_enabled = 0;
|
|
|
|
|
2008-01-30 07:32:57 -05:00
|
|
|
pteval_t __supported_pte_mask __read_mostly = ~_PAGE_NX;
|
|
|
|
EXPORT_SYMBOL_GPL(__supported_pte_mask);
|
|
|
|
|
2007-07-21 11:10:26 -04:00
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
static int disable_nx __initdata = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* noexec = on|off
|
|
|
|
*
|
|
|
|
* Control non executable mappings.
|
|
|
|
*
|
|
|
|
* on Enable
|
|
|
|
* off Disable
|
|
|
|
*/
|
2006-09-26 04:52:32 -04:00
|
|
|
static int __init noexec_setup(char *str)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2006-09-26 04:52:32 -04:00
|
|
|
if (!str || !strcmp(str, "on")) {
|
|
|
|
if (cpu_has_nx) {
|
|
|
|
__supported_pte_mask |= _PAGE_NX;
|
|
|
|
disable_nx = 0;
|
|
|
|
}
|
|
|
|
} else if (!strcmp(str,"off")) {
|
2005-04-16 18:20:36 -04:00
|
|
|
disable_nx = 1;
|
|
|
|
__supported_pte_mask &= ~_PAGE_NX;
|
2006-09-26 04:52:32 -04:00
|
|
|
} else
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
return 0;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
2006-09-26 04:52:32 -04:00
|
|
|
early_param("noexec", noexec_setup);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
static void __init set_nx(void)
|
|
|
|
{
|
|
|
|
unsigned int v[4], l, h;
|
|
|
|
|
|
|
|
if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
|
|
|
|
cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
|
|
|
|
if ((v[3] & (1 << 20)) && !disable_nx) {
|
|
|
|
rdmsr(MSR_EFER, l, h);
|
|
|
|
l |= EFER_NX;
|
|
|
|
wrmsr(MSR_EFER, l, h);
|
|
|
|
nx_enabled = 1;
|
|
|
|
__supported_pte_mask |= _PAGE_NX;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
* paging_init() sets up the page tables - note that the first 8MB are
|
|
|
|
* already mapped by head.S.
|
|
|
|
*
|
|
|
|
* This routines 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)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
set_nx();
|
|
|
|
if (nx_enabled)
|
|
|
|
printk("NX (Execute Disable) protection: active\n");
|
|
|
|
#endif
|
|
|
|
|
|
|
|
pagetable_init();
|
|
|
|
|
|
|
|
load_cr3(swapper_pg_dir);
|
|
|
|
|
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
/*
|
|
|
|
* We will bail out later - printk doesn't work right now so
|
|
|
|
* the user would just see a hanging kernel.
|
|
|
|
*/
|
|
|
|
if (cpu_has_pae)
|
|
|
|
set_in_cr4(X86_CR4_PAE);
|
|
|
|
#endif
|
|
|
|
__flush_tlb_all();
|
|
|
|
|
|
|
|
kmap_init();
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Test if the WP bit works in supervisor mode. It isn't supported on 386's
|
|
|
|
* and also on some strange 486's (NexGen etc.). All 586+'s are OK. This
|
|
|
|
* used to involve black magic jumps to work around some nasty CPU bugs,
|
|
|
|
* but fortunately the switch to using exceptions got rid of all that.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static void __init test_wp_bit(void)
|
|
|
|
{
|
|
|
|
printk("Checking if this processor honours the WP bit even in supervisor mode... ");
|
|
|
|
|
|
|
|
/* Any page-aligned address will do, the test is non-destructive */
|
|
|
|
__set_fixmap(FIX_WP_TEST, __pa(&swapper_pg_dir), PAGE_READONLY);
|
|
|
|
boot_cpu_data.wp_works_ok = do_test_wp_bit();
|
|
|
|
clear_fixmap(FIX_WP_TEST);
|
|
|
|
|
|
|
|
if (!boot_cpu_data.wp_works_ok) {
|
|
|
|
printk("No.\n");
|
|
|
|
#ifdef CONFIG_X86_WP_WORKS_OK
|
|
|
|
panic("This kernel doesn't support CPU's with broken WP. Recompile it for a 386!");
|
|
|
|
#endif
|
|
|
|
} else {
|
|
|
|
printk("Ok.\n");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct kcore_list kcore_mem, kcore_vmalloc;
|
|
|
|
|
|
|
|
void __init mem_init(void)
|
|
|
|
{
|
|
|
|
extern int ppro_with_ram_bug(void);
|
|
|
|
int codesize, reservedpages, datasize, initsize;
|
|
|
|
int tmp;
|
|
|
|
int bad_ppro;
|
|
|
|
|
2005-06-23 03:07:57 -04:00
|
|
|
#ifdef CONFIG_FLATMEM
|
2006-10-03 17:34:58 -04:00
|
|
|
BUG_ON(!mem_map);
|
2005-04-16 18:20:36 -04:00
|
|
|
#endif
|
|
|
|
|
|
|
|
bad_ppro = ppro_with_ram_bug();
|
|
|
|
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
/* check that fixmap and pkmap do not overlap */
|
|
|
|
if (PKMAP_BASE+LAST_PKMAP*PAGE_SIZE >= FIXADDR_START) {
|
|
|
|
printk(KERN_ERR "fixmap and kmap areas overlap - this will crash\n");
|
|
|
|
printk(KERN_ERR "pkstart: %lxh pkend: %lxh fixstart %lxh\n",
|
|
|
|
PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE, FIXADDR_START);
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* this will put all low memory onto the freelists */
|
|
|
|
totalram_pages += free_all_bootmem();
|
|
|
|
|
|
|
|
reservedpages = 0;
|
|
|
|
for (tmp = 0; tmp < max_low_pfn; tmp++)
|
|
|
|
/*
|
|
|
|
* Only count reserved RAM pages
|
|
|
|
*/
|
|
|
|
if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp)))
|
|
|
|
reservedpages++;
|
|
|
|
|
|
|
|
set_highmem_pages_init(bad_ppro);
|
|
|
|
|
|
|
|
codesize = (unsigned long) &_etext - (unsigned long) &_text;
|
|
|
|
datasize = (unsigned long) &_edata - (unsigned long) &_etext;
|
|
|
|
initsize = (unsigned long) &__init_end - (unsigned long) &__init_begin;
|
|
|
|
|
|
|
|
kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT);
|
|
|
|
kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
|
|
|
|
VMALLOC_END-VMALLOC_START);
|
|
|
|
|
|
|
|
printk(KERN_INFO "Memory: %luk/%luk available (%dk kernel code, %dk reserved, %dk data, %dk init, %ldk highmem)\n",
|
|
|
|
(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
|
|
|
|
num_physpages << (PAGE_SHIFT-10),
|
|
|
|
codesize >> 10,
|
|
|
|
reservedpages << (PAGE_SHIFT-10),
|
|
|
|
datasize >> 10,
|
|
|
|
initsize >> 10,
|
|
|
|
(unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))
|
|
|
|
);
|
|
|
|
|
2006-09-26 02:32:25 -04:00
|
|
|
#if 1 /* double-sanity-check paranoia */
|
|
|
|
printk("virtual kernel memory layout:\n"
|
|
|
|
" fixmap : 0x%08lx - 0x%08lx (%4ld kB)\n"
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
" pkmap : 0x%08lx - 0x%08lx (%4ld kB)\n"
|
|
|
|
#endif
|
|
|
|
" vmalloc : 0x%08lx - 0x%08lx (%4ld MB)\n"
|
|
|
|
" lowmem : 0x%08lx - 0x%08lx (%4ld MB)\n"
|
|
|
|
" .init : 0x%08lx - 0x%08lx (%4ld kB)\n"
|
|
|
|
" .data : 0x%08lx - 0x%08lx (%4ld kB)\n"
|
|
|
|
" .text : 0x%08lx - 0x%08lx (%4ld kB)\n",
|
|
|
|
FIXADDR_START, FIXADDR_TOP,
|
|
|
|
(FIXADDR_TOP - FIXADDR_START) >> 10,
|
|
|
|
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE,
|
|
|
|
(LAST_PKMAP*PAGE_SIZE) >> 10,
|
|
|
|
#endif
|
|
|
|
|
|
|
|
VMALLOC_START, VMALLOC_END,
|
|
|
|
(VMALLOC_END - VMALLOC_START) >> 20,
|
|
|
|
|
|
|
|
(unsigned long)__va(0), (unsigned long)high_memory,
|
|
|
|
((unsigned long)high_memory - (unsigned long)__va(0)) >> 20,
|
|
|
|
|
|
|
|
(unsigned long)&__init_begin, (unsigned long)&__init_end,
|
|
|
|
((unsigned long)&__init_end - (unsigned long)&__init_begin) >> 10,
|
|
|
|
|
|
|
|
(unsigned long)&_etext, (unsigned long)&_edata,
|
|
|
|
((unsigned long)&_edata - (unsigned long)&_etext) >> 10,
|
|
|
|
|
|
|
|
(unsigned long)&_text, (unsigned long)&_etext,
|
|
|
|
((unsigned long)&_etext - (unsigned long)&_text) >> 10);
|
|
|
|
|
|
|
|
#ifdef CONFIG_HIGHMEM
|
|
|
|
BUG_ON(PKMAP_BASE+LAST_PKMAP*PAGE_SIZE > FIXADDR_START);
|
|
|
|
BUG_ON(VMALLOC_END > PKMAP_BASE);
|
|
|
|
#endif
|
|
|
|
BUG_ON(VMALLOC_START > VMALLOC_END);
|
|
|
|
BUG_ON((unsigned long)high_memory > VMALLOC_START);
|
|
|
|
#endif /* double-sanity-check paranoia */
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
#ifdef CONFIG_X86_PAE
|
|
|
|
if (!cpu_has_pae)
|
|
|
|
panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
|
|
|
|
#endif
|
|
|
|
if (boot_cpu_data.wp_works_ok < 0)
|
|
|
|
test_wp_bit();
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Subtle. SMP is doing it's boot stuff late (because it has to
|
|
|
|
* fork idle threads) - but it also needs low mappings for the
|
|
|
|
* protected-mode entry to work. We zap these entries only after
|
|
|
|
* the WP-bit has been tested.
|
|
|
|
*/
|
|
|
|
#ifndef CONFIG_SMP
|
|
|
|
zap_low_mappings();
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2006-05-20 18:00:03 -04:00
|
|
|
#ifdef CONFIG_MEMORY_HOTPLUG
|
2006-06-27 05:53:30 -04:00
|
|
|
int arch_add_memory(int nid, u64 start, u64 size)
|
2005-10-29 21:16:57 -04:00
|
|
|
{
|
2006-12-22 04:11:13 -05:00
|
|
|
struct pglist_data *pgdata = NODE_DATA(nid);
|
2006-09-26 02:31:09 -04:00
|
|
|
struct zone *zone = pgdata->node_zones + ZONE_HIGHMEM;
|
2005-10-29 21:16:57 -04:00
|
|
|
unsigned long start_pfn = start >> PAGE_SHIFT;
|
|
|
|
unsigned long nr_pages = size >> PAGE_SHIFT;
|
|
|
|
|
|
|
|
return __add_pages(zone, start_pfn, nr_pages);
|
|
|
|
}
|
|
|
|
|
2006-04-07 13:49:15 -04:00
|
|
|
#endif
|
2005-10-29 21:16:57 -04:00
|
|
|
|
2006-12-06 23:33:20 -05:00
|
|
|
struct kmem_cache *pmd_cache;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
void __init pgtable_cache_init(void)
|
|
|
|
{
|
2007-10-16 14:51:29 -04:00
|
|
|
if (PTRS_PER_PMD > 1)
|
2005-04-16 18:20:36 -04:00
|
|
|
pmd_cache = kmem_cache_create("pmd",
|
2007-10-16 14:51:29 -04:00
|
|
|
PTRS_PER_PMD*sizeof(pmd_t),
|
|
|
|
PTRS_PER_PMD*sizeof(pmd_t),
|
|
|
|
SLAB_PANIC,
|
|
|
|
pmd_ctor);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This function cannot be __init, since exceptions don't work in that
|
|
|
|
* section. Put this after the callers, so that it cannot be inlined.
|
|
|
|
*/
|
|
|
|
static int noinline do_test_wp_bit(void)
|
|
|
|
{
|
|
|
|
char tmp_reg;
|
|
|
|
int flag;
|
|
|
|
|
|
|
|
__asm__ __volatile__(
|
|
|
|
" movb %0,%1 \n"
|
|
|
|
"1: movb %1,%0 \n"
|
|
|
|
" xorl %2,%2 \n"
|
|
|
|
"2: \n"
|
|
|
|
".section __ex_table,\"a\"\n"
|
|
|
|
" .align 4 \n"
|
|
|
|
" .long 1b,2b \n"
|
|
|
|
".previous \n"
|
|
|
|
:"=m" (*(char *)fix_to_virt(FIX_WP_TEST)),
|
|
|
|
"=q" (tmp_reg),
|
|
|
|
"=r" (flag)
|
|
|
|
:"2" (1)
|
|
|
|
:"memory");
|
|
|
|
|
|
|
|
return flag;
|
|
|
|
}
|
|
|
|
|
2006-01-06 03:12:02 -05:00
|
|
|
#ifdef CONFIG_DEBUG_RODATA
|
|
|
|
|
|
|
|
void mark_rodata_ro(void)
|
|
|
|
{
|
[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>
2007-05-02 13:27:10 -04:00
|
|
|
unsigned long start = PFN_ALIGN(_text);
|
|
|
|
unsigned long size = PFN_ALIGN(_etext) - start;
|
2006-01-06 03:12:02 -05:00
|
|
|
|
2007-07-26 15:07:21 -04:00
|
|
|
#ifndef CONFIG_KPROBES
|
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
|
|
/* It must still be possible to apply SMP alternatives. */
|
|
|
|
if (num_possible_cpus() <= 1)
|
|
|
|
#endif
|
|
|
|
{
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT);
|
2007-07-26 15:07:21 -04:00
|
|
|
printk("Write protecting the kernel text: %luk\n", size >> 10);
|
2008-01-30 07:33:42 -05:00
|
|
|
|
|
|
|
#ifdef CONFIG_CPA_DEBUG
|
|
|
|
printk("Testing CPA: Reverting %lx-%lx\n", start, start+size);
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_rw(virt_to_page(start), size>>PAGE_SHIFT);
|
2008-01-30 07:33:42 -05:00
|
|
|
|
|
|
|
printk("Testing CPA: write protecting again\n");
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_ro(virt_to_page(start), size>>PAGE_SHIFT);
|
2008-01-30 07:33:42 -05:00
|
|
|
#endif
|
2007-07-26 15:07:21 -04:00
|
|
|
}
|
|
|
|
#endif
|
[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>
2007-05-02 13:27:10 -04:00
|
|
|
start += size;
|
|
|
|
size = (unsigned long)__end_rodata - start;
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT);
|
[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>
2007-05-02 13:27:10 -04:00
|
|
|
printk("Write protecting the kernel read-only data: %luk\n",
|
|
|
|
size >> 10);
|
2006-01-06 03:12:02 -05:00
|
|
|
|
2008-01-30 07:33:42 -05:00
|
|
|
#ifdef CONFIG_CPA_DEBUG
|
|
|
|
printk("Testing CPA: undo %lx-%lx\n", start, start + size);
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_rw(virt_to_page(start), size >> PAGE_SHIFT);
|
2008-01-30 07:33:42 -05:00
|
|
|
|
|
|
|
printk("Testing CPA: write protecting again\n");
|
2008-01-30 07:34:06 -05:00
|
|
|
set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT);
|
2008-01-30 07:33:42 -05:00
|
|
|
#endif
|
2006-01-06 03:12:02 -05:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2006-03-23 05:59:32 -05:00
|
|
|
void free_init_pages(char *what, unsigned long begin, unsigned long end)
|
|
|
|
{
|
|
|
|
unsigned long addr;
|
|
|
|
|
2008-01-30 07:34:07 -05:00
|
|
|
/*
|
|
|
|
* We just marked the kernel text read only above, now that
|
|
|
|
* we are going to free part of that, we need to make that
|
|
|
|
* writeable first.
|
|
|
|
*/
|
|
|
|
set_memory_rw(begin, (end - begin) >> PAGE_SHIFT);
|
|
|
|
|
2006-03-23 05:59:32 -05:00
|
|
|
for (addr = begin; addr < end; addr += PAGE_SIZE) {
|
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 11:44:24 -04:00
|
|
|
ClearPageReserved(virt_to_page(addr));
|
|
|
|
init_page_count(virt_to_page(addr));
|
|
|
|
memset((void *)addr, POISON_FREE_INITMEM, PAGE_SIZE);
|
|
|
|
free_page(addr);
|
2006-03-23 05:59:32 -05:00
|
|
|
totalram_pages++;
|
|
|
|
}
|
[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>
2007-05-02 13:27:10 -04:00
|
|
|
printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
|
2006-03-23 05:59:32 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
void free_initmem(void)
|
|
|
|
{
|
|
|
|
free_init_pages("unused kernel memory",
|
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 11:44:24 -04:00
|
|
|
(unsigned long)(&__init_begin),
|
|
|
|
(unsigned long)(&__init_end));
|
2006-03-23 05:59:32 -05:00
|
|
|
}
|
2006-01-06 03:12:02 -05:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
#ifdef CONFIG_BLK_DEV_INITRD
|
|
|
|
void free_initrd_mem(unsigned long start, unsigned long end)
|
|
|
|
{
|
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 11:44:24 -04:00
|
|
|
free_init_pages("initrd memory", start, end);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
#endif
|
2006-03-23 05:59:32 -05:00
|
|
|
|