[deliverable/linux.git] / arch / x86 / kernel / machine_kexec_64.c

/*
 * handle transition of Linux booting another kernel
 * Copyright (C) 2002-2005 Eric Biederman  <ebiederm@xmission.com>
 *
 * This source code is licensed under the GNU General Public License,
 * Version 2.  See the file COPYING for more details.
 */

#include <linux/mm.h>
#include <linux/kexec.h>
#include <linux/string.h>
#include <linux/reboot.h>
#include <linux/numa.h>
#include <linux/ftrace.h>

#include <asm/pgtable.h>
#include <asm/tlbflush.h>
#include <asm/mmu_context.h>
#include <asm/io.h>

#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
static u64 kexec_pgd[512] PAGE_ALIGNED;
static u64 kexec_pud0[512] PAGE_ALIGNED;
static u64 kexec_pmd0[512] PAGE_ALIGNED;
static u64 kexec_pte0[512] PAGE_ALIGNED;
static u64 kexec_pud1[512] PAGE_ALIGNED;
static u64 kexec_pmd1[512] PAGE_ALIGNED;
static u64 kexec_pte1[512] PAGE_ALIGNED;

static void init_level2_page(pmd_t *level2p, unsigned long addr)
{
	unsigned long end_addr;

	addr &= PAGE_MASK;
	end_addr = addr + PUD_SIZE;
	while (addr < end_addr) {
		set_pmd(level2p++, __pmd(addr | __PAGE_KERNEL_LARGE_EXEC));
		addr += PMD_SIZE;
	}
}

static int init_level3_page(struct kimage *image, pud_t *level3p,
				unsigned long addr, unsigned long last_addr)
{
	unsigned long end_addr;
	int result;

	result = 0;
	addr &= PAGE_MASK;
	end_addr = addr + PGDIR_SIZE;
	while ((addr < last_addr) && (addr < end_addr)) {
		struct page *page;
		pmd_t *level2p;

		page = kimage_alloc_control_pages(image, 0);
		if (!page) {
			result = -ENOMEM;
			goto out;
		}
		level2p = (pmd_t *)page_address(page);
		init_level2_page(level2p, addr);
		set_pud(level3p++, __pud(__pa(level2p) | _KERNPG_TABLE));
		addr += PUD_SIZE;
	}
	/* clear the unused entries */
	while (addr < end_addr) {
		pud_clear(level3p++);
		addr += PUD_SIZE;
	}
out:
	return result;
}


static int init_level4_page(struct kimage *image, pgd_t *level4p,
				unsigned long addr, unsigned long last_addr)
{
	unsigned long end_addr;
	int result;

	result = 0;
	addr &= PAGE_MASK;
	end_addr = addr + (PTRS_PER_PGD * PGDIR_SIZE);
	while ((addr < last_addr) && (addr < end_addr)) {
		struct page *page;
		pud_t *level3p;

		page = kimage_alloc_control_pages(image, 0);
		if (!page) {
			result = -ENOMEM;
			goto out;
		}
		level3p = (pud_t *)page_address(page);
		result = init_level3_page(image, level3p, addr, last_addr);
		if (result) {
			goto out;
		}
		set_pgd(level4p++, __pgd(__pa(level3p) | _KERNPG_TABLE));
		addr += PGDIR_SIZE;
	}
	/* clear the unused entries */
	while (addr < end_addr) {
		pgd_clear(level4p++);
		addr += PGDIR_SIZE;
	}
out:
	return result;
}


static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
{
	pgd_t *level4p;
	level4p = (pgd_t *)__va(start_pgtable);
	return init_level4_page(image, level4p, 0, max_pfn << PAGE_SHIFT);
}

static void set_idt(void *newidt, u16 limit)
{
	struct desc_ptr curidt;

	/* x86-64 supports unaliged loads & stores */
	curidt.size    = limit;
	curidt.address = (unsigned long)newidt;

	__asm__ __volatile__ (
		"lidtq %0\n"
		: : "m" (curidt)
		);
};


static void set_gdt(void *newgdt, u16 limit)
{
	struct desc_ptr curgdt;

	/* x86-64 supports unaligned loads & stores */
	curgdt.size    = limit;
	curgdt.address = (unsigned long)newgdt;

	__asm__ __volatile__ (
		"lgdtq %0\n"
		: : "m" (curgdt)
		);
};

static void load_segments(void)
{
	__asm__ __volatile__ (
		"\tmovl %0,%%ds\n"
		"\tmovl %0,%%es\n"
		"\tmovl %0,%%ss\n"
		"\tmovl %0,%%fs\n"
		"\tmovl %0,%%gs\n"
		: : "a" (__KERNEL_DS) : "memory"
		);
}

int machine_kexec_prepare(struct kimage *image)
{
	unsigned long start_pgtable;
	int result;

	/* Calculate the offsets */
	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;

	/* Setup the identity mapped 64bit page table */
	result = init_pgtable(image, start_pgtable);
	if (result)
		return result;

	return 0;
}

void machine_kexec_cleanup(struct kimage *image)
{
	return;
}

/*
 * Do not allocate memory (or fail in any way) in machine_kexec().
 * We are past the point of no return, committed to rebooting now.
 */
void machine_kexec(struct kimage *image)
{
	unsigned long page_list[PAGES_NR];
	void *control_page;

	tracer_disable();

	/* Interrupts aren't acceptable while we reboot */
	local_irq_disable();

	control_page = page_address(image->control_code_page) + PAGE_SIZE;
	memcpy(control_page, relocate_kernel, PAGE_SIZE);

	page_list[PA_CONTROL_PAGE] = virt_to_phys(control_page);
	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
	page_list[PA_PGD] = virt_to_phys(&kexec_pgd);
	page_list[VA_PGD] = (unsigned long)kexec_pgd;
	page_list[PA_PUD_0] = virt_to_phys(&kexec_pud0);
	page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
	page_list[PA_PMD_0] = virt_to_phys(&kexec_pmd0);
	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
	page_list[PA_PTE_0] = virt_to_phys(&kexec_pte0);
	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
	page_list[PA_PUD_1] = virt_to_phys(&kexec_pud1);
	page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
	page_list[PA_PMD_1] = virt_to_phys(&kexec_pmd1);
	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
	page_list[PA_PTE_1] = virt_to_phys(&kexec_pte1);
	page_list[VA_PTE_1] = (unsigned long)kexec_pte1;

	page_list[PA_TABLE_PAGE] =
	  (unsigned long)__pa(page_address(image->control_code_page));

	/* The segment registers are funny things, they have both a
	 * visible and an invisible part.  Whenever the visible part is
	 * set to a specific selector, the invisible part is loaded
	 * with from a table in memory.  At no other time is the
	 * descriptor table in memory accessed.
	 *
	 * I take advantage of this here by force loading the
	 * segments, before I zap the gdt with an invalid value.
	 */
	load_segments();
	/* The gdt & idt are now invalid.
	 * If you want to load them you must set up your own idt & gdt.
	 */
	set_gdt(phys_to_virt(0),0);
	set_idt(phys_to_virt(0),0);

	/* now call it */
	relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
			image->start);
}

void arch_crash_save_vmcoreinfo(void)
{
	VMCOREINFO_SYMBOL(phys_base);
	VMCOREINFO_SYMBOL(init_level4_pgt);

#ifdef CONFIG_NUMA
	VMCOREINFO_SYMBOL(node_data);
	VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
#endif
}
Commit	Line	Data
5234f5eb	1	/*
835c34a1	2	* handle transition of Linux booting another kernel
5234f5eb EB	3	* Copyright (C) 2002-2005 Eric Biederman <ebiederm@xmission.com>
	4	*
	5	* This source code is licensed under the GNU General Public License,
	6	* Version 2. See the file COPYING for more details.
	7	*/
	8
	9	#include <linux/mm.h>
	10	#include <linux/kexec.h>
5234f5eb EB	11	#include <linux/string.h>
5234f5eb EB	12	#include <linux/reboot.h>
fd59d231	13	#include <linux/numa.h>
f43fdad8 IM	14	#include <linux/ftrace.h>
f43fdad8 IM	15
5234f5eb	16	#include <asm/pgtable.h>
5234f5eb EB	17	#include <asm/tlbflush.h>
	18	#include <asm/mmu_context.h>
	19	#include <asm/io.h>
8bf27556	20
4bfaaef0 MD	21	#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
	22	static u64 kexec_pgd[512] PAGE_ALIGNED;
	23	static u64 kexec_pud0[512] PAGE_ALIGNED;
	24	static u64 kexec_pmd0[512] PAGE_ALIGNED;
	25	static u64 kexec_pte0[512] PAGE_ALIGNED;
	26	static u64 kexec_pud1[512] PAGE_ALIGNED;
	27	static u64 kexec_pmd1[512] PAGE_ALIGNED;
	28	static u64 kexec_pte1[512] PAGE_ALIGNED;
	29
8bf27556	30	static void init_level2_page(pmd_t *level2p, unsigned long addr)
5234f5eb EB	31	{
5234f5eb EB	32	unsigned long end_addr;
72414d3f	33
5234f5eb	34	addr &= PAGE_MASK;
8bf27556	35	end_addr = addr + PUD_SIZE;
72414d3f	36	while (addr < end_addr) {
8bf27556 EB	37	set_pmd(level2p++, __pmd(addr \| __PAGE_KERNEL_LARGE_EXEC));
8bf27556 EB	38	addr += PMD_SIZE;
5234f5eb EB	39	}
	40	}
	41
8bf27556	42	static int init_level3_page(struct kimage image, pud_t level3p,
72414d3f	43	unsigned long addr, unsigned long last_addr)
5234f5eb EB	44	{
	45	unsigned long end_addr;
	46	int result;
72414d3f	47
5234f5eb EB	48	result = 0;
5234f5eb EB	49	addr &= PAGE_MASK;
8bf27556	50	end_addr = addr + PGDIR_SIZE;
72414d3f	51	while ((addr < last_addr) && (addr < end_addr)) {
5234f5eb	52	struct page *page;
8bf27556	53	pmd_t *level2p;
72414d3f	54
5234f5eb EB	55	page = kimage_alloc_control_pages(image, 0);
	56	if (!page) {
	57	result = -ENOMEM;
	58	goto out;
	59	}
8bf27556	60	level2p = (pmd_t *)page_address(page);
5234f5eb	61	init_level2_page(level2p, addr);
8bf27556 EB	62	set_pud(level3p++, __pud(__pa(level2p) \| _KERNPG_TABLE));
8bf27556 EB	63	addr += PUD_SIZE;
5234f5eb EB	64	}
5234f5eb EB	65	/* clear the unused entries */
72414d3f	66	while (addr < end_addr) {
8bf27556 EB	67	pud_clear(level3p++);
8bf27556 EB	68	addr += PUD_SIZE;
5234f5eb EB	69	}
	70	out:
	71	return result;
	72	}
	73
	74
8bf27556	75	static int init_level4_page(struct kimage image, pgd_t level4p,
72414d3f	76	unsigned long addr, unsigned long last_addr)
5234f5eb EB	77	{
	78	unsigned long end_addr;
	79	int result;
72414d3f	80
5234f5eb EB	81	result = 0;
5234f5eb EB	82	addr &= PAGE_MASK;
8bf27556	83	end_addr = addr + (PTRS_PER_PGD * PGDIR_SIZE);
72414d3f	84	while ((addr < last_addr) && (addr < end_addr)) {
5234f5eb	85	struct page *page;
8bf27556	86	pud_t *level3p;
72414d3f	87
5234f5eb EB	88	page = kimage_alloc_control_pages(image, 0);
	89	if (!page) {
	90	result = -ENOMEM;
	91	goto out;
	92	}
8bf27556	93	level3p = (pud_t *)page_address(page);
5234f5eb EB	94	result = init_level3_page(image, level3p, addr, last_addr);
	95	if (result) {
	96	goto out;
	97	}
8bf27556 EB	98	set_pgd(level4p++, __pgd(__pa(level3p) \| _KERNPG_TABLE));
8bf27556 EB	99	addr += PGDIR_SIZE;
5234f5eb EB	100	}
5234f5eb EB	101	/* clear the unused entries */
72414d3f	102	while (addr < end_addr) {
8bf27556 EB	103	pgd_clear(level4p++);
8bf27556 EB	104	addr += PGDIR_SIZE;
5234f5eb	105	}
72414d3f	106	out:
5234f5eb EB	107	return result;
	108	}
	109
	110
	111	static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
	112	{
8bf27556 EB	113	pgd_t *level4p;
8bf27556 EB	114	level4p = (pgd_t *)__va(start_pgtable);
c987d12f	115	return init_level4_page(image, level4p, 0, max_pfn << PAGE_SHIFT);
5234f5eb EB	116	}
	117
	118	static void set_idt(void *newidt, u16 limit)
	119	{
36c4fd23	120	struct desc_ptr curidt;
5234f5eb EB	121
5234f5eb EB	122	/* x86-64 supports unaliged loads & stores */
36c4fd23 EB	123	curidt.size = limit;
36c4fd23 EB	124	curidt.address = (unsigned long)newidt;
5234f5eb EB	125
5234f5eb EB	126	__asm__ __volatile__ (
36c4fd23 EB	127	"lidtq %0\n"
36c4fd23 EB	128	: : "m" (curidt)
5234f5eb EB	129	);
	130	};
	131
	132
	133	static void set_gdt(void *newgdt, u16 limit)
	134	{
36c4fd23	135	struct desc_ptr curgdt;
5234f5eb EB	136
5234f5eb EB	137	/* x86-64 supports unaligned loads & stores */
36c4fd23 EB	138	curgdt.size = limit;
36c4fd23 EB	139	curgdt.address = (unsigned long)newgdt;
5234f5eb EB	140
5234f5eb EB	141	__asm__ __volatile__ (
36c4fd23 EB	142	"lgdtq %0\n"
36c4fd23 EB	143	: : "m" (curgdt)
5234f5eb EB	144	);
	145	};
	146
	147	static void load_segments(void)
	148	{
	149	__asm__ __volatile__ (
36c4fd23 EB	150	"\tmovl %0,%%ds\n"
	151	"\tmovl %0,%%es\n"
	152	"\tmovl %0,%%ss\n"
	153	"\tmovl %0,%%fs\n"
	154	"\tmovl %0,%%gs\n"
2ec5e3a8	155	: : "a" (__KERNEL_DS) : "memory"
5234f5eb	156	);
5234f5eb EB	157	}
5234f5eb EB	158
5234f5eb EB	159	int machine_kexec_prepare(struct kimage *image)
5234f5eb EB	160	{
4bfaaef0	161	unsigned long start_pgtable;
5234f5eb EB	162	int result;
	163
	164	/* Calculate the offsets */
72414d3f	165	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;
5234f5eb EB	166
	167	/* Setup the identity mapped 64bit page table */
	168	result = init_pgtable(image, start_pgtable);
72414d3f	169	if (result)
5234f5eb	170	return result;
5234f5eb	171
5234f5eb EB	172	return 0;
	173	}
	174
	175	void machine_kexec_cleanup(struct kimage *image)
	176	{
	177	return;
	178	}
	179
	180	/*
	181	* Do not allocate memory (or fail in any way) in machine_kexec().
	182	* We are past the point of no return, committed to rebooting now.
	183	*/
3ab83521	184	void machine_kexec(struct kimage *image)
5234f5eb	185	{
4bfaaef0 MD	186	unsigned long page_list[PAGES_NR];
4bfaaef0 MD	187	void *control_page;
5234f5eb	188
f43fdad8 IM	189	tracer_disable();
f43fdad8 IM	190
5234f5eb EB	191	/* Interrupts aren't acceptable while we reboot */
	192	local_irq_disable();
	193
4bfaaef0 MD	194	control_page = page_address(image->control_code_page) + PAGE_SIZE;
	195	memcpy(control_page, relocate_kernel, PAGE_SIZE);
	196
e3ebadd9	197	page_list[PA_CONTROL_PAGE] = virt_to_phys(control_page);
4bfaaef0	198	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
e3ebadd9	199	page_list[PA_PGD] = virt_to_phys(&kexec_pgd);
4bfaaef0	200	page_list[VA_PGD] = (unsigned long)kexec_pgd;
e3ebadd9	201	page_list[PA_PUD_0] = virt_to_phys(&kexec_pud0);
4bfaaef0	202	page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
e3ebadd9	203	page_list[PA_PMD_0] = virt_to_phys(&kexec_pmd0);
4bfaaef0	204	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
e3ebadd9	205	page_list[PA_PTE_0] = virt_to_phys(&kexec_pte0);
4bfaaef0	206	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
e3ebadd9	207	page_list[PA_PUD_1] = virt_to_phys(&kexec_pud1);
4bfaaef0	208	page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
e3ebadd9	209	page_list[PA_PMD_1] = virt_to_phys(&kexec_pmd1);
4bfaaef0	210	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
e3ebadd9	211	page_list[PA_PTE_1] = virt_to_phys(&kexec_pte1);
4bfaaef0 MD	212	page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
	213
	214	page_list[PA_TABLE_PAGE] =
	215	(unsigned long)__pa(page_address(image->control_code_page));
5234f5eb	216
2a8a3d5b EB	217	/* The segment registers are funny things, they have both a
	218	* visible and an invisible part. Whenever the visible part is
	219	* set to a specific selector, the invisible part is loaded
	220	* with from a table in memory. At no other time is the
	221	* descriptor table in memory accessed.
5234f5eb EB	222	*
	223	* I take advantage of this here by force loading the
	224	* segments, before I zap the gdt with an invalid value.
	225	*/
	226	load_segments();
	227	/* The gdt & idt are now invalid.
	228	* If you want to load them you must set up your own idt & gdt.
	229	*/
	230	set_gdt(phys_to_virt(0),0);
	231	set_idt(phys_to_virt(0),0);
4bfaaef0	232
5234f5eb	233	/* now call it */
4bfaaef0 MD	234	relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
4bfaaef0 MD	235	image->start);
5234f5eb	236	}
2c8c0e6b	237
fd59d231 KO	238	void arch_crash_save_vmcoreinfo(void)
fd59d231 KO	239	{
629c8b4c	240	VMCOREINFO_SYMBOL(phys_base);
69243f91	241	VMCOREINFO_SYMBOL(init_level4_pgt);
92df5c3e KO	242
	243	#ifdef CONFIG_NUMA
	244	VMCOREINFO_SYMBOL(node_data);
	245	VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
	246	#endif
fd59d231 KO	247	}
fd59d231 KO	248