[deliverable/linux.git] / arch / x86_64 / kernel / machine_kexec.c

/*
 * machine_kexec.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 <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, end_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.
 */
NORET_TYPE void machine_kexec(struct kimage *image)
{
	unsigned long page_list[PAGES_NR];
	void *control_page;

	/* 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] = __pa(control_page);
	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
	page_list[PA_PGD] = __pa_symbol(&kexec_pgd);
	page_list[VA_PGD] = (unsigned long)kexec_pgd;
	page_list[PA_PUD_0] = __pa_symbol(&kexec_pud0);
	page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
	page_list[PA_PMD_0] = __pa_symbol(&kexec_pmd0);
	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
	page_list[PA_PTE_0] = __pa_symbol(&kexec_pte0);
	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
	page_list[PA_PUD_1] = __pa_symbol(&kexec_pud1);
	page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
	page_list[PA_PMD_1] = __pa_symbol(&kexec_pmd1);
	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
	page_list[PA_PTE_1] = __pa_symbol(&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);
}

/* crashkernel=size@addr specifies the location to reserve for
 * a crash kernel.  By reserving this memory we guarantee
 * that linux never set's it up as a DMA target.
 * Useful for holding code to do something appropriate
 * after a kernel panic.
 */
static int __init setup_crashkernel(char *arg)
{
	unsigned long size, base;
	char *p;
	if (!arg)
		return -EINVAL;
	size = memparse(arg, &p);
	if (arg == p)
		return -EINVAL;
	if (*p == '@') {
		base = memparse(p+1, &p);
		/* FIXME: Do I want a sanity check to validate the
		 * memory range?  Yes you do, but it's too early for
		 * e820 -AK */
		crashk_res.start = base;
		crashk_res.end   = base + size - 1;
	}
	return 0;
}
early_param("crashkernel", setup_crashkernel);
Commit	Line	Data
5234f5eb EB	1	/*
	2	* machine_kexec.c - handle transition of Linux booting another kernel
	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>
5234f5eb	13	#include <asm/pgtable.h>
5234f5eb EB	14	#include <asm/tlbflush.h>
	15	#include <asm/mmu_context.h>
	16	#include <asm/io.h>
8bf27556	17
4bfaaef0 MD	18	#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
	19	static u64 kexec_pgd[512] PAGE_ALIGNED;
	20	static u64 kexec_pud0[512] PAGE_ALIGNED;
	21	static u64 kexec_pmd0[512] PAGE_ALIGNED;
	22	static u64 kexec_pte0[512] PAGE_ALIGNED;
	23	static u64 kexec_pud1[512] PAGE_ALIGNED;
	24	static u64 kexec_pmd1[512] PAGE_ALIGNED;
	25	static u64 kexec_pte1[512] PAGE_ALIGNED;
	26
8bf27556	27	static void init_level2_page(pmd_t *level2p, unsigned long addr)
5234f5eb EB	28	{
5234f5eb EB	29	unsigned long end_addr;
72414d3f	30
5234f5eb	31	addr &= PAGE_MASK;
8bf27556	32	end_addr = addr + PUD_SIZE;
72414d3f	33	while (addr < end_addr) {
8bf27556 EB	34	set_pmd(level2p++, __pmd(addr \| __PAGE_KERNEL_LARGE_EXEC));
8bf27556 EB	35	addr += PMD_SIZE;
5234f5eb EB	36	}
	37	}
	38
8bf27556	39	static int init_level3_page(struct kimage image, pud_t level3p,
72414d3f	40	unsigned long addr, unsigned long last_addr)
5234f5eb EB	41	{
	42	unsigned long end_addr;
	43	int result;
72414d3f	44
5234f5eb EB	45	result = 0;
5234f5eb EB	46	addr &= PAGE_MASK;
8bf27556	47	end_addr = addr + PGDIR_SIZE;
72414d3f	48	while ((addr < last_addr) && (addr < end_addr)) {
5234f5eb	49	struct page *page;
8bf27556	50	pmd_t *level2p;
72414d3f	51
5234f5eb EB	52	page = kimage_alloc_control_pages(image, 0);
	53	if (!page) {
	54	result = -ENOMEM;
	55	goto out;
	56	}
8bf27556	57	level2p = (pmd_t *)page_address(page);
5234f5eb	58	init_level2_page(level2p, addr);
8bf27556 EB	59	set_pud(level3p++, __pud(__pa(level2p) \| _KERNPG_TABLE));
8bf27556 EB	60	addr += PUD_SIZE;
5234f5eb EB	61	}
5234f5eb EB	62	/* clear the unused entries */
72414d3f	63	while (addr < end_addr) {
8bf27556 EB	64	pud_clear(level3p++);
8bf27556 EB	65	addr += PUD_SIZE;
5234f5eb EB	66	}
	67	out:
	68	return result;
	69	}
	70
	71
8bf27556	72	static int init_level4_page(struct kimage image, pgd_t level4p,
72414d3f	73	unsigned long addr, unsigned long last_addr)
5234f5eb EB	74	{
	75	unsigned long end_addr;
	76	int result;
72414d3f	77
5234f5eb EB	78	result = 0;
5234f5eb EB	79	addr &= PAGE_MASK;
8bf27556	80	end_addr = addr + (PTRS_PER_PGD * PGDIR_SIZE);
72414d3f	81	while ((addr < last_addr) && (addr < end_addr)) {
5234f5eb	82	struct page *page;
8bf27556	83	pud_t *level3p;
72414d3f	84
5234f5eb EB	85	page = kimage_alloc_control_pages(image, 0);
	86	if (!page) {
	87	result = -ENOMEM;
	88	goto out;
	89	}
8bf27556	90	level3p = (pud_t *)page_address(page);
5234f5eb EB	91	result = init_level3_page(image, level3p, addr, last_addr);
	92	if (result) {
	93	goto out;
	94	}
8bf27556 EB	95	set_pgd(level4p++, __pgd(__pa(level3p) \| _KERNPG_TABLE));
8bf27556 EB	96	addr += PGDIR_SIZE;
5234f5eb EB	97	}
5234f5eb EB	98	/* clear the unused entries */
72414d3f	99	while (addr < end_addr) {
8bf27556 EB	100	pgd_clear(level4p++);
8bf27556 EB	101	addr += PGDIR_SIZE;
5234f5eb	102	}
72414d3f	103	out:
5234f5eb EB	104	return result;
	105	}
	106
	107
	108	static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
	109	{
8bf27556 EB	110	pgd_t *level4p;
8bf27556 EB	111	level4p = (pgd_t *)__va(start_pgtable);
72414d3f	112	return init_level4_page(image, level4p, 0, end_pfn << PAGE_SHIFT);
5234f5eb EB	113	}
	114
	115	static void set_idt(void *newidt, u16 limit)
	116	{
36c4fd23	117	struct desc_ptr curidt;
5234f5eb EB	118
5234f5eb EB	119	/* x86-64 supports unaliged loads & stores */
36c4fd23 EB	120	curidt.size = limit;
36c4fd23 EB	121	curidt.address = (unsigned long)newidt;
5234f5eb EB	122
5234f5eb EB	123	__asm__ __volatile__ (
36c4fd23 EB	124	"lidtq %0\n"
36c4fd23 EB	125	: : "m" (curidt)
5234f5eb EB	126	);
	127	};
	128
	129
	130	static void set_gdt(void *newgdt, u16 limit)
	131	{
36c4fd23	132	struct desc_ptr curgdt;
5234f5eb EB	133
5234f5eb EB	134	/* x86-64 supports unaligned loads & stores */
36c4fd23 EB	135	curgdt.size = limit;
36c4fd23 EB	136	curgdt.address = (unsigned long)newgdt;
5234f5eb EB	137
5234f5eb EB	138	__asm__ __volatile__ (
36c4fd23 EB	139	"lgdtq %0\n"
36c4fd23 EB	140	: : "m" (curgdt)
5234f5eb EB	141	);
	142	};
	143
	144	static void load_segments(void)
	145	{
	146	__asm__ __volatile__ (
36c4fd23 EB	147	"\tmovl %0,%%ds\n"
	148	"\tmovl %0,%%es\n"
	149	"\tmovl %0,%%ss\n"
	150	"\tmovl %0,%%fs\n"
	151	"\tmovl %0,%%gs\n"
2ec5e3a8	152	: : "a" (__KERNEL_DS) : "memory"
5234f5eb	153	);
5234f5eb EB	154	}
5234f5eb EB	155
5234f5eb EB	156	int machine_kexec_prepare(struct kimage *image)
5234f5eb EB	157	{
4bfaaef0	158	unsigned long start_pgtable;
5234f5eb EB	159	int result;
	160
	161	/* Calculate the offsets */
72414d3f	162	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;
5234f5eb EB	163
	164	/* Setup the identity mapped 64bit page table */
	165	result = init_pgtable(image, start_pgtable);
72414d3f	166	if (result)
5234f5eb	167	return result;
5234f5eb	168
5234f5eb EB	169	return 0;
	170	}
	171
	172	void machine_kexec_cleanup(struct kimage *image)
	173	{
	174	return;
	175	}
	176
	177	/*
	178	* Do not allocate memory (or fail in any way) in machine_kexec().
	179	* We are past the point of no return, committed to rebooting now.
	180	*/
	181	NORET_TYPE void machine_kexec(struct kimage *image)
	182	{
4bfaaef0 MD	183	unsigned long page_list[PAGES_NR];
4bfaaef0 MD	184	void *control_page;
5234f5eb EB	185
	186	/* Interrupts aren't acceptable while we reboot */
	187	local_irq_disable();
	188
4bfaaef0 MD	189	control_page = page_address(image->control_code_page) + PAGE_SIZE;
	190	memcpy(control_page, relocate_kernel, PAGE_SIZE);
	191
	192	page_list[PA_CONTROL_PAGE] = __pa(control_page);
	193	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
0dbf7028	194	page_list[PA_PGD] = __pa_symbol(&kexec_pgd);
4bfaaef0	195	page_list[VA_PGD] = (unsigned long)kexec_pgd;
0dbf7028	196	page_list[PA_PUD_0] = __pa_symbol(&kexec_pud0);
4bfaaef0	197	page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
0dbf7028	198	page_list[PA_PMD_0] = __pa_symbol(&kexec_pmd0);
4bfaaef0	199	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
0dbf7028	200	page_list[PA_PTE_0] = __pa_symbol(&kexec_pte0);
4bfaaef0	201	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
0dbf7028	202	page_list[PA_PUD_1] = __pa_symbol(&kexec_pud1);
4bfaaef0	203	page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
0dbf7028	204	page_list[PA_PMD_1] = __pa_symbol(&kexec_pmd1);
4bfaaef0	205	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
0dbf7028	206	page_list[PA_PTE_1] = __pa_symbol(&kexec_pte1);
4bfaaef0 MD	207	page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
	208
	209	page_list[PA_TABLE_PAGE] =
	210	(unsigned long)__pa(page_address(image->control_code_page));
5234f5eb	211
2a8a3d5b EB	212	/* The segment registers are funny things, they have both a
	213	* visible and an invisible part. Whenever the visible part is
	214	* set to a specific selector, the invisible part is loaded
	215	* with from a table in memory. At no other time is the
	216	* descriptor table in memory accessed.
5234f5eb EB	217	*
	218	* I take advantage of this here by force loading the
	219	* segments, before I zap the gdt with an invalid value.
	220	*/
	221	load_segments();
	222	/* The gdt & idt are now invalid.
	223	* If you want to load them you must set up your own idt & gdt.
	224	*/
	225	set_gdt(phys_to_virt(0),0);
	226	set_idt(phys_to_virt(0),0);
4bfaaef0	227
5234f5eb	228	/* now call it */
4bfaaef0 MD	229	relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
4bfaaef0 MD	230	image->start);
5234f5eb	231	}
2c8c0e6b AK	232
	233	/* crashkernel=size@addr specifies the location to reserve for
	234	* a crash kernel. By reserving this memory we guarantee
	235	* that linux never set's it up as a DMA target.
	236	* Useful for holding code to do something appropriate
	237	* after a kernel panic.
	238	*/
	239	static int __init setup_crashkernel(char *arg)
	240	{
	241	unsigned long size, base;
	242	char *p;
	243	if (!arg)
	244	return -EINVAL;
	245	size = memparse(arg, &p);
	246	if (arg == p)
	247	return -EINVAL;
	248	if (*p == '@') {
	249	base = memparse(p+1, &p);
	250	/* FIXME: Do I want a sanity check to validate the
	251	* memory range? Yes you do, but it's too early for
	252	* e820 -AK */
	253	crashk_res.start = base;
	254	crashk_res.end = base + size - 1;
	255	}
	256	return 0;
	257	}
	258	early_param("crashkernel", setup_crashkernel);
	259