machine_kexec_32.c 6.8 KB
Newer Older
1
/*
D
Dave Jones 已提交
2
 * handle transition of Linux booting another kernel
3 4 5 6 7 8 9 10 11
 * 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/delay.h>
12
#include <linux/init.h>
K
Ken'ichi Ohmichi 已提交
13
#include <linux/numa.h>
I
Ingo Molnar 已提交
14
#include <linux/ftrace.h>
H
Huang Ying 已提交
15
#include <linux/suspend.h>
16
#include <linux/gfp.h>
17
#include <linux/io.h>
I
Ingo Molnar 已提交
18

19 20 21 22 23 24
#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/tlbflush.h>
#include <asm/mmu_context.h>
#include <asm/apic.h>
#include <asm/cpufeature.h>
25
#include <asm/desc.h>
26
#include <asm/system.h>
H
Huang Ying 已提交
27
#include <asm/cacheflush.h>
28 29 30

static void set_idt(void *newidt, __u16 limit)
{
31
	struct desc_ptr curidt;
32 33

	/* ia32 supports unaliged loads & stores */
34 35
	curidt.size    = limit;
	curidt.address = (unsigned long)newidt;
36

Z
Zachary Amsden 已提交
37
	load_idt(&curidt);
38
}
39 40 41 42


static void set_gdt(void *newgdt, __u16 limit)
{
43
	struct desc_ptr curgdt;
44 45

	/* ia32 supports unaligned loads & stores */
46 47
	curgdt.size    = limit;
	curgdt.address = (unsigned long)newgdt;
48

Z
Zachary Amsden 已提交
49
	load_gdt(&curgdt);
50
}
51 52 53 54 55 56 57 58 59

static void load_segments(void)
{
#define __STR(X) #X
#define STR(X) __STR(X)

	__asm__ __volatile__ (
		"\tljmp $"STR(__KERNEL_CS)",$1f\n"
		"\t1:\n"
M
Michael Matz 已提交
60 61 62 63 64 65
		"\tmovl $"STR(__KERNEL_DS)",%%eax\n"
		"\tmovl %%eax,%%ds\n"
		"\tmovl %%eax,%%es\n"
		"\tmovl %%eax,%%fs\n"
		"\tmovl %%eax,%%gs\n"
		"\tmovl %%eax,%%ss\n"
66
		: : : "eax", "memory");
67 68 69 70
#undef STR
#undef __STR
}

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
static void machine_kexec_free_page_tables(struct kimage *image)
{
	free_page((unsigned long)image->arch.pgd);
#ifdef CONFIG_X86_PAE
	free_page((unsigned long)image->arch.pmd0);
	free_page((unsigned long)image->arch.pmd1);
#endif
	free_page((unsigned long)image->arch.pte0);
	free_page((unsigned long)image->arch.pte1);
}

static int machine_kexec_alloc_page_tables(struct kimage *image)
{
	image->arch.pgd = (pgd_t *)get_zeroed_page(GFP_KERNEL);
#ifdef CONFIG_X86_PAE
	image->arch.pmd0 = (pmd_t *)get_zeroed_page(GFP_KERNEL);
	image->arch.pmd1 = (pmd_t *)get_zeroed_page(GFP_KERNEL);
#endif
	image->arch.pte0 = (pte_t *)get_zeroed_page(GFP_KERNEL);
	image->arch.pte1 = (pte_t *)get_zeroed_page(GFP_KERNEL);
	if (!image->arch.pgd ||
#ifdef CONFIG_X86_PAE
	    !image->arch.pmd0 || !image->arch.pmd1 ||
#endif
	    !image->arch.pte0 || !image->arch.pte1) {
		machine_kexec_free_page_tables(image);
		return -ENOMEM;
	}
	return 0;
}

102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
static void machine_kexec_page_table_set_one(
	pgd_t *pgd, pmd_t *pmd, pte_t *pte,
	unsigned long vaddr, unsigned long paddr)
{
	pud_t *pud;

	pgd += pgd_index(vaddr);
#ifdef CONFIG_X86_PAE
	if (!(pgd_val(*pgd) & _PAGE_PRESENT))
		set_pgd(pgd, __pgd(__pa(pmd) | _PAGE_PRESENT));
#endif
	pud = pud_offset(pgd, vaddr);
	pmd = pmd_offset(pud, vaddr);
	if (!(pmd_val(*pmd) & _PAGE_PRESENT))
		set_pmd(pmd, __pmd(__pa(pte) | _PAGE_TABLE));
	pte = pte_offset_kernel(pmd, vaddr);
	set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC));
}

static void machine_kexec_prepare_page_tables(struct kimage *image)
{
	void *control_page;
124
	pmd_t *pmd = NULL;
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140

	control_page = page_address(image->control_code_page);
#ifdef CONFIG_X86_PAE
	pmd = image->arch.pmd0;
#endif
	machine_kexec_page_table_set_one(
		image->arch.pgd, pmd, image->arch.pte0,
		(unsigned long)control_page, __pa(control_page));
#ifdef CONFIG_X86_PAE
	pmd = image->arch.pmd1;
#endif
	machine_kexec_page_table_set_one(
		image->arch.pgd, pmd, image->arch.pte1,
		__pa(control_page), __pa(control_page));
}

141 142 143
/*
 * A architecture hook called to validate the
 * proposed image and prepare the control pages
144
 * as needed.  The pages for KEXEC_CONTROL_PAGE_SIZE
145 146 147 148 149 150 151
 * have been allocated, but the segments have yet
 * been copied into the kernel.
 *
 * Do what every setup is needed on image and the
 * reboot code buffer to allow us to avoid allocations
 * later.
 *
152 153
 * - Make control page executable.
 * - Allocate page tables
154
 * - Setup page tables
155 156 157
 */
int machine_kexec_prepare(struct kimage *image)
{
158 159
	int error;

160
	set_pages_x(image->control_code_page, 1);
161 162 163 164 165
	error = machine_kexec_alloc_page_tables(image);
	if (error)
		return error;
	machine_kexec_prepare_page_tables(image);
	return 0;
166 167 168 169 170 171 172 173
}

/*
 * Undo anything leftover by machine_kexec_prepare
 * when an image is freed.
 */
void machine_kexec_cleanup(struct kimage *image)
{
174
	set_pages_nx(image->control_code_page, 1);
175
	machine_kexec_free_page_tables(image);
176 177 178 179 180 181
}

/*
 * Do not allocate memory (or fail in any way) in machine_kexec().
 * We are past the point of no return, committed to rebooting now.
 */
H
Huang Ying 已提交
182
void machine_kexec(struct kimage *image)
183
{
184 185
	unsigned long page_list[PAGES_NR];
	void *control_page;
H
Huang Ying 已提交
186
	int save_ftrace_enabled;
H
Huang Ying 已提交
187 188 189 190 191 192
	asmlinkage unsigned long
		(*relocate_kernel_ptr)(unsigned long indirection_page,
				       unsigned long control_page,
				       unsigned long start_address,
				       unsigned int has_pae,
				       unsigned int preserve_context);
193

H
Huang Ying 已提交
194
#ifdef CONFIG_KEXEC_JUMP
195
	if (image->preserve_context)
H
Huang Ying 已提交
196 197 198 199
		save_processor_state();
#endif

	save_ftrace_enabled = __ftrace_enabled_save();
I
Ingo Molnar 已提交
200

201 202 203
	/* Interrupts aren't acceptable while we reboot */
	local_irq_disable();

204 205
	if (image->preserve_context) {
#ifdef CONFIG_X86_IO_APIC
206 207
		/*
		 * We need to put APICs in legacy mode so that we can
208 209 210 211 212 213 214 215 216
		 * get timer interrupts in second kernel. kexec/kdump
		 * paths already have calls to disable_IO_APIC() in
		 * one form or other. kexec jump path also need
		 * one.
		 */
		disable_IO_APIC();
#endif
	}

217
	control_page = page_address(image->control_code_page);
218
	memcpy(control_page, relocate_kernel, KEXEC_CONTROL_CODE_MAX_SIZE);
219

H
Huang Ying 已提交
220
	relocate_kernel_ptr = control_page;
221
	page_list[PA_CONTROL_PAGE] = __pa(control_page);
H
Huang Ying 已提交
222
	page_list[VA_CONTROL_PAGE] = (unsigned long)control_page;
223
	page_list[PA_PGD] = __pa(image->arch.pgd);
224 225 226 227

	if (image->type == KEXEC_TYPE_DEFAULT)
		page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page)
						<< PAGE_SHIFT);
228

229 230
	/*
	 * The segment registers are funny things, they have both a
231 232 233 234
	 * 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.
235 236 237 238 239
	 *
	 * I take advantage of this here by force loading the
	 * segments, before I zap the gdt with an invalid value.
	 */
	load_segments();
240 241
	/*
	 * The gdt & idt are now invalid.
242 243
	 * If you want to load them you must set up your own idt & gdt.
	 */
244 245
	set_gdt(phys_to_virt(0), 0);
	set_idt(phys_to_virt(0), 0);
246 247

	/* now call it */
H
Huang Ying 已提交
248 249 250 251
	image->start = relocate_kernel_ptr((unsigned long)image->head,
					   (unsigned long)page_list,
					   image->start, cpu_has_pae,
					   image->preserve_context);
H
Huang Ying 已提交
252 253

#ifdef CONFIG_KEXEC_JUMP
254
	if (image->preserve_context)
H
Huang Ying 已提交
255 256 257 258
		restore_processor_state();
#endif

	__ftrace_enabled_restore(save_ftrace_enabled);
259
}
260

K
Ken'ichi Ohmichi 已提交
261 262
void arch_crash_save_vmcoreinfo(void)
{
263
#ifdef CONFIG_NUMA
264 265
	VMCOREINFO_SYMBOL(node_data);
	VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
K
Ken'ichi Ohmichi 已提交
266 267
#endif
#ifdef CONFIG_X86_PAE
268
	VMCOREINFO_CONFIG(X86_PAE);
K
Ken'ichi Ohmichi 已提交
269 270 271
#endif
}