machine_kexec_64.c 6.7 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/string.h>
12
#include <linux/gfp.h>
13
#include <linux/reboot.h>
K
Ken'ichi Ohmichi 已提交
14
#include <linux/numa.h>
I
Ingo Molnar 已提交
15
#include <linux/ftrace.h>
16
#include <linux/io.h>
17
#include <linux/suspend.h>
I
Ingo Molnar 已提交
18

19
#include <asm/init.h>
20 21 22
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
#include <asm/mmu_context.h>
23
#include <asm/debugreg.h>
24

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
static void free_transition_pgtable(struct kimage *image)
{
	free_page((unsigned long)image->arch.pud);
	free_page((unsigned long)image->arch.pmd);
	free_page((unsigned long)image->arch.pte);
}

static int init_transition_pgtable(struct kimage *image, pgd_t *pgd)
{
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;
	unsigned long vaddr, paddr;
	int result = -ENOMEM;

	vaddr = (unsigned long)relocate_kernel;
	paddr = __pa(page_address(image->control_code_page)+PAGE_SIZE);
	pgd += pgd_index(vaddr);
	if (!pgd_present(*pgd)) {
		pud = (pud_t *)get_zeroed_page(GFP_KERNEL);
		if (!pud)
			goto err;
		image->arch.pud = pud;
		set_pgd(pgd, __pgd(__pa(pud) | _KERNPG_TABLE));
	}
	pud = pud_offset(pgd, vaddr);
	if (!pud_present(*pud)) {
		pmd = (pmd_t *)get_zeroed_page(GFP_KERNEL);
		if (!pmd)
			goto err;
		image->arch.pmd = pmd;
		set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE));
	}
	pmd = pmd_offset(pud, vaddr);
	if (!pmd_present(*pmd)) {
		pte = (pte_t *)get_zeroed_page(GFP_KERNEL);
		if (!pte)
			goto err;
		image->arch.pte = pte;
		set_pmd(pmd, __pmd(__pa(pte) | _KERNPG_TABLE));
	}
	pte = pte_offset_kernel(pmd, vaddr);
	set_pte(pte, pfn_pte(paddr >> PAGE_SHIFT, PAGE_KERNEL_EXEC));
	return 0;
err:
	free_transition_pgtable(image);
	return result;
}

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
static void *alloc_pgt_page(void *data)
{
	struct kimage *image = (struct kimage *)data;
	struct page *page;
	void *p = NULL;

	page = kimage_alloc_control_pages(image, 0);
	if (page) {
		p = page_address(page);
		clear_page(p);
	}

	return p;
}

89 90
static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
{
91 92 93 94 95
	struct x86_mapping_info info = {
		.alloc_pgt_page	= alloc_pgt_page,
		.context	= image,
		.pmd_flag	= __PAGE_KERNEL_LARGE_EXEC,
	};
96
	unsigned long mstart, mend;
97
	pgd_t *level4p;
98
	int result;
99 100
	int i;

101
	level4p = (pgd_t *)__va(start_pgtable);
102
	clear_page(level4p);
103 104 105 106 107 108 109 110 111
	for (i = 0; i < nr_pfn_mapped; i++) {
		mstart = pfn_mapped[i].start << PAGE_SHIFT;
		mend   = pfn_mapped[i].end << PAGE_SHIFT;

		result = kernel_ident_mapping_init(&info,
						 level4p, mstart, mend);
		if (result)
			return result;
	}
112

113
	/*
114 115 116 117
	 * segments's mem ranges could be outside 0 ~ max_pfn,
	 * for example when jump back to original kernel from kexeced kernel.
	 * or first kernel is booted with user mem map, and second kernel
	 * could be loaded out of that range.
118
	 */
119 120 121 122
	for (i = 0; i < image->nr_segments; i++) {
		mstart = image->segment[i].mem;
		mend   = mstart + image->segment[i].memsz;

123 124
		result = kernel_ident_mapping_init(&info,
						 level4p, mstart, mend);
125 126 127 128 129

		if (result)
			return result;
	}

130
	return init_transition_pgtable(image, level4p);
131 132 133 134
}

static void set_idt(void *newidt, u16 limit)
{
135
	struct desc_ptr curidt;
136 137

	/* x86-64 supports unaliged loads & stores */
138 139
	curidt.size    = limit;
	curidt.address = (unsigned long)newidt;
140 141

	__asm__ __volatile__ (
142 143
		"lidtq %0\n"
		: : "m" (curidt)
144 145 146 147 148 149
		);
};


static void set_gdt(void *newgdt, u16 limit)
{
150
	struct desc_ptr curgdt;
151 152

	/* x86-64 supports unaligned loads & stores */
153 154
	curgdt.size    = limit;
	curgdt.address = (unsigned long)newgdt;
155 156

	__asm__ __volatile__ (
157 158
		"lgdtq %0\n"
		: : "m" (curgdt)
159 160 161 162 163 164
		);
};

static void load_segments(void)
{
	__asm__ __volatile__ (
165 166 167 168 169
		"\tmovl %0,%%ds\n"
		"\tmovl %0,%%es\n"
		"\tmovl %0,%%ss\n"
		"\tmovl %0,%%fs\n"
		"\tmovl %0,%%gs\n"
M
Michael Matz 已提交
170
		: : "a" (__KERNEL_DS) : "memory"
171 172 173 174 175
		);
}

int machine_kexec_prepare(struct kimage *image)
{
176
	unsigned long start_pgtable;
177 178 179
	int result;

	/* Calculate the offsets */
M
Maneesh Soni 已提交
180
	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;
181 182 183

	/* Setup the identity mapped 64bit page table */
	result = init_pgtable(image, start_pgtable);
M
Maneesh Soni 已提交
184
	if (result)
185 186 187 188 189 190 191
		return result;

	return 0;
}

void machine_kexec_cleanup(struct kimage *image)
{
192
	free_transition_pgtable(image);
193 194 195 196 197 198
}

/*
 * 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
199
void machine_kexec(struct kimage *image)
200
{
201 202
	unsigned long page_list[PAGES_NR];
	void *control_page;
203
	int save_ftrace_enabled;
204

205
#ifdef CONFIG_KEXEC_JUMP
206
	if (image->preserve_context)
207 208 209 210
		save_processor_state();
#endif

	save_ftrace_enabled = __ftrace_enabled_save();
I
Ingo Molnar 已提交
211

212 213
	/* Interrupts aren't acceptable while we reboot */
	local_irq_disable();
214
	hw_breakpoint_disable();
215

216 217 218 219 220 221 222 223 224 225 226 227 228
	if (image->preserve_context) {
#ifdef CONFIG_X86_IO_APIC
		/*
		 * We need to put APICs in legacy mode so that we can
		 * 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
	}

229
	control_page = page_address(image->control_code_page) + PAGE_SIZE;
230
	memcpy(control_page, relocate_kernel, KEXEC_CONTROL_CODE_MAX_SIZE);
231

232
	page_list[PA_CONTROL_PAGE] = virt_to_phys(control_page);
233
	page_list[VA_CONTROL_PAGE] = (unsigned long)control_page;
234 235
	page_list[PA_TABLE_PAGE] =
	  (unsigned long)__pa(page_address(image->control_code_page));
236

237 238 239 240
	if (image->type == KEXEC_TYPE_DEFAULT)
		page_list[PA_SWAP_PAGE] = (page_to_pfn(image->swap_page)
						<< PAGE_SHIFT);

241 242
	/*
	 * The segment registers are funny things, they have both a
243 244 245 246
	 * 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.
247 248 249 250 251
	 *
	 * I take advantage of this here by force loading the
	 * segments, before I zap the gdt with an invalid value.
	 */
	load_segments();
252 253
	/*
	 * The gdt & idt are now invalid.
254 255
	 * If you want to load them you must set up your own idt & gdt.
	 */
256 257
	set_gdt(phys_to_virt(0), 0);
	set_idt(phys_to_virt(0), 0);
258

259
	/* now call it */
260 261 262 263 264 265
	image->start = relocate_kernel((unsigned long)image->head,
				       (unsigned long)page_list,
				       image->start,
				       image->preserve_context);

#ifdef CONFIG_KEXEC_JUMP
266
	if (image->preserve_context)
267 268 269 270
		restore_processor_state();
#endif

	__ftrace_enabled_restore(save_ftrace_enabled);
271
}
272

K
Ken'ichi Ohmichi 已提交
273 274
void arch_crash_save_vmcoreinfo(void)
{
275
	VMCOREINFO_SYMBOL(phys_base);
276
	VMCOREINFO_SYMBOL(init_level4_pgt);
277 278 279 280 281

#ifdef CONFIG_NUMA
	VMCOREINFO_SYMBOL(node_data);
	VMCOREINFO_LENGTH(node_data, MAX_NUMNODES);
#endif
282 283
	vmcoreinfo_append_str("KERNELOFFSET=%lx\n",
			      (unsigned long)&_text - __START_KERNEL);
K
Ken'ichi Ohmichi 已提交
284 285
}