kcore.c 15.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 *	fs/proc/kcore.c kernel ELF core dumper
 *
 *	Modelled on fs/exec.c:aout_core_dump()
 *	Jeremy Fitzhardinge <jeremy@sw.oz.au>
 *	ELF version written by David Howells <David.Howells@nexor.co.uk>
 *	Modified and incorporated into 2.3.x by Tigran Aivazian <tigran@veritas.com>
 *	Support to dump vmalloc'd areas (ELF only), Tigran Aivazian <tigran@veritas.com>
 *	Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <kanoj@sgi.com>
 */

#include <linux/mm.h>
#include <linux/proc_fs.h>
#include <linux/user.h>
15
#include <linux/capability.h>
L
Linus Torvalds 已提交
16 17 18 19
#include <linux/elf.h>
#include <linux/elfcore.h>
#include <linux/vmalloc.h>
#include <linux/highmem.h>
20
#include <linux/bootmem.h>
L
Linus Torvalds 已提交
21
#include <linux/init.h>
22
#include <linux/slab.h>
L
Linus Torvalds 已提交
23 24
#include <asm/uaccess.h>
#include <asm/io.h>
25
#include <linux/list.h>
26 27
#include <linux/ioport.h>
#include <linux/memory.h>
28
#include <asm/sections.h>
L
Linus Torvalds 已提交
29

30
#define CORE_STR "CORE"
L
Linus Torvalds 已提交
31

32 33 34 35
#ifndef ELF_CORE_EFLAGS
#define ELF_CORE_EFLAGS	0
#endif

36 37
static struct proc_dir_entry *proc_root_kcore;

L
Linus Torvalds 已提交
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

#ifndef kc_vaddr_to_offset
#define	kc_vaddr_to_offset(v) ((v) - PAGE_OFFSET)
#endif
#ifndef	kc_offset_to_vaddr
#define	kc_offset_to_vaddr(o) ((o) + PAGE_OFFSET)
#endif

/* An ELF note in memory */
struct memelfnote
{
	const char *name;
	int type;
	unsigned int datasz;
	void *data;
};

55
static LIST_HEAD(kclist_head);
L
Linus Torvalds 已提交
56
static DEFINE_RWLOCK(kclist_lock);
57
static int kcore_need_update = 1;
L
Linus Torvalds 已提交
58 59

void
K
KAMEZAWA Hiroyuki 已提交
60
kclist_add(struct kcore_list *new, void *addr, size_t size, int type)
L
Linus Torvalds 已提交
61 62 63
{
	new->addr = (unsigned long)addr;
	new->size = size;
K
KAMEZAWA Hiroyuki 已提交
64
	new->type = type;
L
Linus Torvalds 已提交
65 66

	write_lock(&kclist_lock);
67
	list_add_tail(&new->list, &kclist_head);
L
Linus Torvalds 已提交
68 69 70 71 72 73 74 75 76 77 78
	write_unlock(&kclist_lock);
}

static size_t get_kcore_size(int *nphdr, size_t *elf_buflen)
{
	size_t try, size;
	struct kcore_list *m;

	*nphdr = 1; /* PT_NOTE */
	size = 0;

79
	list_for_each_entry(m, &kclist_head, list) {
L
Linus Torvalds 已提交
80 81 82 83 84 85 86
		try = kc_vaddr_to_offset((size_t)m->addr + m->size);
		if (try > size)
			size = try;
		*nphdr = *nphdr + 1;
	}
	*elf_buflen =	sizeof(struct elfhdr) + 
			(*nphdr + 2)*sizeof(struct elf_phdr) + 
87 88 89 90 91
			3 * ((sizeof(struct elf_note)) +
			     roundup(sizeof(CORE_STR), 4)) +
			roundup(sizeof(struct elf_prstatus), 4) +
			roundup(sizeof(struct elf_prpsinfo), 4) +
			roundup(sizeof(struct task_struct), 4);
L
Linus Torvalds 已提交
92 93 94 95
	*elf_buflen = PAGE_ALIGN(*elf_buflen);
	return size + *elf_buflen;
}

96 97 98 99 100 101 102 103 104 105
static void free_kclist_ents(struct list_head *head)
{
	struct kcore_list *tmp, *pos;

	list_for_each_entry_safe(pos, tmp, head, list) {
		list_del(&pos->list);
		kfree(pos);
	}
}
/*
106
 * Replace all KCORE_RAM/KCORE_VMEMMAP information with passed list.
107 108 109
 */
static void __kcore_update_ram(struct list_head *list)
{
110 111
	int nphdr;
	size_t size;
112 113 114 115 116 117
	struct kcore_list *tmp, *pos;
	LIST_HEAD(garbage);

	write_lock(&kclist_lock);
	if (kcore_need_update) {
		list_for_each_entry_safe(pos, tmp, &kclist_head, list) {
118 119
			if (pos->type == KCORE_RAM
				|| pos->type == KCORE_VMEMMAP)
120 121 122 123 124 125
				list_move(&pos->list, &garbage);
		}
		list_splice_tail(list, &kclist_head);
	} else
		list_splice(list, &garbage);
	kcore_need_update = 0;
126
	proc_root_kcore->size = get_kcore_size(&nphdr, &size);
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157
	write_unlock(&kclist_lock);

	free_kclist_ents(&garbage);
}


#ifdef CONFIG_HIGHMEM
/*
 * If no highmem, we can assume [0...max_low_pfn) continuous range of memory
 * because memory hole is not as big as !HIGHMEM case.
 * (HIGHMEM is special because part of memory is _invisible_ from the kernel.)
 */
static int kcore_update_ram(void)
{
	LIST_HEAD(head);
	struct kcore_list *ent;
	int ret = 0;

	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	if (!ent)
		return -ENOMEM;
	ent->addr = (unsigned long)__va(0);
	ent->size = max_low_pfn << PAGE_SHIFT;
	ent->type = KCORE_RAM;
	list_add(&ent->list, &head);
	__kcore_update_ram(&head);
	return ret;
}

#else /* !CONFIG_HIGHMEM */

158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
#ifdef CONFIG_SPARSEMEM_VMEMMAP
/* calculate vmemmap's address from given system ram pfn and register it */
int get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
{
	unsigned long pfn = __pa(ent->addr) >> PAGE_SHIFT;
	unsigned long nr_pages = ent->size >> PAGE_SHIFT;
	unsigned long start, end;
	struct kcore_list *vmm, *tmp;


	start = ((unsigned long)pfn_to_page(pfn)) & PAGE_MASK;
	end = ((unsigned long)pfn_to_page(pfn + nr_pages)) - 1;
	end = ALIGN(end, PAGE_SIZE);
	/* overlap check (because we have to align page */
	list_for_each_entry(tmp, head, list) {
		if (tmp->type != KCORE_VMEMMAP)
			continue;
		if (start < tmp->addr + tmp->size)
			if (end > tmp->addr)
				end = tmp->addr;
	}
	if (start < end) {
		vmm = kmalloc(sizeof(*vmm), GFP_KERNEL);
		if (!vmm)
			return 0;
		vmm->addr = start;
		vmm->size = end - start;
		vmm->type = KCORE_VMEMMAP;
		list_add_tail(&vmm->list, head);
	}
	return 1;

}
#else
int get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
{
	return 1;
}

#endif

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
static int
kclist_add_private(unsigned long pfn, unsigned long nr_pages, void *arg)
{
	struct list_head *head = (struct list_head *)arg;
	struct kcore_list *ent;

	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	if (!ent)
		return -ENOMEM;
	ent->addr = (unsigned long)__va((pfn << PAGE_SHIFT));
	ent->size = nr_pages << PAGE_SHIFT;

	/* Sanity check: Can happen in 32bit arch...maybe */
	if (ent->addr < (unsigned long) __va(0))
		goto free_out;

	/* cut not-mapped area. ....from ppc-32 code. */
	if (ULONG_MAX - ent->addr < ent->size)
		ent->size = ULONG_MAX - ent->addr;

	/* cut when vmalloc() area is higher than direct-map area */
	if (VMALLOC_START > (unsigned long)__va(0)) {
		if (ent->addr > VMALLOC_START)
			goto free_out;
		if (VMALLOC_START - ent->addr < ent->size)
			ent->size = VMALLOC_START - ent->addr;
	}

	ent->type = KCORE_RAM;
	list_add_tail(&ent->list, head);
229 230 231 232 233 234

	if (!get_sparsemem_vmemmap_info(ent, head)) {
		list_del(&ent->list);
		goto free_out;
	}

235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
	return 0;
free_out:
	kfree(ent);
	return 1;
}

static int kcore_update_ram(void)
{
	int nid, ret;
	unsigned long end_pfn;
	LIST_HEAD(head);

	/* Not inialized....update now */
	/* find out "max pfn" */
	end_pfn = 0;
	for_each_node_state(nid, N_HIGH_MEMORY) {
		unsigned long node_end;
		node_end  = NODE_DATA(nid)->node_start_pfn +
			NODE_DATA(nid)->node_spanned_pages;
		if (end_pfn < node_end)
			end_pfn = node_end;
	}
	/* scan 0 to max_pfn */
	ret = walk_system_ram_range(0, end_pfn, &head, kclist_add_private);
	if (ret) {
		free_kclist_ents(&head);
		return -ENOMEM;
	}
	__kcore_update_ram(&head);
	return ret;
}
#endif /* CONFIG_HIGHMEM */
L
Linus Torvalds 已提交
267 268 269 270 271 272 273 274 275 276

/*****************************************************************************/
/*
 * determine size of ELF note
 */
static int notesize(struct memelfnote *en)
{
	int sz;

	sz = sizeof(struct elf_note);
277
	sz += roundup((strlen(en->name) + 1), 4);
L
Linus Torvalds 已提交
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292
	sz += roundup(en->datasz, 4);

	return sz;
} /* end notesize() */

/*****************************************************************************/
/*
 * store a note in the header buffer
 */
static char *storenote(struct memelfnote *men, char *bufp)
{
	struct elf_note en;

#define DUMP_WRITE(addr,nr) do { memcpy(bufp,addr,nr); bufp += nr; } while(0)

293
	en.n_namesz = strlen(men->name) + 1;
L
Linus Torvalds 已提交
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339
	en.n_descsz = men->datasz;
	en.n_type = men->type;

	DUMP_WRITE(&en, sizeof(en));
	DUMP_WRITE(men->name, en.n_namesz);

	/* XXX - cast from long long to long to avoid need for libgcc.a */
	bufp = (char*) roundup((unsigned long)bufp,4);
	DUMP_WRITE(men->data, men->datasz);
	bufp = (char*) roundup((unsigned long)bufp,4);

#undef DUMP_WRITE

	return bufp;
} /* end storenote() */

/*
 * store an ELF coredump header in the supplied buffer
 * nphdr is the number of elf_phdr to insert
 */
static void elf_kcore_store_hdr(char *bufp, int nphdr, int dataoff)
{
	struct elf_prstatus prstatus;	/* NT_PRSTATUS */
	struct elf_prpsinfo prpsinfo;	/* NT_PRPSINFO */
	struct elf_phdr *nhdr, *phdr;
	struct elfhdr *elf;
	struct memelfnote notes[3];
	off_t offset = 0;
	struct kcore_list *m;

	/* setup ELF header */
	elf = (struct elfhdr *) bufp;
	bufp += sizeof(struct elfhdr);
	offset += sizeof(struct elfhdr);
	memcpy(elf->e_ident, ELFMAG, SELFMAG);
	elf->e_ident[EI_CLASS]	= ELF_CLASS;
	elf->e_ident[EI_DATA]	= ELF_DATA;
	elf->e_ident[EI_VERSION]= EV_CURRENT;
	elf->e_ident[EI_OSABI] = ELF_OSABI;
	memset(elf->e_ident+EI_PAD, 0, EI_NIDENT-EI_PAD);
	elf->e_type	= ET_CORE;
	elf->e_machine	= ELF_ARCH;
	elf->e_version	= EV_CURRENT;
	elf->e_entry	= 0;
	elf->e_phoff	= sizeof(struct elfhdr);
	elf->e_shoff	= 0;
340
	elf->e_flags	= ELF_CORE_EFLAGS;
L
Linus Torvalds 已提交
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361
	elf->e_ehsize	= sizeof(struct elfhdr);
	elf->e_phentsize= sizeof(struct elf_phdr);
	elf->e_phnum	= nphdr;
	elf->e_shentsize= 0;
	elf->e_shnum	= 0;
	elf->e_shstrndx	= 0;

	/* setup ELF PT_NOTE program header */
	nhdr = (struct elf_phdr *) bufp;
	bufp += sizeof(struct elf_phdr);
	offset += sizeof(struct elf_phdr);
	nhdr->p_type	= PT_NOTE;
	nhdr->p_offset	= 0;
	nhdr->p_vaddr	= 0;
	nhdr->p_paddr	= 0;
	nhdr->p_filesz	= 0;
	nhdr->p_memsz	= 0;
	nhdr->p_flags	= 0;
	nhdr->p_align	= 0;

	/* setup ELF PT_LOAD program header for every area */
362
	list_for_each_entry(m, &kclist_head, list) {
L
Linus Torvalds 已提交
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
		phdr = (struct elf_phdr *) bufp;
		bufp += sizeof(struct elf_phdr);
		offset += sizeof(struct elf_phdr);

		phdr->p_type	= PT_LOAD;
		phdr->p_flags	= PF_R|PF_W|PF_X;
		phdr->p_offset	= kc_vaddr_to_offset(m->addr) + dataoff;
		phdr->p_vaddr	= (size_t)m->addr;
		phdr->p_paddr	= 0;
		phdr->p_filesz	= phdr->p_memsz	= m->size;
		phdr->p_align	= PAGE_SIZE;
	}

	/*
	 * Set up the notes in similar form to SVR4 core dumps made
	 * with info from their /proc.
	 */
	nhdr->p_offset	= offset;

	/* set up the process status */
383
	notes[0].name = CORE_STR;
L
Linus Torvalds 已提交
384 385 386 387 388 389 390 391 392 393
	notes[0].type = NT_PRSTATUS;
	notes[0].datasz = sizeof(struct elf_prstatus);
	notes[0].data = &prstatus;

	memset(&prstatus, 0, sizeof(struct elf_prstatus));

	nhdr->p_filesz	= notesize(&notes[0]);
	bufp = storenote(&notes[0], bufp);

	/* set up the process info */
394
	notes[1].name	= CORE_STR;
L
Linus Torvalds 已提交
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
	notes[1].type	= NT_PRPSINFO;
	notes[1].datasz	= sizeof(struct elf_prpsinfo);
	notes[1].data	= &prpsinfo;

	memset(&prpsinfo, 0, sizeof(struct elf_prpsinfo));
	prpsinfo.pr_state	= 0;
	prpsinfo.pr_sname	= 'R';
	prpsinfo.pr_zomb	= 0;

	strcpy(prpsinfo.pr_fname, "vmlinux");
	strncpy(prpsinfo.pr_psargs, saved_command_line, ELF_PRARGSZ);

	nhdr->p_filesz	+= notesize(&notes[1]);
	bufp = storenote(&notes[1], bufp);

	/* set up the task structure */
411
	notes[2].name	= CORE_STR;
L
Linus Torvalds 已提交
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434
	notes[2].type	= NT_TASKSTRUCT;
	notes[2].datasz	= sizeof(struct task_struct);
	notes[2].data	= current;

	nhdr->p_filesz	+= notesize(&notes[2]);
	bufp = storenote(&notes[2], bufp);

} /* end elf_kcore_store_hdr() */

/*****************************************************************************/
/*
 * read from the ELF header and then kernel memory
 */
static ssize_t
read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
{
	ssize_t acc = 0;
	size_t size, tsz;
	size_t elf_buflen;
	int nphdr;
	unsigned long start;

	read_lock(&kclist_lock);
435 436
	size = get_kcore_size(&nphdr, &elf_buflen);

L
Linus Torvalds 已提交
437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452
	if (buflen == 0 || *fpos >= size) {
		read_unlock(&kclist_lock);
		return 0;
	}

	/* trim buflen to not go beyond EOF */
	if (buflen > size - *fpos)
		buflen = size - *fpos;

	/* construct an ELF core header if we'll need some of it */
	if (*fpos < elf_buflen) {
		char * elf_buf;

		tsz = elf_buflen - *fpos;
		if (buflen < tsz)
			tsz = buflen;
453
		elf_buf = kzalloc(elf_buflen, GFP_ATOMIC);
L
Linus Torvalds 已提交
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487
		if (!elf_buf) {
			read_unlock(&kclist_lock);
			return -ENOMEM;
		}
		elf_kcore_store_hdr(elf_buf, nphdr, elf_buflen);
		read_unlock(&kclist_lock);
		if (copy_to_user(buffer, elf_buf + *fpos, tsz)) {
			kfree(elf_buf);
			return -EFAULT;
		}
		kfree(elf_buf);
		buflen -= tsz;
		*fpos += tsz;
		buffer += tsz;
		acc += tsz;

		/* leave now if filled buffer already */
		if (buflen == 0)
			return acc;
	} else
		read_unlock(&kclist_lock);

	/*
	 * Check to see if our file offset matches with any of
	 * the addresses in the elf_phdr on our list.
	 */
	start = kc_offset_to_vaddr(*fpos - elf_buflen);
	if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen)
		tsz = buflen;
		
	while (buflen) {
		struct kcore_list *m;

		read_lock(&kclist_lock);
488
		list_for_each_entry(m, &kclist_head, list) {
L
Linus Torvalds 已提交
489 490 491 492 493
			if (start >= m->addr && start < (m->addr+m->size))
				break;
		}
		read_unlock(&kclist_lock);

D
Dan Carpenter 已提交
494
		if (&m->list == &kclist_head) {
L
Linus Torvalds 已提交
495 496
			if (clear_user(buffer, tsz))
				return -EFAULT;
497
		} else if (is_vmalloc_or_module_addr((void *)start)) {
L
Linus Torvalds 已提交
498 499
			char * elf_buf;

500
			elf_buf = kzalloc(tsz, GFP_KERNEL);
L
Linus Torvalds 已提交
501 502
			if (!elf_buf)
				return -ENOMEM;
503 504
			vread(elf_buf, (char *)start, tsz);
			/* we have to zero-fill user buffer even if no read */
L
Linus Torvalds 已提交
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
			if (copy_to_user(buffer, elf_buf, tsz)) {
				kfree(elf_buf);
				return -EFAULT;
			}
			kfree(elf_buf);
		} else {
			if (kern_addr_valid(start)) {
				unsigned long n;

				n = copy_to_user(buffer, (char *)start, tsz);
				/*
				 * We cannot distingush between fault on source
				 * and fault on destination. When this happens
				 * we clear too and hope it will trigger the
				 * EFAULT again.
				 */
				if (n) { 
					if (clear_user(buffer + tsz - n,
523
								n))
L
Linus Torvalds 已提交
524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
						return -EFAULT;
				}
			} else {
				if (clear_user(buffer, tsz))
					return -EFAULT;
			}
		}
		buflen -= tsz;
		*fpos += tsz;
		buffer += tsz;
		acc += tsz;
		start += tsz;
		tsz = (buflen > PAGE_SIZE ? PAGE_SIZE : buflen);
	}

	return acc;
}
541

542 543 544 545 546 547 548

static int open_kcore(struct inode *inode, struct file *filp)
{
	if (!capable(CAP_SYS_RAWIO))
		return -EPERM;
	if (kcore_need_update)
		kcore_update_ram();
549 550 551 552 553
	if (i_size_read(inode) != proc_root_kcore->size) {
		mutex_lock(&inode->i_mutex);
		i_size_write(inode, proc_root_kcore->size);
		mutex_unlock(&inode->i_mutex);
	}
554 555 556 557 558 559 560
	return 0;
}


static const struct file_operations proc_kcore_operations = {
	.read		= read_kcore,
	.open		= open_kcore,
D
Dave Anderson 已提交
561
	.llseek		= default_llseek,
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
};

#ifdef CONFIG_MEMORY_HOTPLUG
/* just remember that we have to update kcore */
static int __meminit kcore_callback(struct notifier_block *self,
				    unsigned long action, void *arg)
{
	switch (action) {
	case MEM_ONLINE:
	case MEM_OFFLINE:
		write_lock(&kclist_lock);
		kcore_need_update = 1;
		write_unlock(&kclist_lock);
	}
	return NOTIFY_OK;
}
#endif


581 582
static struct kcore_list kcore_vmalloc;

583 584 585 586 587 588 589 590
#ifdef CONFIG_ARCH_PROC_KCORE_TEXT
static struct kcore_list kcore_text;
/*
 * If defined, special segment is used for mapping kernel text instead of
 * direct-map area. We need to create special TEXT section.
 */
static void __init proc_kcore_text_init(void)
{
W
Wu Fengguang 已提交
591
	kclist_add(&kcore_text, _text, _end - _text, KCORE_TEXT);
592 593 594 595 596 597 598
}
#else
static void __init proc_kcore_text_init(void)
{
}
#endif

599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614
#if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
/*
 * MODULES_VADDR has no intersection with VMALLOC_ADDR.
 */
struct kcore_list kcore_modules;
static void __init add_modules_range(void)
{
	kclist_add(&kcore_modules, (void *)MODULES_VADDR,
			MODULES_END - MODULES_VADDR, KCORE_VMALLOC);
}
#else
static void __init add_modules_range(void)
{
}
#endif

615 616
static int __init proc_kcore_init(void)
{
617 618
	proc_root_kcore = proc_create("kcore", S_IRUSR, NULL,
				      &proc_kcore_operations);
K
KAMEZAWA Hiroyuki 已提交
619 620 621 622
	if (!proc_root_kcore) {
		printk(KERN_ERR "couldn't create /proc/kcore\n");
		return 0; /* Always returns 0. */
	}
623
	/* Store text area if it's special */
624
	proc_kcore_text_init();
625
	/* Store vmalloc area */
626 627
	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
		VMALLOC_END - VMALLOC_START, KCORE_VMALLOC);
628
	add_modules_range();
629 630 631 632
	/* Store direct-map area from physical memory map */
	kcore_update_ram();
	hotplug_memory_notifier(kcore_callback, 0);

633 634 635
	return 0;
}
module_init(proc_kcore_init);