exec.c 48.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
/*
 *  linux/fs/exec.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

/*
 * #!-checking implemented by tytso.
 */
/*
 * Demand-loading implemented 01.12.91 - no need to read anything but
 * the header into memory. The inode of the executable is put into
 * "current->executable", and page faults do the actual loading. Clean.
 *
 * Once more I can proudly say that linux stood up to being changed: it
 * was less than 2 hours work to get demand-loading completely implemented.
 *
 * Demand loading changed July 1993 by Eric Youngdale.   Use mmap instead,
 * current->executable is only used by the procfs.  This allows a dispatch
 * table to check for several different types  of binary formats.  We keep
 * trying until we recognize the file or we run out of supported binary
 * formats. 
 */

#include <linux/slab.h>
#include <linux/file.h>
A
Al Viro 已提交
27
#include <linux/fdtable.h>
H
Hugh Dickins 已提交
28
#include <linux/mm.h>
L
Linus Torvalds 已提交
29 30
#include <linux/stat.h>
#include <linux/fcntl.h>
H
Hugh Dickins 已提交
31
#include <linux/swap.h>
32
#include <linux/string.h>
L
Linus Torvalds 已提交
33
#include <linux/init.h>
34
#include <linux/pagemap.h>
35
#include <linux/perf_event.h>
L
Linus Torvalds 已提交
36 37 38 39 40 41
#include <linux/highmem.h>
#include <linux/spinlock.h>
#include <linux/key.h>
#include <linux/personality.h>
#include <linux/binfmts.h>
#include <linux/utsname.h>
42
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
43 44 45 46 47 48
#include <linux/module.h>
#include <linux/namei.h>
#include <linux/proc_fs.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/syscalls.h>
49
#include <linux/tsacct_kern.h>
M
Matt Helsley 已提交
50
#include <linux/cn_proc.h>
A
Al Viro 已提交
51
#include <linux/audit.h>
R
Roland McGrath 已提交
52
#include <linux/tracehook.h>
J
Johannes Berg 已提交
53
#include <linux/kmod.h>
54
#include <linux/fsnotify.h>
55
#include <linux/fs_struct.h>
56
#include <linux/pipe_fs_i.h>
Y
Ying Han 已提交
57
#include <linux/oom.h>
L
Linus Torvalds 已提交
58 59 60

#include <asm/uaccess.h>
#include <asm/mmu_context.h>
61
#include <asm/tlb.h>
62
#include "internal.h"
L
Linus Torvalds 已提交
63 64

int core_uses_pid;
65
char core_pattern[CORENAME_MAX_SIZE] = "core";
66
unsigned int core_pipe_limit;
A
Alan Cox 已提交
67 68
int suid_dumpable = 0;

69 70 71 72 73 74
struct core_name {
	char *corename;
	int used, size;
};
static atomic_t call_count = ATOMIC_INIT(1);

L
Linus Torvalds 已提交
75 76
/* The maximal length of core_pattern is also specified in sysctl.c */

A
Alexey Dobriyan 已提交
77
static LIST_HEAD(formats);
L
Linus Torvalds 已提交
78 79
static DEFINE_RWLOCK(binfmt_lock);

I
Ivan Kokshaysky 已提交
80
int __register_binfmt(struct linux_binfmt * fmt, int insert)
L
Linus Torvalds 已提交
81 82 83 84
{
	if (!fmt)
		return -EINVAL;
	write_lock(&binfmt_lock);
I
Ivan Kokshaysky 已提交
85 86
	insert ? list_add(&fmt->lh, &formats) :
		 list_add_tail(&fmt->lh, &formats);
L
Linus Torvalds 已提交
87 88 89 90
	write_unlock(&binfmt_lock);
	return 0;	
}

I
Ivan Kokshaysky 已提交
91
EXPORT_SYMBOL(__register_binfmt);
L
Linus Torvalds 已提交
92

93
void unregister_binfmt(struct linux_binfmt * fmt)
L
Linus Torvalds 已提交
94 95
{
	write_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
96
	list_del(&fmt->lh);
L
Linus Torvalds 已提交
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
	write_unlock(&binfmt_lock);
}

EXPORT_SYMBOL(unregister_binfmt);

static inline void put_binfmt(struct linux_binfmt * fmt)
{
	module_put(fmt->module);
}

/*
 * Note that a shared library must be both readable and executable due to
 * security reasons.
 *
 * Also note that we take the address to load from from the file itself.
 */
113
SYSCALL_DEFINE1(uselib, const char __user *, library)
L
Linus Torvalds 已提交
114
{
115 116 117
	struct file *file;
	char *tmp = getname(library);
	int error = PTR_ERR(tmp);
118 119 120 121 122
	static const struct open_flags uselib_flags = {
		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
		.acc_mode = MAY_READ | MAY_EXEC | MAY_OPEN,
		.intent = LOOKUP_OPEN
	};
123

124 125 126
	if (IS_ERR(tmp))
		goto out;

127
	file = do_filp_open(AT_FDCWD, tmp, &uselib_flags, LOOKUP_FOLLOW);
128 129 130
	putname(tmp);
	error = PTR_ERR(file);
	if (IS_ERR(file))
L
Linus Torvalds 已提交
131 132 133
		goto out;

	error = -EINVAL;
134
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
L
Linus Torvalds 已提交
135 136
		goto exit;

137
	error = -EACCES;
138
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
L
Linus Torvalds 已提交
139 140
		goto exit;

141
	fsnotify_open(file);
142

L
Linus Torvalds 已提交
143 144 145 146 147
	error = -ENOEXEC;
	if(file->f_op) {
		struct linux_binfmt * fmt;

		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
148
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
149 150 151 152 153 154 155 156 157 158 159 160 161
			if (!fmt->load_shlib)
				continue;
			if (!try_module_get(fmt->module))
				continue;
			read_unlock(&binfmt_lock);
			error = fmt->load_shlib(file);
			read_lock(&binfmt_lock);
			put_binfmt(fmt);
			if (error != -ENOEXEC)
				break;
		}
		read_unlock(&binfmt_lock);
	}
162
exit:
L
Linus Torvalds 已提交
163 164 165 166 167
	fput(file);
out:
  	return error;
}

168 169
#ifdef CONFIG_MMU

170
void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188
{
	struct mm_struct *mm = current->mm;
	long diff = (long)(pages - bprm->vma_pages);

	if (!mm || !diff)
		return;

	bprm->vma_pages = pages;

#ifdef SPLIT_RSS_COUNTING
	add_mm_counter(mm, MM_ANONPAGES, diff);
#else
	spin_lock(&mm->page_table_lock);
	add_mm_counter(mm, MM_ANONPAGES, diff);
	spin_unlock(&mm->page_table_lock);
#endif
}

189
struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
		int write)
{
	struct page *page;
	int ret;

#ifdef CONFIG_STACK_GROWSUP
	if (write) {
		ret = expand_stack_downwards(bprm->vma, pos);
		if (ret < 0)
			return NULL;
	}
#endif
	ret = get_user_pages(current, bprm->mm, pos,
			1, write, 1, &page, NULL);
	if (ret <= 0)
		return NULL;

	if (write) {
		unsigned long size = bprm->vma->vm_end - bprm->vma->vm_start;
209 210
		struct rlimit *rlim;

211 212
		acct_arg_size(bprm, size / PAGE_SIZE);

213 214 215 216 217 218
		/*
		 * We've historically supported up to 32 pages (ARG_MAX)
		 * of argument strings even with small stacks
		 */
		if (size <= ARG_MAX)
			return page;
219 220 221 222 223 224 225 226

		/*
		 * Limit to 1/4-th the stack size for the argv+env strings.
		 * This ensures that:
		 *  - the remaining binfmt code will not run out of stack space,
		 *  - the program will have a reasonable amount of stack left
		 *    to work from.
		 */
227
		rlim = current->signal->rlim;
J
Jiri Slaby 已提交
228
		if (size > ACCESS_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4) {
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
			put_page(page);
			return NULL;
		}
	}

	return page;
}

static void put_arg_page(struct page *page)
{
	put_page(page);
}

static void free_arg_page(struct linux_binprm *bprm, int i)
{
}

static void free_arg_pages(struct linux_binprm *bprm)
{
}

static void flush_arg_page(struct linux_binprm *bprm, unsigned long pos,
		struct page *page)
{
	flush_cache_page(bprm->vma, pos, page_to_pfn(page));
}

static int __bprm_mm_init(struct linux_binprm *bprm)
{
258
	int err;
259 260 261 262 263
	struct vm_area_struct *vma = NULL;
	struct mm_struct *mm = bprm->mm;

	bprm->vma = vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
	if (!vma)
264
		return -ENOMEM;
265 266 267 268 269 270 271 272 273 274

	down_write(&mm->mmap_sem);
	vma->vm_mm = mm;

	/*
	 * Place the stack at the largest stack address the architecture
	 * supports. Later, we'll move this to an appropriate place. We don't
	 * use STACK_TOP because that can depend on attributes which aren't
	 * configured yet.
	 */
275
	BUG_ON(VM_STACK_FLAGS & VM_STACK_INCOMPLETE_SETUP);
276 277
	vma->vm_end = STACK_TOP_MAX;
	vma->vm_start = vma->vm_end - PAGE_SIZE;
278
	vma->vm_flags = VM_STACK_FLAGS | VM_STACK_INCOMPLETE_SETUP;
279
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
280
	INIT_LIST_HEAD(&vma->anon_vma_chain);
281 282 283 284 285

	err = security_file_mmap(NULL, 0, 0, 0, vma->vm_start, 1);
	if (err)
		goto err;

286
	err = insert_vm_struct(mm, vma);
287
	if (err)
288 289 290 291 292 293 294
		goto err;

	mm->stack_vm = mm->total_vm = 1;
	up_write(&mm->mmap_sem);
	bprm->p = vma->vm_end - sizeof(void *);
	return 0;
err:
295 296 297
	up_write(&mm->mmap_sem);
	bprm->vma = NULL;
	kmem_cache_free(vm_area_cachep, vma);
298 299 300 301 302 303 304 305 306 307
	return err;
}

static bool valid_arg_len(struct linux_binprm *bprm, long len)
{
	return len <= MAX_ARG_STRLEN;
}

#else

308
void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
309 310 311
{
}

312
struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400
		int write)
{
	struct page *page;

	page = bprm->page[pos / PAGE_SIZE];
	if (!page && write) {
		page = alloc_page(GFP_HIGHUSER|__GFP_ZERO);
		if (!page)
			return NULL;
		bprm->page[pos / PAGE_SIZE] = page;
	}

	return page;
}

static void put_arg_page(struct page *page)
{
}

static void free_arg_page(struct linux_binprm *bprm, int i)
{
	if (bprm->page[i]) {
		__free_page(bprm->page[i]);
		bprm->page[i] = NULL;
	}
}

static void free_arg_pages(struct linux_binprm *bprm)
{
	int i;

	for (i = 0; i < MAX_ARG_PAGES; i++)
		free_arg_page(bprm, i);
}

static void flush_arg_page(struct linux_binprm *bprm, unsigned long pos,
		struct page *page)
{
}

static int __bprm_mm_init(struct linux_binprm *bprm)
{
	bprm->p = PAGE_SIZE * MAX_ARG_PAGES - sizeof(void *);
	return 0;
}

static bool valid_arg_len(struct linux_binprm *bprm, long len)
{
	return len <= bprm->p;
}

#endif /* CONFIG_MMU */

/*
 * Create a new mm_struct and populate it with a temporary stack
 * vm_area_struct.  We don't have enough context at this point to set the stack
 * flags, permissions, and offset, so we use temporary values.  We'll update
 * them later in setup_arg_pages().
 */
int bprm_mm_init(struct linux_binprm *bprm)
{
	int err;
	struct mm_struct *mm = NULL;

	bprm->mm = mm = mm_alloc();
	err = -ENOMEM;
	if (!mm)
		goto err;

	err = init_new_context(current, mm);
	if (err)
		goto err;

	err = __bprm_mm_init(bprm);
	if (err)
		goto err;

	return 0;

err:
	if (mm) {
		bprm->mm = NULL;
		mmdrop(mm);
	}

	return err;
}

L
Linus Torvalds 已提交
401 402 403
/*
 * count() counts the number of strings in array ARGV.
 */
404
static int count(const char __user * const __user * argv, int max)
L
Linus Torvalds 已提交
405 406 407 408 409
{
	int i = 0;

	if (argv != NULL) {
		for (;;) {
410
			const char __user * p;
L
Linus Torvalds 已提交
411 412 413 414 415 416

			if (get_user(p, argv))
				return -EFAULT;
			if (!p)
				break;
			argv++;
417
			if (i++ >= max)
L
Linus Torvalds 已提交
418
				return -E2BIG;
419 420 421

			if (fatal_signal_pending(current))
				return -ERESTARTNOHAND;
L
Linus Torvalds 已提交
422 423 424 425 426 427 428
			cond_resched();
		}
	}
	return i;
}

/*
429 430 431
 * 'copy_strings()' copies argument/environment strings from the old
 * processes's memory to the new process's stack.  The call to get_user_pages()
 * ensures the destination page is created and not swapped out.
L
Linus Torvalds 已提交
432
 */
433
static int copy_strings(int argc, const char __user *const __user *argv,
A
Adrian Bunk 已提交
434
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
435 436 437
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
438
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
439 440 441
	int ret;

	while (argc-- > 0) {
442
		const char __user *str;
L
Linus Torvalds 已提交
443 444 445 446
		int len;
		unsigned long pos;

		if (get_user(str, argv+argc) ||
447
				!(len = strnlen_user(str, MAX_ARG_STRLEN))) {
L
Linus Torvalds 已提交
448 449 450 451
			ret = -EFAULT;
			goto out;
		}

452
		if (!valid_arg_len(bprm, len)) {
L
Linus Torvalds 已提交
453 454 455 456
			ret = -E2BIG;
			goto out;
		}

457
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
458
		pos = bprm->p;
459 460
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
461 462 463 464

		while (len > 0) {
			int offset, bytes_to_copy;

465 466 467 468
			if (fatal_signal_pending(current)) {
				ret = -ERESTARTNOHAND;
				goto out;
			}
469 470
			cond_resched();

L
Linus Torvalds 已提交
471
			offset = pos % PAGE_SIZE;
472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487
			if (offset == 0)
				offset = PAGE_SIZE;

			bytes_to_copy = offset;
			if (bytes_to_copy > len)
				bytes_to_copy = len;

			offset -= bytes_to_copy;
			pos -= bytes_to_copy;
			str -= bytes_to_copy;
			len -= bytes_to_copy;

			if (!kmapped_page || kpos != (pos & PAGE_MASK)) {
				struct page *page;

				page = get_arg_page(bprm, pos, 1);
L
Linus Torvalds 已提交
488
				if (!page) {
489
					ret = -E2BIG;
L
Linus Torvalds 已提交
490 491 492
					goto out;
				}

493 494
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
495
					kunmap(kmapped_page);
496 497
					put_arg_page(kmapped_page);
				}
L
Linus Torvalds 已提交
498 499
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
500 501
				kpos = pos & PAGE_MASK;
				flush_arg_page(bprm, kpos, kmapped_page);
L
Linus Torvalds 已提交
502
			}
503
			if (copy_from_user(kaddr+offset, str, bytes_to_copy)) {
L
Linus Torvalds 已提交
504 505 506 507 508 509 510
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
511 512
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
513
		kunmap(kmapped_page);
514 515
		put_arg_page(kmapped_page);
	}
L
Linus Torvalds 已提交
516 517 518 519 520 521
	return ret;
}

/*
 * Like copy_strings, but get argv and its values from kernel memory.
 */
522 523
int copy_strings_kernel(int argc, const char *const *argv,
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
524 525 526 527
{
	int r;
	mm_segment_t oldfs = get_fs();
	set_fs(KERNEL_DS);
528
	r = copy_strings(argc, (const char __user *const  __user *)argv, bprm);
L
Linus Torvalds 已提交
529 530 531 532 533 534
	set_fs(oldfs);
	return r;
}
EXPORT_SYMBOL(copy_strings_kernel);

#ifdef CONFIG_MMU
535

L
Linus Torvalds 已提交
536
/*
537 538 539
 * During bprm_mm_init(), we create a temporary stack at STACK_TOP_MAX.  Once
 * the binfmt code determines where the new stack should reside, we shift it to
 * its final location.  The process proceeds as follows:
L
Linus Torvalds 已提交
540
 *
541 542 543 544 545 546
 * 1) Use shift to calculate the new vma endpoints.
 * 2) Extend vma to cover both the old and new ranges.  This ensures the
 *    arguments passed to subsequent functions are consistent.
 * 3) Move vma's page tables to the new range.
 * 4) Free up any cleared pgd range.
 * 5) Shrink the vma to cover only the new range.
L
Linus Torvalds 已提交
547
 */
548
static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift)
L
Linus Torvalds 已提交
549 550
{
	struct mm_struct *mm = vma->vm_mm;
551 552 553 554 555 556
	unsigned long old_start = vma->vm_start;
	unsigned long old_end = vma->vm_end;
	unsigned long length = old_end - old_start;
	unsigned long new_start = old_start - shift;
	unsigned long new_end = old_end - shift;
	struct mmu_gather *tlb;
L
Linus Torvalds 已提交
557

558
	BUG_ON(new_start > new_end);
L
Linus Torvalds 已提交
559

560 561 562 563 564 565 566 567 568 569
	/*
	 * ensure there are no vmas between where we want to go
	 * and where we are
	 */
	if (vma != find_vma(mm, new_start))
		return -EFAULT;

	/*
	 * cover the whole range: [new_start, old_end)
	 */
570 571
	if (vma_adjust(vma, new_start, old_end, vma->vm_pgoff, NULL))
		return -ENOMEM;
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586

	/*
	 * move the page tables downwards, on failure we rely on
	 * process cleanup to remove whatever mess we made.
	 */
	if (length != move_page_tables(vma, old_start,
				       vma, new_start, length))
		return -ENOMEM;

	lru_add_drain();
	tlb = tlb_gather_mmu(mm, 0);
	if (new_end > old_start) {
		/*
		 * when the old and new regions overlap clear from new_end.
		 */
587
		free_pgd_range(tlb, new_end, old_end, new_end,
588 589 590 591 592 593 594 595
			vma->vm_next ? vma->vm_next->vm_start : 0);
	} else {
		/*
		 * otherwise, clean from old_start; this is done to not touch
		 * the address space in [new_end, old_start) some architectures
		 * have constraints on va-space that make this illegal (IA64) -
		 * for the others its just a little faster.
		 */
596
		free_pgd_range(tlb, old_start, old_end, new_end,
597
			vma->vm_next ? vma->vm_next->vm_start : 0);
L
Linus Torvalds 已提交
598
	}
599 600 601
	tlb_finish_mmu(tlb, new_end, old_end);

	/*
602
	 * Shrink the vma to just the new range.  Always succeeds.
603 604 605 606
	 */
	vma_adjust(vma, new_start, new_end, vma->vm_pgoff, NULL);

	return 0;
L
Linus Torvalds 已提交
607 608
}

609 610 611 612
/*
 * Finalizes the stack vm_area_struct. The flags and permissions are updated,
 * the stack is optionally relocated, and some extra space is added.
 */
L
Linus Torvalds 已提交
613 614 615 616
int setup_arg_pages(struct linux_binprm *bprm,
		    unsigned long stack_top,
		    int executable_stack)
{
617 618
	unsigned long ret;
	unsigned long stack_shift;
L
Linus Torvalds 已提交
619
	struct mm_struct *mm = current->mm;
620 621 622 623
	struct vm_area_struct *vma = bprm->vma;
	struct vm_area_struct *prev = NULL;
	unsigned long vm_flags;
	unsigned long stack_base;
624 625 626
	unsigned long stack_size;
	unsigned long stack_expand;
	unsigned long rlim_stack;
L
Linus Torvalds 已提交
627 628 629

#ifdef CONFIG_STACK_GROWSUP
	/* Limit stack size to 1GB */
J
Jiri Slaby 已提交
630
	stack_base = rlimit_max(RLIMIT_STACK);
L
Linus Torvalds 已提交
631 632 633
	if (stack_base > (1 << 30))
		stack_base = 1 << 30;

634 635 636
	/* Make sure we didn't let the argument array grow too large. */
	if (vma->vm_end - vma->vm_start > stack_base)
		return -ENOMEM;
L
Linus Torvalds 已提交
637

638
	stack_base = PAGE_ALIGN(stack_top - stack_base);
L
Linus Torvalds 已提交
639

640 641 642
	stack_shift = vma->vm_start - stack_base;
	mm->arg_start = bprm->p - stack_shift;
	bprm->p = vma->vm_end - stack_shift;
L
Linus Torvalds 已提交
643
#else
644 645
	stack_top = arch_align_stack(stack_top);
	stack_top = PAGE_ALIGN(stack_top);
646 647 648 649 650

	if (unlikely(stack_top < mmap_min_addr) ||
	    unlikely(vma->vm_end - vma->vm_start >= stack_top - mmap_min_addr))
		return -ENOMEM;

651 652 653
	stack_shift = vma->vm_end - stack_top;

	bprm->p -= stack_shift;
L
Linus Torvalds 已提交
654 655 656 657
	mm->arg_start = bprm->p;
#endif

	if (bprm->loader)
658 659
		bprm->loader -= stack_shift;
	bprm->exec -= stack_shift;
L
Linus Torvalds 已提交
660 661

	down_write(&mm->mmap_sem);
662
	vm_flags = VM_STACK_FLAGS;
663 664 665 666 667 668 669 670 671 672 673

	/*
	 * Adjust stack execute permissions; explicitly enable for
	 * EXSTACK_ENABLE_X, disable for EXSTACK_DISABLE_X and leave alone
	 * (arch default) otherwise.
	 */
	if (unlikely(executable_stack == EXSTACK_ENABLE_X))
		vm_flags |= VM_EXEC;
	else if (executable_stack == EXSTACK_DISABLE_X)
		vm_flags &= ~VM_EXEC;
	vm_flags |= mm->def_flags;
674
	vm_flags |= VM_STACK_INCOMPLETE_SETUP;
675 676 677 678 679 680 681 682 683 684

	ret = mprotect_fixup(vma, &prev, vma->vm_start, vma->vm_end,
			vm_flags);
	if (ret)
		goto out_unlock;
	BUG_ON(prev != vma);

	/* Move stack pages down in memory. */
	if (stack_shift) {
		ret = shift_arg_pages(vma, stack_shift);
685 686
		if (ret)
			goto out_unlock;
L
Linus Torvalds 已提交
687 688
	}

689 690 691
	/* mprotect_fixup is overkill to remove the temporary stack flags */
	vma->vm_flags &= ~VM_STACK_INCOMPLETE_SETUP;

692
	stack_expand = 131072UL; /* randomly 32*4k (or 2*64k) pages */
693 694 695 696 697 698
	stack_size = vma->vm_end - vma->vm_start;
	/*
	 * Align this down to a page boundary as expand_stack
	 * will align it up.
	 */
	rlim_stack = rlimit(RLIMIT_STACK) & PAGE_MASK;
699
#ifdef CONFIG_STACK_GROWSUP
700 701 702 703
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_start + rlim_stack;
	else
		stack_base = vma->vm_end + stack_expand;
704
#else
705 706 707 708
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_end - rlim_stack;
	else
		stack_base = vma->vm_start - stack_expand;
709
#endif
710
	current->mm->start_stack = bprm->p;
711 712 713 714 715
	ret = expand_stack(vma, stack_base);
	if (ret)
		ret = -EFAULT;

out_unlock:
L
Linus Torvalds 已提交
716
	up_write(&mm->mmap_sem);
717
	return ret;
L
Linus Torvalds 已提交
718 719 720 721 722 723 724 725
}
EXPORT_SYMBOL(setup_arg_pages);

#endif /* CONFIG_MMU */

struct file *open_exec(const char *name)
{
	struct file *file;
726
	int err;
727 728 729 730 731
	static const struct open_flags open_exec_flags = {
		.open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
		.acc_mode = MAY_EXEC | MAY_OPEN,
		.intent = LOOKUP_OPEN
	};
L
Linus Torvalds 已提交
732

733
	file = do_filp_open(AT_FDCWD, name, &open_exec_flags, LOOKUP_FOLLOW);
734
	if (IS_ERR(file))
735 736 737
		goto out;

	err = -EACCES;
738 739
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
		goto exit;
740

741 742
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
		goto exit;
743

744
	fsnotify_open(file);
745

746
	err = deny_write_access(file);
747 748
	if (err)
		goto exit;
L
Linus Torvalds 已提交
749

750
out:
751 752
	return file;

753 754
exit:
	fput(file);
755 756
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
757 758
EXPORT_SYMBOL(open_exec);

M
Mimi Zohar 已提交
759 760
int kernel_read(struct file *file, loff_t offset,
		char *addr, unsigned long count)
L
Linus Torvalds 已提交
761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783
{
	mm_segment_t old_fs;
	loff_t pos = offset;
	int result;

	old_fs = get_fs();
	set_fs(get_ds());
	/* The cast to a user pointer is valid due to the set_fs() */
	result = vfs_read(file, (void __user *)addr, count, &pos);
	set_fs(old_fs);
	return result;
}

EXPORT_SYMBOL(kernel_read);

static int exec_mmap(struct mm_struct *mm)
{
	struct task_struct *tsk;
	struct mm_struct * old_mm, *active_mm;

	/* Notify parent that we're no longer interested in the old VM */
	tsk = current;
	old_mm = current->mm;
784
	sync_mm_rss(tsk, old_mm);
L
Linus Torvalds 已提交
785 786 787 788 789 790 791
	mm_release(tsk, old_mm);

	if (old_mm) {
		/*
		 * Make sure that if there is a core dump in progress
		 * for the old mm, we get out and die instead of going
		 * through with the exec.  We must hold mmap_sem around
792
		 * checking core_state and changing tsk->mm.
L
Linus Torvalds 已提交
793 794
		 */
		down_read(&old_mm->mmap_sem);
795
		if (unlikely(old_mm->core_state)) {
L
Linus Torvalds 已提交
796 797 798 799 800 801 802 803 804
			up_read(&old_mm->mmap_sem);
			return -EINTR;
		}
	}
	task_lock(tsk);
	active_mm = tsk->active_mm;
	tsk->mm = mm;
	tsk->active_mm = mm;
	activate_mm(active_mm, mm);
Y
Ying Han 已提交
805 806 807 808
	if (old_mm && tsk->signal->oom_score_adj == OOM_SCORE_ADJ_MIN) {
		atomic_dec(&old_mm->oom_disable_count);
		atomic_inc(&tsk->mm->oom_disable_count);
	}
L
Linus Torvalds 已提交
809 810 811 812
	task_unlock(tsk);
	arch_pick_mmap_layout(mm);
	if (old_mm) {
		up_read(&old_mm->mmap_sem);
813
		BUG_ON(active_mm != old_mm);
814
		mm_update_next_owner(old_mm);
L
Linus Torvalds 已提交
815 816 817 818 819 820 821 822 823 824 825 826 827
		mmput(old_mm);
		return 0;
	}
	mmdrop(active_mm);
	return 0;
}

/*
 * This function makes sure the current process has its own signal table,
 * so that flush_signal_handlers can later reset the handlers without
 * disturbing other processes.  (Other processes might share the signal
 * table via the CLONE_SIGHAND option to clone().)
 */
828
static int de_thread(struct task_struct *tsk)
L
Linus Torvalds 已提交
829 830
{
	struct signal_struct *sig = tsk->signal;
831
	struct sighand_struct *oldsighand = tsk->sighand;
L
Linus Torvalds 已提交
832 833
	spinlock_t *lock = &oldsighand->siglock;

834
	if (thread_group_empty(tsk))
L
Linus Torvalds 已提交
835 836 837 838 839 840
		goto no_thread_group;

	/*
	 * Kill all other threads in the thread group.
	 */
	spin_lock_irq(lock);
841
	if (signal_group_exit(sig)) {
L
Linus Torvalds 已提交
842 843 844 845 846 847 848
		/*
		 * Another group action in progress, just
		 * return so that the signal is processed.
		 */
		spin_unlock_irq(lock);
		return -EAGAIN;
	}
849

850
	sig->group_exit_task = tsk;
851 852 853
	sig->notify_count = zap_other_threads(tsk);
	if (!thread_group_leader(tsk))
		sig->notify_count--;
L
Linus Torvalds 已提交
854

855
	while (sig->notify_count) {
L
Linus Torvalds 已提交
856 857 858 859 860 861 862 863 864 865 866 867
		__set_current_state(TASK_UNINTERRUPTIBLE);
		spin_unlock_irq(lock);
		schedule();
		spin_lock_irq(lock);
	}
	spin_unlock_irq(lock);

	/*
	 * At this point all other threads have exited, all we have to
	 * do is to wait for the thread group leader to become inactive,
	 * and to assume its PID:
	 */
868
	if (!thread_group_leader(tsk)) {
869
		struct task_struct *leader = tsk->group_leader;
870

871
		sig->notify_count = -1;	/* for exit_notify() */
872 873 874 875 876 877 878 879
		for (;;) {
			write_lock_irq(&tasklist_lock);
			if (likely(leader->exit_state))
				break;
			__set_current_state(TASK_UNINTERRUPTIBLE);
			write_unlock_irq(&tasklist_lock);
			schedule();
		}
L
Linus Torvalds 已提交
880

881 882 883 884 885 886 887 888 889 890
		/*
		 * The only record we have of the real-time age of a
		 * process, regardless of execs it's done, is start_time.
		 * All the past CPU time is accumulated in signal_struct
		 * from sister threads now dead.  But in this non-leader
		 * exec, nothing survives from the original leader thread,
		 * whose birth marks the true age of this process now.
		 * When we take on its identity by switching to its PID, we
		 * also take its birthdate (always earlier than our own).
		 */
891
		tsk->start_time = leader->start_time;
892

893 894
		BUG_ON(!same_thread_group(leader, tsk));
		BUG_ON(has_group_leader_pid(tsk));
L
Linus Torvalds 已提交
895 896 897 898 899 900
		/*
		 * An exec() starts a new thread group with the
		 * TGID of the previous thread group. Rehash the
		 * two threads with a switched PID, and release
		 * the former thread group leader:
		 */
901 902

		/* Become a process group leader with the old leader's pid.
903 904
		 * The old leader becomes a thread of the this thread group.
		 * Note: The old leader also uses this pid until release_task
905 906
		 *       is called.  Odd but simple and correct.
		 */
907 908
		detach_pid(tsk, PIDTYPE_PID);
		tsk->pid = leader->pid;
909
		attach_pid(tsk, PIDTYPE_PID,  task_pid(leader));
910 911
		transfer_pid(leader, tsk, PIDTYPE_PGID);
		transfer_pid(leader, tsk, PIDTYPE_SID);
912

913
		list_replace_rcu(&leader->tasks, &tsk->tasks);
914
		list_replace_init(&leader->sibling, &tsk->sibling);
L
Linus Torvalds 已提交
915

916 917
		tsk->group_leader = tsk;
		leader->group_leader = tsk;
918

919
		tsk->exit_signal = SIGCHLD;
920 921 922

		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
		leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
923
		write_unlock_irq(&tasklist_lock);
924 925

		release_task(leader);
926
	}
L
Linus Torvalds 已提交
927

928 929
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
L
Linus Torvalds 已提交
930 931

no_thread_group:
J
Jiri Pirko 已提交
932 933 934
	if (current->mm)
		setmax_mm_hiwater_rss(&sig->maxrss, current->mm);

L
Linus Torvalds 已提交
935
	exit_itimers(sig);
936
	flush_itimer_signals();
937

938 939
	if (atomic_read(&oldsighand->count) != 1) {
		struct sighand_struct *newsighand;
L
Linus Torvalds 已提交
940
		/*
941 942
		 * This ->sighand is shared with the CLONE_SIGHAND
		 * but not CLONE_THREAD task, switch to the new one.
L
Linus Torvalds 已提交
943
		 */
944 945 946 947
		newsighand = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
		if (!newsighand)
			return -ENOMEM;

L
Linus Torvalds 已提交
948 949 950 951 952 953
		atomic_set(&newsighand->count, 1);
		memcpy(newsighand->action, oldsighand->action,
		       sizeof(newsighand->action));

		write_lock_irq(&tasklist_lock);
		spin_lock(&oldsighand->siglock);
954
		rcu_assign_pointer(tsk->sighand, newsighand);
L
Linus Torvalds 已提交
955 956 957
		spin_unlock(&oldsighand->siglock);
		write_unlock_irq(&tasklist_lock);

958
		__cleanup_sighand(oldsighand);
L
Linus Torvalds 已提交
959 960
	}

961
	BUG_ON(!thread_group_leader(tsk));
L
Linus Torvalds 已提交
962 963
	return 0;
}
O
Oleg Nesterov 已提交
964

L
Linus Torvalds 已提交
965 966 967 968
/*
 * These functions flushes out all traces of the currently running executable
 * so that a new one can be started
 */
969
static void flush_old_files(struct files_struct * files)
L
Linus Torvalds 已提交
970 971
{
	long j = -1;
972
	struct fdtable *fdt;
L
Linus Torvalds 已提交
973 974 975 976 977 978 979

	spin_lock(&files->file_lock);
	for (;;) {
		unsigned long set, i;

		j++;
		i = j * __NFDBITS;
980
		fdt = files_fdtable(files);
981
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
982
			break;
983
		set = fdt->close_on_exec->fds_bits[j];
L
Linus Torvalds 已提交
984 985
		if (!set)
			continue;
986
		fdt->close_on_exec->fds_bits[j] = 0;
L
Linus Torvalds 已提交
987 988 989 990 991 992 993 994 995 996 997 998
		spin_unlock(&files->file_lock);
		for ( ; set ; i++,set >>= 1) {
			if (set & 1) {
				sys_close(i);
			}
		}
		spin_lock(&files->file_lock);

	}
	spin_unlock(&files->file_lock);
}

999
char *get_task_comm(char *buf, struct task_struct *tsk)
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004
{
	/* buf must be at least sizeof(tsk->comm) in size */
	task_lock(tsk);
	strncpy(buf, tsk->comm, sizeof(tsk->comm));
	task_unlock(tsk);
1005
	return buf;
L
Linus Torvalds 已提交
1006 1007 1008 1009 1010
}

void set_task_comm(struct task_struct *tsk, char *buf)
{
	task_lock(tsk);
1011 1012 1013 1014 1015 1016 1017 1018 1019

	/*
	 * Threads may access current->comm without holding
	 * the task lock, so write the string carefully.
	 * Readers without a lock may see incomplete new
	 * names but are safe from non-terminating string reads.
	 */
	memset(tsk->comm, 0, TASK_COMM_LEN);
	wmb();
L
Linus Torvalds 已提交
1020 1021
	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
	task_unlock(tsk);
1022
	perf_event_comm(tsk);
L
Linus Torvalds 已提交
1023 1024 1025 1026
}

int flush_old_exec(struct linux_binprm * bprm)
{
1027
	int retval;
L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036

	/*
	 * Make sure we have a private signal table and that
	 * we are unassociated from the previous thread group.
	 */
	retval = de_thread(current);
	if (retval)
		goto out;

M
Matt Helsley 已提交
1037 1038
	set_mm_exe_file(bprm->mm, bprm->file);

L
Linus Torvalds 已提交
1039 1040 1041
	/*
	 * Release all of the old mmap stuff
	 */
1042
	acct_arg_size(bprm, 0);
L
Linus Torvalds 已提交
1043 1044
	retval = exec_mmap(bprm->mm);
	if (retval)
1045
		goto out;
L
Linus Torvalds 已提交
1046 1047

	bprm->mm = NULL;		/* We're using it now */
1048

1049
	current->flags &= ~(PF_RANDOMIZE | PF_KTHREAD);
1050 1051 1052
	flush_thread();
	current->personality &= ~bprm->per_clear;

1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
	return 0;

out:
	return retval;
}
EXPORT_SYMBOL(flush_old_exec);

void setup_new_exec(struct linux_binprm * bprm)
{
	int i, ch;
1063
	const char *name;
1064 1065 1066
	char tcomm[sizeof(current->comm)];

	arch_pick_mmap_layout(current->mm);
L
Linus Torvalds 已提交
1067 1068 1069 1070

	/* This is the point of no return */
	current->sas_ss_sp = current->sas_ss_size = 0;

1071
	if (current_euid() == current_uid() && current_egid() == current_gid())
1072
		set_dumpable(current->mm, 1);
A
Alan Cox 已提交
1073
	else
1074
		set_dumpable(current->mm, suid_dumpable);
A
Alan Cox 已提交
1075

L
Linus Torvalds 已提交
1076
	name = bprm->filename;
1077 1078

	/* Copies the binary name from after last slash */
L
Linus Torvalds 已提交
1079 1080
	for (i=0; (ch = *(name++)) != '\0';) {
		if (ch == '/')
1081
			i = 0; /* overwrite what we wrote */
L
Linus Torvalds 已提交
1082 1083 1084 1085 1086 1087 1088
		else
			if (i < (sizeof(tcomm) - 1))
				tcomm[i++] = ch;
	}
	tcomm[i] = '\0';
	set_task_comm(current, tcomm);

1089 1090 1091 1092 1093 1094
	/* Set the new mm task size. We have to do that late because it may
	 * depend on TIF_32BIT which is only updated in flush_thread() on
	 * some architectures like powerpc
	 */
	current->mm->task_size = TASK_SIZE;

1095 1096 1097
	/* install the new credentials */
	if (bprm->cred->uid != current_euid() ||
	    bprm->cred->gid != current_egid()) {
1098 1099
		current->pdeath_signal = 0;
	} else if (file_permission(bprm->file, MAY_READ) ||
1100
		   bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP) {
1101
		set_dumpable(current->mm, suid_dumpable);
L
Linus Torvalds 已提交
1102 1103
	}

I
Ingo Molnar 已提交
1104 1105 1106 1107 1108
	/*
	 * Flush performance counters when crossing a
	 * security domain:
	 */
	if (!get_dumpable(current->mm))
1109
		perf_event_exit_task(current);
I
Ingo Molnar 已提交
1110

L
Linus Torvalds 已提交
1111 1112 1113 1114 1115 1116 1117 1118
	/* An exec changes our domain. We are no longer part of the thread
	   group */

	current->self_exec_id++;
			
	flush_signal_handlers(current, 0);
	flush_old_files(current->files);
}
1119
EXPORT_SYMBOL(setup_new_exec);
L
Linus Torvalds 已提交
1120

1121 1122 1123 1124 1125 1126 1127 1128
/*
 * Prepare credentials and lock ->cred_guard_mutex.
 * install_exec_creds() commits the new creds and drops the lock.
 * Or, if exec fails before, free_bprm() should release ->cred and
 * and unlock.
 */
int prepare_bprm_creds(struct linux_binprm *bprm)
{
1129
	if (mutex_lock_interruptible(&current->signal->cred_guard_mutex))
1130 1131 1132 1133 1134 1135
		return -ERESTARTNOINTR;

	bprm->cred = prepare_exec_creds();
	if (likely(bprm->cred))
		return 0;

1136
	mutex_unlock(&current->signal->cred_guard_mutex);
1137 1138 1139 1140 1141 1142 1143
	return -ENOMEM;
}

void free_bprm(struct linux_binprm *bprm)
{
	free_arg_pages(bprm);
	if (bprm->cred) {
1144
		mutex_unlock(&current->signal->cred_guard_mutex);
1145 1146 1147 1148 1149
		abort_creds(bprm->cred);
	}
	kfree(bprm);
}

1150 1151 1152 1153 1154 1155 1156 1157 1158
/*
 * install the new credentials for this executable
 */
void install_exec_creds(struct linux_binprm *bprm)
{
	security_bprm_committing_creds(bprm);

	commit_creds(bprm->cred);
	bprm->cred = NULL;
1159 1160
	/*
	 * cred_guard_mutex must be held at least to this point to prevent
1161
	 * ptrace_attach() from altering our determination of the task's
1162 1163
	 * credentials; any time after this it may be unlocked.
	 */
1164
	security_bprm_committed_creds(bprm);
1165
	mutex_unlock(&current->signal->cred_guard_mutex);
1166 1167 1168 1169 1170
}
EXPORT_SYMBOL(install_exec_creds);

/*
 * determine how safe it is to execute the proposed program
1171
 * - the caller must hold ->cred_guard_mutex to protect against
1172 1173
 *   PTRACE_ATTACH
 */
A
Al Viro 已提交
1174
int check_unsafe_exec(struct linux_binprm *bprm)
1175
{
D
David Howells 已提交
1176
	struct task_struct *p = current, *t;
1177
	unsigned n_fs;
A
Al Viro 已提交
1178
	int res = 0;
1179 1180 1181

	bprm->unsafe = tracehook_unsafe_exec(p);

D
David Howells 已提交
1182
	n_fs = 1;
N
Nick Piggin 已提交
1183
	spin_lock(&p->fs->lock);
1184
	rcu_read_lock();
D
David Howells 已提交
1185 1186 1187 1188
	for (t = next_thread(p); t != p; t = next_thread(t)) {
		if (t->fs == p->fs)
			n_fs++;
	}
1189
	rcu_read_unlock();
D
David Howells 已提交
1190

1191
	if (p->fs->users > n_fs) {
1192
		bprm->unsafe |= LSM_UNSAFE_SHARE;
A
Al Viro 已提交
1193
	} else {
1194 1195 1196 1197 1198
		res = -EAGAIN;
		if (!p->fs->in_exec) {
			p->fs->in_exec = 1;
			res = 1;
		}
A
Al Viro 已提交
1199
	}
N
Nick Piggin 已提交
1200
	spin_unlock(&p->fs->lock);
A
Al Viro 已提交
1201 1202

	return res;
1203 1204
}

L
Linus Torvalds 已提交
1205 1206 1207
/* 
 * Fill the binprm structure from the inode. 
 * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes
1208 1209
 *
 * This may be called multiple times for binary chains (scripts for example).
L
Linus Torvalds 已提交
1210 1211 1212
 */
int prepare_binprm(struct linux_binprm *bprm)
{
1213
	umode_t mode;
1214
	struct inode * inode = bprm->file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1215 1216 1217 1218 1219 1220
	int retval;

	mode = inode->i_mode;
	if (bprm->file->f_op == NULL)
		return -EACCES;

1221 1222 1223
	/* clear any previous set[ug]id data from a previous binary */
	bprm->cred->euid = current_euid();
	bprm->cred->egid = current_egid();
L
Linus Torvalds 已提交
1224

1225
	if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) {
L
Linus Torvalds 已提交
1226 1227
		/* Set-uid? */
		if (mode & S_ISUID) {
1228 1229
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->euid = inode->i_uid;
L
Linus Torvalds 已提交
1230 1231 1232 1233 1234 1235 1236 1237 1238
		}

		/* Set-gid? */
		/*
		 * If setgid is set but no group execute bit then this
		 * is a candidate for mandatory locking, not a setgid
		 * executable.
		 */
		if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
1239 1240
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->egid = inode->i_gid;
L
Linus Torvalds 已提交
1241 1242 1243 1244
		}
	}

	/* fill in binprm security blob */
1245
	retval = security_bprm_set_creds(bprm);
L
Linus Torvalds 已提交
1246 1247
	if (retval)
		return retval;
1248
	bprm->cred_prepared = 1;
L
Linus Torvalds 已提交
1249

1250 1251
	memset(bprm->buf, 0, BINPRM_BUF_SIZE);
	return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE);
L
Linus Torvalds 已提交
1252 1253 1254 1255
}

EXPORT_SYMBOL(prepare_binprm);

N
Nick Piggin 已提交
1256 1257 1258 1259 1260
/*
 * Arguments are '\0' separated strings found at the location bprm->p
 * points to; chop off the first by relocating brpm->p to right after
 * the first '\0' encountered.
 */
1261
int remove_arg_zero(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1262
{
1263 1264 1265 1266
	int ret = 0;
	unsigned long offset;
	char *kaddr;
	struct page *page;
N
Nick Piggin 已提交
1267

1268 1269
	if (!bprm->argc)
		return 0;
L
Linus Torvalds 已提交
1270

1271 1272 1273 1274 1275 1276 1277 1278
	do {
		offset = bprm->p & ~PAGE_MASK;
		page = get_arg_page(bprm, bprm->p, 0);
		if (!page) {
			ret = -EFAULT;
			goto out;
		}
		kaddr = kmap_atomic(page, KM_USER0);
N
Nick Piggin 已提交
1279

1280 1281 1282
		for (; offset < PAGE_SIZE && kaddr[offset];
				offset++, bprm->p++)
			;
N
Nick Piggin 已提交
1283

1284 1285
		kunmap_atomic(kaddr, KM_USER0);
		put_arg_page(page);
N
Nick Piggin 已提交
1286

1287 1288 1289
		if (offset == PAGE_SIZE)
			free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1);
	} while (offset == PAGE_SIZE);
N
Nick Piggin 已提交
1290

1291 1292 1293
	bprm->p++;
	bprm->argc--;
	ret = 0;
N
Nick Piggin 已提交
1294

1295 1296
out:
	return ret;
L
Linus Torvalds 已提交
1297 1298 1299 1300 1301 1302 1303 1304
}
EXPORT_SYMBOL(remove_arg_zero);

/*
 * cycle the list of binary formats handler, until one recognizes the image
 */
int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs)
{
1305
	unsigned int depth = bprm->recursion_depth;
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
	int try,retval;
	struct linux_binfmt *fmt;

	retval = security_bprm_check(bprm);
	if (retval)
		return retval;

	/* kernel module loader fixup */
	/* so we don't try to load run modprobe in kernel space. */
	set_fs(USER_DS);
A
Al Viro 已提交
1316 1317 1318 1319 1320

	retval = audit_bprm(bprm);
	if (retval)
		return retval;

L
Linus Torvalds 已提交
1321 1322 1323
	retval = -ENOENT;
	for (try=0; try<2; try++) {
		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
1324
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330 1331
			int (*fn)(struct linux_binprm *, struct pt_regs *) = fmt->load_binary;
			if (!fn)
				continue;
			if (!try_module_get(fmt->module))
				continue;
			read_unlock(&binfmt_lock);
			retval = fn(bprm, regs);
1332 1333 1334 1335 1336 1337
			/*
			 * Restore the depth counter to its starting value
			 * in this call, so we don't have to rely on every
			 * load_binary function to restore it on return.
			 */
			bprm->recursion_depth = depth;
L
Linus Torvalds 已提交
1338
			if (retval >= 0) {
1339 1340
				if (depth == 0)
					tracehook_report_exec(fmt, bprm, regs);
L
Linus Torvalds 已提交
1341 1342 1343 1344 1345 1346
				put_binfmt(fmt);
				allow_write_access(bprm->file);
				if (bprm->file)
					fput(bprm->file);
				bprm->file = NULL;
				current->did_exec = 1;
M
Matt Helsley 已提交
1347
				proc_exec_connector(current);
L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361
				return retval;
			}
			read_lock(&binfmt_lock);
			put_binfmt(fmt);
			if (retval != -ENOEXEC || bprm->mm == NULL)
				break;
			if (!bprm->file) {
				read_unlock(&binfmt_lock);
				return retval;
			}
		}
		read_unlock(&binfmt_lock);
		if (retval != -ENOEXEC || bprm->mm == NULL) {
			break;
J
Johannes Berg 已提交
1362 1363
#ifdef CONFIG_MODULES
		} else {
L
Linus Torvalds 已提交
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381
#define printable(c) (((c)=='\t') || ((c)=='\n') || (0x20<=(c) && (c)<=0x7e))
			if (printable(bprm->buf[0]) &&
			    printable(bprm->buf[1]) &&
			    printable(bprm->buf[2]) &&
			    printable(bprm->buf[3]))
				break; /* -ENOEXEC */
			request_module("binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
#endif
		}
	}
	return retval;
}

EXPORT_SYMBOL(search_binary_handler);

/*
 * sys_execve() executes a new program.
 */
1382 1383 1384
int do_execve(const char * filename,
	const char __user *const __user *argv,
	const char __user *const __user *envp,
L
Linus Torvalds 已提交
1385 1386 1387 1388
	struct pt_regs * regs)
{
	struct linux_binprm *bprm;
	struct file *file;
1389
	struct files_struct *displaced;
1390
	bool clear_in_exec;
L
Linus Torvalds 已提交
1391 1392
	int retval;

1393
	retval = unshare_files(&displaced);
1394 1395 1396
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1397
	retval = -ENOMEM;
1398
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1399
	if (!bprm)
1400
		goto out_files;
L
Linus Torvalds 已提交
1401

1402 1403
	retval = prepare_bprm_creds(bprm);
	if (retval)
1404
		goto out_free;
A
Al Viro 已提交
1405 1406

	retval = check_unsafe_exec(bprm);
1407
	if (retval < 0)
1408
		goto out_free;
1409
	clear_in_exec = retval;
1410
	current->in_execve = 1;
1411

L
Linus Torvalds 已提交
1412 1413 1414
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1415
		goto out_unmark;
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420 1421 1422

	sched_exec();

	bprm->file = file;
	bprm->filename = filename;
	bprm->interp = filename;

1423 1424 1425
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1426

1427
	bprm->argc = count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1428
	if ((retval = bprm->argc) < 0)
1429
		goto out;
L
Linus Torvalds 已提交
1430

1431
	bprm->envc = count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
	if ((retval = bprm->envc) < 0)
		goto out;

	retval = prepare_binprm(bprm);
	if (retval < 0)
		goto out;

	retval = copy_strings_kernel(1, &bprm->filename, bprm);
	if (retval < 0)
		goto out;

	bprm->exec = bprm->p;
	retval = copy_strings(bprm->envc, envp, bprm);
	if (retval < 0)
		goto out;

	retval = copy_strings(bprm->argc, argv, bprm);
	if (retval < 0)
		goto out;

	retval = search_binary_handler(bprm,regs);
1453 1454
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1455

1456
	/* execve succeeded */
A
Al Viro 已提交
1457
	current->fs->in_exec = 0;
1458
	current->in_execve = 0;
1459 1460 1461 1462 1463
	acct_update_integrals(current);
	free_bprm(bprm);
	if (displaced)
		put_files_struct(displaced);
	return retval;
L
Linus Torvalds 已提交
1464

1465
out:
1466 1467 1468 1469
	if (bprm->mm) {
		acct_arg_size(bprm, 0);
		mmput(bprm->mm);
	}
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}
1476

A
Al Viro 已提交
1477
out_unmark:
1478 1479
	if (clear_in_exec)
		current->fs->in_exec = 0;
1480
	current->in_execve = 0;
1481 1482

out_free:
1483
	free_bprm(bprm);
L
Linus Torvalds 已提交
1484

1485
out_files:
1486 1487
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1488 1489 1490 1491
out_ret:
	return retval;
}

1492
void set_binfmt(struct linux_binfmt *new)
L
Linus Torvalds 已提交
1493
{
1494 1495 1496 1497
	struct mm_struct *mm = current->mm;

	if (mm->binfmt)
		module_put(mm->binfmt->module);
L
Linus Torvalds 已提交
1498

1499
	mm->binfmt = new;
1500 1501
	if (new)
		__module_get(new->module);
L
Linus Torvalds 已提交
1502 1503 1504 1505
}

EXPORT_SYMBOL(set_binfmt);

1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550
static int expand_corename(struct core_name *cn)
{
	char *old_corename = cn->corename;

	cn->size = CORENAME_MAX_SIZE * atomic_inc_return(&call_count);
	cn->corename = krealloc(old_corename, cn->size, GFP_KERNEL);

	if (!cn->corename) {
		kfree(old_corename);
		return -ENOMEM;
	}

	return 0;
}

static int cn_printf(struct core_name *cn, const char *fmt, ...)
{
	char *cur;
	int need;
	int ret;
	va_list arg;

	va_start(arg, fmt);
	need = vsnprintf(NULL, 0, fmt, arg);
	va_end(arg);

	if (likely(need < cn->size - cn->used - 1))
		goto out_printf;

	ret = expand_corename(cn);
	if (ret)
		goto expand_fail;

out_printf:
	cur = cn->corename + cn->used;
	va_start(arg, fmt);
	vsnprintf(cur, need + 1, fmt, arg);
	va_end(arg);
	cn->used += need;
	return 0;

expand_fail:
	return ret;
}

L
Linus Torvalds 已提交
1551 1552 1553 1554
/* format_corename will inspect the pattern parameter, and output a
 * name into corename, which must have space for at least
 * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator.
 */
1555
static int format_corename(struct core_name *cn, long signr)
L
Linus Torvalds 已提交
1556
{
1557
	const struct cred *cred = current_cred();
1558 1559
	const char *pat_ptr = core_pattern;
	int ispipe = (*pat_ptr == '|');
L
Linus Torvalds 已提交
1560
	int pid_in_pattern = 0;
1561 1562 1563 1564 1565 1566 1567 1568
	int err = 0;

	cn->size = CORENAME_MAX_SIZE * atomic_read(&call_count);
	cn->corename = kmalloc(cn->size, GFP_KERNEL);
	cn->used = 0;

	if (!cn->corename)
		return -ENOMEM;
L
Linus Torvalds 已提交
1569 1570 1571 1572 1573

	/* Repeat as long as we have more pattern to process and more output
	   space */
	while (*pat_ptr) {
		if (*pat_ptr != '%') {
1574
			if (*pat_ptr == 0)
L
Linus Torvalds 已提交
1575
				goto out;
1576
			err = cn_printf(cn, "%c", *pat_ptr++);
L
Linus Torvalds 已提交
1577 1578
		} else {
			switch (*++pat_ptr) {
1579
			/* single % at the end, drop that */
L
Linus Torvalds 已提交
1580 1581 1582 1583
			case 0:
				goto out;
			/* Double percent, output one percent */
			case '%':
1584
				err = cn_printf(cn, "%c", '%');
L
Linus Torvalds 已提交
1585 1586 1587 1588
				break;
			/* pid */
			case 'p':
				pid_in_pattern = 1;
1589 1590
				err = cn_printf(cn, "%d",
					      task_tgid_vnr(current));
L
Linus Torvalds 已提交
1591 1592 1593
				break;
			/* uid */
			case 'u':
1594
				err = cn_printf(cn, "%d", cred->uid);
L
Linus Torvalds 已提交
1595 1596 1597
				break;
			/* gid */
			case 'g':
1598
				err = cn_printf(cn, "%d", cred->gid);
L
Linus Torvalds 已提交
1599 1600 1601
				break;
			/* signal that caused the coredump */
			case 's':
1602
				err = cn_printf(cn, "%ld", signr);
L
Linus Torvalds 已提交
1603 1604 1605 1606 1607
				break;
			/* UNIX time of coredump */
			case 't': {
				struct timeval tv;
				do_gettimeofday(&tv);
1608
				err = cn_printf(cn, "%lu", tv.tv_sec);
L
Linus Torvalds 已提交
1609 1610 1611 1612 1613
				break;
			}
			/* hostname */
			case 'h':
				down_read(&uts_sem);
1614 1615
				err = cn_printf(cn, "%s",
					      utsname()->nodename);
L
Linus Torvalds 已提交
1616 1617 1618 1619
				up_read(&uts_sem);
				break;
			/* executable */
			case 'e':
1620
				err = cn_printf(cn, "%s", current->comm);
L
Linus Torvalds 已提交
1621
				break;
1622 1623
			/* core limit size */
			case 'c':
1624 1625
				err = cn_printf(cn, "%lu",
					      rlimit(RLIMIT_CORE));
1626
				break;
L
Linus Torvalds 已提交
1627 1628 1629 1630 1631
			default:
				break;
			}
			++pat_ptr;
		}
1632 1633 1634

		if (err)
			return err;
L
Linus Torvalds 已提交
1635
	}
1636

L
Linus Torvalds 已提交
1637 1638 1639 1640
	/* Backward compatibility with core_uses_pid:
	 *
	 * If core_pattern does not include a %p (as is the default)
	 * and core_uses_pid is set, then .%pid will be appended to
1641
	 * the filename. Do not do this for piped commands. */
1642
	if (!ispipe && !pid_in_pattern && core_uses_pid) {
1643 1644 1645
		err = cn_printf(cn, ".%d", task_tgid_vnr(current));
		if (err)
			return err;
L
Linus Torvalds 已提交
1646
	}
1647 1648
out:
	return ispipe;
L
Linus Torvalds 已提交
1649 1650
}

1651
static int zap_process(struct task_struct *start, int exit_code)
1652 1653
{
	struct task_struct *t;
1654
	int nr = 0;
1655

1656
	start->signal->flags = SIGNAL_GROUP_EXIT;
1657
	start->signal->group_exit_code = exit_code;
1658
	start->signal->group_stop_count = 0;
1659 1660 1661 1662

	t = start;
	do {
		if (t != current && t->mm) {
1663 1664
			sigaddset(&t->pending.signal, SIGKILL);
			signal_wake_up(t, 1);
1665
			nr++;
1666
		}
1667
	} while_each_thread(start, t);
1668 1669

	return nr;
1670 1671
}

1672
static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm,
1673
				struct core_state *core_state, int exit_code)
L
Linus Torvalds 已提交
1674 1675
{
	struct task_struct *g, *p;
1676
	unsigned long flags;
1677
	int nr = -EAGAIN;
1678 1679

	spin_lock_irq(&tsk->sighand->siglock);
1680
	if (!signal_group_exit(tsk->signal)) {
1681
		mm->core_state = core_state;
1682
		nr = zap_process(tsk, exit_code);
L
Linus Torvalds 已提交
1683
	}
1684
	spin_unlock_irq(&tsk->sighand->siglock);
1685 1686
	if (unlikely(nr < 0))
		return nr;
L
Linus Torvalds 已提交
1687

1688
	if (atomic_read(&mm->mm_users) == nr + 1)
1689
		goto done;
1690 1691
	/*
	 * We should find and kill all tasks which use this mm, and we should
1692
	 * count them correctly into ->nr_threads. We don't take tasklist
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
	 * lock, but this is safe wrt:
	 *
	 * fork:
	 *	None of sub-threads can fork after zap_process(leader). All
	 *	processes which were created before this point should be
	 *	visible to zap_threads() because copy_process() adds the new
	 *	process to the tail of init_task.tasks list, and lock/unlock
	 *	of ->siglock provides a memory barrier.
	 *
	 * do_exit:
	 *	The caller holds mm->mmap_sem. This means that the task which
	 *	uses this mm can't pass exit_mm(), so it can't exit or clear
	 *	its ->mm.
	 *
	 * de_thread:
	 *	It does list_replace_rcu(&leader->tasks, &current->tasks),
	 *	we must see either old or new leader, this does not matter.
	 *	However, it can change p->sighand, so lock_task_sighand(p)
	 *	must be used. Since p->mm != NULL and we hold ->mmap_sem
	 *	it can't fail.
	 *
	 *	Note also that "g" can be the old leader with ->mm == NULL
	 *	and already unhashed and thus removed from ->thread_group.
	 *	This is OK, __unhash_process()->list_del_rcu() does not
	 *	clear the ->next pointer, we will find the new leader via
	 *	next_thread().
	 */
1720
	rcu_read_lock();
1721
	for_each_process(g) {
1722 1723
		if (g == tsk->group_leader)
			continue;
1724 1725
		if (g->flags & PF_KTHREAD)
			continue;
1726 1727 1728
		p = g;
		do {
			if (p->mm) {
1729
				if (unlikely(p->mm == mm)) {
1730
					lock_task_sighand(p, &flags);
1731
					nr += zap_process(p, exit_code);
1732 1733
					unlock_task_sighand(p, &flags);
				}
1734 1735
				break;
			}
1736
		} while_each_thread(g, p);
1737
	}
1738
	rcu_read_unlock();
1739
done:
1740
	atomic_set(&core_state->nr_threads, nr);
1741
	return nr;
L
Linus Torvalds 已提交
1742 1743
}

1744
static int coredump_wait(int exit_code, struct core_state *core_state)
L
Linus Torvalds 已提交
1745
{
1746 1747 1748
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;
	struct completion *vfork_done;
1749
	int core_waiters = -EBUSY;
L
Linus Torvalds 已提交
1750

1751
	init_completion(&core_state->startup);
1752 1753
	core_state->dumper.task = tsk;
	core_state->dumper.next = NULL;
1754 1755 1756 1757

	down_write(&mm->mmap_sem);
	if (!mm->core_state)
		core_waiters = zap_threads(tsk, mm, core_state, exit_code);
O
Oleg Nesterov 已提交
1758 1759
	up_write(&mm->mmap_sem);

1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772
	if (unlikely(core_waiters < 0))
		goto fail;

	/*
	 * Make sure nobody is waiting for us to release the VM,
	 * otherwise we can deadlock when we wait on each other
	 */
	vfork_done = tsk->vfork_done;
	if (vfork_done) {
		tsk->vfork_done = NULL;
		complete(vfork_done);
	}

O
Oleg Nesterov 已提交
1773
	if (core_waiters)
1774
		wait_for_completion(&core_state->startup);
1775 1776
fail:
	return core_waiters;
L
Linus Torvalds 已提交
1777 1778
}

O
Oleg Nesterov 已提交
1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799
static void coredump_finish(struct mm_struct *mm)
{
	struct core_thread *curr, *next;
	struct task_struct *task;

	next = mm->core_state->dumper.next;
	while ((curr = next) != NULL) {
		next = curr->next;
		task = curr->task;
		/*
		 * see exit_mm(), curr->task must not see
		 * ->task == NULL before we read ->next.
		 */
		smp_mb();
		curr->task = NULL;
		wake_up_process(task);
	}

	mm->core_state = NULL;
}

1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840
/*
 * set_dumpable converts traditional three-value dumpable to two flags and
 * stores them into mm->flags.  It modifies lower two bits of mm->flags, but
 * these bits are not changed atomically.  So get_dumpable can observe the
 * intermediate state.  To avoid doing unexpected behavior, get get_dumpable
 * return either old dumpable or new one by paying attention to the order of
 * modifying the bits.
 *
 * dumpable |   mm->flags (binary)
 * old  new | initial interim  final
 * ---------+-----------------------
 *  0    1  |   00      01      01
 *  0    2  |   00      10(*)   11
 *  1    0  |   01      00      00
 *  1    2  |   01      11      11
 *  2    0  |   11      10(*)   00
 *  2    1  |   11      11      01
 *
 * (*) get_dumpable regards interim value of 10 as 11.
 */
void set_dumpable(struct mm_struct *mm, int value)
{
	switch (value) {
	case 0:
		clear_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
	case 1:
		set_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
	case 2:
		set_bit(MMF_DUMP_SECURELY, &mm->flags);
		smp_wmb();
		set_bit(MMF_DUMPABLE, &mm->flags);
		break;
	}
}

1841
static int __get_dumpable(unsigned long mm_flags)
1842 1843 1844
{
	int ret;

1845
	ret = mm_flags & MMF_DUMPABLE_MASK;
1846 1847 1848
	return (ret >= 2) ? 2 : ret;
}

1849 1850 1851 1852 1853
int get_dumpable(struct mm_struct *mm)
{
	return __get_dumpable(mm->flags);
}

1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876
static void wait_for_dump_helpers(struct file *file)
{
	struct pipe_inode_info *pipe;

	pipe = file->f_path.dentry->d_inode->i_pipe;

	pipe_lock(pipe);
	pipe->readers++;
	pipe->writers--;

	while ((pipe->readers > 1) && (!signal_pending(current))) {
		wake_up_interruptible_sync(&pipe->wait);
		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
		pipe_wait(pipe);
	}

	pipe->readers--;
	pipe->writers++;
	pipe_unlock(pipe);

}


1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920
/*
 * uhm_pipe_setup
 * helper function to customize the process used
 * to collect the core in userspace.  Specifically
 * it sets up a pipe and installs it as fd 0 (stdin)
 * for the process.  Returns 0 on success, or
 * PTR_ERR on failure.
 * Note that it also sets the core limit to 1.  This
 * is a special value that we use to trap recursive
 * core dumps
 */
static int umh_pipe_setup(struct subprocess_info *info)
{
	struct file *rp, *wp;
	struct fdtable *fdt;
	struct coredump_params *cp = (struct coredump_params *)info->data;
	struct files_struct *cf = current->files;

	wp = create_write_pipe(0);
	if (IS_ERR(wp))
		return PTR_ERR(wp);

	rp = create_read_pipe(wp, 0);
	if (IS_ERR(rp)) {
		free_write_pipe(wp);
		return PTR_ERR(rp);
	}

	cp->file = wp;

	sys_close(0);
	fd_install(0, rp);
	spin_lock(&cf->file_lock);
	fdt = files_fdtable(cf);
	FD_SET(0, fdt->open_fds);
	FD_CLR(0, fdt->close_on_exec);
	spin_unlock(&cf->file_lock);

	/* and disallow core files too */
	current->signal->rlim[RLIMIT_CORE] = (struct rlimit){1, 1};

	return 0;
}

W
WANG Cong 已提交
1921
void do_coredump(long signr, int exit_code, struct pt_regs *regs)
L
Linus Torvalds 已提交
1922
{
1923
	struct core_state core_state;
1924
	struct core_name cn;
L
Linus Torvalds 已提交
1925 1926
	struct mm_struct *mm = current->mm;
	struct linux_binfmt * binfmt;
D
David Howells 已提交
1927 1928
	const struct cred *old_cred;
	struct cred *cred;
L
Linus Torvalds 已提交
1929
	int retval = 0;
A
Alan Cox 已提交
1930
	int flag = 0;
O
Oleg Nesterov 已提交
1931
	int ispipe;
1932
	static atomic_t core_dump_count = ATOMIC_INIT(0);
1933 1934 1935
	struct coredump_params cprm = {
		.signr = signr,
		.regs = regs,
J
Jiri Slaby 已提交
1936
		.limit = rlimit(RLIMIT_CORE),
1937 1938 1939 1940 1941 1942
		/*
		 * We must use the same mm->flags while dumping core to avoid
		 * inconsistency of bit flags, since this flag is not protected
		 * by any locks.
		 */
		.mm_flags = mm->flags,
1943
	};
L
Linus Torvalds 已提交
1944

S
Steve Grubb 已提交
1945 1946
	audit_core_dumps(signr);

1947
	binfmt = mm->binfmt;
L
Linus Torvalds 已提交
1948 1949
	if (!binfmt || !binfmt->core_dump)
		goto fail;
1950 1951
	if (!__get_dumpable(cprm.mm_flags))
		goto fail;
D
David Howells 已提交
1952 1953

	cred = prepare_creds();
1954
	if (!cred)
D
David Howells 已提交
1955
		goto fail;
A
Alan Cox 已提交
1956 1957 1958 1959 1960
	/*
	 *	We cannot trust fsuid as being the "true" uid of the
	 *	process nor do we know its entire history. We only know it
	 *	was tainted so we dump it as root in mode 2.
	 */
1961 1962
	if (__get_dumpable(cprm.mm_flags) == 2) {
		/* Setuid core dump mode */
A
Alan Cox 已提交
1963
		flag = O_EXCL;		/* Stop rewrite attacks */
D
David Howells 已提交
1964
		cred->fsuid = 0;	/* Dump root private */
A
Alan Cox 已提交
1965
	}
1966

1967
	retval = coredump_wait(exit_code, &core_state);
1968 1969
	if (retval < 0)
		goto fail_creds;
D
David Howells 已提交
1970 1971

	old_cred = override_creds(cred);
L
Linus Torvalds 已提交
1972 1973 1974 1975 1976 1977 1978

	/*
	 * Clear any false indication of pending signals that might
	 * be seen by the filesystem code called to write the core file.
	 */
	clear_thread_flag(TIF_SIGPENDING);

1979 1980 1981 1982 1983 1984 1985
	ispipe = format_corename(&cn, signr);

	if (ispipe == -ENOMEM) {
		printk(KERN_WARNING "format_corename failed\n");
		printk(KERN_WARNING "Aborting core\n");
		goto fail_corename;
	}
1986

1987
 	if (ispipe) {
O
Oleg Nesterov 已提交
1988 1989 1990
		int dump_count;
		char **helper_argv;

1991
		if (cprm.limit == 1) {
1992 1993 1994
			/*
			 * Normally core limits are irrelevant to pipes, since
			 * we're not writing to the file system, but we use
1995 1996
			 * cprm.limit of 1 here as a speacial value. Any
			 * non-1 limit gets set to RLIM_INFINITY below, but
1997 1998
			 * a limit of 0 skips the dump.  This is a consistent
			 * way to catch recursive crashes.  We can still crash
1999
			 * if the core_pattern binary sets RLIM_CORE =  !1
2000 2001 2002 2003 2004 2005 2006
			 * but it runs as root, and can do lots of stupid things
			 * Note that we use task_tgid_vnr here to grab the pid
			 * of the process group leader.  That way we get the
			 * right pid if a thread in a multi-threaded
			 * core_pattern process dies.
			 */
			printk(KERN_WARNING
2007
				"Process %d(%s) has RLIMIT_CORE set to 1\n",
2008 2009 2010 2011
				task_tgid_vnr(current), current->comm);
			printk(KERN_WARNING "Aborting core\n");
			goto fail_unlock;
		}
O
Oleg Nesterov 已提交
2012
		cprm.limit = RLIM_INFINITY;
2013

2014 2015 2016 2017 2018 2019 2020 2021
		dump_count = atomic_inc_return(&core_dump_count);
		if (core_pipe_limit && (core_pipe_limit < dump_count)) {
			printk(KERN_WARNING "Pid %d(%s) over core_pipe_limit\n",
			       task_tgid_vnr(current), current->comm);
			printk(KERN_WARNING "Skipping core dump\n");
			goto fail_dropcount;
		}

2022
		helper_argv = argv_split(GFP_KERNEL, cn.corename+1, NULL);
2023 2024 2025
		if (!helper_argv) {
			printk(KERN_WARNING "%s failed to allocate memory\n",
			       __func__);
2026
			goto fail_dropcount;
2027
		}
2028

O
Oleg Nesterov 已提交
2029 2030 2031 2032 2033
		retval = call_usermodehelper_fns(helper_argv[0], helper_argv,
					NULL, UMH_WAIT_EXEC, umh_pipe_setup,
					NULL, &cprm);
		argv_free(helper_argv);
		if (retval) {
2034
 			printk(KERN_INFO "Core dump to %s pipe failed\n",
2035
			       cn.corename);
O
Oleg Nesterov 已提交
2036
			goto close_fail;
2037
 		}
2038 2039 2040 2041 2042 2043
	} else {
		struct inode *inode;

		if (cprm.limit < binfmt->min_coredump)
			goto fail_unlock;

2044
		cprm.file = filp_open(cn.corename,
2045 2046
				 O_CREAT | 2 | O_NOFOLLOW | O_LARGEFILE | flag,
				 0600);
2047 2048
		if (IS_ERR(cprm.file))
			goto fail_unlock;
L
Linus Torvalds 已提交
2049

2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
		inode = cprm.file->f_path.dentry->d_inode;
		if (inode->i_nlink > 1)
			goto close_fail;
		if (d_unhashed(cprm.file->f_path.dentry))
			goto close_fail;
		/*
		 * AK: actually i see no reason to not allow this for named
		 * pipes etc, but keep the previous behaviour for now.
		 */
		if (!S_ISREG(inode->i_mode))
			goto close_fail;
		/*
		 * Dont allow local users get cute and trick others to coredump
		 * into their pre-created files.
		 */
		if (inode->i_uid != current_fsuid())
			goto close_fail;
		if (!cprm.file->f_op || !cprm.file->f_op->write)
			goto close_fail;
		if (do_truncate(cprm.file->f_path.dentry, 0, 0, cprm.file))
			goto close_fail;
	}
L
Linus Torvalds 已提交
2072

2073
	retval = binfmt->core_dump(&cprm);
L
Linus Torvalds 已提交
2074 2075
	if (retval)
		current->signal->group_exit_code |= 0x80;
O
Oleg Nesterov 已提交
2076

2077
	if (ispipe && core_pipe_limit)
2078
		wait_for_dump_helpers(cprm.file);
O
Oleg Nesterov 已提交
2079 2080 2081
close_fail:
	if (cprm.file)
		filp_close(cprm.file, NULL);
2082
fail_dropcount:
O
Oleg Nesterov 已提交
2083
	if (ispipe)
2084
		atomic_dec(&core_dump_count);
L
Linus Torvalds 已提交
2085
fail_unlock:
2086 2087
	kfree(cn.corename);
fail_corename:
2088
	coredump_finish(mm);
D
David Howells 已提交
2089
	revert_creds(old_cred);
2090
fail_creds:
D
David Howells 已提交
2091
	put_cred(cred);
L
Linus Torvalds 已提交
2092
fail:
W
WANG Cong 已提交
2093
	return;
L
Linus Torvalds 已提交
2094
}
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104

/*
 * Core dumping helper functions.  These are the only things you should
 * do on a core-file: use only these functions to write out all the
 * necessary info.
 */
int dump_write(struct file *file, const void *addr, int nr)
{
	return access_ok(VERIFY_READ, addr, nr) && file->f_op->write(file, addr, nr, &file->f_pos) == nr;
}
2105
EXPORT_SYMBOL(dump_write);
2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133

int dump_seek(struct file *file, loff_t off)
{
	int ret = 1;

	if (file->f_op->llseek && file->f_op->llseek != no_llseek) {
		if (file->f_op->llseek(file, off, SEEK_CUR) < 0)
			return 0;
	} else {
		char *buf = (char *)get_zeroed_page(GFP_KERNEL);

		if (!buf)
			return 0;
		while (off > 0) {
			unsigned long n = off;

			if (n > PAGE_SIZE)
				n = PAGE_SIZE;
			if (!dump_write(file, buf, n)) {
				ret = 0;
				break;
			}
			off -= n;
		}
		free_page((unsigned long)buf);
	}
	return ret;
}
2134
EXPORT_SYMBOL(dump_seek);