exec.c 52.7 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
#include <linux/module.h>
#include <linux/namei.h>
#include <linux/mount.h>
#include <linux/security.h>
#include <linux/syscalls.h>
48
#include <linux/tsacct_kern.h>
M
Matt Helsley 已提交
49
#include <linux/cn_proc.h>
A
Al Viro 已提交
50
#include <linux/audit.h>
R
Roland McGrath 已提交
51
#include <linux/tracehook.h>
J
Johannes Berg 已提交
52
#include <linux/kmod.h>
53
#include <linux/fsnotify.h>
54
#include <linux/fs_struct.h>
55
#include <linux/pipe_fs_i.h>
Y
Ying Han 已提交
56
#include <linux/oom.h>
57
#include <linux/compat.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 63

#include <trace/events/task.h>
64
#include "internal.h"
L
Linus Torvalds 已提交
65 66

int core_uses_pid;
67
char core_pattern[CORENAME_MAX_SIZE] = "core";
68
unsigned int core_pipe_limit;
A
Alan Cox 已提交
69 70
int suid_dumpable = 0;

71 72 73 74 75 76
struct core_name {
	char *corename;
	int used, size;
};
static atomic_t call_count = ATOMIC_INIT(1);

L
Linus Torvalds 已提交
77 78
/* The maximal length of core_pattern is also specified in sysctl.c */

A
Alexey Dobriyan 已提交
79
static LIST_HEAD(formats);
L
Linus Torvalds 已提交
80 81
static DEFINE_RWLOCK(binfmt_lock);

I
Ivan Kokshaysky 已提交
82
int __register_binfmt(struct linux_binfmt * fmt, int insert)
L
Linus Torvalds 已提交
83 84 85 86
{
	if (!fmt)
		return -EINVAL;
	write_lock(&binfmt_lock);
I
Ivan Kokshaysky 已提交
87 88
	insert ? list_add(&fmt->lh, &formats) :
		 list_add_tail(&fmt->lh, &formats);
L
Linus Torvalds 已提交
89 90 91 92
	write_unlock(&binfmt_lock);
	return 0;	
}

I
Ivan Kokshaysky 已提交
93
EXPORT_SYMBOL(__register_binfmt);
L
Linus Torvalds 已提交
94

95
void unregister_binfmt(struct linux_binfmt * fmt)
L
Linus Torvalds 已提交
96 97
{
	write_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
98
	list_del(&fmt->lh);
L
Linus Torvalds 已提交
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
	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.
 */
115
SYSCALL_DEFINE1(uselib, const char __user *, library)
L
Linus Torvalds 已提交
116
{
117 118 119
	struct file *file;
	char *tmp = getname(library);
	int error = PTR_ERR(tmp);
120 121 122 123 124
	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
	};
125

126 127 128
	if (IS_ERR(tmp))
		goto out;

129
	file = do_filp_open(AT_FDCWD, tmp, &uselib_flags, LOOKUP_FOLLOW);
130 131 132
	putname(tmp);
	error = PTR_ERR(file);
	if (IS_ERR(file))
L
Linus Torvalds 已提交
133 134 135
		goto out;

	error = -EINVAL;
136
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
L
Linus Torvalds 已提交
137 138
		goto exit;

139
	error = -EACCES;
140
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
L
Linus Torvalds 已提交
141 142
		goto exit;

143
	fsnotify_open(file);
144

L
Linus Torvalds 已提交
145 146 147 148 149
	error = -ENOEXEC;
	if(file->f_op) {
		struct linux_binfmt * fmt;

		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
150
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
151 152 153 154 155 156 157 158 159 160 161 162 163
			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);
	}
164
exit:
L
Linus Torvalds 已提交
165 166 167 168 169
	fput(file);
out:
  	return error;
}

170
#ifdef CONFIG_MMU
O
Oleg Nesterov 已提交
171 172 173 174 175 176
/*
 * The nascent bprm->mm is not visible until exec_mmap() but it can
 * use a lot of memory, account these pages in current->mm temporary
 * for oom_badness()->get_mm_rss(). Once exec succeeds or fails, we
 * change the counter back via acct_arg_size(0).
 */
177
static void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
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;
	add_mm_counter(mm, MM_ANONPAGES, diff);
}

189
static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
190 191 192 193 194 195 196
		int write)
{
	struct page *page;
	int ret;

#ifdef CONFIG_STACK_GROWSUP
	if (write) {
197
		ret = expand_downwards(bprm->vma, pos);
198 199 200 201 202 203 204 205 206 207 208
		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
	BUILD_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
static inline void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
309 310 311
{
}

312
static 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;
}

401
struct user_arg_ptr {
402 403 404 405 406 407 408 409 410
#ifdef CONFIG_COMPAT
	bool is_compat;
#endif
	union {
		const char __user *const __user *native;
#ifdef CONFIG_COMPAT
		compat_uptr_t __user *compat;
#endif
	} ptr;
411 412 413
};

static const char __user *get_user_arg_ptr(struct user_arg_ptr argv, int nr)
414
{
415 416 417 418 419 420 421 422
	const char __user *native;

#ifdef CONFIG_COMPAT
	if (unlikely(argv.is_compat)) {
		compat_uptr_t compat;

		if (get_user(compat, argv.ptr.compat + nr))
			return ERR_PTR(-EFAULT);
423

424 425 426 427 428
		return compat_ptr(compat);
	}
#endif

	if (get_user(native, argv.ptr.native + nr))
429 430
		return ERR_PTR(-EFAULT);

431
	return native;
432 433
}

L
Linus Torvalds 已提交
434 435 436
/*
 * count() counts the number of strings in array ARGV.
 */
437
static int count(struct user_arg_ptr argv, int max)
L
Linus Torvalds 已提交
438 439 440
{
	int i = 0;

441
	if (argv.ptr.native != NULL) {
L
Linus Torvalds 已提交
442
		for (;;) {
443
			const char __user *p = get_user_arg_ptr(argv, i);
L
Linus Torvalds 已提交
444 445 446

			if (!p)
				break;
447 448 449 450

			if (IS_ERR(p))
				return -EFAULT;

451
			if (i++ >= max)
L
Linus Torvalds 已提交
452
				return -E2BIG;
453 454 455

			if (fatal_signal_pending(current))
				return -ERESTARTNOHAND;
L
Linus Torvalds 已提交
456 457 458 459 460 461 462
			cond_resched();
		}
	}
	return i;
}

/*
463 464 465
 * '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
466
 */
467
static int copy_strings(int argc, struct user_arg_ptr argv,
A
Adrian Bunk 已提交
468
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
469 470 471
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
472
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
473 474 475
	int ret;

	while (argc-- > 0) {
476
		const char __user *str;
L
Linus Torvalds 已提交
477 478 479
		int len;
		unsigned long pos;

480 481 482
		ret = -EFAULT;
		str = get_user_arg_ptr(argv, argc);
		if (IS_ERR(str))
L
Linus Torvalds 已提交
483 484
			goto out;

485 486 487 488 489 490
		len = strnlen_user(str, MAX_ARG_STRLEN);
		if (!len)
			goto out;

		ret = -E2BIG;
		if (!valid_arg_len(bprm, len))
L
Linus Torvalds 已提交
491 492
			goto out;

493
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
494
		pos = bprm->p;
495 496
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
497 498 499 500

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

501 502 503 504
			if (fatal_signal_pending(current)) {
				ret = -ERESTARTNOHAND;
				goto out;
			}
505 506
			cond_resched();

L
Linus Torvalds 已提交
507
			offset = pos % PAGE_SIZE;
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
			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
524
				if (!page) {
525
					ret = -E2BIG;
L
Linus Torvalds 已提交
526 527 528
					goto out;
				}

529 530
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
531
					kunmap(kmapped_page);
532 533
					put_arg_page(kmapped_page);
				}
L
Linus Torvalds 已提交
534 535
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
536 537
				kpos = pos & PAGE_MASK;
				flush_arg_page(bprm, kpos, kmapped_page);
L
Linus Torvalds 已提交
538
			}
539
			if (copy_from_user(kaddr+offset, str, bytes_to_copy)) {
L
Linus Torvalds 已提交
540 541 542 543 544 545 546
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
547 548
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
549
		kunmap(kmapped_page);
550 551
		put_arg_page(kmapped_page);
	}
L
Linus Torvalds 已提交
552 553 554 555 556 557
	return ret;
}

/*
 * Like copy_strings, but get argv and its values from kernel memory.
 */
558
int copy_strings_kernel(int argc, const char *const *__argv,
559
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
560 561 562
{
	int r;
	mm_segment_t oldfs = get_fs();
563
	struct user_arg_ptr argv = {
564
		.ptr.native = (const char __user *const  __user *)__argv,
565 566
	};

L
Linus Torvalds 已提交
567
	set_fs(KERNEL_DS);
568
	r = copy_strings(argc, argv, bprm);
L
Linus Torvalds 已提交
569
	set_fs(oldfs);
570

L
Linus Torvalds 已提交
571 572 573 574 575
	return r;
}
EXPORT_SYMBOL(copy_strings_kernel);

#ifdef CONFIG_MMU
576

L
Linus Torvalds 已提交
577
/*
578 579 580
 * 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
581
 *
582 583 584 585 586 587
 * 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
588
 */
589
static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift)
L
Linus Torvalds 已提交
590 591
{
	struct mm_struct *mm = vma->vm_mm;
592 593 594 595 596
	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;
P
Peter Zijlstra 已提交
597
	struct mmu_gather tlb;
L
Linus Torvalds 已提交
598

599
	BUG_ON(new_start > new_end);
L
Linus Torvalds 已提交
600

601 602 603 604 605 606 607 608 609 610
	/*
	 * 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)
	 */
611 612
	if (vma_adjust(vma, new_start, old_end, vma->vm_pgoff, NULL))
		return -ENOMEM;
613 614 615 616 617 618 619 620 621 622

	/*
	 * 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();
P
Peter Zijlstra 已提交
623
	tlb_gather_mmu(&tlb, mm, 0);
624 625 626 627
	if (new_end > old_start) {
		/*
		 * when the old and new regions overlap clear from new_end.
		 */
P
Peter Zijlstra 已提交
628
		free_pgd_range(&tlb, new_end, old_end, new_end,
629 630 631 632 633 634 635 636
			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.
		 */
P
Peter Zijlstra 已提交
637
		free_pgd_range(&tlb, old_start, old_end, new_end,
638
			vma->vm_next ? vma->vm_next->vm_start : 0);
L
Linus Torvalds 已提交
639
	}
P
Peter Zijlstra 已提交
640
	tlb_finish_mmu(&tlb, new_end, old_end);
641 642

	/*
643
	 * Shrink the vma to just the new range.  Always succeeds.
644 645 646 647
	 */
	vma_adjust(vma, new_start, new_end, vma->vm_pgoff, NULL);

	return 0;
L
Linus Torvalds 已提交
648 649
}

650 651 652 653
/*
 * 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
654 655 656 657
int setup_arg_pages(struct linux_binprm *bprm,
		    unsigned long stack_top,
		    int executable_stack)
{
658 659
	unsigned long ret;
	unsigned long stack_shift;
L
Linus Torvalds 已提交
660
	struct mm_struct *mm = current->mm;
661 662 663 664
	struct vm_area_struct *vma = bprm->vma;
	struct vm_area_struct *prev = NULL;
	unsigned long vm_flags;
	unsigned long stack_base;
665 666 667
	unsigned long stack_size;
	unsigned long stack_expand;
	unsigned long rlim_stack;
L
Linus Torvalds 已提交
668 669 670

#ifdef CONFIG_STACK_GROWSUP
	/* Limit stack size to 1GB */
J
Jiri Slaby 已提交
671
	stack_base = rlimit_max(RLIMIT_STACK);
L
Linus Torvalds 已提交
672 673 674
	if (stack_base > (1 << 30))
		stack_base = 1 << 30;

675 676 677
	/* 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
678

679
	stack_base = PAGE_ALIGN(stack_top - stack_base);
L
Linus Torvalds 已提交
680

681 682 683
	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
684
#else
685 686
	stack_top = arch_align_stack(stack_top);
	stack_top = PAGE_ALIGN(stack_top);
687 688 689 690 691

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

692 693 694
	stack_shift = vma->vm_end - stack_top;

	bprm->p -= stack_shift;
L
Linus Torvalds 已提交
695 696 697 698
	mm->arg_start = bprm->p;
#endif

	if (bprm->loader)
699 700
		bprm->loader -= stack_shift;
	bprm->exec -= stack_shift;
L
Linus Torvalds 已提交
701 702

	down_write(&mm->mmap_sem);
703
	vm_flags = VM_STACK_FLAGS;
704 705 706 707 708 709 710 711 712 713 714

	/*
	 * 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;
715
	vm_flags |= VM_STACK_INCOMPLETE_SETUP;
716 717 718 719 720 721 722 723 724 725

	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);
726 727
		if (ret)
			goto out_unlock;
L
Linus Torvalds 已提交
728 729
	}

730 731 732
	/* mprotect_fixup is overkill to remove the temporary stack flags */
	vma->vm_flags &= ~VM_STACK_INCOMPLETE_SETUP;

733
	stack_expand = 131072UL; /* randomly 32*4k (or 2*64k) pages */
734 735 736 737 738 739
	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;
740
#ifdef CONFIG_STACK_GROWSUP
741 742 743 744
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_start + rlim_stack;
	else
		stack_base = vma->vm_end + stack_expand;
745
#else
746 747 748 749
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_end - rlim_stack;
	else
		stack_base = vma->vm_start - stack_expand;
750
#endif
751
	current->mm->start_stack = bprm->p;
752 753 754 755 756
	ret = expand_stack(vma, stack_base);
	if (ret)
		ret = -EFAULT;

out_unlock:
L
Linus Torvalds 已提交
757
	up_write(&mm->mmap_sem);
758
	return ret;
L
Linus Torvalds 已提交
759 760 761 762 763 764 765 766
}
EXPORT_SYMBOL(setup_arg_pages);

#endif /* CONFIG_MMU */

struct file *open_exec(const char *name)
{
	struct file *file;
767
	int err;
768 769 770 771 772
	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
773

774
	file = do_filp_open(AT_FDCWD, name, &open_exec_flags, LOOKUP_FOLLOW);
775
	if (IS_ERR(file))
776 777 778
		goto out;

	err = -EACCES;
779 780
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
		goto exit;
781

782 783
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
		goto exit;
784

785
	fsnotify_open(file);
786

787
	err = deny_write_access(file);
788 789
	if (err)
		goto exit;
L
Linus Torvalds 已提交
790

791
out:
792 793
	return file;

794 795
exit:
	fput(file);
796 797
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
798 799
EXPORT_SYMBOL(open_exec);

M
Mimi Zohar 已提交
800 801
int kernel_read(struct file *file, loff_t offset,
		char *addr, unsigned long count)
L
Linus Torvalds 已提交
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
{
	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;
825
	sync_mm_rss(tsk, old_mm);
L
Linus Torvalds 已提交
826 827 828 829 830 831 832
	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
833
		 * checking core_state and changing tsk->mm.
L
Linus Torvalds 已提交
834 835
		 */
		down_read(&old_mm->mmap_sem);
836
		if (unlikely(old_mm->core_state)) {
L
Linus Torvalds 已提交
837 838 839 840 841 842 843 844 845 846 847 848 849
			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);
	task_unlock(tsk);
	arch_pick_mmap_layout(mm);
	if (old_mm) {
		up_read(&old_mm->mmap_sem);
850
		BUG_ON(active_mm != old_mm);
851
		mm_update_next_owner(old_mm);
L
Linus Torvalds 已提交
852 853 854 855 856 857 858 859 860 861 862 863 864
		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().)
 */
865
static int de_thread(struct task_struct *tsk)
L
Linus Torvalds 已提交
866 867
{
	struct signal_struct *sig = tsk->signal;
868
	struct sighand_struct *oldsighand = tsk->sighand;
L
Linus Torvalds 已提交
869 870
	spinlock_t *lock = &oldsighand->siglock;

871
	if (thread_group_empty(tsk))
L
Linus Torvalds 已提交
872 873 874 875 876 877
		goto no_thread_group;

	/*
	 * Kill all other threads in the thread group.
	 */
	spin_lock_irq(lock);
878
	if (signal_group_exit(sig)) {
L
Linus Torvalds 已提交
879 880 881 882 883 884 885
		/*
		 * Another group action in progress, just
		 * return so that the signal is processed.
		 */
		spin_unlock_irq(lock);
		return -EAGAIN;
	}
886

887
	sig->group_exit_task = tsk;
888 889 890
	sig->notify_count = zap_other_threads(tsk);
	if (!thread_group_leader(tsk))
		sig->notify_count--;
L
Linus Torvalds 已提交
891

892
	while (sig->notify_count) {
L
Linus Torvalds 已提交
893 894 895 896 897 898 899 900 901 902 903 904
		__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:
	 */
905
	if (!thread_group_leader(tsk)) {
906
		struct task_struct *leader = tsk->group_leader;
907

908
		sig->notify_count = -1;	/* for exit_notify() */
909 910 911 912 913 914 915 916
		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
917

918 919 920 921 922 923 924 925 926 927
		/*
		 * 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).
		 */
928
		tsk->start_time = leader->start_time;
929

930 931
		BUG_ON(!same_thread_group(leader, tsk));
		BUG_ON(has_group_leader_pid(tsk));
L
Linus Torvalds 已提交
932 933 934 935 936 937
		/*
		 * 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:
		 */
938 939

		/* Become a process group leader with the old leader's pid.
940 941
		 * The old leader becomes a thread of the this thread group.
		 * Note: The old leader also uses this pid until release_task
942 943
		 *       is called.  Odd but simple and correct.
		 */
944 945
		detach_pid(tsk, PIDTYPE_PID);
		tsk->pid = leader->pid;
946
		attach_pid(tsk, PIDTYPE_PID,  task_pid(leader));
947 948
		transfer_pid(leader, tsk, PIDTYPE_PGID);
		transfer_pid(leader, tsk, PIDTYPE_SID);
949

950
		list_replace_rcu(&leader->tasks, &tsk->tasks);
951
		list_replace_init(&leader->sibling, &tsk->sibling);
L
Linus Torvalds 已提交
952

953 954
		tsk->group_leader = tsk;
		leader->group_leader = tsk;
955

956
		tsk->exit_signal = SIGCHLD;
957
		leader->exit_signal = -1;
958 959 960

		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
		leader->exit_state = EXIT_DEAD;
961 962 963 964 965 966 967 968

		/*
		 * We are going to release_task()->ptrace_unlink() silently,
		 * the tracer can sleep in do_wait(). EXIT_DEAD guarantees
		 * the tracer wont't block again waiting for this thread.
		 */
		if (unlikely(leader->ptrace))
			__wake_up_parent(leader, leader->parent);
L
Linus Torvalds 已提交
969
		write_unlock_irq(&tasklist_lock);
970 971

		release_task(leader);
972
	}
L
Linus Torvalds 已提交
973

974 975
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
L
Linus Torvalds 已提交
976 977

no_thread_group:
J
Jiri Pirko 已提交
978 979 980
	if (current->mm)
		setmax_mm_hiwater_rss(&sig->maxrss, current->mm);

L
Linus Torvalds 已提交
981
	exit_itimers(sig);
982
	flush_itimer_signals();
983

984 985
	if (atomic_read(&oldsighand->count) != 1) {
		struct sighand_struct *newsighand;
L
Linus Torvalds 已提交
986
		/*
987 988
		 * This ->sighand is shared with the CLONE_SIGHAND
		 * but not CLONE_THREAD task, switch to the new one.
L
Linus Torvalds 已提交
989
		 */
990 991 992 993
		newsighand = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
		if (!newsighand)
			return -ENOMEM;

L
Linus Torvalds 已提交
994 995 996 997 998 999
		atomic_set(&newsighand->count, 1);
		memcpy(newsighand->action, oldsighand->action,
		       sizeof(newsighand->action));

		write_lock_irq(&tasklist_lock);
		spin_lock(&oldsighand->siglock);
1000
		rcu_assign_pointer(tsk->sighand, newsighand);
L
Linus Torvalds 已提交
1001 1002 1003
		spin_unlock(&oldsighand->siglock);
		write_unlock_irq(&tasklist_lock);

1004
		__cleanup_sighand(oldsighand);
L
Linus Torvalds 已提交
1005 1006
	}

1007
	BUG_ON(!thread_group_leader(tsk));
L
Linus Torvalds 已提交
1008 1009
	return 0;
}
O
Oleg Nesterov 已提交
1010

L
Linus Torvalds 已提交
1011 1012 1013 1014
/*
 * These functions flushes out all traces of the currently running executable
 * so that a new one can be started
 */
1015
static void flush_old_files(struct files_struct * files)
L
Linus Torvalds 已提交
1016 1017
{
	long j = -1;
1018
	struct fdtable *fdt;
L
Linus Torvalds 已提交
1019 1020 1021 1022 1023 1024 1025

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

		j++;
		i = j * __NFDBITS;
1026
		fdt = files_fdtable(files);
1027
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
1028
			break;
1029
		set = fdt->close_on_exec->fds_bits[j];
L
Linus Torvalds 已提交
1030 1031
		if (!set)
			continue;
1032
		fdt->close_on_exec->fds_bits[j] = 0;
L
Linus Torvalds 已提交
1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
		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);
}

1045
char *get_task_comm(char *buf, struct task_struct *tsk)
L
Linus Torvalds 已提交
1046 1047 1048 1049 1050
{
	/* buf must be at least sizeof(tsk->comm) in size */
	task_lock(tsk);
	strncpy(buf, tsk->comm, sizeof(tsk->comm));
	task_unlock(tsk);
1051
	return buf;
L
Linus Torvalds 已提交
1052
}
1053
EXPORT_SYMBOL_GPL(get_task_comm);
L
Linus Torvalds 已提交
1054 1055 1056 1057

void set_task_comm(struct task_struct *tsk, char *buf)
{
	task_lock(tsk);
1058

1059 1060
	trace_task_rename(tsk, buf);

1061 1062 1063 1064 1065 1066 1067 1068
	/*
	 * 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
1069 1070
	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
	task_unlock(tsk);
1071
	perf_event_comm(tsk);
L
Linus Torvalds 已提交
1072 1073
}

1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088
static void filename_to_taskname(char *tcomm, const char *fn, unsigned int len)
{
	int i, ch;

	/* Copies the binary name from after last slash */
	for (i = 0; (ch = *(fn++)) != '\0';) {
		if (ch == '/')
			i = 0; /* overwrite what we wrote */
		else
			if (i < len - 1)
				tcomm[i++] = ch;
	}
	tcomm[i] = '\0';
}

L
Linus Torvalds 已提交
1089 1090
int flush_old_exec(struct linux_binprm * bprm)
{
1091
	int retval;
L
Linus Torvalds 已提交
1092 1093 1094 1095 1096 1097 1098 1099 1100

	/*
	 * 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
1101 1102
	set_mm_exe_file(bprm->mm, bprm->file);

1103
	filename_to_taskname(bprm->tcomm, bprm->filename, sizeof(bprm->tcomm));
L
Linus Torvalds 已提交
1104 1105 1106
	/*
	 * Release all of the old mmap stuff
	 */
1107
	acct_arg_size(bprm, 0);
L
Linus Torvalds 已提交
1108 1109
	retval = exec_mmap(bprm->mm);
	if (retval)
1110
		goto out;
L
Linus Torvalds 已提交
1111 1112

	bprm->mm = NULL;		/* We're using it now */
1113

1114
	set_fs(USER_DS);
1115
	current->flags &= ~(PF_RANDOMIZE | PF_KTHREAD);
1116 1117 1118
	flush_thread();
	current->personality &= ~bprm->per_clear;

1119 1120 1121 1122 1123 1124 1125
	return 0;

out:
	return retval;
}
EXPORT_SYMBOL(flush_old_exec);

1126 1127 1128 1129 1130 1131 1132
void would_dump(struct linux_binprm *bprm, struct file *file)
{
	if (inode_permission(file->f_path.dentry->d_inode, MAY_READ) < 0)
		bprm->interp_flags |= BINPRM_FLAGS_ENFORCE_NONDUMP;
}
EXPORT_SYMBOL(would_dump);

1133 1134 1135
void setup_new_exec(struct linux_binprm * bprm)
{
	arch_pick_mmap_layout(current->mm);
L
Linus Torvalds 已提交
1136 1137 1138 1139

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

1140
	if (current_euid() == current_uid() && current_egid() == current_gid())
1141
		set_dumpable(current->mm, 1);
A
Alan Cox 已提交
1142
	else
1143
		set_dumpable(current->mm, suid_dumpable);
A
Alan Cox 已提交
1144

1145
	set_task_comm(current, bprm->tcomm);
L
Linus Torvalds 已提交
1146

1147 1148 1149 1150 1151 1152
	/* 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;

1153 1154 1155
	/* install the new credentials */
	if (bprm->cred->uid != current_euid() ||
	    bprm->cred->gid != current_egid()) {
1156
		current->pdeath_signal = 0;
1157 1158 1159 1160
	} else {
		would_dump(bprm, bprm->file);
		if (bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP)
			set_dumpable(current->mm, suid_dumpable);
L
Linus Torvalds 已提交
1161 1162
	}

I
Ingo Molnar 已提交
1163 1164 1165 1166 1167
	/*
	 * Flush performance counters when crossing a
	 * security domain:
	 */
	if (!get_dumpable(current->mm))
1168
		perf_event_exit_task(current);
I
Ingo Molnar 已提交
1169

L
Linus Torvalds 已提交
1170 1171 1172 1173 1174 1175 1176 1177
	/* 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);
}
1178
EXPORT_SYMBOL(setup_new_exec);
L
Linus Torvalds 已提交
1179

1180 1181 1182 1183 1184 1185 1186 1187
/*
 * 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)
{
1188
	if (mutex_lock_interruptible(&current->signal->cred_guard_mutex))
1189 1190 1191 1192 1193 1194
		return -ERESTARTNOINTR;

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

1195
	mutex_unlock(&current->signal->cred_guard_mutex);
1196 1197 1198 1199 1200 1201 1202
	return -ENOMEM;
}

void free_bprm(struct linux_binprm *bprm)
{
	free_arg_pages(bprm);
	if (bprm->cred) {
1203
		mutex_unlock(&current->signal->cred_guard_mutex);
1204 1205 1206 1207 1208
		abort_creds(bprm->cred);
	}
	kfree(bprm);
}

1209 1210 1211 1212 1213 1214 1215 1216 1217
/*
 * 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;
1218 1219
	/*
	 * cred_guard_mutex must be held at least to this point to prevent
1220
	 * ptrace_attach() from altering our determination of the task's
1221 1222
	 * credentials; any time after this it may be unlocked.
	 */
1223
	security_bprm_committed_creds(bprm);
1224
	mutex_unlock(&current->signal->cred_guard_mutex);
1225 1226 1227 1228 1229
}
EXPORT_SYMBOL(install_exec_creds);

/*
 * determine how safe it is to execute the proposed program
1230
 * - the caller must hold ->cred_guard_mutex to protect against
1231 1232
 *   PTRACE_ATTACH
 */
A
Al Viro 已提交
1233
static int check_unsafe_exec(struct linux_binprm *bprm)
1234
{
D
David Howells 已提交
1235
	struct task_struct *p = current, *t;
1236
	unsigned n_fs;
A
Al Viro 已提交
1237
	int res = 0;
1238

T
Tejun Heo 已提交
1239 1240 1241 1242 1243 1244
	if (p->ptrace) {
		if (p->ptrace & PT_PTRACE_CAP)
			bprm->unsafe |= LSM_UNSAFE_PTRACE_CAP;
		else
			bprm->unsafe |= LSM_UNSAFE_PTRACE;
	}
1245

D
David Howells 已提交
1246
	n_fs = 1;
N
Nick Piggin 已提交
1247
	spin_lock(&p->fs->lock);
1248
	rcu_read_lock();
D
David Howells 已提交
1249 1250 1251 1252
	for (t = next_thread(p); t != p; t = next_thread(t)) {
		if (t->fs == p->fs)
			n_fs++;
	}
1253
	rcu_read_unlock();
D
David Howells 已提交
1254

1255
	if (p->fs->users > n_fs) {
1256
		bprm->unsafe |= LSM_UNSAFE_SHARE;
A
Al Viro 已提交
1257
	} else {
1258 1259 1260 1261 1262
		res = -EAGAIN;
		if (!p->fs->in_exec) {
			p->fs->in_exec = 1;
			res = 1;
		}
A
Al Viro 已提交
1263
	}
N
Nick Piggin 已提交
1264
	spin_unlock(&p->fs->lock);
A
Al Viro 已提交
1265 1266

	return res;
1267 1268
}

L
Linus Torvalds 已提交
1269 1270 1271
/* 
 * Fill the binprm structure from the inode. 
 * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes
1272 1273
 *
 * This may be called multiple times for binary chains (scripts for example).
L
Linus Torvalds 已提交
1274 1275 1276
 */
int prepare_binprm(struct linux_binprm *bprm)
{
1277
	umode_t mode;
1278
	struct inode * inode = bprm->file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283 1284
	int retval;

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

1285 1286 1287
	/* 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
1288

1289
	if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) {
L
Linus Torvalds 已提交
1290 1291
		/* Set-uid? */
		if (mode & S_ISUID) {
1292 1293
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->euid = inode->i_uid;
L
Linus Torvalds 已提交
1294 1295 1296 1297 1298 1299 1300 1301 1302
		}

		/* 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)) {
1303 1304
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->egid = inode->i_gid;
L
Linus Torvalds 已提交
1305 1306 1307 1308
		}
	}

	/* fill in binprm security blob */
1309
	retval = security_bprm_set_creds(bprm);
L
Linus Torvalds 已提交
1310 1311
	if (retval)
		return retval;
1312
	bprm->cred_prepared = 1;
L
Linus Torvalds 已提交
1313

1314 1315
	memset(bprm->buf, 0, BINPRM_BUF_SIZE);
	return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE);
L
Linus Torvalds 已提交
1316 1317 1318 1319
}

EXPORT_SYMBOL(prepare_binprm);

N
Nick Piggin 已提交
1320 1321 1322 1323 1324
/*
 * 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.
 */
1325
int remove_arg_zero(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1326
{
1327 1328 1329 1330
	int ret = 0;
	unsigned long offset;
	char *kaddr;
	struct page *page;
N
Nick Piggin 已提交
1331

1332 1333
	if (!bprm->argc)
		return 0;
L
Linus Torvalds 已提交
1334

1335 1336 1337 1338 1339 1340 1341 1342
	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
1343

1344 1345 1346
		for (; offset < PAGE_SIZE && kaddr[offset];
				offset++, bprm->p++)
			;
N
Nick Piggin 已提交
1347

1348 1349
		kunmap_atomic(kaddr, KM_USER0);
		put_arg_page(page);
N
Nick Piggin 已提交
1350

1351 1352 1353
		if (offset == PAGE_SIZE)
			free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1);
	} while (offset == PAGE_SIZE);
N
Nick Piggin 已提交
1354

1355 1356 1357
	bprm->p++;
	bprm->argc--;
	ret = 0;
N
Nick Piggin 已提交
1358

1359 1360
out:
	return ret;
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368
}
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)
{
1369
	unsigned int depth = bprm->recursion_depth;
L
Linus Torvalds 已提交
1370 1371
	int try,retval;
	struct linux_binfmt *fmt;
1372
	pid_t old_pid;
L
Linus Torvalds 已提交
1373 1374 1375 1376 1377

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

A
Al Viro 已提交
1378 1379 1380 1381
	retval = audit_bprm(bprm);
	if (retval)
		return retval;

1382 1383 1384 1385 1386
	/* Need to fetch pid before load_binary changes it */
	rcu_read_lock();
	old_pid = task_pid_nr_ns(current, task_active_pid_ns(current->parent));
	rcu_read_unlock();

L
Linus Torvalds 已提交
1387 1388 1389
	retval = -ENOENT;
	for (try=0; try<2; try++) {
		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
1390
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
1391 1392 1393 1394 1395 1396 1397
			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);
1398 1399 1400 1401 1402 1403
			/*
			 * 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
1404
			if (retval >= 0) {
1405
				if (depth == 0)
1406 1407
					ptrace_event(PTRACE_EVENT_EXEC,
							old_pid);
L
Linus Torvalds 已提交
1408 1409 1410 1411 1412 1413
				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
1414
				proc_exec_connector(current);
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
				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);
1427
#ifdef CONFIG_MODULES
L
Linus Torvalds 已提交
1428 1429
		if (retval != -ENOEXEC || bprm->mm == NULL) {
			break;
J
Johannes Berg 已提交
1430
		} else {
L
Linus Torvalds 已提交
1431 1432 1433 1434 1435 1436
#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 */
1437 1438
			if (try)
				break; /* -ENOEXEC */
L
Linus Torvalds 已提交
1439 1440
			request_module("binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
		}
1441 1442 1443
#else
		break;
#endif
L
Linus Torvalds 已提交
1444 1445 1446 1447 1448 1449 1450 1451 1452
	}
	return retval;
}

EXPORT_SYMBOL(search_binary_handler);

/*
 * sys_execve() executes a new program.
 */
1453 1454 1455 1456
static int do_execve_common(const char *filename,
				struct user_arg_ptr argv,
				struct user_arg_ptr envp,
				struct pt_regs *regs)
L
Linus Torvalds 已提交
1457 1458 1459
{
	struct linux_binprm *bprm;
	struct file *file;
1460
	struct files_struct *displaced;
1461
	bool clear_in_exec;
L
Linus Torvalds 已提交
1462
	int retval;
1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479
	const struct cred *cred = current_cred();

	/*
	 * We move the actual failure in case of RLIMIT_NPROC excess from
	 * set*uid() to execve() because too many poorly written programs
	 * don't check setuid() return code.  Here we additionally recheck
	 * whether NPROC limit is still exceeded.
	 */
	if ((current->flags & PF_NPROC_EXCEEDED) &&
	    atomic_read(&cred->user->processes) > rlimit(RLIMIT_NPROC)) {
		retval = -EAGAIN;
		goto out_ret;
	}

	/* We're below the limit (still or again), so we don't want to make
	 * further execve() calls fail. */
	current->flags &= ~PF_NPROC_EXCEEDED;
L
Linus Torvalds 已提交
1480

1481
	retval = unshare_files(&displaced);
1482 1483 1484
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1485
	retval = -ENOMEM;
1486
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1487
	if (!bprm)
1488
		goto out_files;
L
Linus Torvalds 已提交
1489

1490 1491
	retval = prepare_bprm_creds(bprm);
	if (retval)
1492
		goto out_free;
A
Al Viro 已提交
1493 1494

	retval = check_unsafe_exec(bprm);
1495
	if (retval < 0)
1496
		goto out_free;
1497
	clear_in_exec = retval;
1498
	current->in_execve = 1;
1499

L
Linus Torvalds 已提交
1500 1501 1502
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1503
		goto out_unmark;
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510

	sched_exec();

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

1511 1512 1513
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1514

1515
	bprm->argc = count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1516
	if ((retval = bprm->argc) < 0)
1517
		goto out;
L
Linus Torvalds 已提交
1518

1519
	bprm->envc = count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
	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);
1541 1542
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1543

1544
	/* execve succeeded */
A
Al Viro 已提交
1545
	current->fs->in_exec = 0;
1546
	current->in_execve = 0;
1547 1548 1549 1550 1551
	acct_update_integrals(current);
	free_bprm(bprm);
	if (displaced)
		put_files_struct(displaced);
	return retval;
L
Linus Torvalds 已提交
1552

1553
out:
1554 1555 1556 1557
	if (bprm->mm) {
		acct_arg_size(bprm, 0);
		mmput(bprm->mm);
	}
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562 1563

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}
1564

A
Al Viro 已提交
1565
out_unmark:
1566 1567
	if (clear_in_exec)
		current->fs->in_exec = 0;
1568
	current->in_execve = 0;
1569 1570

out_free:
1571
	free_bprm(bprm);
L
Linus Torvalds 已提交
1572

1573
out_files:
1574 1575
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1576 1577 1578 1579
out_ret:
	return retval;
}

1580 1581 1582 1583 1584
int do_execve(const char *filename,
	const char __user *const __user *__argv,
	const char __user *const __user *__envp,
	struct pt_regs *regs)
{
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603
	struct user_arg_ptr argv = { .ptr.native = __argv };
	struct user_arg_ptr envp = { .ptr.native = __envp };
	return do_execve_common(filename, argv, envp, regs);
}

#ifdef CONFIG_COMPAT
int compat_do_execve(char *filename,
	compat_uptr_t __user *__argv,
	compat_uptr_t __user *__envp,
	struct pt_regs *regs)
{
	struct user_arg_ptr argv = {
		.is_compat = true,
		.ptr.compat = __argv,
	};
	struct user_arg_ptr envp = {
		.is_compat = true,
		.ptr.compat = __envp,
	};
1604 1605
	return do_execve_common(filename, argv, envp, regs);
}
1606
#endif
1607

1608
void set_binfmt(struct linux_binfmt *new)
L
Linus Torvalds 已提交
1609
{
1610 1611 1612 1613
	struct mm_struct *mm = current->mm;

	if (mm->binfmt)
		module_put(mm->binfmt->module);
L
Linus Torvalds 已提交
1614

1615
	mm->binfmt = new;
1616 1617
	if (new)
		__module_get(new->module);
L
Linus Torvalds 已提交
1618 1619 1620 1621
}

EXPORT_SYMBOL(set_binfmt);

1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
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;
}

1667 1668 1669 1670 1671 1672 1673
static void cn_escape(char *str)
{
	for (; *str; str++)
		if (*str == '/')
			*str = '!';
}

1674 1675 1676
static int cn_print_exe_file(struct core_name *cn)
{
	struct file *exe_file;
1677
	char *pathbuf, *path;
1678 1679 1680
	int ret;

	exe_file = get_mm_exe_file(current->mm);
1681 1682 1683 1684 1685 1686
	if (!exe_file) {
		char *commstart = cn->corename + cn->used;
		ret = cn_printf(cn, "%s (path unknown)", current->comm);
		cn_escape(commstart);
		return ret;
	}
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699

	pathbuf = kmalloc(PATH_MAX, GFP_TEMPORARY);
	if (!pathbuf) {
		ret = -ENOMEM;
		goto put_exe_file;
	}

	path = d_path(&exe_file->f_path, pathbuf, PATH_MAX);
	if (IS_ERR(path)) {
		ret = PTR_ERR(path);
		goto free_buf;
	}

1700
	cn_escape(path);
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710

	ret = cn_printf(cn, "%s", path);

free_buf:
	kfree(pathbuf);
put_exe_file:
	fput(exe_file);
	return ret;
}

L
Linus Torvalds 已提交
1711 1712 1713 1714
/* 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.
 */
1715
static int format_corename(struct core_name *cn, long signr)
L
Linus Torvalds 已提交
1716
{
1717
	const struct cred *cred = current_cred();
1718 1719
	const char *pat_ptr = core_pattern;
	int ispipe = (*pat_ptr == '|');
L
Linus Torvalds 已提交
1720
	int pid_in_pattern = 0;
1721 1722 1723 1724 1725 1726 1727 1728
	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
1729 1730 1731 1732 1733

	/* Repeat as long as we have more pattern to process and more output
	   space */
	while (*pat_ptr) {
		if (*pat_ptr != '%') {
1734
			if (*pat_ptr == 0)
L
Linus Torvalds 已提交
1735
				goto out;
1736
			err = cn_printf(cn, "%c", *pat_ptr++);
L
Linus Torvalds 已提交
1737 1738
		} else {
			switch (*++pat_ptr) {
1739
			/* single % at the end, drop that */
L
Linus Torvalds 已提交
1740 1741 1742 1743
			case 0:
				goto out;
			/* Double percent, output one percent */
			case '%':
1744
				err = cn_printf(cn, "%c", '%');
L
Linus Torvalds 已提交
1745 1746 1747 1748
				break;
			/* pid */
			case 'p':
				pid_in_pattern = 1;
1749 1750
				err = cn_printf(cn, "%d",
					      task_tgid_vnr(current));
L
Linus Torvalds 已提交
1751 1752 1753
				break;
			/* uid */
			case 'u':
1754
				err = cn_printf(cn, "%d", cred->uid);
L
Linus Torvalds 已提交
1755 1756 1757
				break;
			/* gid */
			case 'g':
1758
				err = cn_printf(cn, "%d", cred->gid);
L
Linus Torvalds 已提交
1759 1760 1761
				break;
			/* signal that caused the coredump */
			case 's':
1762
				err = cn_printf(cn, "%ld", signr);
L
Linus Torvalds 已提交
1763 1764 1765 1766 1767
				break;
			/* UNIX time of coredump */
			case 't': {
				struct timeval tv;
				do_gettimeofday(&tv);
1768
				err = cn_printf(cn, "%lu", tv.tv_sec);
L
Linus Torvalds 已提交
1769 1770 1771
				break;
			}
			/* hostname */
1772 1773
			case 'h': {
				char *namestart = cn->corename + cn->used;
L
Linus Torvalds 已提交
1774
				down_read(&uts_sem);
1775 1776
				err = cn_printf(cn, "%s",
					      utsname()->nodename);
L
Linus Torvalds 已提交
1777
				up_read(&uts_sem);
1778
				cn_escape(namestart);
L
Linus Torvalds 已提交
1779
				break;
1780
			}
L
Linus Torvalds 已提交
1781
			/* executable */
1782 1783
			case 'e': {
				char *commstart = cn->corename + cn->used;
1784
				err = cn_printf(cn, "%s", current->comm);
1785
				cn_escape(commstart);
L
Linus Torvalds 已提交
1786
				break;
1787
			}
1788 1789 1790
			case 'E':
				err = cn_print_exe_file(cn);
				break;
1791 1792
			/* core limit size */
			case 'c':
1793 1794
				err = cn_printf(cn, "%lu",
					      rlimit(RLIMIT_CORE));
1795
				break;
L
Linus Torvalds 已提交
1796 1797 1798 1799 1800
			default:
				break;
			}
			++pat_ptr;
		}
1801 1802 1803

		if (err)
			return err;
L
Linus Torvalds 已提交
1804
	}
1805

L
Linus Torvalds 已提交
1806 1807 1808 1809
	/* 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
1810
	 * the filename. Do not do this for piped commands. */
1811
	if (!ispipe && !pid_in_pattern && core_uses_pid) {
1812 1813 1814
		err = cn_printf(cn, ".%d", task_tgid_vnr(current));
		if (err)
			return err;
L
Linus Torvalds 已提交
1815
	}
1816 1817
out:
	return ispipe;
L
Linus Torvalds 已提交
1818 1819
}

1820
static int zap_process(struct task_struct *start, int exit_code)
1821 1822
{
	struct task_struct *t;
1823
	int nr = 0;
1824

1825
	start->signal->flags = SIGNAL_GROUP_EXIT;
1826
	start->signal->group_exit_code = exit_code;
1827
	start->signal->group_stop_count = 0;
1828 1829 1830

	t = start;
	do {
1831
		task_clear_jobctl_pending(t, JOBCTL_PENDING_MASK);
1832
		if (t != current && t->mm) {
1833 1834
			sigaddset(&t->pending.signal, SIGKILL);
			signal_wake_up(t, 1);
1835
			nr++;
1836
		}
1837
	} while_each_thread(start, t);
1838 1839

	return nr;
1840 1841
}

1842
static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm,
1843
				struct core_state *core_state, int exit_code)
L
Linus Torvalds 已提交
1844 1845
{
	struct task_struct *g, *p;
1846
	unsigned long flags;
1847
	int nr = -EAGAIN;
1848 1849

	spin_lock_irq(&tsk->sighand->siglock);
1850
	if (!signal_group_exit(tsk->signal)) {
1851
		mm->core_state = core_state;
1852
		nr = zap_process(tsk, exit_code);
L
Linus Torvalds 已提交
1853
	}
1854
	spin_unlock_irq(&tsk->sighand->siglock);
1855 1856
	if (unlikely(nr < 0))
		return nr;
L
Linus Torvalds 已提交
1857

1858
	if (atomic_read(&mm->mm_users) == nr + 1)
1859
		goto done;
1860 1861
	/*
	 * We should find and kill all tasks which use this mm, and we should
1862
	 * count them correctly into ->nr_threads. We don't take tasklist
1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889
	 * 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().
	 */
1890
	rcu_read_lock();
1891
	for_each_process(g) {
1892 1893
		if (g == tsk->group_leader)
			continue;
1894 1895
		if (g->flags & PF_KTHREAD)
			continue;
1896 1897 1898
		p = g;
		do {
			if (p->mm) {
1899
				if (unlikely(p->mm == mm)) {
1900
					lock_task_sighand(p, &flags);
1901
					nr += zap_process(p, exit_code);
1902 1903
					unlock_task_sighand(p, &flags);
				}
1904 1905
				break;
			}
1906
		} while_each_thread(g, p);
1907
	}
1908
	rcu_read_unlock();
1909
done:
1910
	atomic_set(&core_state->nr_threads, nr);
1911
	return nr;
L
Linus Torvalds 已提交
1912 1913
}

1914
static int coredump_wait(int exit_code, struct core_state *core_state)
L
Linus Torvalds 已提交
1915
{
1916 1917 1918
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;
	struct completion *vfork_done;
1919
	int core_waiters = -EBUSY;
L
Linus Torvalds 已提交
1920

1921
	init_completion(&core_state->startup);
1922 1923
	core_state->dumper.task = tsk;
	core_state->dumper.next = NULL;
1924 1925 1926 1927

	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
1928 1929
	up_write(&mm->mmap_sem);

1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942
	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
1943
	if (core_waiters)
1944
		wait_for_completion(&core_state->startup);
1945 1946
fail:
	return core_waiters;
L
Linus Torvalds 已提交
1947 1948
}

O
Oleg Nesterov 已提交
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969
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;
}

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
/*
 * 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;
	}
}

2011
static int __get_dumpable(unsigned long mm_flags)
2012 2013 2014
{
	int ret;

2015
	ret = mm_flags & MMF_DUMPABLE_MASK;
2016 2017 2018
	return (ret >= 2) ? 2 : ret;
}

2019 2020 2021 2022 2023
int get_dumpable(struct mm_struct *mm)
{
	return __get_dumpable(mm->flags);
}

2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046
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);

}


2047
/*
H
Holger Hans Peter Freyther 已提交
2048
 * umh_pipe_setup
2049 2050 2051 2052 2053 2054 2055 2056 2057
 * 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
 */
2058
static int umh_pipe_setup(struct subprocess_info *info, struct cred *new)
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
{
	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
2091
void do_coredump(long signr, int exit_code, struct pt_regs *regs)
L
Linus Torvalds 已提交
2092
{
2093
	struct core_state core_state;
2094
	struct core_name cn;
L
Linus Torvalds 已提交
2095 2096
	struct mm_struct *mm = current->mm;
	struct linux_binfmt * binfmt;
D
David Howells 已提交
2097 2098
	const struct cred *old_cred;
	struct cred *cred;
L
Linus Torvalds 已提交
2099
	int retval = 0;
A
Alan Cox 已提交
2100
	int flag = 0;
O
Oleg Nesterov 已提交
2101
	int ispipe;
2102
	static atomic_t core_dump_count = ATOMIC_INIT(0);
2103 2104 2105
	struct coredump_params cprm = {
		.signr = signr,
		.regs = regs,
J
Jiri Slaby 已提交
2106
		.limit = rlimit(RLIMIT_CORE),
2107 2108 2109 2110 2111 2112
		/*
		 * 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,
2113
	};
L
Linus Torvalds 已提交
2114

S
Steve Grubb 已提交
2115 2116
	audit_core_dumps(signr);

2117
	binfmt = mm->binfmt;
L
Linus Torvalds 已提交
2118 2119
	if (!binfmt || !binfmt->core_dump)
		goto fail;
2120 2121
	if (!__get_dumpable(cprm.mm_flags))
		goto fail;
D
David Howells 已提交
2122 2123

	cred = prepare_creds();
2124
	if (!cred)
D
David Howells 已提交
2125
		goto fail;
A
Alan Cox 已提交
2126 2127 2128 2129 2130
	/*
	 *	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.
	 */
2131 2132
	if (__get_dumpable(cprm.mm_flags) == 2) {
		/* Setuid core dump mode */
A
Alan Cox 已提交
2133
		flag = O_EXCL;		/* Stop rewrite attacks */
D
David Howells 已提交
2134
		cred->fsuid = 0;	/* Dump root private */
A
Alan Cox 已提交
2135
	}
2136

2137
	retval = coredump_wait(exit_code, &core_state);
2138 2139
	if (retval < 0)
		goto fail_creds;
D
David Howells 已提交
2140 2141

	old_cred = override_creds(cred);
L
Linus Torvalds 已提交
2142 2143 2144 2145 2146 2147 2148

	/*
	 * 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);

2149 2150
	ispipe = format_corename(&cn, signr);

2151
 	if (ispipe) {
O
Oleg Nesterov 已提交
2152 2153 2154
		int dump_count;
		char **helper_argv;

2155 2156 2157 2158 2159 2160
		if (ispipe < 0) {
			printk(KERN_WARNING "format_corename failed\n");
			printk(KERN_WARNING "Aborting core\n");
			goto fail_corename;
		}

2161
		if (cprm.limit == 1) {
2162 2163 2164
			/*
			 * Normally core limits are irrelevant to pipes, since
			 * we're not writing to the file system, but we use
2165 2166
			 * cprm.limit of 1 here as a speacial value. Any
			 * non-1 limit gets set to RLIM_INFINITY below, but
2167 2168
			 * a limit of 0 skips the dump.  This is a consistent
			 * way to catch recursive crashes.  We can still crash
2169
			 * if the core_pattern binary sets RLIM_CORE =  !1
2170 2171 2172 2173 2174 2175 2176
			 * 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
2177
				"Process %d(%s) has RLIMIT_CORE set to 1\n",
2178 2179 2180 2181
				task_tgid_vnr(current), current->comm);
			printk(KERN_WARNING "Aborting core\n");
			goto fail_unlock;
		}
O
Oleg Nesterov 已提交
2182
		cprm.limit = RLIM_INFINITY;
2183

2184 2185 2186 2187 2188 2189 2190 2191
		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;
		}

2192
		helper_argv = argv_split(GFP_KERNEL, cn.corename+1, NULL);
2193 2194 2195
		if (!helper_argv) {
			printk(KERN_WARNING "%s failed to allocate memory\n",
			       __func__);
2196
			goto fail_dropcount;
2197
		}
2198

O
Oleg Nesterov 已提交
2199 2200 2201 2202 2203
		retval = call_usermodehelper_fns(helper_argv[0], helper_argv,
					NULL, UMH_WAIT_EXEC, umh_pipe_setup,
					NULL, &cprm);
		argv_free(helper_argv);
		if (retval) {
2204
 			printk(KERN_INFO "Core dump to %s pipe failed\n",
2205
			       cn.corename);
O
Oleg Nesterov 已提交
2206
			goto close_fail;
2207
 		}
2208 2209 2210 2211 2212 2213
	} else {
		struct inode *inode;

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

2214
		cprm.file = filp_open(cn.corename,
2215 2216
				 O_CREAT | 2 | O_NOFOLLOW | O_LARGEFILE | flag,
				 0600);
2217 2218
		if (IS_ERR(cprm.file))
			goto fail_unlock;
L
Linus Torvalds 已提交
2219

2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241
		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
2242

2243
	retval = binfmt->core_dump(&cprm);
L
Linus Torvalds 已提交
2244 2245
	if (retval)
		current->signal->group_exit_code |= 0x80;
O
Oleg Nesterov 已提交
2246

2247
	if (ispipe && core_pipe_limit)
2248
		wait_for_dump_helpers(cprm.file);
O
Oleg Nesterov 已提交
2249 2250 2251
close_fail:
	if (cprm.file)
		filp_close(cprm.file, NULL);
2252
fail_dropcount:
O
Oleg Nesterov 已提交
2253
	if (ispipe)
2254
		atomic_dec(&core_dump_count);
L
Linus Torvalds 已提交
2255
fail_unlock:
2256 2257
	kfree(cn.corename);
fail_corename:
2258
	coredump_finish(mm);
D
David Howells 已提交
2259
	revert_creds(old_cred);
2260
fail_creds:
D
David Howells 已提交
2261
	put_cred(cred);
L
Linus Torvalds 已提交
2262
fail:
W
WANG Cong 已提交
2263
	return;
L
Linus Torvalds 已提交
2264
}
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274

/*
 * 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;
}
2275
EXPORT_SYMBOL(dump_write);
2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303

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;
}
2304
EXPORT_SYMBOL(dump_seek);