exec.c 50.2 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>
58
#include <linux/compat.h>
L
Linus Torvalds 已提交
59 60 61

#include <asm/uaccess.h>
#include <asm/mmu_context.h>
62
#include <asm/tlb.h>
63
#include "internal.h"
L
Linus Torvalds 已提交
64 65

int core_uses_pid;
66
char core_pattern[CORENAME_MAX_SIZE] = "core";
67
unsigned int core_pipe_limit;
A
Alan Cox 已提交
68 69
int suid_dumpable = 0;

70 71 72 73 74 75
struct core_name {
	char *corename;
	int used, size;
};
static atomic_t call_count = ATOMIC_INIT(1);

L
Linus Torvalds 已提交
76 77
/* The maximal length of core_pattern is also specified in sysctl.c */

A
Alexey Dobriyan 已提交
78
static LIST_HEAD(formats);
L
Linus Torvalds 已提交
79 80
static DEFINE_RWLOCK(binfmt_lock);

I
Ivan Kokshaysky 已提交
81
int __register_binfmt(struct linux_binfmt * fmt, int insert)
L
Linus Torvalds 已提交
82 83 84 85
{
	if (!fmt)
		return -EINVAL;
	write_lock(&binfmt_lock);
I
Ivan Kokshaysky 已提交
86 87
	insert ? list_add(&fmt->lh, &formats) :
		 list_add_tail(&fmt->lh, &formats);
L
Linus Torvalds 已提交
88 89 90 91
	write_unlock(&binfmt_lock);
	return 0;	
}

I
Ivan Kokshaysky 已提交
92
EXPORT_SYMBOL(__register_binfmt);
L
Linus Torvalds 已提交
93

94
void unregister_binfmt(struct linux_binfmt * fmt)
L
Linus Torvalds 已提交
95 96
{
	write_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
97
	list_del(&fmt->lh);
L
Linus Torvalds 已提交
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
	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.
 */
114
SYSCALL_DEFINE1(uselib, const char __user *, library)
L
Linus Torvalds 已提交
115
{
116 117 118
	struct file *file;
	char *tmp = getname(library);
	int error = PTR_ERR(tmp);
119 120 121 122 123
	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
	};
124

125 126 127
	if (IS_ERR(tmp))
		goto out;

128
	file = do_filp_open(AT_FDCWD, tmp, &uselib_flags, LOOKUP_FOLLOW);
129 130 131
	putname(tmp);
	error = PTR_ERR(file);
	if (IS_ERR(file))
L
Linus Torvalds 已提交
132 133 134
		goto out;

	error = -EINVAL;
135
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
L
Linus Torvalds 已提交
136 137
		goto exit;

138
	error = -EACCES;
139
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
L
Linus Torvalds 已提交
140 141
		goto exit;

142
	fsnotify_open(file);
143

L
Linus Torvalds 已提交
144 145 146 147 148
	error = -ENOEXEC;
	if(file->f_op) {
		struct linux_binfmt * fmt;

		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
149
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
150 151 152 153 154 155 156 157 158 159 160 161 162
			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);
	}
163
exit:
L
Linus Torvalds 已提交
164 165 166 167 168
	fput(file);
out:
  	return error;
}

169 170
#ifdef CONFIG_MMU

171
static void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
{
	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
}

190
static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
		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;
210 211
		struct rlimit *rlim;

212 213
		acct_arg_size(bprm, size / PAGE_SIZE);

214 215 216 217 218 219
		/*
		 * We've historically supported up to 32 pages (ARG_MAX)
		 * of argument strings even with small stacks
		 */
		if (size <= ARG_MAX)
			return page;
220 221 222 223 224 225 226 227

		/*
		 * 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.
		 */
228
		rlim = current->signal->rlim;
J
Jiri Slaby 已提交
229
		if (size > ACCESS_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4) {
230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
			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)
{
259
	int err;
260 261 262 263 264
	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)
265
		return -ENOMEM;
266 267 268 269 270 271 272 273 274 275

	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.
	 */
276
	BUG_ON(VM_STACK_FLAGS & VM_STACK_INCOMPLETE_SETUP);
277 278
	vma->vm_end = STACK_TOP_MAX;
	vma->vm_start = vma->vm_end - PAGE_SIZE;
279
	vma->vm_flags = VM_STACK_FLAGS | VM_STACK_INCOMPLETE_SETUP;
280
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
281
	INIT_LIST_HEAD(&vma->anon_vma_chain);
282 283 284 285 286

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

287
	err = insert_vm_struct(mm, vma);
288
	if (err)
289 290 291 292 293 294 295
		goto err;

	mm->stack_vm = mm->total_vm = 1;
	up_write(&mm->mmap_sem);
	bprm->p = vma->vm_end - sizeof(void *);
	return 0;
err:
296 297 298
	up_write(&mm->mmap_sem);
	bprm->vma = NULL;
	kmem_cache_free(vm_area_cachep, vma);
299 300 301 302 303 304 305 306 307 308
	return err;
}

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

#else

309
static inline void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
310 311 312
{
}

313
static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
		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;
}

402
struct user_arg_ptr {
403 404 405 406 407 408 409 410 411
#ifdef CONFIG_COMPAT
	bool is_compat;
#endif
	union {
		const char __user *const __user *native;
#ifdef CONFIG_COMPAT
		compat_uptr_t __user *compat;
#endif
	} ptr;
412 413 414
};

static const char __user *get_user_arg_ptr(struct user_arg_ptr argv, int nr)
415
{
416 417 418 419 420 421 422 423
	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);
424

425 426 427 428 429
		return compat_ptr(compat);
	}
#endif

	if (get_user(native, argv.ptr.native + nr))
430 431
		return ERR_PTR(-EFAULT);

432
	return native;
433 434
}

L
Linus Torvalds 已提交
435 436 437
/*
 * count() counts the number of strings in array ARGV.
 */
438
static int count(struct user_arg_ptr argv, int max)
L
Linus Torvalds 已提交
439 440 441
{
	int i = 0;

442
	if (argv.ptr.native != NULL) {
L
Linus Torvalds 已提交
443
		for (;;) {
444
			const char __user *p = get_user_arg_ptr(argv, i);
L
Linus Torvalds 已提交
445 446 447

			if (!p)
				break;
448 449 450 451

			if (IS_ERR(p))
				return -EFAULT;

452
			if (i++ >= max)
L
Linus Torvalds 已提交
453
				return -E2BIG;
454 455 456

			if (fatal_signal_pending(current))
				return -ERESTARTNOHAND;
L
Linus Torvalds 已提交
457 458 459 460 461 462 463
			cond_resched();
		}
	}
	return i;
}

/*
464 465 466
 * '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
467
 */
468
static int copy_strings(int argc, struct user_arg_ptr argv,
A
Adrian Bunk 已提交
469
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
470 471 472
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
473
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
474 475 476
	int ret;

	while (argc-- > 0) {
477
		const char __user *str;
L
Linus Torvalds 已提交
478 479 480
		int len;
		unsigned long pos;

481 482 483
		ret = -EFAULT;
		str = get_user_arg_ptr(argv, argc);
		if (IS_ERR(str))
L
Linus Torvalds 已提交
484 485
			goto out;

486 487 488 489 490 491
		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
492 493
			goto out;

494
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
495
		pos = bprm->p;
496 497
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
498 499 500 501

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

502 503 504 505
			if (fatal_signal_pending(current)) {
				ret = -ERESTARTNOHAND;
				goto out;
			}
506 507
			cond_resched();

L
Linus Torvalds 已提交
508
			offset = pos % PAGE_SIZE;
509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
			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
525
				if (!page) {
526
					ret = -E2BIG;
L
Linus Torvalds 已提交
527 528 529
					goto out;
				}

530 531
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
532
					kunmap(kmapped_page);
533 534
					put_arg_page(kmapped_page);
				}
L
Linus Torvalds 已提交
535 536
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
537 538
				kpos = pos & PAGE_MASK;
				flush_arg_page(bprm, kpos, kmapped_page);
L
Linus Torvalds 已提交
539
			}
540
			if (copy_from_user(kaddr+offset, str, bytes_to_copy)) {
L
Linus Torvalds 已提交
541 542 543 544 545 546 547
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
548 549
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
550
		kunmap(kmapped_page);
551 552
		put_arg_page(kmapped_page);
	}
L
Linus Torvalds 已提交
553 554 555 556 557 558
	return ret;
}

/*
 * Like copy_strings, but get argv and its values from kernel memory.
 */
559
int copy_strings_kernel(int argc, const char *const *__argv,
560
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
561 562 563
{
	int r;
	mm_segment_t oldfs = get_fs();
564
	struct user_arg_ptr argv = {
565
		.ptr.native = (const char __user *const  __user *)__argv,
566 567
	};

L
Linus Torvalds 已提交
568
	set_fs(KERNEL_DS);
569
	r = copy_strings(argc, argv, bprm);
L
Linus Torvalds 已提交
570
	set_fs(oldfs);
571

L
Linus Torvalds 已提交
572 573 574 575 576
	return r;
}
EXPORT_SYMBOL(copy_strings_kernel);

#ifdef CONFIG_MMU
577

L
Linus Torvalds 已提交
578
/*
579 580 581
 * 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
582
 *
583 584 585 586 587 588
 * 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
589
 */
590
static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift)
L
Linus Torvalds 已提交
591 592
{
	struct mm_struct *mm = vma->vm_mm;
593 594 595 596 597 598
	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
599

600
	BUG_ON(new_start > new_end);
L
Linus Torvalds 已提交
601

602 603 604 605 606 607 608 609 610 611
	/*
	 * 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)
	 */
612 613
	if (vma_adjust(vma, new_start, old_end, vma->vm_pgoff, NULL))
		return -ENOMEM;
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628

	/*
	 * 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.
		 */
629
		free_pgd_range(tlb, new_end, old_end, new_end,
630 631 632 633 634 635 636 637
			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.
		 */
638
		free_pgd_range(tlb, old_start, old_end, new_end,
639
			vma->vm_next ? vma->vm_next->vm_start : 0);
L
Linus Torvalds 已提交
640
	}
641 642 643
	tlb_finish_mmu(tlb, new_end, old_end);

	/*
644
	 * Shrink the vma to just the new range.  Always succeeds.
645 646 647 648
	 */
	vma_adjust(vma, new_start, new_end, vma->vm_pgoff, NULL);

	return 0;
L
Linus Torvalds 已提交
649 650
}

651 652 653 654
/*
 * 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
655 656 657 658
int setup_arg_pages(struct linux_binprm *bprm,
		    unsigned long stack_top,
		    int executable_stack)
{
659 660
	unsigned long ret;
	unsigned long stack_shift;
L
Linus Torvalds 已提交
661
	struct mm_struct *mm = current->mm;
662 663 664 665
	struct vm_area_struct *vma = bprm->vma;
	struct vm_area_struct *prev = NULL;
	unsigned long vm_flags;
	unsigned long stack_base;
666 667 668
	unsigned long stack_size;
	unsigned long stack_expand;
	unsigned long rlim_stack;
L
Linus Torvalds 已提交
669 670 671

#ifdef CONFIG_STACK_GROWSUP
	/* Limit stack size to 1GB */
J
Jiri Slaby 已提交
672
	stack_base = rlimit_max(RLIMIT_STACK);
L
Linus Torvalds 已提交
673 674 675
	if (stack_base > (1 << 30))
		stack_base = 1 << 30;

676 677 678
	/* 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
679

680
	stack_base = PAGE_ALIGN(stack_top - stack_base);
L
Linus Torvalds 已提交
681

682 683 684
	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
685
#else
686 687
	stack_top = arch_align_stack(stack_top);
	stack_top = PAGE_ALIGN(stack_top);
688 689 690 691 692

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

693 694 695
	stack_shift = vma->vm_end - stack_top;

	bprm->p -= stack_shift;
L
Linus Torvalds 已提交
696 697 698 699
	mm->arg_start = bprm->p;
#endif

	if (bprm->loader)
700 701
		bprm->loader -= stack_shift;
	bprm->exec -= stack_shift;
L
Linus Torvalds 已提交
702 703

	down_write(&mm->mmap_sem);
704
	vm_flags = VM_STACK_FLAGS;
705 706 707 708 709 710 711 712 713 714 715

	/*
	 * 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;
716
	vm_flags |= VM_STACK_INCOMPLETE_SETUP;
717 718 719 720 721 722 723 724 725 726

	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);
727 728
		if (ret)
			goto out_unlock;
L
Linus Torvalds 已提交
729 730
	}

731 732 733
	/* mprotect_fixup is overkill to remove the temporary stack flags */
	vma->vm_flags &= ~VM_STACK_INCOMPLETE_SETUP;

734
	stack_expand = 131072UL; /* randomly 32*4k (or 2*64k) pages */
735 736 737 738 739 740
	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;
741
#ifdef CONFIG_STACK_GROWSUP
742 743 744 745
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_start + rlim_stack;
	else
		stack_base = vma->vm_end + stack_expand;
746
#else
747 748 749 750
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_end - rlim_stack;
	else
		stack_base = vma->vm_start - stack_expand;
751
#endif
752
	current->mm->start_stack = bprm->p;
753 754 755 756 757
	ret = expand_stack(vma, stack_base);
	if (ret)
		ret = -EFAULT;

out_unlock:
L
Linus Torvalds 已提交
758
	up_write(&mm->mmap_sem);
759
	return ret;
L
Linus Torvalds 已提交
760 761 762 763 764 765 766 767
}
EXPORT_SYMBOL(setup_arg_pages);

#endif /* CONFIG_MMU */

struct file *open_exec(const char *name)
{
	struct file *file;
768
	int err;
769 770 771 772 773
	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
774

775
	file = do_filp_open(AT_FDCWD, name, &open_exec_flags, LOOKUP_FOLLOW);
776
	if (IS_ERR(file))
777 778 779
		goto out;

	err = -EACCES;
780 781
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
		goto exit;
782

783 784
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
		goto exit;
785

786
	fsnotify_open(file);
787

788
	err = deny_write_access(file);
789 790
	if (err)
		goto exit;
L
Linus Torvalds 已提交
791

792
out:
793 794
	return file;

795 796
exit:
	fput(file);
797 798
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
799 800
EXPORT_SYMBOL(open_exec);

M
Mimi Zohar 已提交
801 802
int kernel_read(struct file *file, loff_t offset,
		char *addr, unsigned long count)
L
Linus Torvalds 已提交
803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
{
	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;
826
	sync_mm_rss(tsk, old_mm);
L
Linus Torvalds 已提交
827 828 829 830 831 832 833
	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
834
		 * checking core_state and changing tsk->mm.
L
Linus Torvalds 已提交
835 836
		 */
		down_read(&old_mm->mmap_sem);
837
		if (unlikely(old_mm->core_state)) {
L
Linus Torvalds 已提交
838 839 840 841 842 843 844 845 846
			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
847 848 849 850
	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
851 852 853 854
	task_unlock(tsk);
	arch_pick_mmap_layout(mm);
	if (old_mm) {
		up_read(&old_mm->mmap_sem);
855
		BUG_ON(active_mm != old_mm);
856
		mm_update_next_owner(old_mm);
L
Linus Torvalds 已提交
857 858 859 860 861 862 863 864 865 866 867 868 869
		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().)
 */
870
static int de_thread(struct task_struct *tsk)
L
Linus Torvalds 已提交
871 872
{
	struct signal_struct *sig = tsk->signal;
873
	struct sighand_struct *oldsighand = tsk->sighand;
L
Linus Torvalds 已提交
874 875
	spinlock_t *lock = &oldsighand->siglock;

876
	if (thread_group_empty(tsk))
L
Linus Torvalds 已提交
877 878 879 880 881 882
		goto no_thread_group;

	/*
	 * Kill all other threads in the thread group.
	 */
	spin_lock_irq(lock);
883
	if (signal_group_exit(sig)) {
L
Linus Torvalds 已提交
884 885 886 887 888 889 890
		/*
		 * Another group action in progress, just
		 * return so that the signal is processed.
		 */
		spin_unlock_irq(lock);
		return -EAGAIN;
	}
891

892
	sig->group_exit_task = tsk;
893 894 895
	sig->notify_count = zap_other_threads(tsk);
	if (!thread_group_leader(tsk))
		sig->notify_count--;
L
Linus Torvalds 已提交
896

897
	while (sig->notify_count) {
L
Linus Torvalds 已提交
898 899 900 901 902 903 904 905 906 907 908 909
		__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:
	 */
910
	if (!thread_group_leader(tsk)) {
911
		struct task_struct *leader = tsk->group_leader;
912

913
		sig->notify_count = -1;	/* for exit_notify() */
914 915 916 917 918 919 920 921
		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
922

923 924 925 926 927 928 929 930 931 932
		/*
		 * 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).
		 */
933
		tsk->start_time = leader->start_time;
934

935 936
		BUG_ON(!same_thread_group(leader, tsk));
		BUG_ON(has_group_leader_pid(tsk));
L
Linus Torvalds 已提交
937 938 939 940 941 942
		/*
		 * 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:
		 */
943 944

		/* Become a process group leader with the old leader's pid.
945 946
		 * The old leader becomes a thread of the this thread group.
		 * Note: The old leader also uses this pid until release_task
947 948
		 *       is called.  Odd but simple and correct.
		 */
949 950
		detach_pid(tsk, PIDTYPE_PID);
		tsk->pid = leader->pid;
951
		attach_pid(tsk, PIDTYPE_PID,  task_pid(leader));
952 953
		transfer_pid(leader, tsk, PIDTYPE_PGID);
		transfer_pid(leader, tsk, PIDTYPE_SID);
954

955
		list_replace_rcu(&leader->tasks, &tsk->tasks);
956
		list_replace_init(&leader->sibling, &tsk->sibling);
L
Linus Torvalds 已提交
957

958 959
		tsk->group_leader = tsk;
		leader->group_leader = tsk;
960

961
		tsk->exit_signal = SIGCHLD;
962 963 964

		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
		leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
965
		write_unlock_irq(&tasklist_lock);
966 967

		release_task(leader);
968
	}
L
Linus Torvalds 已提交
969

970 971
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
L
Linus Torvalds 已提交
972 973

no_thread_group:
J
Jiri Pirko 已提交
974 975 976
	if (current->mm)
		setmax_mm_hiwater_rss(&sig->maxrss, current->mm);

L
Linus Torvalds 已提交
977
	exit_itimers(sig);
978
	flush_itimer_signals();
979

980 981
	if (atomic_read(&oldsighand->count) != 1) {
		struct sighand_struct *newsighand;
L
Linus Torvalds 已提交
982
		/*
983 984
		 * This ->sighand is shared with the CLONE_SIGHAND
		 * but not CLONE_THREAD task, switch to the new one.
L
Linus Torvalds 已提交
985
		 */
986 987 988 989
		newsighand = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
		if (!newsighand)
			return -ENOMEM;

L
Linus Torvalds 已提交
990 991 992 993 994 995
		atomic_set(&newsighand->count, 1);
		memcpy(newsighand->action, oldsighand->action,
		       sizeof(newsighand->action));

		write_lock_irq(&tasklist_lock);
		spin_lock(&oldsighand->siglock);
996
		rcu_assign_pointer(tsk->sighand, newsighand);
L
Linus Torvalds 已提交
997 998 999
		spin_unlock(&oldsighand->siglock);
		write_unlock_irq(&tasklist_lock);

1000
		__cleanup_sighand(oldsighand);
L
Linus Torvalds 已提交
1001 1002
	}

1003
	BUG_ON(!thread_group_leader(tsk));
L
Linus Torvalds 已提交
1004 1005
	return 0;
}
O
Oleg Nesterov 已提交
1006

L
Linus Torvalds 已提交
1007 1008 1009 1010
/*
 * These functions flushes out all traces of the currently running executable
 * so that a new one can be started
 */
1011
static void flush_old_files(struct files_struct * files)
L
Linus Torvalds 已提交
1012 1013
{
	long j = -1;
1014
	struct fdtable *fdt;
L
Linus Torvalds 已提交
1015 1016 1017 1018 1019 1020 1021

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

		j++;
		i = j * __NFDBITS;
1022
		fdt = files_fdtable(files);
1023
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
1024
			break;
1025
		set = fdt->close_on_exec->fds_bits[j];
L
Linus Torvalds 已提交
1026 1027
		if (!set)
			continue;
1028
		fdt->close_on_exec->fds_bits[j] = 0;
L
Linus Torvalds 已提交
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
		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);
}

1041
char *get_task_comm(char *buf, struct task_struct *tsk)
L
Linus Torvalds 已提交
1042 1043 1044 1045 1046
{
	/* buf must be at least sizeof(tsk->comm) in size */
	task_lock(tsk);
	strncpy(buf, tsk->comm, sizeof(tsk->comm));
	task_unlock(tsk);
1047
	return buf;
L
Linus Torvalds 已提交
1048 1049 1050 1051 1052
}

void set_task_comm(struct task_struct *tsk, char *buf)
{
	task_lock(tsk);
1053 1054 1055 1056 1057 1058 1059 1060 1061

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

int flush_old_exec(struct linux_binprm * bprm)
{
1069
	int retval;
L
Linus Torvalds 已提交
1070 1071 1072 1073 1074 1075 1076 1077 1078

	/*
	 * 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
1079 1080
	set_mm_exe_file(bprm->mm, bprm->file);

L
Linus Torvalds 已提交
1081 1082 1083
	/*
	 * Release all of the old mmap stuff
	 */
1084
	acct_arg_size(bprm, 0);
L
Linus Torvalds 已提交
1085 1086
	retval = exec_mmap(bprm->mm);
	if (retval)
1087
		goto out;
L
Linus Torvalds 已提交
1088 1089

	bprm->mm = NULL;		/* We're using it now */
1090

1091
	current->flags &= ~(PF_RANDOMIZE | PF_KTHREAD);
1092 1093 1094
	flush_thread();
	current->personality &= ~bprm->per_clear;

1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
	return 0;

out:
	return retval;
}
EXPORT_SYMBOL(flush_old_exec);

void setup_new_exec(struct linux_binprm * bprm)
{
	int i, ch;
1105
	const char *name;
1106 1107 1108
	char tcomm[sizeof(current->comm)];

	arch_pick_mmap_layout(current->mm);
L
Linus Torvalds 已提交
1109 1110 1111 1112

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

1113
	if (current_euid() == current_uid() && current_egid() == current_gid())
1114
		set_dumpable(current->mm, 1);
A
Alan Cox 已提交
1115
	else
1116
		set_dumpable(current->mm, suid_dumpable);
A
Alan Cox 已提交
1117

L
Linus Torvalds 已提交
1118
	name = bprm->filename;
1119 1120

	/* Copies the binary name from after last slash */
L
Linus Torvalds 已提交
1121 1122
	for (i=0; (ch = *(name++)) != '\0';) {
		if (ch == '/')
1123
			i = 0; /* overwrite what we wrote */
L
Linus Torvalds 已提交
1124 1125 1126 1127 1128 1129 1130
		else
			if (i < (sizeof(tcomm) - 1))
				tcomm[i++] = ch;
	}
	tcomm[i] = '\0';
	set_task_comm(current, tcomm);

1131 1132 1133 1134 1135 1136
	/* 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;

1137 1138 1139
	/* install the new credentials */
	if (bprm->cred->uid != current_euid() ||
	    bprm->cred->gid != current_egid()) {
1140 1141
		current->pdeath_signal = 0;
	} else if (file_permission(bprm->file, MAY_READ) ||
1142
		   bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP) {
1143
		set_dumpable(current->mm, suid_dumpable);
L
Linus Torvalds 已提交
1144 1145
	}

I
Ingo Molnar 已提交
1146 1147 1148 1149 1150
	/*
	 * Flush performance counters when crossing a
	 * security domain:
	 */
	if (!get_dumpable(current->mm))
1151
		perf_event_exit_task(current);
I
Ingo Molnar 已提交
1152

L
Linus Torvalds 已提交
1153 1154 1155 1156 1157 1158 1159 1160
	/* 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);
}
1161
EXPORT_SYMBOL(setup_new_exec);
L
Linus Torvalds 已提交
1162

1163 1164 1165 1166 1167 1168 1169 1170
/*
 * 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)
{
1171
	if (mutex_lock_interruptible(&current->signal->cred_guard_mutex))
1172 1173 1174 1175 1176 1177
		return -ERESTARTNOINTR;

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

1178
	mutex_unlock(&current->signal->cred_guard_mutex);
1179 1180 1181 1182 1183 1184 1185
	return -ENOMEM;
}

void free_bprm(struct linux_binprm *bprm)
{
	free_arg_pages(bprm);
	if (bprm->cred) {
1186
		mutex_unlock(&current->signal->cred_guard_mutex);
1187 1188 1189 1190 1191
		abort_creds(bprm->cred);
	}
	kfree(bprm);
}

1192 1193 1194 1195 1196 1197 1198 1199 1200
/*
 * 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;
1201 1202
	/*
	 * cred_guard_mutex must be held at least to this point to prevent
1203
	 * ptrace_attach() from altering our determination of the task's
1204 1205
	 * credentials; any time after this it may be unlocked.
	 */
1206
	security_bprm_committed_creds(bprm);
1207
	mutex_unlock(&current->signal->cred_guard_mutex);
1208 1209 1210 1211 1212
}
EXPORT_SYMBOL(install_exec_creds);

/*
 * determine how safe it is to execute the proposed program
1213
 * - the caller must hold ->cred_guard_mutex to protect against
1214 1215
 *   PTRACE_ATTACH
 */
A
Al Viro 已提交
1216
int check_unsafe_exec(struct linux_binprm *bprm)
1217
{
D
David Howells 已提交
1218
	struct task_struct *p = current, *t;
1219
	unsigned n_fs;
A
Al Viro 已提交
1220
	int res = 0;
1221 1222 1223

	bprm->unsafe = tracehook_unsafe_exec(p);

D
David Howells 已提交
1224
	n_fs = 1;
N
Nick Piggin 已提交
1225
	spin_lock(&p->fs->lock);
1226
	rcu_read_lock();
D
David Howells 已提交
1227 1228 1229 1230
	for (t = next_thread(p); t != p; t = next_thread(t)) {
		if (t->fs == p->fs)
			n_fs++;
	}
1231
	rcu_read_unlock();
D
David Howells 已提交
1232

1233
	if (p->fs->users > n_fs) {
1234
		bprm->unsafe |= LSM_UNSAFE_SHARE;
A
Al Viro 已提交
1235
	} else {
1236 1237 1238 1239 1240
		res = -EAGAIN;
		if (!p->fs->in_exec) {
			p->fs->in_exec = 1;
			res = 1;
		}
A
Al Viro 已提交
1241
	}
N
Nick Piggin 已提交
1242
	spin_unlock(&p->fs->lock);
A
Al Viro 已提交
1243 1244

	return res;
1245 1246
}

L
Linus Torvalds 已提交
1247 1248 1249
/* 
 * Fill the binprm structure from the inode. 
 * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes
1250 1251
 *
 * This may be called multiple times for binary chains (scripts for example).
L
Linus Torvalds 已提交
1252 1253 1254
 */
int prepare_binprm(struct linux_binprm *bprm)
{
1255
	umode_t mode;
1256
	struct inode * inode = bprm->file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262
	int retval;

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

1263 1264 1265
	/* 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
1266

1267
	if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) {
L
Linus Torvalds 已提交
1268 1269
		/* Set-uid? */
		if (mode & S_ISUID) {
1270 1271
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->euid = inode->i_uid;
L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280
		}

		/* 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)) {
1281 1282
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->egid = inode->i_gid;
L
Linus Torvalds 已提交
1283 1284 1285 1286
		}
	}

	/* fill in binprm security blob */
1287
	retval = security_bprm_set_creds(bprm);
L
Linus Torvalds 已提交
1288 1289
	if (retval)
		return retval;
1290
	bprm->cred_prepared = 1;
L
Linus Torvalds 已提交
1291

1292 1293
	memset(bprm->buf, 0, BINPRM_BUF_SIZE);
	return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE);
L
Linus Torvalds 已提交
1294 1295 1296 1297
}

EXPORT_SYMBOL(prepare_binprm);

N
Nick Piggin 已提交
1298 1299 1300 1301 1302
/*
 * 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.
 */
1303
int remove_arg_zero(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1304
{
1305 1306 1307 1308
	int ret = 0;
	unsigned long offset;
	char *kaddr;
	struct page *page;
N
Nick Piggin 已提交
1309

1310 1311
	if (!bprm->argc)
		return 0;
L
Linus Torvalds 已提交
1312

1313 1314 1315 1316 1317 1318 1319 1320
	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
1321

1322 1323 1324
		for (; offset < PAGE_SIZE && kaddr[offset];
				offset++, bprm->p++)
			;
N
Nick Piggin 已提交
1325

1326 1327
		kunmap_atomic(kaddr, KM_USER0);
		put_arg_page(page);
N
Nick Piggin 已提交
1328

1329 1330 1331
		if (offset == PAGE_SIZE)
			free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1);
	} while (offset == PAGE_SIZE);
N
Nick Piggin 已提交
1332

1333 1334 1335
	bprm->p++;
	bprm->argc--;
	ret = 0;
N
Nick Piggin 已提交
1336

1337 1338
out:
	return ret;
L
Linus Torvalds 已提交
1339 1340 1341 1342 1343 1344 1345 1346
}
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)
{
1347
	unsigned int depth = bprm->recursion_depth;
L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
	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
1358 1359 1360 1361 1362

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

L
Linus Torvalds 已提交
1363 1364 1365
	retval = -ENOENT;
	for (try=0; try<2; try++) {
		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
1366
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
1367 1368 1369 1370 1371 1372 1373
			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);
1374 1375 1376 1377 1378 1379
			/*
			 * 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
1380
			if (retval >= 0) {
1381 1382
				if (depth == 0)
					tracehook_report_exec(fmt, bprm, regs);
L
Linus Torvalds 已提交
1383 1384 1385 1386 1387 1388
				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
1389
				proc_exec_connector(current);
L
Linus Torvalds 已提交
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403
				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
1404 1405
#ifdef CONFIG_MODULES
		} else {
L
Linus Torvalds 已提交
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
#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.
 */
1424 1425 1426 1427
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
1428 1429 1430
{
	struct linux_binprm *bprm;
	struct file *file;
1431
	struct files_struct *displaced;
1432
	bool clear_in_exec;
L
Linus Torvalds 已提交
1433 1434
	int retval;

1435
	retval = unshare_files(&displaced);
1436 1437 1438
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1439
	retval = -ENOMEM;
1440
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1441
	if (!bprm)
1442
		goto out_files;
L
Linus Torvalds 已提交
1443

1444 1445
	retval = prepare_bprm_creds(bprm);
	if (retval)
1446
		goto out_free;
A
Al Viro 已提交
1447 1448

	retval = check_unsafe_exec(bprm);
1449
	if (retval < 0)
1450
		goto out_free;
1451
	clear_in_exec = retval;
1452
	current->in_execve = 1;
1453

L
Linus Torvalds 已提交
1454 1455 1456
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1457
		goto out_unmark;
L
Linus Torvalds 已提交
1458 1459 1460 1461 1462 1463 1464

	sched_exec();

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

1465 1466 1467
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1468

1469
	bprm->argc = count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1470
	if ((retval = bprm->argc) < 0)
1471
		goto out;
L
Linus Torvalds 已提交
1472

1473
	bprm->envc = count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494
	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);
1495 1496
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1497

1498
	/* execve succeeded */
A
Al Viro 已提交
1499
	current->fs->in_exec = 0;
1500
	current->in_execve = 0;
1501 1502 1503 1504 1505
	acct_update_integrals(current);
	free_bprm(bprm);
	if (displaced)
		put_files_struct(displaced);
	return retval;
L
Linus Torvalds 已提交
1506

1507
out:
1508 1509 1510 1511
	if (bprm->mm) {
		acct_arg_size(bprm, 0);
		mmput(bprm->mm);
	}
L
Linus Torvalds 已提交
1512 1513 1514 1515 1516 1517

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}
1518

A
Al Viro 已提交
1519
out_unmark:
1520 1521
	if (clear_in_exec)
		current->fs->in_exec = 0;
1522
	current->in_execve = 0;
1523 1524

out_free:
1525
	free_bprm(bprm);
L
Linus Torvalds 已提交
1526

1527
out_files:
1528 1529
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1530 1531 1532 1533
out_ret:
	return retval;
}

1534 1535 1536 1537 1538
int do_execve(const char *filename,
	const char __user *const __user *__argv,
	const char __user *const __user *__envp,
	struct pt_regs *regs)
{
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
	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,
	};
1558 1559
	return do_execve_common(filename, argv, envp, regs);
}
1560
#endif
1561

1562
void set_binfmt(struct linux_binfmt *new)
L
Linus Torvalds 已提交
1563
{
1564 1565 1566 1567
	struct mm_struct *mm = current->mm;

	if (mm->binfmt)
		module_put(mm->binfmt->module);
L
Linus Torvalds 已提交
1568

1569
	mm->binfmt = new;
1570 1571
	if (new)
		__module_get(new->module);
L
Linus Torvalds 已提交
1572 1573 1574 1575
}

EXPORT_SYMBOL(set_binfmt);

1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
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
1621 1622 1623 1624
/* 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.
 */
1625
static int format_corename(struct core_name *cn, long signr)
L
Linus Torvalds 已提交
1626
{
1627
	const struct cred *cred = current_cred();
1628 1629
	const char *pat_ptr = core_pattern;
	int ispipe = (*pat_ptr == '|');
L
Linus Torvalds 已提交
1630
	int pid_in_pattern = 0;
1631 1632 1633 1634 1635 1636 1637 1638
	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
1639 1640 1641 1642 1643

	/* Repeat as long as we have more pattern to process and more output
	   space */
	while (*pat_ptr) {
		if (*pat_ptr != '%') {
1644
			if (*pat_ptr == 0)
L
Linus Torvalds 已提交
1645
				goto out;
1646
			err = cn_printf(cn, "%c", *pat_ptr++);
L
Linus Torvalds 已提交
1647 1648
		} else {
			switch (*++pat_ptr) {
1649
			/* single % at the end, drop that */
L
Linus Torvalds 已提交
1650 1651 1652 1653
			case 0:
				goto out;
			/* Double percent, output one percent */
			case '%':
1654
				err = cn_printf(cn, "%c", '%');
L
Linus Torvalds 已提交
1655 1656 1657 1658
				break;
			/* pid */
			case 'p':
				pid_in_pattern = 1;
1659 1660
				err = cn_printf(cn, "%d",
					      task_tgid_vnr(current));
L
Linus Torvalds 已提交
1661 1662 1663
				break;
			/* uid */
			case 'u':
1664
				err = cn_printf(cn, "%d", cred->uid);
L
Linus Torvalds 已提交
1665 1666 1667
				break;
			/* gid */
			case 'g':
1668
				err = cn_printf(cn, "%d", cred->gid);
L
Linus Torvalds 已提交
1669 1670 1671
				break;
			/* signal that caused the coredump */
			case 's':
1672
				err = cn_printf(cn, "%ld", signr);
L
Linus Torvalds 已提交
1673 1674 1675 1676 1677
				break;
			/* UNIX time of coredump */
			case 't': {
				struct timeval tv;
				do_gettimeofday(&tv);
1678
				err = cn_printf(cn, "%lu", tv.tv_sec);
L
Linus Torvalds 已提交
1679 1680 1681 1682 1683
				break;
			}
			/* hostname */
			case 'h':
				down_read(&uts_sem);
1684 1685
				err = cn_printf(cn, "%s",
					      utsname()->nodename);
L
Linus Torvalds 已提交
1686 1687 1688 1689
				up_read(&uts_sem);
				break;
			/* executable */
			case 'e':
1690
				err = cn_printf(cn, "%s", current->comm);
L
Linus Torvalds 已提交
1691
				break;
1692 1693
			/* core limit size */
			case 'c':
1694 1695
				err = cn_printf(cn, "%lu",
					      rlimit(RLIMIT_CORE));
1696
				break;
L
Linus Torvalds 已提交
1697 1698 1699 1700 1701
			default:
				break;
			}
			++pat_ptr;
		}
1702 1703 1704

		if (err)
			return err;
L
Linus Torvalds 已提交
1705
	}
1706

L
Linus Torvalds 已提交
1707 1708 1709 1710
	/* 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
1711
	 * the filename. Do not do this for piped commands. */
1712
	if (!ispipe && !pid_in_pattern && core_uses_pid) {
1713 1714 1715
		err = cn_printf(cn, ".%d", task_tgid_vnr(current));
		if (err)
			return err;
L
Linus Torvalds 已提交
1716
	}
1717 1718
out:
	return ispipe;
L
Linus Torvalds 已提交
1719 1720
}

1721
static int zap_process(struct task_struct *start, int exit_code)
1722 1723
{
	struct task_struct *t;
1724
	int nr = 0;
1725

1726
	start->signal->flags = SIGNAL_GROUP_EXIT;
1727
	start->signal->group_exit_code = exit_code;
1728
	start->signal->group_stop_count = 0;
1729 1730 1731 1732

	t = start;
	do {
		if (t != current && t->mm) {
1733 1734
			sigaddset(&t->pending.signal, SIGKILL);
			signal_wake_up(t, 1);
1735
			nr++;
1736
		}
1737
	} while_each_thread(start, t);
1738 1739

	return nr;
1740 1741
}

1742
static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm,
1743
				struct core_state *core_state, int exit_code)
L
Linus Torvalds 已提交
1744 1745
{
	struct task_struct *g, *p;
1746
	unsigned long flags;
1747
	int nr = -EAGAIN;
1748 1749

	spin_lock_irq(&tsk->sighand->siglock);
1750
	if (!signal_group_exit(tsk->signal)) {
1751
		mm->core_state = core_state;
1752
		nr = zap_process(tsk, exit_code);
L
Linus Torvalds 已提交
1753
	}
1754
	spin_unlock_irq(&tsk->sighand->siglock);
1755 1756
	if (unlikely(nr < 0))
		return nr;
L
Linus Torvalds 已提交
1757

1758
	if (atomic_read(&mm->mm_users) == nr + 1)
1759
		goto done;
1760 1761
	/*
	 * We should find and kill all tasks which use this mm, and we should
1762
	 * count them correctly into ->nr_threads. We don't take tasklist
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789
	 * 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().
	 */
1790
	rcu_read_lock();
1791
	for_each_process(g) {
1792 1793
		if (g == tsk->group_leader)
			continue;
1794 1795
		if (g->flags & PF_KTHREAD)
			continue;
1796 1797 1798
		p = g;
		do {
			if (p->mm) {
1799
				if (unlikely(p->mm == mm)) {
1800
					lock_task_sighand(p, &flags);
1801
					nr += zap_process(p, exit_code);
1802 1803
					unlock_task_sighand(p, &flags);
				}
1804 1805
				break;
			}
1806
		} while_each_thread(g, p);
1807
	}
1808
	rcu_read_unlock();
1809
done:
1810
	atomic_set(&core_state->nr_threads, nr);
1811
	return nr;
L
Linus Torvalds 已提交
1812 1813
}

1814
static int coredump_wait(int exit_code, struct core_state *core_state)
L
Linus Torvalds 已提交
1815
{
1816 1817 1818
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;
	struct completion *vfork_done;
1819
	int core_waiters = -EBUSY;
L
Linus Torvalds 已提交
1820

1821
	init_completion(&core_state->startup);
1822 1823
	core_state->dumper.task = tsk;
	core_state->dumper.next = NULL;
1824 1825 1826 1827

	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
1828 1829
	up_write(&mm->mmap_sem);

1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842
	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
1843
	if (core_waiters)
1844
		wait_for_completion(&core_state->startup);
1845 1846
fail:
	return core_waiters;
L
Linus Torvalds 已提交
1847 1848
}

O
Oleg Nesterov 已提交
1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869
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;
}

1870 1871 1872 1873 1874 1875 1876 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
/*
 * 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;
	}
}

1911
static int __get_dumpable(unsigned long mm_flags)
1912 1913 1914
{
	int ret;

1915
	ret = mm_flags & MMF_DUMPABLE_MASK;
1916 1917 1918
	return (ret >= 2) ? 2 : ret;
}

1919 1920 1921 1922 1923
int get_dumpable(struct mm_struct *mm)
{
	return __get_dumpable(mm->flags);
}

1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946
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);

}


1947
/*
H
Holger Hans Peter Freyther 已提交
1948
 * umh_pipe_setup
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
 * 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
1991
void do_coredump(long signr, int exit_code, struct pt_regs *regs)
L
Linus Torvalds 已提交
1992
{
1993
	struct core_state core_state;
1994
	struct core_name cn;
L
Linus Torvalds 已提交
1995 1996
	struct mm_struct *mm = current->mm;
	struct linux_binfmt * binfmt;
D
David Howells 已提交
1997 1998
	const struct cred *old_cred;
	struct cred *cred;
L
Linus Torvalds 已提交
1999
	int retval = 0;
A
Alan Cox 已提交
2000
	int flag = 0;
O
Oleg Nesterov 已提交
2001
	int ispipe;
2002
	static atomic_t core_dump_count = ATOMIC_INIT(0);
2003 2004 2005
	struct coredump_params cprm = {
		.signr = signr,
		.regs = regs,
J
Jiri Slaby 已提交
2006
		.limit = rlimit(RLIMIT_CORE),
2007 2008 2009 2010 2011 2012
		/*
		 * 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,
2013
	};
L
Linus Torvalds 已提交
2014

S
Steve Grubb 已提交
2015 2016
	audit_core_dumps(signr);

2017
	binfmt = mm->binfmt;
L
Linus Torvalds 已提交
2018 2019
	if (!binfmt || !binfmt->core_dump)
		goto fail;
2020 2021
	if (!__get_dumpable(cprm.mm_flags))
		goto fail;
D
David Howells 已提交
2022 2023

	cred = prepare_creds();
2024
	if (!cred)
D
David Howells 已提交
2025
		goto fail;
A
Alan Cox 已提交
2026 2027 2028 2029 2030
	/*
	 *	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.
	 */
2031 2032
	if (__get_dumpable(cprm.mm_flags) == 2) {
		/* Setuid core dump mode */
A
Alan Cox 已提交
2033
		flag = O_EXCL;		/* Stop rewrite attacks */
D
David Howells 已提交
2034
		cred->fsuid = 0;	/* Dump root private */
A
Alan Cox 已提交
2035
	}
2036

2037
	retval = coredump_wait(exit_code, &core_state);
2038 2039
	if (retval < 0)
		goto fail_creds;
D
David Howells 已提交
2040 2041

	old_cred = override_creds(cred);
L
Linus Torvalds 已提交
2042 2043 2044 2045 2046 2047 2048

	/*
	 * 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);

2049 2050 2051 2052 2053 2054 2055
	ispipe = format_corename(&cn, signr);

	if (ispipe == -ENOMEM) {
		printk(KERN_WARNING "format_corename failed\n");
		printk(KERN_WARNING "Aborting core\n");
		goto fail_corename;
	}
2056

2057
 	if (ispipe) {
O
Oleg Nesterov 已提交
2058 2059 2060
		int dump_count;
		char **helper_argv;

2061
		if (cprm.limit == 1) {
2062 2063 2064
			/*
			 * Normally core limits are irrelevant to pipes, since
			 * we're not writing to the file system, but we use
2065 2066
			 * cprm.limit of 1 here as a speacial value. Any
			 * non-1 limit gets set to RLIM_INFINITY below, but
2067 2068
			 * a limit of 0 skips the dump.  This is a consistent
			 * way to catch recursive crashes.  We can still crash
2069
			 * if the core_pattern binary sets RLIM_CORE =  !1
2070 2071 2072 2073 2074 2075 2076
			 * 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
2077
				"Process %d(%s) has RLIMIT_CORE set to 1\n",
2078 2079 2080 2081
				task_tgid_vnr(current), current->comm);
			printk(KERN_WARNING "Aborting core\n");
			goto fail_unlock;
		}
O
Oleg Nesterov 已提交
2082
		cprm.limit = RLIM_INFINITY;
2083

2084 2085 2086 2087 2088 2089 2090 2091
		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;
		}

2092
		helper_argv = argv_split(GFP_KERNEL, cn.corename+1, NULL);
2093 2094 2095
		if (!helper_argv) {
			printk(KERN_WARNING "%s failed to allocate memory\n",
			       __func__);
2096
			goto fail_dropcount;
2097
		}
2098

O
Oleg Nesterov 已提交
2099 2100 2101 2102 2103
		retval = call_usermodehelper_fns(helper_argv[0], helper_argv,
					NULL, UMH_WAIT_EXEC, umh_pipe_setup,
					NULL, &cprm);
		argv_free(helper_argv);
		if (retval) {
2104
 			printk(KERN_INFO "Core dump to %s pipe failed\n",
2105
			       cn.corename);
O
Oleg Nesterov 已提交
2106
			goto close_fail;
2107
 		}
2108 2109 2110 2111 2112 2113
	} else {
		struct inode *inode;

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

2114
		cprm.file = filp_open(cn.corename,
2115 2116
				 O_CREAT | 2 | O_NOFOLLOW | O_LARGEFILE | flag,
				 0600);
2117 2118
		if (IS_ERR(cprm.file))
			goto fail_unlock;
L
Linus Torvalds 已提交
2119

2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
		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
2142

2143
	retval = binfmt->core_dump(&cprm);
L
Linus Torvalds 已提交
2144 2145
	if (retval)
		current->signal->group_exit_code |= 0x80;
O
Oleg Nesterov 已提交
2146

2147
	if (ispipe && core_pipe_limit)
2148
		wait_for_dump_helpers(cprm.file);
O
Oleg Nesterov 已提交
2149 2150 2151
close_fail:
	if (cprm.file)
		filp_close(cprm.file, NULL);
2152
fail_dropcount:
O
Oleg Nesterov 已提交
2153
	if (ispipe)
2154
		atomic_dec(&core_dump_count);
L
Linus Torvalds 已提交
2155
fail_unlock:
2156 2157
	kfree(cn.corename);
fail_corename:
2158
	coredump_finish(mm);
D
David Howells 已提交
2159
	revert_creds(old_cred);
2160
fail_creds:
D
David Howells 已提交
2161
	put_cred(cred);
L
Linus Torvalds 已提交
2162
fail:
W
WANG Cong 已提交
2163
	return;
L
Linus Torvalds 已提交
2164
}
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174

/*
 * 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;
}
2175
EXPORT_SYMBOL(dump_write);
2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203

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;
}
2204
EXPORT_SYMBOL(dump_seek);