exec.c 39.3 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"
65
#include "coredump.h"
L
Linus Torvalds 已提交
66

67 68
#include <trace/events/sched.h>

A
Alan Cox 已提交
69 70
int suid_dumpable = 0;

A
Alexey Dobriyan 已提交
71
static LIST_HEAD(formats);
L
Linus Torvalds 已提交
72 73
static DEFINE_RWLOCK(binfmt_lock);

A
Al Viro 已提交
74
void __register_binfmt(struct linux_binfmt * fmt, int insert)
L
Linus Torvalds 已提交
75
{
A
Al Viro 已提交
76
	BUG_ON(!fmt);
L
Linus Torvalds 已提交
77
	write_lock(&binfmt_lock);
I
Ivan Kokshaysky 已提交
78 79
	insert ? list_add(&fmt->lh, &formats) :
		 list_add_tail(&fmt->lh, &formats);
L
Linus Torvalds 已提交
80 81 82
	write_unlock(&binfmt_lock);
}

I
Ivan Kokshaysky 已提交
83
EXPORT_SYMBOL(__register_binfmt);
L
Linus Torvalds 已提交
84

85
void unregister_binfmt(struct linux_binfmt * fmt)
L
Linus Torvalds 已提交
86 87
{
	write_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
88
	list_del(&fmt->lh);
L
Linus Torvalds 已提交
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
	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.
 */
105
SYSCALL_DEFINE1(uselib, const char __user *, library)
L
Linus Torvalds 已提交
106
{
107
	struct file *file;
108
	struct filename *tmp = getname(library);
109
	int error = PTR_ERR(tmp);
110 111 112 113 114
	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
	};
115

116 117 118
	if (IS_ERR(tmp))
		goto out;

119
	file = do_filp_open(AT_FDCWD, tmp, &uselib_flags, LOOKUP_FOLLOW);
120 121 122
	putname(tmp);
	error = PTR_ERR(file);
	if (IS_ERR(file))
L
Linus Torvalds 已提交
123 124 125
		goto out;

	error = -EINVAL;
126
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
L
Linus Torvalds 已提交
127 128
		goto exit;

129
	error = -EACCES;
130
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
L
Linus Torvalds 已提交
131 132
		goto exit;

133
	fsnotify_open(file);
134

L
Linus Torvalds 已提交
135 136 137 138 139
	error = -ENOEXEC;
	if(file->f_op) {
		struct linux_binfmt * fmt;

		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
140
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
141 142 143 144 145 146 147 148 149 150 151 152 153
			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);
	}
154
exit:
L
Linus Torvalds 已提交
155 156 157 158 159
	fput(file);
out:
  	return error;
}

160
#ifdef CONFIG_MMU
O
Oleg Nesterov 已提交
161 162 163 164 165 166
/*
 * 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).
 */
167
static void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
168 169 170 171 172 173 174 175 176 177 178
{
	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);
}

179
static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
180 181 182 183 184 185 186
		int write)
{
	struct page *page;
	int ret;

#ifdef CONFIG_STACK_GROWSUP
	if (write) {
187
		ret = expand_downwards(bprm->vma, pos);
188 189 190 191 192 193 194 195 196 197 198
		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;
199 200
		struct rlimit *rlim;

201 202
		acct_arg_size(bprm, size / PAGE_SIZE);

203 204 205 206 207 208
		/*
		 * We've historically supported up to 32 pages (ARG_MAX)
		 * of argument strings even with small stacks
		 */
		if (size <= ARG_MAX)
			return page;
209 210 211 212 213 214 215 216

		/*
		 * 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.
		 */
217
		rlim = current->signal->rlim;
J
Jiri Slaby 已提交
218
		if (size > ACCESS_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4) {
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
			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)
{
248
	int err;
249 250 251 252 253
	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)
254
		return -ENOMEM;
255 256 257 258 259 260 261 262 263 264

	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.
	 */
265
	BUILD_BUG_ON(VM_STACK_FLAGS & VM_STACK_INCOMPLETE_SETUP);
266 267
	vma->vm_end = STACK_TOP_MAX;
	vma->vm_start = vma->vm_end - PAGE_SIZE;
268
	vma->vm_flags = VM_STACK_FLAGS | VM_STACK_INCOMPLETE_SETUP;
269
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
270
	INIT_LIST_HEAD(&vma->anon_vma_chain);
271

272
	err = insert_vm_struct(mm, vma);
273
	if (err)
274 275 276 277 278 279 280
		goto err;

	mm->stack_vm = mm->total_vm = 1;
	up_write(&mm->mmap_sem);
	bprm->p = vma->vm_end - sizeof(void *);
	return 0;
err:
281 282 283
	up_write(&mm->mmap_sem);
	bprm->vma = NULL;
	kmem_cache_free(vm_area_cachep, vma);
284 285 286 287 288 289 290 291 292 293
	return err;
}

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

#else

294
static inline void acct_arg_size(struct linux_binprm *bprm, unsigned long pages)
295 296 297
{
}

298
static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
		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;
}

387
struct user_arg_ptr {
388 389 390 391 392 393
#ifdef CONFIG_COMPAT
	bool is_compat;
#endif
	union {
		const char __user *const __user *native;
#ifdef CONFIG_COMPAT
A
Al Viro 已提交
394
		const compat_uptr_t __user *compat;
395 396
#endif
	} ptr;
397 398 399
};

static const char __user *get_user_arg_ptr(struct user_arg_ptr argv, int nr)
400
{
401 402 403 404 405 406 407 408
	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);
409

410 411 412 413 414
		return compat_ptr(compat);
	}
#endif

	if (get_user(native, argv.ptr.native + nr))
415 416
		return ERR_PTR(-EFAULT);

417
	return native;
418 419
}

L
Linus Torvalds 已提交
420 421 422
/*
 * count() counts the number of strings in array ARGV.
 */
423
static int count(struct user_arg_ptr argv, int max)
L
Linus Torvalds 已提交
424 425 426
{
	int i = 0;

427
	if (argv.ptr.native != NULL) {
L
Linus Torvalds 已提交
428
		for (;;) {
429
			const char __user *p = get_user_arg_ptr(argv, i);
L
Linus Torvalds 已提交
430 431 432

			if (!p)
				break;
433 434 435 436

			if (IS_ERR(p))
				return -EFAULT;

437
			if (i++ >= max)
L
Linus Torvalds 已提交
438
				return -E2BIG;
439 440 441

			if (fatal_signal_pending(current))
				return -ERESTARTNOHAND;
L
Linus Torvalds 已提交
442 443 444 445 446 447 448
			cond_resched();
		}
	}
	return i;
}

/*
449 450 451
 * '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
452
 */
453
static int copy_strings(int argc, struct user_arg_ptr argv,
A
Adrian Bunk 已提交
454
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
455 456 457
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
458
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
459 460 461
	int ret;

	while (argc-- > 0) {
462
		const char __user *str;
L
Linus Torvalds 已提交
463 464 465
		int len;
		unsigned long pos;

466 467 468
		ret = -EFAULT;
		str = get_user_arg_ptr(argv, argc);
		if (IS_ERR(str))
L
Linus Torvalds 已提交
469 470
			goto out;

471 472 473 474 475 476
		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
477 478
			goto out;

479
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
480
		pos = bprm->p;
481 482
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
483 484 485 486

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

487 488 489 490
			if (fatal_signal_pending(current)) {
				ret = -ERESTARTNOHAND;
				goto out;
			}
491 492
			cond_resched();

L
Linus Torvalds 已提交
493
			offset = pos % PAGE_SIZE;
494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
			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
510
				if (!page) {
511
					ret = -E2BIG;
L
Linus Torvalds 已提交
512 513 514
					goto out;
				}

515 516
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
517
					kunmap(kmapped_page);
518 519
					put_arg_page(kmapped_page);
				}
L
Linus Torvalds 已提交
520 521
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
522 523
				kpos = pos & PAGE_MASK;
				flush_arg_page(bprm, kpos, kmapped_page);
L
Linus Torvalds 已提交
524
			}
525
			if (copy_from_user(kaddr+offset, str, bytes_to_copy)) {
L
Linus Torvalds 已提交
526 527 528 529 530 531 532
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
533 534
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
535
		kunmap(kmapped_page);
536 537
		put_arg_page(kmapped_page);
	}
L
Linus Torvalds 已提交
538 539 540 541 542 543
	return ret;
}

/*
 * Like copy_strings, but get argv and its values from kernel memory.
 */
544
int copy_strings_kernel(int argc, const char *const *__argv,
545
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
546 547 548
{
	int r;
	mm_segment_t oldfs = get_fs();
549
	struct user_arg_ptr argv = {
550
		.ptr.native = (const char __user *const  __user *)__argv,
551 552
	};

L
Linus Torvalds 已提交
553
	set_fs(KERNEL_DS);
554
	r = copy_strings(argc, argv, bprm);
L
Linus Torvalds 已提交
555
	set_fs(oldfs);
556

L
Linus Torvalds 已提交
557 558 559 560 561
	return r;
}
EXPORT_SYMBOL(copy_strings_kernel);

#ifdef CONFIG_MMU
562

L
Linus Torvalds 已提交
563
/*
564 565 566
 * 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
567
 *
568 569 570 571 572 573
 * 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
574
 */
575
static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift)
L
Linus Torvalds 已提交
576 577
{
	struct mm_struct *mm = vma->vm_mm;
578 579 580 581 582
	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
583
	struct mmu_gather tlb;
L
Linus Torvalds 已提交
584

585
	BUG_ON(new_start > new_end);
L
Linus Torvalds 已提交
586

587 588 589 590 591 592 593 594 595 596
	/*
	 * 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)
	 */
597 598
	if (vma_adjust(vma, new_start, old_end, vma->vm_pgoff, NULL))
		return -ENOMEM;
599 600 601 602 603 604

	/*
	 * 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,
605
				       vma, new_start, length, false))
606 607 608
		return -ENOMEM;

	lru_add_drain();
P
Peter Zijlstra 已提交
609
	tlb_gather_mmu(&tlb, mm, 0);
610 611 612 613
	if (new_end > old_start) {
		/*
		 * when the old and new regions overlap clear from new_end.
		 */
P
Peter Zijlstra 已提交
614
		free_pgd_range(&tlb, new_end, old_end, new_end,
615 616 617 618 619 620 621 622
			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
623
		free_pgd_range(&tlb, old_start, old_end, new_end,
624
			vma->vm_next ? vma->vm_next->vm_start : 0);
L
Linus Torvalds 已提交
625
	}
P
Peter Zijlstra 已提交
626
	tlb_finish_mmu(&tlb, new_end, old_end);
627 628

	/*
629
	 * Shrink the vma to just the new range.  Always succeeds.
630 631 632 633
	 */
	vma_adjust(vma, new_start, new_end, vma->vm_pgoff, NULL);

	return 0;
L
Linus Torvalds 已提交
634 635
}

636 637 638 639
/*
 * 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
640 641 642 643
int setup_arg_pages(struct linux_binprm *bprm,
		    unsigned long stack_top,
		    int executable_stack)
{
644 645
	unsigned long ret;
	unsigned long stack_shift;
L
Linus Torvalds 已提交
646
	struct mm_struct *mm = current->mm;
647 648 649 650
	struct vm_area_struct *vma = bprm->vma;
	struct vm_area_struct *prev = NULL;
	unsigned long vm_flags;
	unsigned long stack_base;
651 652 653
	unsigned long stack_size;
	unsigned long stack_expand;
	unsigned long rlim_stack;
L
Linus Torvalds 已提交
654 655 656

#ifdef CONFIG_STACK_GROWSUP
	/* Limit stack size to 1GB */
J
Jiri Slaby 已提交
657
	stack_base = rlimit_max(RLIMIT_STACK);
L
Linus Torvalds 已提交
658 659 660
	if (stack_base > (1 << 30))
		stack_base = 1 << 30;

661 662 663
	/* 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
664

665
	stack_base = PAGE_ALIGN(stack_top - stack_base);
L
Linus Torvalds 已提交
666

667 668 669
	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
670
#else
671 672
	stack_top = arch_align_stack(stack_top);
	stack_top = PAGE_ALIGN(stack_top);
673 674 675 676 677

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

678 679 680
	stack_shift = vma->vm_end - stack_top;

	bprm->p -= stack_shift;
L
Linus Torvalds 已提交
681 682 683 684
	mm->arg_start = bprm->p;
#endif

	if (bprm->loader)
685 686
		bprm->loader -= stack_shift;
	bprm->exec -= stack_shift;
L
Linus Torvalds 已提交
687 688

	down_write(&mm->mmap_sem);
689
	vm_flags = VM_STACK_FLAGS;
690 691 692 693 694 695 696 697 698 699 700

	/*
	 * 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;
701
	vm_flags |= VM_STACK_INCOMPLETE_SETUP;
702 703 704 705 706 707 708 709 710 711

	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);
712 713
		if (ret)
			goto out_unlock;
L
Linus Torvalds 已提交
714 715
	}

716 717 718
	/* mprotect_fixup is overkill to remove the temporary stack flags */
	vma->vm_flags &= ~VM_STACK_INCOMPLETE_SETUP;

719
	stack_expand = 131072UL; /* randomly 32*4k (or 2*64k) pages */
720 721 722 723 724 725
	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;
726
#ifdef CONFIG_STACK_GROWSUP
727 728 729 730
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_start + rlim_stack;
	else
		stack_base = vma->vm_end + stack_expand;
731
#else
732 733 734 735
	if (stack_size + stack_expand > rlim_stack)
		stack_base = vma->vm_end - rlim_stack;
	else
		stack_base = vma->vm_start - stack_expand;
736
#endif
737
	current->mm->start_stack = bprm->p;
738 739 740 741 742
	ret = expand_stack(vma, stack_base);
	if (ret)
		ret = -EFAULT;

out_unlock:
L
Linus Torvalds 已提交
743
	up_write(&mm->mmap_sem);
744
	return ret;
L
Linus Torvalds 已提交
745 746 747 748 749 750 751 752
}
EXPORT_SYMBOL(setup_arg_pages);

#endif /* CONFIG_MMU */

struct file *open_exec(const char *name)
{
	struct file *file;
753
	int err;
754
	struct filename tmp = { .name = name };
755 756 757 758 759
	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
760

761
	file = do_filp_open(AT_FDCWD, &tmp, &open_exec_flags, LOOKUP_FOLLOW);
762
	if (IS_ERR(file))
763 764 765
		goto out;

	err = -EACCES;
766 767
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
		goto exit;
768

769 770
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
		goto exit;
771

772
	fsnotify_open(file);
773

774
	err = deny_write_access(file);
775 776
	if (err)
		goto exit;
L
Linus Torvalds 已提交
777

778
out:
779 780
	return file;

781 782
exit:
	fput(file);
783 784
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
785 786
EXPORT_SYMBOL(open_exec);

M
Mimi Zohar 已提交
787 788
int kernel_read(struct file *file, loff_t offset,
		char *addr, unsigned long count)
L
Linus Torvalds 已提交
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814
{
	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;
	mm_release(tsk, old_mm);

	if (old_mm) {
815
		sync_mm_rss(old_mm);
L
Linus Torvalds 已提交
816 817 818 819
		/*
		 * 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
820
		 * checking core_state and changing tsk->mm.
L
Linus Torvalds 已提交
821 822
		 */
		down_read(&old_mm->mmap_sem);
823
		if (unlikely(old_mm->core_state)) {
L
Linus Torvalds 已提交
824 825 826 827 828 829 830 831 832 833 834 835 836
			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);
837
		BUG_ON(active_mm != old_mm);
838
		setmax_mm_hiwater_rss(&tsk->signal->maxrss, old_mm);
839
		mm_update_next_owner(old_mm);
L
Linus Torvalds 已提交
840 841 842 843 844 845 846 847 848 849 850 851 852
		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().)
 */
853
static int de_thread(struct task_struct *tsk)
L
Linus Torvalds 已提交
854 855
{
	struct signal_struct *sig = tsk->signal;
856
	struct sighand_struct *oldsighand = tsk->sighand;
L
Linus Torvalds 已提交
857 858
	spinlock_t *lock = &oldsighand->siglock;

859
	if (thread_group_empty(tsk))
L
Linus Torvalds 已提交
860 861 862 863 864 865
		goto no_thread_group;

	/*
	 * Kill all other threads in the thread group.
	 */
	spin_lock_irq(lock);
866
	if (signal_group_exit(sig)) {
L
Linus Torvalds 已提交
867 868 869 870 871 872 873
		/*
		 * Another group action in progress, just
		 * return so that the signal is processed.
		 */
		spin_unlock_irq(lock);
		return -EAGAIN;
	}
874

875
	sig->group_exit_task = tsk;
876 877 878
	sig->notify_count = zap_other_threads(tsk);
	if (!thread_group_leader(tsk))
		sig->notify_count--;
L
Linus Torvalds 已提交
879

880
	while (sig->notify_count) {
O
Oleg Nesterov 已提交
881
		__set_current_state(TASK_KILLABLE);
L
Linus Torvalds 已提交
882 883
		spin_unlock_irq(lock);
		schedule();
O
Oleg Nesterov 已提交
884 885
		if (unlikely(__fatal_signal_pending(tsk)))
			goto killed;
L
Linus Torvalds 已提交
886 887 888 889 890 891 892 893 894
		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:
	 */
895
	if (!thread_group_leader(tsk)) {
896
		struct task_struct *leader = tsk->group_leader;
897

898
		sig->notify_count = -1;	/* for exit_notify() */
899 900 901 902
		for (;;) {
			write_lock_irq(&tasklist_lock);
			if (likely(leader->exit_state))
				break;
O
Oleg Nesterov 已提交
903
			__set_current_state(TASK_KILLABLE);
904 905
			write_unlock_irq(&tasklist_lock);
			schedule();
O
Oleg Nesterov 已提交
906 907
			if (unlikely(__fatal_signal_pending(tsk)))
				goto killed;
908
		}
L
Linus Torvalds 已提交
909

910 911 912 913 914 915 916 917 918 919
		/*
		 * 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).
		 */
920
		tsk->start_time = leader->start_time;
921

922 923
		BUG_ON(!same_thread_group(leader, tsk));
		BUG_ON(has_group_leader_pid(tsk));
L
Linus Torvalds 已提交
924 925 926 927 928 929
		/*
		 * 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:
		 */
930 931

		/* Become a process group leader with the old leader's pid.
932 933
		 * The old leader becomes a thread of the this thread group.
		 * Note: The old leader also uses this pid until release_task
934 935
		 *       is called.  Odd but simple and correct.
		 */
936 937
		detach_pid(tsk, PIDTYPE_PID);
		tsk->pid = leader->pid;
938
		attach_pid(tsk, PIDTYPE_PID,  task_pid(leader));
939 940
		transfer_pid(leader, tsk, PIDTYPE_PGID);
		transfer_pid(leader, tsk, PIDTYPE_SID);
941

942
		list_replace_rcu(&leader->tasks, &tsk->tasks);
943
		list_replace_init(&leader->sibling, &tsk->sibling);
L
Linus Torvalds 已提交
944

945 946
		tsk->group_leader = tsk;
		leader->group_leader = tsk;
947

948
		tsk->exit_signal = SIGCHLD;
949
		leader->exit_signal = -1;
950 951 952

		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
		leader->exit_state = EXIT_DEAD;
953 954 955 956 957 958 959 960

		/*
		 * 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
961
		write_unlock_irq(&tasklist_lock);
962 963

		release_task(leader);
964
	}
L
Linus Torvalds 已提交
965

966 967
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
L
Linus Torvalds 已提交
968 969

no_thread_group:
970 971 972
	/* we have changed execution domain */
	tsk->exit_signal = SIGCHLD;

L
Linus Torvalds 已提交
973
	exit_itimers(sig);
974
	flush_itimer_signals();
975

976 977
	if (atomic_read(&oldsighand->count) != 1) {
		struct sighand_struct *newsighand;
L
Linus Torvalds 已提交
978
		/*
979 980
		 * This ->sighand is shared with the CLONE_SIGHAND
		 * but not CLONE_THREAD task, switch to the new one.
L
Linus Torvalds 已提交
981
		 */
982 983 984 985
		newsighand = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
		if (!newsighand)
			return -ENOMEM;

L
Linus Torvalds 已提交
986 987 988 989 990 991
		atomic_set(&newsighand->count, 1);
		memcpy(newsighand->action, oldsighand->action,
		       sizeof(newsighand->action));

		write_lock_irq(&tasklist_lock);
		spin_lock(&oldsighand->siglock);
992
		rcu_assign_pointer(tsk->sighand, newsighand);
L
Linus Torvalds 已提交
993 994 995
		spin_unlock(&oldsighand->siglock);
		write_unlock_irq(&tasklist_lock);

996
		__cleanup_sighand(oldsighand);
L
Linus Torvalds 已提交
997 998
	}

999
	BUG_ON(!thread_group_leader(tsk));
L
Linus Torvalds 已提交
1000
	return 0;
O
Oleg Nesterov 已提交
1001 1002 1003 1004 1005 1006 1007 1008

killed:
	/* protects against exit_notify() and __exit_signal() */
	read_lock(&tasklist_lock);
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
	read_unlock(&tasklist_lock);
	return -EAGAIN;
L
Linus Torvalds 已提交
1009
}
O
Oleg Nesterov 已提交
1010

1011
char *get_task_comm(char *buf, struct task_struct *tsk)
L
Linus Torvalds 已提交
1012 1013 1014 1015 1016
{
	/* buf must be at least sizeof(tsk->comm) in size */
	task_lock(tsk);
	strncpy(buf, tsk->comm, sizeof(tsk->comm));
	task_unlock(tsk);
1017
	return buf;
L
Linus Torvalds 已提交
1018
}
1019
EXPORT_SYMBOL_GPL(get_task_comm);
L
Linus Torvalds 已提交
1020

1021 1022 1023 1024 1025
/*
 * These functions flushes out all traces of the currently running executable
 * so that a new one can be started
 */

L
Linus Torvalds 已提交
1026 1027 1028
void set_task_comm(struct task_struct *tsk, char *buf)
{
	task_lock(tsk);
1029

1030 1031
	trace_task_rename(tsk, buf);

1032 1033 1034 1035 1036 1037 1038 1039
	/*
	 * 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
1040 1041
	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
	task_unlock(tsk);
1042
	perf_event_comm(tsk);
L
Linus Torvalds 已提交
1043 1044
}

1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
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
1060 1061
int flush_old_exec(struct linux_binprm * bprm)
{
1062
	int retval;
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069 1070 1071

	/*
	 * 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
1072 1073
	set_mm_exe_file(bprm->mm, bprm->file);

1074
	filename_to_taskname(bprm->tcomm, bprm->filename, sizeof(bprm->tcomm));
L
Linus Torvalds 已提交
1075 1076 1077
	/*
	 * Release all of the old mmap stuff
	 */
1078
	acct_arg_size(bprm, 0);
L
Linus Torvalds 已提交
1079 1080
	retval = exec_mmap(bprm->mm);
	if (retval)
1081
		goto out;
L
Linus Torvalds 已提交
1082 1083

	bprm->mm = NULL;		/* We're using it now */
1084

1085
	set_fs(USER_DS);
1086 1087
	current->flags &=
		~(PF_RANDOMIZE | PF_FORKNOEXEC | PF_KTHREAD | PF_NOFREEZE);
1088 1089 1090
	flush_thread();
	current->personality &= ~bprm->per_clear;

1091 1092 1093 1094 1095 1096 1097
	return 0;

out:
	return retval;
}
EXPORT_SYMBOL(flush_old_exec);

1098 1099 1100 1101 1102 1103 1104
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);

1105 1106 1107
void setup_new_exec(struct linux_binprm * bprm)
{
	arch_pick_mmap_layout(current->mm);
L
Linus Torvalds 已提交
1108 1109 1110 1111

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

1112
	if (uid_eq(current_euid(), current_uid()) && gid_eq(current_egid(), current_gid()))
1113
		set_dumpable(current->mm, SUID_DUMPABLE_ENABLED);
A
Alan Cox 已提交
1114
	else
1115
		set_dumpable(current->mm, suid_dumpable);
A
Alan Cox 已提交
1116

1117
	set_task_comm(current, bprm->tcomm);
L
Linus Torvalds 已提交
1118

1119 1120 1121 1122 1123 1124
	/* 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;

1125
	/* install the new credentials */
1126 1127
	if (!uid_eq(bprm->cred->uid, current_euid()) ||
	    !gid_eq(bprm->cred->gid, current_egid())) {
1128
		current->pdeath_signal = 0;
1129 1130 1131 1132
	} 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
1133 1134
	}

I
Ingo Molnar 已提交
1135 1136 1137 1138 1139
	/*
	 * Flush performance counters when crossing a
	 * security domain:
	 */
	if (!get_dumpable(current->mm))
1140
		perf_event_exit_task(current);
I
Ingo Molnar 已提交
1141

L
Linus Torvalds 已提交
1142 1143 1144 1145 1146 1147
	/* An exec changes our domain. We are no longer part of the thread
	   group */

	current->self_exec_id++;
			
	flush_signal_handlers(current, 0);
1148
	do_close_on_exec(current->files);
L
Linus Torvalds 已提交
1149
}
1150
EXPORT_SYMBOL(setup_new_exec);
L
Linus Torvalds 已提交
1151

1152 1153 1154 1155 1156 1157 1158 1159
/*
 * 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)
{
1160
	if (mutex_lock_interruptible(&current->signal->cred_guard_mutex))
1161 1162 1163 1164 1165 1166
		return -ERESTARTNOINTR;

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

1167
	mutex_unlock(&current->signal->cred_guard_mutex);
1168 1169 1170 1171 1172 1173 1174
	return -ENOMEM;
}

void free_bprm(struct linux_binprm *bprm)
{
	free_arg_pages(bprm);
	if (bprm->cred) {
1175
		mutex_unlock(&current->signal->cred_guard_mutex);
1176 1177 1178 1179 1180
		abort_creds(bprm->cred);
	}
	kfree(bprm);
}

1181 1182 1183 1184 1185 1186 1187 1188 1189
/*
 * 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;
1190 1191
	/*
	 * cred_guard_mutex must be held at least to this point to prevent
1192
	 * ptrace_attach() from altering our determination of the task's
1193 1194
	 * credentials; any time after this it may be unlocked.
	 */
1195
	security_bprm_committed_creds(bprm);
1196
	mutex_unlock(&current->signal->cred_guard_mutex);
1197 1198 1199 1200 1201
}
EXPORT_SYMBOL(install_exec_creds);

/*
 * determine how safe it is to execute the proposed program
1202
 * - the caller must hold ->cred_guard_mutex to protect against
1203 1204
 *   PTRACE_ATTACH
 */
A
Al Viro 已提交
1205
static int check_unsafe_exec(struct linux_binprm *bprm)
1206
{
D
David Howells 已提交
1207
	struct task_struct *p = current, *t;
1208
	unsigned n_fs;
A
Al Viro 已提交
1209
	int res = 0;
1210

T
Tejun Heo 已提交
1211 1212 1213 1214 1215 1216
	if (p->ptrace) {
		if (p->ptrace & PT_PTRACE_CAP)
			bprm->unsafe |= LSM_UNSAFE_PTRACE_CAP;
		else
			bprm->unsafe |= LSM_UNSAFE_PTRACE;
	}
1217

1218 1219 1220 1221 1222 1223 1224
	/*
	 * This isn't strictly necessary, but it makes it harder for LSMs to
	 * mess up.
	 */
	if (current->no_new_privs)
		bprm->unsafe |= LSM_UNSAFE_NO_NEW_PRIVS;

D
David Howells 已提交
1225
	n_fs = 1;
N
Nick Piggin 已提交
1226
	spin_lock(&p->fs->lock);
1227
	rcu_read_lock();
D
David Howells 已提交
1228 1229 1230 1231
	for (t = next_thread(p); t != p; t = next_thread(t)) {
		if (t->fs == p->fs)
			n_fs++;
	}
1232
	rcu_read_unlock();
D
David Howells 已提交
1233

1234
	if (p->fs->users > n_fs) {
1235
		bprm->unsafe |= LSM_UNSAFE_SHARE;
A
Al Viro 已提交
1236
	} else {
1237 1238 1239 1240 1241
		res = -EAGAIN;
		if (!p->fs->in_exec) {
			p->fs->in_exec = 1;
			res = 1;
		}
A
Al Viro 已提交
1242
	}
N
Nick Piggin 已提交
1243
	spin_unlock(&p->fs->lock);
A
Al Viro 已提交
1244 1245

	return res;
1246 1247
}

L
Linus Torvalds 已提交
1248 1249 1250
/* 
 * Fill the binprm structure from the inode. 
 * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes
1251 1252
 *
 * This may be called multiple times for binary chains (scripts for example).
L
Linus Torvalds 已提交
1253 1254 1255
 */
int prepare_binprm(struct linux_binprm *bprm)
{
1256
	umode_t mode;
1257
	struct inode * inode = bprm->file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1258 1259 1260 1261 1262 1263
	int retval;

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

1264 1265 1266
	/* 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
1267

1268 1269
	if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID) &&
	    !current->no_new_privs) {
L
Linus Torvalds 已提交
1270 1271
		/* Set-uid? */
		if (mode & S_ISUID) {
1272 1273
			if (!kuid_has_mapping(bprm->cred->user_ns, inode->i_uid))
				return -EPERM;
1274 1275
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->euid = inode->i_uid;
1276

L
Linus Torvalds 已提交
1277 1278 1279 1280 1281 1282 1283 1284 1285
		}

		/* 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)) {
1286 1287
			if (!kgid_has_mapping(bprm->cred->user_ns, inode->i_gid))
				return -EPERM;
1288 1289
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->egid = inode->i_gid;
L
Linus Torvalds 已提交
1290 1291 1292 1293
		}
	}

	/* fill in binprm security blob */
1294
	retval = security_bprm_set_creds(bprm);
L
Linus Torvalds 已提交
1295 1296
	if (retval)
		return retval;
1297
	bprm->cred_prepared = 1;
L
Linus Torvalds 已提交
1298

1299 1300
	memset(bprm->buf, 0, BINPRM_BUF_SIZE);
	return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE);
L
Linus Torvalds 已提交
1301 1302 1303 1304
}

EXPORT_SYMBOL(prepare_binprm);

N
Nick Piggin 已提交
1305 1306 1307 1308 1309
/*
 * 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.
 */
1310
int remove_arg_zero(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1311
{
1312 1313 1314 1315
	int ret = 0;
	unsigned long offset;
	char *kaddr;
	struct page *page;
N
Nick Piggin 已提交
1316

1317 1318
	if (!bprm->argc)
		return 0;
L
Linus Torvalds 已提交
1319

1320 1321 1322 1323 1324 1325 1326
	do {
		offset = bprm->p & ~PAGE_MASK;
		page = get_arg_page(bprm, bprm->p, 0);
		if (!page) {
			ret = -EFAULT;
			goto out;
		}
1327
		kaddr = kmap_atomic(page);
N
Nick Piggin 已提交
1328

1329 1330 1331
		for (; offset < PAGE_SIZE && kaddr[offset];
				offset++, bprm->p++)
			;
N
Nick Piggin 已提交
1332

1333
		kunmap_atomic(kaddr);
1334
		put_arg_page(page);
N
Nick Piggin 已提交
1335

1336 1337 1338
		if (offset == PAGE_SIZE)
			free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1);
	} while (offset == PAGE_SIZE);
N
Nick Piggin 已提交
1339

1340 1341 1342
	bprm->p++;
	bprm->argc--;
	ret = 0;
N
Nick Piggin 已提交
1343

1344 1345
out:
	return ret;
L
Linus Torvalds 已提交
1346 1347 1348 1349 1350 1351
}
EXPORT_SYMBOL(remove_arg_zero);

/*
 * cycle the list of binary formats handler, until one recognizes the image
 */
1352
int search_binary_handler(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1353
{
1354
	unsigned int depth = bprm->recursion_depth;
L
Linus Torvalds 已提交
1355 1356
	int try,retval;
	struct linux_binfmt *fmt;
1357
	pid_t old_pid, old_vpid;
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362

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

A
Al Viro 已提交
1363 1364 1365 1366
	retval = audit_bprm(bprm);
	if (retval)
		return retval;

1367
	/* Need to fetch pid before load_binary changes it */
1368
	old_pid = current->pid;
1369
	rcu_read_lock();
1370
	old_vpid = task_pid_nr_ns(current, task_active_pid_ns(current->parent));
1371 1372
	rcu_read_unlock();

L
Linus Torvalds 已提交
1373 1374 1375
	retval = -ENOENT;
	for (try=0; try<2; try++) {
		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
1376
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
1377 1378 1379 1380 1381 1382
			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);
1383
			retval = fn(bprm, current_pt_regs());
1384 1385 1386 1387 1388 1389
			/*
			 * 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
1390
			if (retval >= 0) {
1391 1392
				if (depth == 0) {
					trace_sched_process_exec(current, old_pid, bprm);
1393
					ptrace_event(PTRACE_EVENT_EXEC, old_vpid);
1394
				}
L
Linus Torvalds 已提交
1395 1396 1397 1398 1399 1400
				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
1401
				proc_exec_connector(current);
L
Linus Torvalds 已提交
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
				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);
1414
#ifdef CONFIG_MODULES
L
Linus Torvalds 已提交
1415 1416
		if (retval != -ENOEXEC || bprm->mm == NULL) {
			break;
J
Johannes Berg 已提交
1417
		} else {
L
Linus Torvalds 已提交
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 */
1424 1425
			if (try)
				break; /* -ENOEXEC */
L
Linus Torvalds 已提交
1426 1427
			request_module("binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
		}
1428 1429 1430
#else
		break;
#endif
L
Linus Torvalds 已提交
1431 1432 1433 1434 1435 1436 1437 1438 1439
	}
	return retval;
}

EXPORT_SYMBOL(search_binary_handler);

/*
 * sys_execve() executes a new program.
 */
1440 1441
static int do_execve_common(const char *filename,
				struct user_arg_ptr argv,
1442
				struct user_arg_ptr envp)
L
Linus Torvalds 已提交
1443 1444 1445
{
	struct linux_binprm *bprm;
	struct file *file;
1446
	struct files_struct *displaced;
1447
	bool clear_in_exec;
L
Linus Torvalds 已提交
1448
	int retval;
1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465
	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
1466

1467
	retval = unshare_files(&displaced);
1468 1469 1470
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1471
	retval = -ENOMEM;
1472
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1473
	if (!bprm)
1474
		goto out_files;
L
Linus Torvalds 已提交
1475

1476 1477
	retval = prepare_bprm_creds(bprm);
	if (retval)
1478
		goto out_free;
A
Al Viro 已提交
1479 1480

	retval = check_unsafe_exec(bprm);
1481
	if (retval < 0)
1482
		goto out_free;
1483
	clear_in_exec = retval;
1484
	current->in_execve = 1;
1485

L
Linus Torvalds 已提交
1486 1487 1488
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1489
		goto out_unmark;
L
Linus Torvalds 已提交
1490 1491 1492 1493 1494 1495 1496

	sched_exec();

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

1497 1498 1499
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1500

1501
	bprm->argc = count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1502
	if ((retval = bprm->argc) < 0)
1503
		goto out;
L
Linus Torvalds 已提交
1504

1505
	bprm->envc = count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525
	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;

1526
	retval = search_binary_handler(bprm);
1527 1528
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1529

1530
	/* execve succeeded */
A
Al Viro 已提交
1531
	current->fs->in_exec = 0;
1532
	current->in_execve = 0;
1533 1534 1535 1536 1537
	acct_update_integrals(current);
	free_bprm(bprm);
	if (displaced)
		put_files_struct(displaced);
	return retval;
L
Linus Torvalds 已提交
1538

1539
out:
1540 1541 1542 1543
	if (bprm->mm) {
		acct_arg_size(bprm, 0);
		mmput(bprm->mm);
	}
L
Linus Torvalds 已提交
1544 1545 1546 1547 1548 1549

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}
1550

A
Al Viro 已提交
1551
out_unmark:
1552 1553
	if (clear_in_exec)
		current->fs->in_exec = 0;
1554
	current->in_execve = 0;
1555 1556

out_free:
1557
	free_bprm(bprm);
L
Linus Torvalds 已提交
1558

1559
out_files:
1560 1561
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1562 1563 1564 1565
out_ret:
	return retval;
}

1566 1567
int do_execve(const char *filename,
	const char __user *const __user *__argv,
1568
	const char __user *const __user *__envp)
1569
{
1570 1571
	struct user_arg_ptr argv = { .ptr.native = __argv };
	struct user_arg_ptr envp = { .ptr.native = __envp };
1572
	return do_execve_common(filename, argv, envp);
1573 1574 1575
}

#ifdef CONFIG_COMPAT
1576
static int compat_do_execve(const char *filename,
A
Al Viro 已提交
1577
	const compat_uptr_t __user *__argv,
1578
	const compat_uptr_t __user *__envp)
1579 1580 1581 1582 1583 1584 1585 1586 1587
{
	struct user_arg_ptr argv = {
		.is_compat = true,
		.ptr.compat = __argv,
	};
	struct user_arg_ptr envp = {
		.is_compat = true,
		.ptr.compat = __envp,
	};
1588
	return do_execve_common(filename, argv, envp);
1589
}
1590
#endif
1591

1592
void set_binfmt(struct linux_binfmt *new)
L
Linus Torvalds 已提交
1593
{
1594 1595 1596 1597
	struct mm_struct *mm = current->mm;

	if (mm->binfmt)
		module_put(mm->binfmt->module);
L
Linus Torvalds 已提交
1598

1599
	mm->binfmt = new;
1600 1601
	if (new)
		__module_get(new->module);
L
Linus Torvalds 已提交
1602 1603 1604 1605
}

EXPORT_SYMBOL(set_binfmt);

1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628
/*
 * 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) {
1629
	case SUID_DUMPABLE_DISABLED:
1630 1631 1632 1633
		clear_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
1634
	case SUID_DUMPABLE_ENABLED:
1635 1636 1637 1638
		set_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
1639
	case SUID_DUMPABLE_SAFE:
1640 1641 1642 1643 1644 1645 1646
		set_bit(MMF_DUMP_SECURELY, &mm->flags);
		smp_wmb();
		set_bit(MMF_DUMPABLE, &mm->flags);
		break;
	}
}

1647
int __get_dumpable(unsigned long mm_flags)
1648 1649 1650
{
	int ret;

1651
	ret = mm_flags & MMF_DUMPABLE_MASK;
1652
	return (ret > SUID_DUMPABLE_ENABLED) ? SUID_DUMPABLE_SAFE : ret;
1653 1654
}

1655 1656 1657 1658 1659
int get_dumpable(struct mm_struct *mm)
{
	return __get_dumpable(mm->flags);
}

A
Al Viro 已提交
1660 1661 1662 1663 1664 1665
#ifdef __ARCH_WANT_SYS_EXECVE
SYSCALL_DEFINE3(execve,
		const char __user *, filename,
		const char __user *const __user *, argv,
		const char __user *const __user *, envp)
{
1666
	struct filename *path = getname(filename);
A
Al Viro 已提交
1667 1668
	int error = PTR_ERR(path);
	if (!IS_ERR(path)) {
1669
		error = do_execve(path->name, argv, envp);
A
Al Viro 已提交
1670 1671 1672 1673 1674 1675 1676 1677 1678
		putname(path);
	}
	return error;
}
#ifdef CONFIG_COMPAT
asmlinkage long compat_sys_execve(const char __user * filename,
	const compat_uptr_t __user * argv,
	const compat_uptr_t __user * envp)
{
1679
	struct filename *path = getname(filename);
A
Al Viro 已提交
1680 1681
	int error = PTR_ERR(path);
	if (!IS_ERR(path)) {
1682
		error = compat_do_execve(path->name, argv, envp);
A
Al Viro 已提交
1683 1684 1685 1686 1687 1688 1689
		putname(path);
	}
	return error;
}
#endif
#endif

A
Al Viro 已提交
1690 1691 1692 1693 1694
#ifdef __ARCH_WANT_KERNEL_EXECVE
int kernel_execve(const char *filename,
		  const char *const argv[],
		  const char *const envp[])
{
1695
	int ret = do_execve(filename,
A
Al Viro 已提交
1696
			(const char __user *const __user *)argv,
1697
			(const char __user *const __user *)envp);
A
Al Viro 已提交
1698 1699 1700 1701 1702 1703 1704
	if (ret < 0)
		return ret;

	/*
	 * We were successful.  We won't be returning to our caller, but
	 * instead to user space by manipulating the kernel stack.
	 */
1705
	ret_from_kernel_execve(current_pt_regs());
A
Al Viro 已提交
1706 1707
}
#endif