fork.c 42.6 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
/*
 *  linux/kernel/fork.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

/*
 *  'fork.c' contains the help-routines for the 'fork' system call
 * (see also entry.S and others).
 * Fork is rather simple, once you get the hang of it, but the memory
 * management can be a bitch. See 'mm/memory.c': 'copy_page_range()'
 */

#include <linux/slab.h>
#include <linux/init.h>
#include <linux/unistd.h>
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/completion.h>
#include <linux/personality.h>
#include <linux/mempolicy.h>
#include <linux/sem.h>
#include <linux/file.h>
A
Al Viro 已提交
24
#include <linux/fdtable.h>
25
#include <linux/iocontext.h>
L
Linus Torvalds 已提交
26 27 28
#include <linux/key.h>
#include <linux/binfmts.h>
#include <linux/mman.h>
A
Andrea Arcangeli 已提交
29
#include <linux/mmu_notifier.h>
L
Linus Torvalds 已提交
30
#include <linux/fs.h>
S
Serge E. Hallyn 已提交
31
#include <linux/nsproxy.h>
32
#include <linux/capability.h>
L
Linus Torvalds 已提交
33
#include <linux/cpu.h>
34
#include <linux/cgroup.h>
L
Linus Torvalds 已提交
35
#include <linux/security.h>
36
#include <linux/hugetlb.h>
L
Linus Torvalds 已提交
37 38 39
#include <linux/swap.h>
#include <linux/syscalls.h>
#include <linux/jiffies.h>
R
Roland McGrath 已提交
40
#include <linux/tracehook.h>
L
Linus Torvalds 已提交
41
#include <linux/futex.h>
42
#include <linux/compat.h>
43
#include <linux/task_io_accounting_ops.h>
44
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
45 46 47
#include <linux/ptrace.h>
#include <linux/mount.h>
#include <linux/audit.h>
48
#include <linux/memcontrol.h>
49
#include <linux/ftrace.h>
L
Linus Torvalds 已提交
50 51
#include <linux/profile.h>
#include <linux/rmap.h>
H
Hugh Dickins 已提交
52
#include <linux/ksm.h>
L
Linus Torvalds 已提交
53
#include <linux/acct.h>
54
#include <linux/tsacct_kern.h>
M
Matt Helsley 已提交
55
#include <linux/cn_proc.h>
R
Rafael J. Wysocki 已提交
56
#include <linux/freezer.h>
57
#include <linux/delayacct.h>
58
#include <linux/taskstats_kern.h>
59
#include <linux/random.h>
M
Miloslav Trmac 已提交
60
#include <linux/tty.h>
61
#include <linux/proc_fs.h>
62
#include <linux/blkdev.h>
63
#include <linux/fs_struct.h>
64
#include <linux/magic.h>
65
#include <linux/perf_event.h>
66
#include <linux/posix-timers.h>
67
#include <linux/user-return-notifier.h>
L
Linus Torvalds 已提交
68 69 70 71 72 73 74 75

#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/uaccess.h>
#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

76 77
#include <trace/events/sched.h>

L
Linus Torvalds 已提交
78 79 80 81 82 83 84 85 86 87
/*
 * Protected counters by write_lock_irq(&tasklist_lock)
 */
unsigned long total_forks;	/* Handle normal Linux uptimes. */
int nr_threads; 		/* The idle threads do not count.. */

int max_threads;		/* tunable limit on nr_threads */

DEFINE_PER_CPU(unsigned long, process_counts) = 0;

88
__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
L
Linus Torvalds 已提交
89 90 91 92 93 94

int nr_processes(void)
{
	int cpu;
	int total = 0;

95
	for_each_possible_cpu(cpu)
L
Linus Torvalds 已提交
96 97 98 99 100 101 102 103
		total += per_cpu(process_counts, cpu);

	return total;
}

#ifndef __HAVE_ARCH_TASK_STRUCT_ALLOCATOR
# define alloc_task_struct()	kmem_cache_alloc(task_struct_cachep, GFP_KERNEL)
# define free_task_struct(tsk)	kmem_cache_free(task_struct_cachep, (tsk))
104
static struct kmem_cache *task_struct_cachep;
L
Linus Torvalds 已提交
105 106
#endif

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
#ifndef __HAVE_ARCH_THREAD_INFO_ALLOCATOR
static inline struct thread_info *alloc_thread_info(struct task_struct *tsk)
{
#ifdef CONFIG_DEBUG_STACK_USAGE
	gfp_t mask = GFP_KERNEL | __GFP_ZERO;
#else
	gfp_t mask = GFP_KERNEL;
#endif
	return (struct thread_info *)__get_free_pages(mask, THREAD_SIZE_ORDER);
}

static inline void free_thread_info(struct thread_info *ti)
{
	free_pages((unsigned long)ti, THREAD_SIZE_ORDER);
}
#endif

L
Linus Torvalds 已提交
124
/* SLAB cache for signal_struct structures (tsk->signal) */
125
static struct kmem_cache *signal_cachep;
L
Linus Torvalds 已提交
126 127

/* SLAB cache for sighand_struct structures (tsk->sighand) */
128
struct kmem_cache *sighand_cachep;
L
Linus Torvalds 已提交
129 130

/* SLAB cache for files_struct structures (tsk->files) */
131
struct kmem_cache *files_cachep;
L
Linus Torvalds 已提交
132 133

/* SLAB cache for fs_struct structures (tsk->fs) */
134
struct kmem_cache *fs_cachep;
L
Linus Torvalds 已提交
135 136

/* SLAB cache for vm_area_struct structures */
137
struct kmem_cache *vm_area_cachep;
L
Linus Torvalds 已提交
138 139

/* SLAB cache for mm_struct structures (tsk->mm) */
140
static struct kmem_cache *mm_cachep;
L
Linus Torvalds 已提交
141

142 143 144 145 146 147 148
static void account_kernel_stack(struct thread_info *ti, int account)
{
	struct zone *zone = page_zone(virt_to_page(ti));

	mod_zone_page_state(zone, NR_KERNEL_STACK, account);
}

L
Linus Torvalds 已提交
149 150
void free_task(struct task_struct *tsk)
{
P
Peter Zijlstra 已提交
151
	prop_local_destroy_single(&tsk->dirties);
152
	account_kernel_stack(tsk->stack, -1);
R
Roman Zippel 已提交
153
	free_thread_info(tsk->stack);
I
Ingo Molnar 已提交
154
	rt_mutex_debug_task_free(tsk);
155
	ftrace_graph_exit_task(tsk);
L
Linus Torvalds 已提交
156 157 158 159
	free_task_struct(tsk);
}
EXPORT_SYMBOL(free_task);

160
void __put_task_struct(struct task_struct *tsk)
L
Linus Torvalds 已提交
161
{
E
Eugene Teo 已提交
162
	WARN_ON(!tsk->exit_state);
L
Linus Torvalds 已提交
163 164 165
	WARN_ON(atomic_read(&tsk->usage));
	WARN_ON(tsk == current);

166
	exit_creds(tsk);
167
	delayacct_tsk_free(tsk);
L
Linus Torvalds 已提交
168 169 170 171 172

	if (!profile_handoff_task(tsk))
		free_task(tsk);
}

S
Suresh Siddha 已提交
173 174 175 176 177 178 179
/*
 * macro override instead of weak attribute alias, to workaround
 * gcc 4.1.0 and 4.1.1 bugs with weak attribute and empty functions.
 */
#ifndef arch_task_cache_init
#define arch_task_cache_init()
#endif
180

L
Linus Torvalds 已提交
181 182 183 184 185 186 187 188 189
void __init fork_init(unsigned long mempages)
{
#ifndef __HAVE_ARCH_TASK_STRUCT_ALLOCATOR
#ifndef ARCH_MIN_TASKALIGN
#define ARCH_MIN_TASKALIGN	L1_CACHE_BYTES
#endif
	/* create a slab on which task_structs can be allocated */
	task_struct_cachep =
		kmem_cache_create("task_struct", sizeof(struct task_struct),
V
Vegard Nossum 已提交
190
			ARCH_MIN_TASKALIGN, SLAB_PANIC | SLAB_NOTRACK, NULL);
L
Linus Torvalds 已提交
191 192
#endif

193 194 195
	/* do the arch specific task caches init */
	arch_task_cache_init();

L
Linus Torvalds 已提交
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214
	/*
	 * The default maximum number of threads is set to a safe
	 * value: the thread structures can take up at most half
	 * of memory.
	 */
	max_threads = mempages / (8 * THREAD_SIZE / PAGE_SIZE);

	/*
	 * we need to allow at least 20 threads to boot a system
	 */
	if(max_threads < 20)
		max_threads = 20;

	init_task.signal->rlim[RLIMIT_NPROC].rlim_cur = max_threads/2;
	init_task.signal->rlim[RLIMIT_NPROC].rlim_max = max_threads/2;
	init_task.signal->rlim[RLIMIT_SIGPENDING] =
		init_task.signal->rlim[RLIMIT_NPROC];
}

215 216 217 218 219 220 221
int __attribute__((weak)) arch_dup_task_struct(struct task_struct *dst,
					       struct task_struct *src)
{
	*dst = *src;
	return 0;
}

L
Linus Torvalds 已提交
222 223 224 225
static struct task_struct *dup_task_struct(struct task_struct *orig)
{
	struct task_struct *tsk;
	struct thread_info *ti;
226 227
	unsigned long *stackend;

P
Peter Zijlstra 已提交
228
	int err;
L
Linus Torvalds 已提交
229 230 231 232 233 234 235 236 237 238 239 240 241

	prepare_to_copy(orig);

	tsk = alloc_task_struct();
	if (!tsk)
		return NULL;

	ti = alloc_thread_info(tsk);
	if (!ti) {
		free_task_struct(tsk);
		return NULL;
	}

242 243 244 245
 	err = arch_dup_task_struct(tsk, orig);
	if (err)
		goto out;

R
Roman Zippel 已提交
246
	tsk->stack = ti;
P
Peter Zijlstra 已提交
247 248

	err = prop_local_init_single(&tsk->dirties);
249 250
	if (err)
		goto out;
P
Peter Zijlstra 已提交
251

252
	setup_thread_stack(tsk, orig);
253
	clear_user_return_notifier(tsk);
254 255
	stackend = end_of_stack(tsk);
	*stackend = STACK_END_MAGIC;	/* for overflow detection */
L
Linus Torvalds 已提交
256

257 258 259 260
#ifdef CONFIG_CC_STACKPROTECTOR
	tsk->stack_canary = get_random_int();
#endif

L
Linus Torvalds 已提交
261 262
	/* One for us, one for whoever does the "release_task()" (usually parent) */
	atomic_set(&tsk->usage,2);
263
	atomic_set(&tsk->fs_excl, 0);
264
#ifdef CONFIG_BLK_DEV_IO_TRACE
265
	tsk->btrace_seq = 0;
266
#endif
267
	tsk->splice_pipe = NULL;
268 269 270

	account_kernel_stack(ti, 1);

L
Linus Torvalds 已提交
271
	return tsk;
272 273 274 275 276

out:
	free_thread_info(ti);
	free_task_struct(tsk);
	return NULL;
L
Linus Torvalds 已提交
277 278 279
}

#ifdef CONFIG_MMU
A
Alexey Dobriyan 已提交
280
static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
L
Linus Torvalds 已提交
281
{
H
Hugh Dickins 已提交
282
	struct vm_area_struct *mpnt, *tmp, **pprev;
L
Linus Torvalds 已提交
283 284 285 286 287 288
	struct rb_node **rb_link, *rb_parent;
	int retval;
	unsigned long charge;
	struct mempolicy *pol;

	down_write(&oldmm->mmap_sem);
289
	flush_cache_dup_mm(oldmm);
290 291 292 293
	/*
	 * Not linked in yet - no deadlock potential:
	 */
	down_write_nested(&mm->mmap_sem, SINGLE_DEPTH_NESTING);
294

L
Linus Torvalds 已提交
295 296 297 298
	mm->locked_vm = 0;
	mm->mmap = NULL;
	mm->mmap_cache = NULL;
	mm->free_area_cache = oldmm->mmap_base;
299
	mm->cached_hole_size = ~0UL;
L
Linus Torvalds 已提交
300
	mm->map_count = 0;
301
	cpumask_clear(mm_cpumask(mm));
L
Linus Torvalds 已提交
302 303 304 305
	mm->mm_rb = RB_ROOT;
	rb_link = &mm->mm_rb.rb_node;
	rb_parent = NULL;
	pprev = &mm->mmap;
H
Hugh Dickins 已提交
306 307 308
	retval = ksm_fork(mm, oldmm);
	if (retval)
		goto out;
L
Linus Torvalds 已提交
309

H
Hugh Dickins 已提交
310
	for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) {
L
Linus Torvalds 已提交
311 312 313
		struct file *file;

		if (mpnt->vm_flags & VM_DONTCOPY) {
314 315
			long pages = vma_pages(mpnt);
			mm->total_vm -= pages;
316
			vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file,
317
								-pages);
L
Linus Torvalds 已提交
318 319 320 321 322 323 324 325 326
			continue;
		}
		charge = 0;
		if (mpnt->vm_flags & VM_ACCOUNT) {
			unsigned int len = (mpnt->vm_end - mpnt->vm_start) >> PAGE_SHIFT;
			if (security_vm_enough_memory(len))
				goto fail_nomem;
			charge = len;
		}
327
		tmp = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
328 329 330
		if (!tmp)
			goto fail_nomem;
		*tmp = *mpnt;
331
		pol = mpol_dup(vma_policy(mpnt));
L
Linus Torvalds 已提交
332 333 334 335 336 337 338 339 340 341
		retval = PTR_ERR(pol);
		if (IS_ERR(pol))
			goto fail_nomem_policy;
		vma_set_policy(tmp, pol);
		tmp->vm_flags &= ~VM_LOCKED;
		tmp->vm_mm = mm;
		tmp->vm_next = NULL;
		anon_vma_link(tmp);
		file = tmp->vm_file;
		if (file) {
342
			struct inode *inode = file->f_path.dentry->d_inode;
H
Hugh Dickins 已提交
343 344
			struct address_space *mapping = file->f_mapping;

L
Linus Torvalds 已提交
345 346 347
			get_file(file);
			if (tmp->vm_flags & VM_DENYWRITE)
				atomic_dec(&inode->i_writecount);
H
Hugh Dickins 已提交
348 349 350
			spin_lock(&mapping->i_mmap_lock);
			if (tmp->vm_flags & VM_SHARED)
				mapping->i_mmap_writable++;
L
Linus Torvalds 已提交
351
			tmp->vm_truncate_count = mpnt->vm_truncate_count;
H
Hugh Dickins 已提交
352 353
			flush_dcache_mmap_lock(mapping);
			/* insert tmp into the share list, just after mpnt */
L
Linus Torvalds 已提交
354
			vma_prio_tree_add(tmp, mpnt);
H
Hugh Dickins 已提交
355 356
			flush_dcache_mmap_unlock(mapping);
			spin_unlock(&mapping->i_mmap_lock);
L
Linus Torvalds 已提交
357 358
		}

359 360 361 362 363 364 365 366
		/*
		 * Clear hugetlb-related page reserves for children. This only
		 * affects MAP_PRIVATE mappings. Faults generated by the child
		 * are not guaranteed to succeed, even if read-only
		 */
		if (is_vm_hugetlb_page(tmp))
			reset_vma_resv_huge_pages(tmp);

L
Linus Torvalds 已提交
367
		/*
368
		 * Link in the new vma and copy the page table entries.
L
Linus Torvalds 已提交
369 370 371 372 373 374 375 376 377
		 */
		*pprev = tmp;
		pprev = &tmp->vm_next;

		__vma_link_rb(mm, tmp, rb_link, rb_parent);
		rb_link = &tmp->vm_rb.rb_right;
		rb_parent = &tmp->vm_rb;

		mm->map_count++;
378
		retval = copy_page_range(mm, oldmm, mpnt);
L
Linus Torvalds 已提交
379 380 381 382 383 384 385

		if (tmp->vm_ops && tmp->vm_ops->open)
			tmp->vm_ops->open(tmp);

		if (retval)
			goto out;
	}
386 387
	/* a new mm has just been created */
	arch_dup_mmap(oldmm, mm);
L
Linus Torvalds 已提交
388 389
	retval = 0;
out:
390
	up_write(&mm->mmap_sem);
H
Hugh Dickins 已提交
391
	flush_tlb_mm(oldmm);
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
	up_write(&oldmm->mmap_sem);
	return retval;
fail_nomem_policy:
	kmem_cache_free(vm_area_cachep, tmp);
fail_nomem:
	retval = -ENOMEM;
	vm_unacct_memory(charge);
	goto out;
}

static inline int mm_alloc_pgd(struct mm_struct * mm)
{
	mm->pgd = pgd_alloc(mm);
	if (unlikely(!mm->pgd))
		return -ENOMEM;
	return 0;
}

static inline void mm_free_pgd(struct mm_struct * mm)
{
412
	pgd_free(mm, mm->pgd);
L
Linus Torvalds 已提交
413 414 415 416 417 418 419
}
#else
#define dup_mmap(mm, oldmm)	(0)
#define mm_alloc_pgd(mm)	(0)
#define mm_free_pgd(mm)
#endif /* CONFIG_MMU */

D
Daniel Walker 已提交
420
__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
L
Linus Torvalds 已提交
421

422
#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
L
Linus Torvalds 已提交
423 424
#define free_mm(mm)	(kmem_cache_free(mm_cachep, (mm)))

425 426 427 428 429 430 431 432 433 434 435 436
static unsigned long default_dump_filter = MMF_DUMP_FILTER_DEFAULT;

static int __init coredump_filter_setup(char *s)
{
	default_dump_filter =
		(simple_strtoul(s, NULL, 0) << MMF_DUMP_FILTER_SHIFT) &
		MMF_DUMP_FILTER_MASK;
	return 1;
}

__setup("coredump_filter=", coredump_filter_setup);

L
Linus Torvalds 已提交
437 438
#include <linux/init_task.h>

A
Alexey Dobriyan 已提交
439 440 441 442 443 444 445 446
static void mm_init_aio(struct mm_struct *mm)
{
#ifdef CONFIG_AIO
	spin_lock_init(&mm->ioctx_lock);
	INIT_HLIST_HEAD(&mm->ioctx_list);
#endif
}

447
static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p)
L
Linus Torvalds 已提交
448 449 450 451 452
{
	atomic_set(&mm->mm_users, 1);
	atomic_set(&mm->mm_count, 1);
	init_rwsem(&mm->mmap_sem);
	INIT_LIST_HEAD(&mm->mmlist);
H
Hugh Dickins 已提交
453 454
	mm->flags = (current->mm) ?
		(current->mm->flags & MMF_INIT_MASK) : default_dump_filter;
455
	mm->core_state = NULL;
L
Linus Torvalds 已提交
456
	mm->nr_ptes = 0;
457
	set_mm_counter(mm, file_rss, 0);
H
Hugh Dickins 已提交
458
	set_mm_counter(mm, anon_rss, 0);
L
Linus Torvalds 已提交
459 460
	spin_lock_init(&mm->page_table_lock);
	mm->free_area_cache = TASK_UNMAPPED_BASE;
461
	mm->cached_hole_size = ~0UL;
A
Alexey Dobriyan 已提交
462
	mm_init_aio(mm);
463
	mm_init_owner(mm, p);
L
Linus Torvalds 已提交
464 465 466

	if (likely(!mm_alloc_pgd(mm))) {
		mm->def_flags = 0;
A
Andrea Arcangeli 已提交
467
		mmu_notifier_mm_init(mm);
L
Linus Torvalds 已提交
468 469
		return mm;
	}
470

L
Linus Torvalds 已提交
471 472 473 474 475 476 477 478 479 480 481 482 483 484
	free_mm(mm);
	return NULL;
}

/*
 * Allocate and initialize an mm_struct.
 */
struct mm_struct * mm_alloc(void)
{
	struct mm_struct * mm;

	mm = allocate_mm();
	if (mm) {
		memset(mm, 0, sizeof(*mm));
485
		mm = mm_init(mm, current);
L
Linus Torvalds 已提交
486 487 488 489 490 491 492 493 494
	}
	return mm;
}

/*
 * Called when the last reference to the mm
 * is dropped: either by a lazy thread or by
 * mmput. Free the page directory and the mm.
 */
495
void __mmdrop(struct mm_struct *mm)
L
Linus Torvalds 已提交
496 497 498 499
{
	BUG_ON(mm == &init_mm);
	mm_free_pgd(mm);
	destroy_context(mm);
A
Andrea Arcangeli 已提交
500
	mmu_notifier_mm_destroy(mm);
L
Linus Torvalds 已提交
501 502
	free_mm(mm);
}
503
EXPORT_SYMBOL_GPL(__mmdrop);
L
Linus Torvalds 已提交
504 505 506 507 508 509

/*
 * Decrement the use count and release all resources for an mm.
 */
void mmput(struct mm_struct *mm)
{
A
Andrew Morton 已提交
510 511
	might_sleep();

L
Linus Torvalds 已提交
512 513
	if (atomic_dec_and_test(&mm->mm_users)) {
		exit_aio(mm);
514
		ksm_exit(mm);
L
Linus Torvalds 已提交
515
		exit_mmap(mm);
M
Matt Helsley 已提交
516
		set_mm_exe_file(mm, NULL);
L
Linus Torvalds 已提交
517 518 519 520 521 522
		if (!list_empty(&mm->mmlist)) {
			spin_lock(&mmlist_lock);
			list_del(&mm->mmlist);
			spin_unlock(&mmlist_lock);
		}
		put_swap_token(mm);
523 524
		if (mm->binfmt)
			module_put(mm->binfmt->module);
L
Linus Torvalds 已提交
525 526 527 528 529 530 531 532
		mmdrop(mm);
	}
}
EXPORT_SYMBOL_GPL(mmput);

/**
 * get_task_mm - acquire a reference to the task's mm
 *
533
 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
L
Linus Torvalds 已提交
534 535 536 537 538 539 540 541 542 543 544 545
 * this kernel workthread has transiently adopted a user mm with use_mm,
 * to do its AIO) is not set and if so returns a reference to it, after
 * bumping up the use count.  User must release the mm via mmput()
 * after use.  Typically used by /proc and ptrace.
 */
struct mm_struct *get_task_mm(struct task_struct *task)
{
	struct mm_struct *mm;

	task_lock(task);
	mm = task->mm;
	if (mm) {
546
		if (task->flags & PF_KTHREAD)
L
Linus Torvalds 已提交
547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572
			mm = NULL;
		else
			atomic_inc(&mm->mm_users);
	}
	task_unlock(task);
	return mm;
}
EXPORT_SYMBOL_GPL(get_task_mm);

/* Please note the differences between mmput and mm_release.
 * mmput is called whenever we stop holding onto a mm_struct,
 * error success whatever.
 *
 * mm_release is called after a mm_struct has been removed
 * from the current process.
 *
 * This difference is important for error handling, when we
 * only half set up a mm_struct for a new process and need to restore
 * the old one.  Because we mmput the new mm_struct before
 * restoring the old one. . .
 * Eric Biederman 10 January 1998
 */
void mm_release(struct task_struct *tsk, struct mm_struct *mm)
{
	struct completion *vfork_done = tsk->vfork_done;

573 574
	/* Get rid of any futexes when releasing the mm */
#ifdef CONFIG_FUTEX
575
	if (unlikely(tsk->robust_list)) {
576
		exit_robust_list(tsk);
577 578
		tsk->robust_list = NULL;
	}
579
#ifdef CONFIG_COMPAT
580
	if (unlikely(tsk->compat_robust_list)) {
581
		compat_exit_robust_list(tsk);
582 583
		tsk->compat_robust_list = NULL;
	}
584
#endif
585 586
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
587 588
#endif

L
Linus Torvalds 已提交
589 590 591 592 593 594 595 596
	/* Get rid of any cached register state */
	deactivate_mm(tsk, mm);

	/* notify parent sleeping on vfork() */
	if (vfork_done) {
		tsk->vfork_done = NULL;
		complete(vfork_done);
	}
597 598 599 600 601 602 603

	/*
	 * If we're exiting normally, clear a user-space tid field if
	 * requested.  We leave this alone when dying by signal, to leave
	 * the value intact in a core dump, and to save the unnecessary
	 * trouble otherwise.  Userland only wants this done for a sys_exit.
	 */
604 605 606 607 608 609 610 611 612 613 614
	if (tsk->clear_child_tid) {
		if (!(tsk->flags & PF_SIGNALED) &&
		    atomic_read(&mm->mm_users) > 1) {
			/*
			 * We don't check the error code - if userspace has
			 * not set up a proper pointer then tough luck.
			 */
			put_user(0, tsk->clear_child_tid);
			sys_futex(tsk->clear_child_tid, FUTEX_WAKE,
					1, NULL, NULL, 0);
		}
L
Linus Torvalds 已提交
615 616 617 618
		tsk->clear_child_tid = NULL;
	}
}

619 620 621 622
/*
 * Allocate a new mm structure and copy contents from the
 * mm structure of the passed in task structure.
 */
623
struct mm_struct *dup_mm(struct task_struct *tsk)
624 625 626 627 628 629 630 631 632 633 634 635 636
{
	struct mm_struct *mm, *oldmm = current->mm;
	int err;

	if (!oldmm)
		return NULL;

	mm = allocate_mm();
	if (!mm)
		goto fail_nomem;

	memcpy(mm, oldmm, sizeof(*mm));

637 638 639 640
	/* Initializing for Swap token stuff */
	mm->token_priority = 0;
	mm->last_interval = 0;

641
	if (!mm_init(mm, tsk))
642 643 644 645 646
		goto fail_nomem;

	if (init_new_context(tsk, mm))
		goto fail_nocontext;

M
Matt Helsley 已提交
647 648
	dup_mm_exe_file(oldmm, mm);

649 650 651 652 653 654 655
	err = dup_mmap(mm, oldmm);
	if (err)
		goto free_pt;

	mm->hiwater_rss = get_mm_rss(mm);
	mm->hiwater_vm = mm->total_vm;

656 657 658
	if (mm->binfmt && !try_module_get(mm->binfmt->module))
		goto free_pt;

659 660 661
	return mm;

free_pt:
662 663
	/* don't put binfmt in mmput, we haven't got module yet */
	mm->binfmt = NULL;
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678
	mmput(mm);

fail_nomem:
	return NULL;

fail_nocontext:
	/*
	 * If init_new_context() failed, we cannot use mmput() to free the mm
	 * because it calls destroy_context()
	 */
	mm_free_pgd(mm);
	free_mm(mm);
	return NULL;
}

L
Linus Torvalds 已提交
679 680 681 682 683 684 685
static int copy_mm(unsigned long clone_flags, struct task_struct * tsk)
{
	struct mm_struct * mm, *oldmm;
	int retval;

	tsk->min_flt = tsk->maj_flt = 0;
	tsk->nvcsw = tsk->nivcsw = 0;
686 687 688
#ifdef CONFIG_DETECT_HUNG_TASK
	tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw;
#endif
L
Linus Torvalds 已提交
689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708

	tsk->mm = NULL;
	tsk->active_mm = NULL;

	/*
	 * Are we cloning a kernel thread?
	 *
	 * We need to steal a active VM for that..
	 */
	oldmm = current->mm;
	if (!oldmm)
		return 0;

	if (clone_flags & CLONE_VM) {
		atomic_inc(&oldmm->mm_users);
		mm = oldmm;
		goto good_mm;
	}

	retval = -ENOMEM;
709
	mm = dup_mm(tsk);
L
Linus Torvalds 已提交
710 711 712 713
	if (!mm)
		goto fail_nomem;

good_mm:
714 715 716 717
	/* Initializing for Swap token stuff */
	mm->token_priority = 0;
	mm->last_interval = 0;

L
Linus Torvalds 已提交
718 719 720 721 722 723 724 725
	tsk->mm = mm;
	tsk->active_mm = mm;
	return 0;

fail_nomem:
	return retval;
}

A
Alexey Dobriyan 已提交
726
static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
727
{
A
Al Viro 已提交
728
	struct fs_struct *fs = current->fs;
L
Linus Torvalds 已提交
729
	if (clone_flags & CLONE_FS) {
A
Al Viro 已提交
730 731 732 733 734 735 736 737
		/* tsk->fs is already what we want */
		write_lock(&fs->lock);
		if (fs->in_exec) {
			write_unlock(&fs->lock);
			return -EAGAIN;
		}
		fs->users++;
		write_unlock(&fs->lock);
L
Linus Torvalds 已提交
738 739
		return 0;
	}
A
Al Viro 已提交
740
	tsk->fs = copy_fs_struct(fs);
L
Linus Torvalds 已提交
741 742 743 744 745
	if (!tsk->fs)
		return -ENOMEM;
	return 0;
}

746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
static int copy_files(unsigned long clone_flags, struct task_struct * tsk)
{
	struct files_struct *oldf, *newf;
	int error = 0;

	/*
	 * A background process may not have any files ...
	 */
	oldf = current->files;
	if (!oldf)
		goto out;

	if (clone_flags & CLONE_FILES) {
		atomic_inc(&oldf->count);
		goto out;
	}

	newf = dup_fd(oldf, &error);
	if (!newf)
		goto out;

	tsk->files = newf;
	error = 0;
out:
	return error;
}

773
static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
774 775 776 777 778 779
{
#ifdef CONFIG_BLOCK
	struct io_context *ioc = current->io_context;

	if (!ioc)
		return 0;
780 781 782 783 784 785 786 787
	/*
	 * Share io context with parent, if CLONE_IO is set
	 */
	if (clone_flags & CLONE_IO) {
		tsk->io_context = ioc_task_link(ioc);
		if (unlikely(!tsk->io_context))
			return -ENOMEM;
	} else if (ioprio_valid(ioc->ioprio)) {
788 789 790 791 792 793 794 795 796 797
		tsk->io_context = alloc_io_context(GFP_KERNEL, -1);
		if (unlikely(!tsk->io_context))
			return -ENOMEM;

		tsk->io_context->ioprio = ioc->ioprio;
	}
#endif
	return 0;
}

A
Alexey Dobriyan 已提交
798
static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
799 800 801
{
	struct sighand_struct *sig;

Z
Zhaolei 已提交
802
	if (clone_flags & CLONE_SIGHAND) {
L
Linus Torvalds 已提交
803 804 805 806
		atomic_inc(&current->sighand->count);
		return 0;
	}
	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
I
Ingo Molnar 已提交
807
	rcu_assign_pointer(tsk->sighand, sig);
L
Linus Torvalds 已提交
808 809 810 811 812 813 814
	if (!sig)
		return -ENOMEM;
	atomic_set(&sig->count, 1);
	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
	return 0;
}

815
void __cleanup_sighand(struct sighand_struct *sighand)
816 817 818 819 820
{
	if (atomic_dec_and_test(&sighand->count))
		kmem_cache_free(sighand_cachep, sighand);
}

821 822 823 824 825 826 827 828 829 830

/*
 * Initialize POSIX timer handling for a thread group.
 */
static void posix_cpu_timers_init_group(struct signal_struct *sig)
{
	/* Thread group counters. */
	thread_group_cputime_init(sig);

	/* Expiration times and increments. */
831 832 833 834
	sig->it[CPUCLOCK_PROF].expires = cputime_zero;
	sig->it[CPUCLOCK_PROF].incr = cputime_zero;
	sig->it[CPUCLOCK_VIRT].expires = cputime_zero;
	sig->it[CPUCLOCK_VIRT].incr = cputime_zero;
835 836 837 838 839 840

	/* Cached expiration times. */
	sig->cputime_expires.prof_exp = cputime_zero;
	sig->cputime_expires.virt_exp = cputime_zero;
	sig->cputime_expires.sched_exp = 0;

841 842 843 844 845 846
	if (sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY) {
		sig->cputime_expires.prof_exp =
			secs_to_cputime(sig->rlim[RLIMIT_CPU].rlim_cur);
		sig->cputimer.running = 1;
	}

847 848 849 850 851 852
	/* The timer lists. */
	INIT_LIST_HEAD(&sig->cpu_timers[0]);
	INIT_LIST_HEAD(&sig->cpu_timers[1]);
	INIT_LIST_HEAD(&sig->cpu_timers[2]);
}

A
Alexey Dobriyan 已提交
853
static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
854 855 856
{
	struct signal_struct *sig;

857
	if (clone_flags & CLONE_THREAD)
858 859
		return 0;

860
	sig = kmem_cache_alloc(signal_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
861 862 863 864 865 866 867 868
	tsk->signal = sig;
	if (!sig)
		return -ENOMEM;

	atomic_set(&sig->count, 1);
	atomic_set(&sig->live, 1);
	init_waitqueue_head(&sig->wait_chldexit);
	sig->flags = 0;
869 870
	if (clone_flags & CLONE_NEWPID)
		sig->flags |= SIGNAL_UNKILLABLE;
L
Linus Torvalds 已提交
871 872 873
	sig->group_exit_code = 0;
	sig->group_exit_task = NULL;
	sig->group_stop_count = 0;
874
	sig->curr_target = tsk;
L
Linus Torvalds 已提交
875 876 877
	init_sigpending(&sig->shared_pending);
	INIT_LIST_HEAD(&sig->posix_timers);

878
	hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
879
	sig->it_real_incr.tv64 = 0;
L
Linus Torvalds 已提交
880 881 882
	sig->real_timer.function = it_real_fn;

	sig->leader = 0;	/* session leadership doesn't inherit */
883
	sig->tty_old_pgrp = NULL;
A
Alan Cox 已提交
884
	sig->tty = NULL;
L
Linus Torvalds 已提交
885

886
	sig->utime = sig->stime = sig->cutime = sig->cstime = cputime_zero;
887 888
	sig->gtime = cputime_zero;
	sig->cgtime = cputime_zero;
889 890 891
#ifndef CONFIG_VIRT_CPU_ACCOUNTING
	sig->prev_utime = sig->prev_stime = cputime_zero;
#endif
L
Linus Torvalds 已提交
892 893
	sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0;
	sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0;
894
	sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0;
J
Jiri Pirko 已提交
895
	sig->maxrss = sig->cmaxrss = 0;
896
	task_io_accounting_init(&sig->ioac);
897
	sig->sum_sched_runtime = 0;
898
	taskstats_tgid_init(sig);
L
Linus Torvalds 已提交
899 900 901 902 903

	task_lock(current->group_leader);
	memcpy(sig->rlim, current->signal->rlim, sizeof sig->rlim);
	task_unlock(current->group_leader);

904 905
	posix_cpu_timers_init_group(sig);

906
	acct_init_pacct(&sig->pacct);
L
Linus Torvalds 已提交
907

M
Miloslav Trmac 已提交
908 909
	tty_audit_fork(sig);

910 911
	sig->oom_adj = current->signal->oom_adj;

L
Linus Torvalds 已提交
912 913 914
	return 0;
}

915 916
void __cleanup_signal(struct signal_struct *sig)
{
917
	thread_group_cputime_free(sig);
A
Alan Cox 已提交
918
	tty_kref_put(sig->tty);
919 920 921
	kmem_cache_free(signal_cachep, sig);
}

A
Alexey Dobriyan 已提交
922
static void copy_flags(unsigned long clone_flags, struct task_struct *p)
L
Linus Torvalds 已提交
923 924 925
{
	unsigned long new_flags = p->flags;

926
	new_flags &= ~PF_SUPERPRIV;
L
Linus Torvalds 已提交
927
	new_flags |= PF_FORKNOEXEC;
R
Roland McGrath 已提交
928
	new_flags |= PF_STARTING;
L
Linus Torvalds 已提交
929
	p->flags = new_flags;
930
	clear_freeze_flag(p);
L
Linus Torvalds 已提交
931 932
}

933
SYSCALL_DEFINE1(set_tid_address, int __user *, tidptr)
L
Linus Torvalds 已提交
934 935 936
{
	current->clear_child_tid = tidptr;

937
	return task_pid_vnr(current);
L
Linus Torvalds 已提交
938 939
}

A
Alexey Dobriyan 已提交
940
static void rt_mutex_init_task(struct task_struct *p)
I
Ingo Molnar 已提交
941
{
942
	raw_spin_lock_init(&p->pi_lock);
943
#ifdef CONFIG_RT_MUTEXES
944
	plist_head_init_raw(&p->pi_waiters, &p->pi_lock);
I
Ingo Molnar 已提交
945 946 947 948
	p->pi_blocked_on = NULL;
#endif
}

949 950 951 952 953 954 955
#ifdef CONFIG_MM_OWNER
void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
{
	mm->owner = p;
}
#endif /* CONFIG_MM_OWNER */

956 957 958 959 960 961 962 963 964 965 966 967 968
/*
 * Initialize POSIX timer handling for a single task.
 */
static void posix_cpu_timers_init(struct task_struct *tsk)
{
	tsk->cputime_expires.prof_exp = cputime_zero;
	tsk->cputime_expires.virt_exp = cputime_zero;
	tsk->cputime_expires.sched_exp = 0;
	INIT_LIST_HEAD(&tsk->cpu_timers[0]);
	INIT_LIST_HEAD(&tsk->cpu_timers[1]);
	INIT_LIST_HEAD(&tsk->cpu_timers[2]);
}

L
Linus Torvalds 已提交
969 970 971 972 973 974 975 976
/*
 * This creates a new process as a copy of the old one,
 * but does not actually start it yet.
 *
 * It copies the registers, and all the appropriate
 * parts of the process environment (as per the clone
 * flags). The actual kick-off is left to the caller.
 */
977 978 979 980 981
static struct task_struct *copy_process(unsigned long clone_flags,
					unsigned long stack_start,
					struct pt_regs *regs,
					unsigned long stack_size,
					int __user *child_tidptr,
R
Roland McGrath 已提交
982 983
					struct pid *pid,
					int trace)
L
Linus Torvalds 已提交
984 985
{
	int retval;
986
	struct task_struct *p;
987
	int cgroup_callbacks_done = 0;
L
Linus Torvalds 已提交
988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006

	if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
		return ERR_PTR(-EINVAL);

	/*
	 * Thread groups must share signals as well, and detached threads
	 * can only be started up within the thread group.
	 */
	if ((clone_flags & CLONE_THREAD) && !(clone_flags & CLONE_SIGHAND))
		return ERR_PTR(-EINVAL);

	/*
	 * Shared signal handlers imply shared VM. By way of the above,
	 * thread groups also imply shared VM. Blocking this case allows
	 * for various simplifications in other code.
	 */
	if ((clone_flags & CLONE_SIGHAND) && !(clone_flags & CLONE_VM))
		return ERR_PTR(-EINVAL);

1007 1008 1009 1010 1011 1012 1013 1014 1015 1016
	/*
	 * Siblings of global init remain as zombies on exit since they are
	 * not reaped by their parent (swapper). To solve this and to avoid
	 * multi-rooted process trees, prevent global and container-inits
	 * from creating siblings.
	 */
	if ((clone_flags & CLONE_PARENT) &&
				current->signal->flags & SIGNAL_UNKILLABLE)
		return ERR_PTR(-EINVAL);

L
Linus Torvalds 已提交
1017 1018 1019 1020 1021 1022 1023 1024 1025
	retval = security_task_create(clone_flags);
	if (retval)
		goto fork_out;

	retval = -ENOMEM;
	p = dup_task_struct(current);
	if (!p)
		goto fork_out;

1026 1027
	ftrace_graph_init_task(p);

1028 1029
	rt_mutex_init_task(p);

I
Ingo Molnar 已提交
1030
#ifdef CONFIG_PROVE_LOCKING
1031 1032 1033
	DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
#endif
L
Linus Torvalds 已提交
1034
	retval = -EAGAIN;
1035
	if (atomic_read(&p->real_cred->user->processes) >=
L
Linus Torvalds 已提交
1036 1037
			p->signal->rlim[RLIMIT_NPROC].rlim_cur) {
		if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
1038
		    p->real_cred->user != INIT_USER)
L
Linus Torvalds 已提交
1039 1040 1041
			goto bad_fork_free;
	}

1042 1043 1044
	retval = copy_creds(p, clone_flags);
	if (retval < 0)
		goto bad_fork_free;
L
Linus Torvalds 已提交
1045 1046 1047 1048 1049 1050

	/*
	 * If multiple threads are within copy_process(), then this check
	 * triggers too late. This doesn't hurt, the check is only there
	 * to stop root fork bombs.
	 */
1051
	retval = -EAGAIN;
L
Linus Torvalds 已提交
1052 1053 1054
	if (nr_threads >= max_threads)
		goto bad_fork_cleanup_count;

A
Al Viro 已提交
1055
	if (!try_module_get(task_thread_info(p)->exec_domain->module))
L
Linus Torvalds 已提交
1056 1057 1058
		goto bad_fork_cleanup_count;

	p->did_exec = 0;
1059
	delayacct_tsk_init(p);	/* Must remain after dup_task_struct() */
L
Linus Torvalds 已提交
1060 1061 1062
	copy_flags(clone_flags, p);
	INIT_LIST_HEAD(&p->children);
	INIT_LIST_HEAD(&p->sibling);
1063
	rcu_copy_process(p);
L
Linus Torvalds 已提交
1064 1065 1066 1067 1068 1069 1070
	p->vfork_done = NULL;
	spin_lock_init(&p->alloc_lock);

	init_sigpending(&p->pending);

	p->utime = cputime_zero;
	p->stime = cputime_zero;
1071
	p->gtime = cputime_zero;
1072 1073
	p->utimescaled = cputime_zero;
	p->stimescaled = cputime_zero;
1074
#ifndef CONFIG_VIRT_CPU_ACCOUNTING
P
Peter Zijlstra 已提交
1075
	p->prev_utime = cputime_zero;
1076
	p->prev_stime = cputime_zero;
1077
#endif
1078

1079 1080
	p->default_timer_slack_ns = current->timer_slack_ns;

1081
	task_io_accounting_init(&p->ioac);
L
Linus Torvalds 已提交
1082 1083
	acct_clear_integrals(p);

1084
	posix_cpu_timers_init(p);
L
Linus Torvalds 已提交
1085 1086 1087

	p->lock_depth = -1;		/* -1 = no lock */
	do_posix_clock_monotonic_gettime(&p->start_time);
1088 1089
	p->real_start_time = p->start_time;
	monotonic_to_bootbased(&p->real_start_time);
L
Linus Torvalds 已提交
1090 1091
	p->io_context = NULL;
	p->audit_context = NULL;
1092
	cgroup_fork(p);
L
Linus Torvalds 已提交
1093
#ifdef CONFIG_NUMA
1094
	p->mempolicy = mpol_dup(p->mempolicy);
L
Linus Torvalds 已提交
1095 1096 1097
 	if (IS_ERR(p->mempolicy)) {
 		retval = PTR_ERR(p->mempolicy);
 		p->mempolicy = NULL;
1098
 		goto bad_fork_cleanup_cgroup;
L
Linus Torvalds 已提交
1099
 	}
1100
	mpol_fix_fork_child_flag(p);
L
Linus Torvalds 已提交
1101
#endif
1102 1103
#ifdef CONFIG_TRACE_IRQFLAGS
	p->irq_events = 0;
1104 1105 1106
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
	p->hardirqs_enabled = 1;
#else
1107
	p->hardirqs_enabled = 0;
1108
#endif
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
	p->hardirq_enable_ip = 0;
	p->hardirq_enable_event = 0;
	p->hardirq_disable_ip = _THIS_IP_;
	p->hardirq_disable_event = 0;
	p->softirqs_enabled = 1;
	p->softirq_enable_ip = _THIS_IP_;
	p->softirq_enable_event = 0;
	p->softirq_disable_ip = 0;
	p->softirq_disable_event = 0;
	p->hardirq_context = 0;
	p->softirq_context = 0;
#endif
I
Ingo Molnar 已提交
1121 1122 1123 1124 1125
#ifdef CONFIG_LOCKDEP
	p->lockdep_depth = 0; /* no locks held yet */
	p->curr_chain_key = 0;
	p->lockdep_recursion = 0;
#endif
L
Linus Torvalds 已提交
1126

1127 1128 1129
#ifdef CONFIG_DEBUG_MUTEXES
	p->blocked_on = NULL; /* not blocked yet */
#endif
1130 1131 1132 1133
#ifdef CONFIG_CGROUP_MEM_RES_CTLR
	p->memcg_batch.do_batch = 0;
	p->memcg_batch.memcg = NULL;
#endif
1134 1135

	p->bts = NULL;
1136

1137 1138
	p->stack_start = stack_start;

1139 1140
	/* Perform scheduler related setup. Assign this task to a CPU. */
	sched_fork(p, clone_flags);
1141

1142
	retval = perf_event_init_task(p);
1143 1144
	if (retval)
		goto bad_fork_cleanup_policy;
1145

L
Linus Torvalds 已提交
1146
	if ((retval = audit_alloc(p)))
1147
		goto bad_fork_cleanup_policy;
L
Linus Torvalds 已提交
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
	/* copy all the process information */
	if ((retval = copy_semundo(clone_flags, p)))
		goto bad_fork_cleanup_audit;
	if ((retval = copy_files(clone_flags, p)))
		goto bad_fork_cleanup_semundo;
	if ((retval = copy_fs(clone_flags, p)))
		goto bad_fork_cleanup_files;
	if ((retval = copy_sighand(clone_flags, p)))
		goto bad_fork_cleanup_fs;
	if ((retval = copy_signal(clone_flags, p)))
		goto bad_fork_cleanup_sighand;
	if ((retval = copy_mm(clone_flags, p)))
		goto bad_fork_cleanup_signal;
S
Serge E. Hallyn 已提交
1161
	if ((retval = copy_namespaces(clone_flags, p)))
D
David Howells 已提交
1162
		goto bad_fork_cleanup_mm;
1163
	if ((retval = copy_io(clone_flags, p)))
1164
		goto bad_fork_cleanup_namespaces;
A
Alexey Dobriyan 已提交
1165
	retval = copy_thread(clone_flags, stack_start, stack_size, p, regs);
L
Linus Torvalds 已提交
1166
	if (retval)
1167
		goto bad_fork_cleanup_io;
L
Linus Torvalds 已提交
1168

1169 1170
	if (pid != &init_struct_pid) {
		retval = -ENOMEM;
1171
		pid = alloc_pid(p->nsproxy->pid_ns);
1172
		if (!pid)
1173
			goto bad_fork_cleanup_io;
1174 1175

		if (clone_flags & CLONE_NEWPID) {
1176
			retval = pid_ns_prepare_proc(p->nsproxy->pid_ns);
1177 1178 1179
			if (retval < 0)
				goto bad_fork_free_pid;
		}
1180 1181 1182 1183 1184 1185 1186
	}

	p->pid = pid_nr(pid);
	p->tgid = p->pid;
	if (clone_flags & CLONE_THREAD)
		p->tgid = current->tgid;

1187 1188 1189
	if (current->nsproxy != p->nsproxy) {
		retval = ns_cgroup_clone(p, pid);
		if (retval)
1190
			goto bad_fork_free_pid;
1191 1192
	}

L
Linus Torvalds 已提交
1193 1194 1195 1196 1197
	p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL;
	/*
	 * Clear TID on mm_release()?
	 */
	p->clear_child_tid = (clone_flags & CLONE_CHILD_CLEARTID) ? child_tidptr: NULL;
1198
#ifdef CONFIG_FUTEX
1199 1200 1201 1202
	p->robust_list = NULL;
#ifdef CONFIG_COMPAT
	p->compat_robust_list = NULL;
#endif
1203 1204
	INIT_LIST_HEAD(&p->pi_state_list);
	p->pi_state_cache = NULL;
1205
#endif
1206 1207 1208 1209 1210 1211
	/*
	 * sigaltstack should be cleared when sharing the same VM
	 */
	if ((clone_flags & (CLONE_VM|CLONE_VFORK)) == CLONE_VM)
		p->sas_ss_sp = p->sas_ss_size = 0;

L
Linus Torvalds 已提交
1212 1213 1214 1215 1216
	/*
	 * Syscall tracing should be turned off in the child regardless
	 * of CLONE_PTRACE.
	 */
	clear_tsk_thread_flag(p, TIF_SYSCALL_TRACE);
1217 1218 1219
#ifdef TIF_SYSCALL_EMU
	clear_tsk_thread_flag(p, TIF_SYSCALL_EMU);
#endif
A
Arjan van de Ven 已提交
1220
	clear_all_latency_tracing(p);
L
Linus Torvalds 已提交
1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231

	/* ok, now we should be set up.. */
	p->exit_signal = (clone_flags & CLONE_THREAD) ? -1 : (clone_flags & CSIGNAL);
	p->pdeath_signal = 0;
	p->exit_state = 0;

	/*
	 * Ok, make it visible to the rest of the system.
	 * We dont wake it up yet.
	 */
	p->group_leader = p;
O
Oleg Nesterov 已提交
1232
	INIT_LIST_HEAD(&p->thread_group);
L
Linus Torvalds 已提交
1233

1234 1235 1236 1237 1238 1239
	/* Now that the task is set up, run cgroup callbacks if
	 * necessary. We need to run them before the task is visible
	 * on the tasklist. */
	cgroup_fork_callbacks(p);
	cgroup_callbacks_done = 1;

L
Linus Torvalds 已提交
1240 1241 1242 1243
	/* Need tasklist lock for parent etc handling! */
	write_lock_irq(&tasklist_lock);

	/*
N
Nick Piggin 已提交
1244 1245 1246 1247 1248 1249 1250
	 * The task hasn't been attached yet, so its cpus_allowed mask will
	 * not be changed, nor will its assigned CPU.
	 *
	 * The cpus_allowed mask of the parent may have changed after it was
	 * copied first time - so re-copy it here, then check the child's CPU
	 * to ensure it is on a valid CPU (and if not, just force it back to
	 * parent's CPU). This avoids alot of nasty races.
L
Linus Torvalds 已提交
1251 1252
	 */
	p->cpus_allowed = current->cpus_allowed;
P
Peter Zijlstra 已提交
1253
	p->rt.nr_cpus_allowed = current->rt.nr_cpus_allowed;
1254 1255
	if (unlikely(!cpu_isset(task_cpu(p), p->cpus_allowed) ||
			!cpu_online(task_cpu(p))))
N
Nick Piggin 已提交
1256
		set_task_cpu(p, smp_processor_id());
L
Linus Torvalds 已提交
1257 1258

	/* CLONE_PARENT re-uses the old parent */
1259
	if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) {
L
Linus Torvalds 已提交
1260
		p->real_parent = current->real_parent;
1261 1262
		p->parent_exec_id = current->parent_exec_id;
	} else {
L
Linus Torvalds 已提交
1263
		p->real_parent = current;
1264 1265
		p->parent_exec_id = current->self_exec_id;
	}
L
Linus Torvalds 已提交
1266

1267
	spin_lock(&current->sighand->siglock);
1268 1269 1270 1271 1272 1273 1274 1275 1276

	/*
	 * Process group and session signals need to be delivered to just the
	 * parent before the fork or both the parent and the child after the
	 * fork. Restart if a signal comes in before we add the new process to
	 * it's process group.
	 * A fatal signal pending means that current will exit, so the new
	 * thread can't slip out of an OOM kill (or normal SIGKILL).
 	 */
D
Daniel Walker 已提交
1277
	recalc_sigpending();
1278 1279 1280 1281
	if (signal_pending(current)) {
		spin_unlock(&current->sighand->siglock);
		write_unlock_irq(&tasklist_lock);
		retval = -ERESTARTNOINTR;
1282
		goto bad_fork_free_pid;
1283 1284
	}

L
Linus Torvalds 已提交
1285
	if (clone_flags & CLONE_THREAD) {
1286 1287
		atomic_inc(&current->signal->count);
		atomic_inc(&current->signal->live);
L
Linus Torvalds 已提交
1288
		p->group_leader = current->group_leader;
O
Oleg Nesterov 已提交
1289
		list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group);
L
Linus Torvalds 已提交
1290 1291
	}

1292
	if (likely(p->pid)) {
R
Roland McGrath 已提交
1293
		list_add_tail(&p->sibling, &p->real_parent->children);
R
Roland McGrath 已提交
1294
		tracehook_finish_clone(p, clone_flags, trace);
1295 1296

		if (thread_group_leader(p)) {
E
Eric W. Biederman 已提交
1297
			if (clone_flags & CLONE_NEWPID)
1298
				p->nsproxy->pid_ns->child_reaper = p;
1299

1300
			p->signal->leader_pid = pid;
A
Alan Cox 已提交
1301 1302
			tty_kref_put(p->signal->tty);
			p->signal->tty = tty_kref_get(current->signal->tty);
E
Eric W. Biederman 已提交
1303 1304
			attach_pid(p, PIDTYPE_PGID, task_pgrp(current));
			attach_pid(p, PIDTYPE_SID, task_session(current));
1305
			list_add_tail_rcu(&p->tasks, &init_task.tasks);
L
Linus Torvalds 已提交
1306
			__get_cpu_var(process_counts)++;
1307
		}
1308
		attach_pid(p, PIDTYPE_PID, pid);
1309
		nr_threads++;
L
Linus Torvalds 已提交
1310 1311 1312
	}

	total_forks++;
1313
	spin_unlock(&current->sighand->siglock);
L
Linus Torvalds 已提交
1314
	write_unlock_irq(&tasklist_lock);
1315
	proc_fork_connector(p);
1316
	cgroup_post_fork(p);
1317
	perf_event_fork(p);
L
Linus Torvalds 已提交
1318 1319
	return p;

1320 1321 1322
bad_fork_free_pid:
	if (pid != &init_struct_pid)
		free_pid(pid);
1323
bad_fork_cleanup_io:
1324 1325
	if (p->io_context)
		exit_io_context(p);
S
Serge E. Hallyn 已提交
1326
bad_fork_cleanup_namespaces:
1327
	exit_task_namespaces(p);
L
Linus Torvalds 已提交
1328 1329 1330 1331
bad_fork_cleanup_mm:
	if (p->mm)
		mmput(p->mm);
bad_fork_cleanup_signal:
1332 1333
	if (!(clone_flags & CLONE_THREAD))
		__cleanup_signal(p->signal);
L
Linus Torvalds 已提交
1334
bad_fork_cleanup_sighand:
1335
	__cleanup_sighand(p->sighand);
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340 1341 1342 1343 1344
bad_fork_cleanup_fs:
	exit_fs(p); /* blocking */
bad_fork_cleanup_files:
	exit_files(p); /* blocking */
bad_fork_cleanup_semundo:
	exit_sem(p);
bad_fork_cleanup_audit:
	audit_free(p);
bad_fork_cleanup_policy:
1345
	perf_event_free_task(p);
L
Linus Torvalds 已提交
1346
#ifdef CONFIG_NUMA
1347
	mpol_put(p->mempolicy);
1348
bad_fork_cleanup_cgroup:
L
Linus Torvalds 已提交
1349
#endif
1350
	cgroup_exit(p, cgroup_callbacks_done);
1351
	delayacct_tsk_free(p);
A
Al Viro 已提交
1352
	module_put(task_thread_info(p)->exec_domain->module);
L
Linus Torvalds 已提交
1353
bad_fork_cleanup_count:
D
David Howells 已提交
1354
	atomic_dec(&p->cred->user->processes);
1355
	exit_creds(p);
L
Linus Torvalds 已提交
1356 1357
bad_fork_free:
	free_task(p);
1358 1359
fork_out:
	return ERR_PTR(retval);
L
Linus Torvalds 已提交
1360 1361
}

A
Adrian Bunk 已提交
1362
noinline struct pt_regs * __cpuinit __attribute__((weak)) idle_regs(struct pt_regs *regs)
L
Linus Torvalds 已提交
1363 1364 1365 1366 1367
{
	memset(regs, 0, sizeof(struct pt_regs));
	return regs;
}

1368
struct task_struct * __cpuinit fork_idle(int cpu)
L
Linus Torvalds 已提交
1369
{
1370
	struct task_struct *task;
L
Linus Torvalds 已提交
1371 1372
	struct pt_regs regs;

1373
	task = copy_process(CLONE_VM, 0, idle_regs(&regs), 0, NULL,
R
Roland McGrath 已提交
1374
			    &init_struct_pid, 0);
1375 1376
	if (!IS_ERR(task))
		init_idle(task, cpu);
1377

L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
	return task;
}

/*
 *  Ok, this is the main fork-routine.
 *
 * It copies the process, and if successful kick-starts
 * it and waits for it to finish using the VM if required.
 */
long do_fork(unsigned long clone_flags,
	      unsigned long stack_start,
	      struct pt_regs *regs,
	      unsigned long stack_size,
	      int __user *parent_tidptr,
	      int __user *child_tidptr)
{
	struct task_struct *p;
	int trace = 0;
1396
	long nr;
L
Linus Torvalds 已提交
1397

1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
	/*
	 * Do some preliminary argument and permissions checking before we
	 * actually start allocating stuff
	 */
	if (clone_flags & CLONE_NEWUSER) {
		if (clone_flags & CLONE_THREAD)
			return -EINVAL;
		/* hopefully this check will go away when userns support is
		 * complete
		 */
1408 1409
		if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SETUID) ||
				!capable(CAP_SETGID))
1410 1411 1412
			return -EPERM;
	}

1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
	/*
	 * We hope to recycle these flags after 2.6.26
	 */
	if (unlikely(clone_flags & CLONE_STOPPED)) {
		static int __read_mostly count = 100;

		if (count > 0 && printk_ratelimit()) {
			char comm[TASK_COMM_LEN];

			count--;
			printk(KERN_INFO "fork(): process `%s' used deprecated "
					"clone flags 0x%lx\n",
				get_task_comm(comm, current),
				clone_flags & CLONE_STOPPED);
		}
	}

R
Roland McGrath 已提交
1430 1431 1432 1433 1434
	/*
	 * When called from kernel_thread, don't do user tracing stuff.
	 */
	if (likely(user_mode(regs)))
		trace = tracehook_prepare_clone(clone_flags);
L
Linus Torvalds 已提交
1435

1436
	p = copy_process(clone_flags, stack_start, regs, stack_size,
R
Roland McGrath 已提交
1437
			 child_tidptr, NULL, trace);
L
Linus Torvalds 已提交
1438 1439 1440 1441 1442 1443 1444
	/*
	 * Do this prior waking up the new thread - the thread pointer
	 * might get invalid after that point, if the thread exits quickly.
	 */
	if (!IS_ERR(p)) {
		struct completion vfork;

1445 1446
		trace_sched_process_fork(current, p);

1447
		nr = task_pid_vnr(p);
1448 1449 1450

		if (clone_flags & CLONE_PARENT_SETTID)
			put_user(nr, parent_tidptr);
1451

L
Linus Torvalds 已提交
1452 1453 1454 1455 1456
		if (clone_flags & CLONE_VFORK) {
			p->vfork_done = &vfork;
			init_completion(&vfork);
		}

1457
		audit_finish_fork(p);
1458
		tracehook_report_clone(regs, clone_flags, nr, p);
R
Roland McGrath 已提交
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468

		/*
		 * We set PF_STARTING at creation in case tracing wants to
		 * use this to distinguish a fully live task from one that
		 * hasn't gotten to tracehook_report_clone() yet.  Now we
		 * clear it and set the child going.
		 */
		p->flags &= ~PF_STARTING;

		if (unlikely(clone_flags & CLONE_STOPPED)) {
L
Linus Torvalds 已提交
1469 1470 1471 1472 1473
			/*
			 * We'll start up with an immediate SIGSTOP.
			 */
			sigaddset(&p->pending.signal, SIGSTOP);
			set_tsk_thread_flag(p, TIF_SIGPENDING);
1474
			__set_task_state(p, TASK_STOPPED);
R
Roland McGrath 已提交
1475 1476
		} else {
			wake_up_new_task(p, clone_flags);
L
Linus Torvalds 已提交
1477 1478
		}

R
Roland McGrath 已提交
1479 1480 1481
		tracehook_report_clone_complete(trace, regs,
						clone_flags, nr, p);

L
Linus Torvalds 已提交
1482
		if (clone_flags & CLONE_VFORK) {
R
Rafael J. Wysocki 已提交
1483
			freezer_do_not_count();
L
Linus Torvalds 已提交
1484
			wait_for_completion(&vfork);
R
Rafael J. Wysocki 已提交
1485
			freezer_count();
R
Roland McGrath 已提交
1486
			tracehook_report_vfork_done(p, nr);
L
Linus Torvalds 已提交
1487 1488
		}
	} else {
1489
		nr = PTR_ERR(p);
L
Linus Torvalds 已提交
1490
	}
1491
	return nr;
L
Linus Torvalds 已提交
1492 1493
}

1494 1495 1496 1497
#ifndef ARCH_MIN_MMSTRUCT_ALIGN
#define ARCH_MIN_MMSTRUCT_ALIGN 0
#endif

1498
static void sighand_ctor(void *data)
1499 1500 1501
{
	struct sighand_struct *sighand = data;

C
Christoph Lameter 已提交
1502
	spin_lock_init(&sighand->siglock);
D
Davide Libenzi 已提交
1503
	init_waitqueue_head(&sighand->signalfd_wqh);
1504 1505
}

L
Linus Torvalds 已提交
1506 1507 1508 1509
void __init proc_caches_init(void)
{
	sighand_cachep = kmem_cache_create("sighand_cache",
			sizeof(struct sighand_struct), 0,
V
Vegard Nossum 已提交
1510 1511
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_DESTROY_BY_RCU|
			SLAB_NOTRACK, sighand_ctor);
L
Linus Torvalds 已提交
1512 1513
	signal_cachep = kmem_cache_create("signal_cache",
			sizeof(struct signal_struct), 0,
V
Vegard Nossum 已提交
1514
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1515
	files_cachep = kmem_cache_create("files_cache",
L
Linus Torvalds 已提交
1516
			sizeof(struct files_struct), 0,
V
Vegard Nossum 已提交
1517
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1518
	fs_cachep = kmem_cache_create("fs_cache",
L
Linus Torvalds 已提交
1519
			sizeof(struct fs_struct), 0,
V
Vegard Nossum 已提交
1520
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
L
Linus Torvalds 已提交
1521
	mm_cachep = kmem_cache_create("mm_struct",
1522
			sizeof(struct mm_struct), ARCH_MIN_MMSTRUCT_ALIGN,
V
Vegard Nossum 已提交
1523
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1524
	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC);
1525
	mmap_init();
L
Linus Torvalds 已提交
1526
}
1527 1528 1529 1530 1531

/*
 * Check constraints on flags passed to the unshare system call and
 * force unsharing of additional process context as appropriate.
 */
A
Alexey Dobriyan 已提交
1532
static void check_unshare_flags(unsigned long *flags_ptr)
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573
{
	/*
	 * If unsharing a thread from a thread group, must also
	 * unshare vm.
	 */
	if (*flags_ptr & CLONE_THREAD)
		*flags_ptr |= CLONE_VM;

	/*
	 * If unsharing vm, must also unshare signal handlers.
	 */
	if (*flags_ptr & CLONE_VM)
		*flags_ptr |= CLONE_SIGHAND;

	/*
	 * If unsharing signal handlers and the task was created
	 * using CLONE_THREAD, then must unshare the thread
	 */
	if ((*flags_ptr & CLONE_SIGHAND) &&
	    (atomic_read(&current->signal->count) > 1))
		*flags_ptr |= CLONE_THREAD;

	/*
	 * If unsharing namespace, must also unshare filesystem information.
	 */
	if (*flags_ptr & CLONE_NEWNS)
		*flags_ptr |= CLONE_FS;
}

/*
 * Unsharing of tasks created with CLONE_THREAD is not supported yet
 */
static int unshare_thread(unsigned long unshare_flags)
{
	if (unshare_flags & CLONE_THREAD)
		return -EINVAL;

	return 0;
}

/*
1574
 * Unshare the filesystem structure if it is being shared
1575 1576 1577 1578 1579
 */
static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp)
{
	struct fs_struct *fs = current->fs;

A
Al Viro 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
	if (!(unshare_flags & CLONE_FS) || !fs)
		return 0;

	/* don't need lock here; in the worst case we'll do useless copy */
	if (fs->users == 1)
		return 0;

	*new_fsp = copy_fs_struct(fs);
	if (!*new_fsp)
		return -ENOMEM;
1590 1591 1592 1593 1594

	return 0;
}

/*
1595
 * Unsharing of sighand is not supported yet
1596 1597 1598 1599 1600
 */
static int unshare_sighand(unsigned long unshare_flags, struct sighand_struct **new_sighp)
{
	struct sighand_struct *sigh = current->sighand;

1601
	if ((unshare_flags & CLONE_SIGHAND) && atomic_read(&sigh->count) > 1)
1602 1603 1604 1605 1606 1607
		return -EINVAL;
	else
		return 0;
}

/*
1608
 * Unshare vm if it is being shared
1609 1610 1611 1612 1613 1614
 */
static int unshare_vm(unsigned long unshare_flags, struct mm_struct **new_mmp)
{
	struct mm_struct *mm = current->mm;

	if ((unshare_flags & CLONE_VM) &&
1615
	    (mm && atomic_read(&mm->mm_users) > 1)) {
1616
		return -EINVAL;
1617
	}
1618 1619 1620 1621 1622

	return 0;
}

/*
1623
 * Unshare file descriptor table if it is being shared
1624 1625 1626 1627
 */
static int unshare_fd(unsigned long unshare_flags, struct files_struct **new_fdp)
{
	struct files_struct *fd = current->files;
1628
	int error = 0;
1629 1630

	if ((unshare_flags & CLONE_FILES) &&
1631 1632 1633 1634 1635
	    (fd && atomic_read(&fd->count) > 1)) {
		*new_fdp = dup_fd(fd, &error);
		if (!*new_fdp)
			return error;
	}
1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647

	return 0;
}

/*
 * unshare allows a process to 'unshare' part of the process
 * context which was originally shared using clone.  copy_*
 * functions used by do_fork() cannot be used here directly
 * because they modify an inactive task_struct that is being
 * constructed. Here we are modifying the current, active,
 * task_struct.
 */
1648
SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags)
1649 1650 1651
{
	int err = 0;
	struct fs_struct *fs, *new_fs = NULL;
1652
	struct sighand_struct *new_sigh = NULL;
1653 1654
	struct mm_struct *mm, *new_mm = NULL, *active_mm = NULL;
	struct files_struct *fd, *new_fd = NULL;
1655
	struct nsproxy *new_nsproxy = NULL;
1656
	int do_sysvsem = 0;
1657 1658 1659

	check_unshare_flags(&unshare_flags);

1660 1661 1662
	/* Return -EINVAL for all unsupported flags */
	err = -EINVAL;
	if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND|
K
Kirill Korotaev 已提交
1663
				CLONE_VM|CLONE_FILES|CLONE_SYSVSEM|
1664
				CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET))
1665 1666
		goto bad_unshare_out;

1667 1668 1669 1670 1671 1672
	/*
	 * CLONE_NEWIPC must also detach from the undolist: after switching
	 * to a new ipc namespace, the semaphore arrays from the old
	 * namespace are unreachable.
	 */
	if (unshare_flags & (CLONE_NEWIPC|CLONE_SYSVSEM))
1673
		do_sysvsem = 1;
1674 1675 1676 1677 1678
	if ((err = unshare_thread(unshare_flags)))
		goto bad_unshare_out;
	if ((err = unshare_fs(unshare_flags, &new_fs)))
		goto bad_unshare_cleanup_thread;
	if ((err = unshare_sighand(unshare_flags, &new_sigh)))
1679
		goto bad_unshare_cleanup_fs;
1680 1681 1682 1683
	if ((err = unshare_vm(unshare_flags, &new_mm)))
		goto bad_unshare_cleanup_sigh;
	if ((err = unshare_fd(unshare_flags, &new_fd)))
		goto bad_unshare_cleanup_vm;
1684 1685
	if ((err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy,
			new_fs)))
1686
		goto bad_unshare_cleanup_fd;
1687

1688 1689 1690 1691 1692 1693 1694
	if (new_fs ||  new_mm || new_fd || do_sysvsem || new_nsproxy) {
		if (do_sysvsem) {
			/*
			 * CLONE_SYSVSEM is equivalent to sys_exit().
			 */
			exit_sem(current);
		}
S
Serge E. Hallyn 已提交
1695

1696
		if (new_nsproxy) {
1697 1698
			switch_task_namespaces(current, new_nsproxy);
			new_nsproxy = NULL;
1699
		}
1700

1701 1702
		task_lock(current);

1703 1704
		if (new_fs) {
			fs = current->fs;
A
Al Viro 已提交
1705
			write_lock(&fs->lock);
1706
			current->fs = new_fs;
A
Al Viro 已提交
1707 1708 1709 1710 1711
			if (--fs->users)
				new_fs = NULL;
			else
				new_fs = fs;
			write_unlock(&fs->lock);
1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731
		}

		if (new_mm) {
			mm = current->mm;
			active_mm = current->active_mm;
			current->mm = new_mm;
			current->active_mm = new_mm;
			activate_mm(active_mm, new_mm);
			new_mm = mm;
		}

		if (new_fd) {
			fd = current->files;
			current->files = new_fd;
			new_fd = fd;
		}

		task_unlock(current);
	}

1732
	if (new_nsproxy)
1733
		put_nsproxy(new_nsproxy);
1734

1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749
bad_unshare_cleanup_fd:
	if (new_fd)
		put_files_struct(new_fd);

bad_unshare_cleanup_vm:
	if (new_mm)
		mmput(new_mm);

bad_unshare_cleanup_sigh:
	if (new_sigh)
		if (atomic_dec_and_test(&new_sigh->count))
			kmem_cache_free(sighand_cachep, new_sigh);

bad_unshare_cleanup_fs:
	if (new_fs)
A
Al Viro 已提交
1750
		free_fs_struct(new_fs);
1751 1752 1753 1754 1755

bad_unshare_cleanup_thread:
bad_unshare_out:
	return err;
}
1756 1757 1758 1759 1760 1761 1762 1763 1764 1765

/*
 *	Helper to unshare the files of the current task.
 *	We don't want to expose copy_files internals to
 *	the exec layer of the kernel.
 */

int unshare_files(struct files_struct **displaced)
{
	struct task_struct *task = current;
1766
	struct files_struct *copy = NULL;
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
	int error;

	error = unshare_fd(CLONE_FILES, &copy);
	if (error || !copy) {
		*displaced = NULL;
		return error;
	}
	*displaced = task->files;
	task_lock(task);
	task->files = copy;
	task_unlock(task);
	return 0;
}