fork.c 41.5 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
/*
 *  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>
20
#include <linux/mnt_namespace.h>
L
Linus Torvalds 已提交
21 22 23 24
#include <linux/personality.h>
#include <linux/mempolicy.h>
#include <linux/sem.h>
#include <linux/file.h>
A
Al Viro 已提交
25
#include <linux/fdtable.h>
26
#include <linux/iocontext.h>
L
Linus Torvalds 已提交
27 28 29
#include <linux/key.h>
#include <linux/binfmts.h>
#include <linux/mman.h>
A
Andrea Arcangeli 已提交
30
#include <linux/mmu_notifier.h>
L
Linus Torvalds 已提交
31
#include <linux/fs.h>
S
Serge E. Hallyn 已提交
32
#include <linux/nsproxy.h>
33
#include <linux/capability.h>
L
Linus Torvalds 已提交
34
#include <linux/cpu.h>
35
#include <linux/cgroup.h>
L
Linus Torvalds 已提交
36
#include <linux/security.h>
37
#include <linux/hugetlb.h>
L
Linus Torvalds 已提交
38 39 40
#include <linux/swap.h>
#include <linux/syscalls.h>
#include <linux/jiffies.h>
R
Roland McGrath 已提交
41
#include <linux/tracehook.h>
L
Linus Torvalds 已提交
42
#include <linux/futex.h>
43
#include <linux/compat.h>
44
#include <linux/task_io_accounting_ops.h>
45
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
46 47 48
#include <linux/ptrace.h>
#include <linux/mount.h>
#include <linux/audit.h>
49
#include <linux/memcontrol.h>
50
#include <linux/ftrace.h>
L
Linus Torvalds 已提交
51 52 53
#include <linux/profile.h>
#include <linux/rmap.h>
#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 <trace/sched.h>
64
#include <linux/magic.h>
L
Linus Torvalds 已提交
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

#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>

/*
 * 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;

83
__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
L
Linus Torvalds 已提交
84

85 86
DEFINE_TRACE(sched_process_fork);

L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96 97 98 99 100
int nr_processes(void)
{
	int cpu;
	int total = 0;

	for_each_online_cpu(cpu)
		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))
101
static struct kmem_cache *task_struct_cachep;
L
Linus Torvalds 已提交
102 103
#endif

104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
#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
121
/* SLAB cache for signal_struct structures (tsk->signal) */
122
static struct kmem_cache *signal_cachep;
L
Linus Torvalds 已提交
123 124

/* SLAB cache for sighand_struct structures (tsk->sighand) */
125
struct kmem_cache *sighand_cachep;
L
Linus Torvalds 已提交
126 127

/* SLAB cache for files_struct structures (tsk->files) */
128
struct kmem_cache *files_cachep;
L
Linus Torvalds 已提交
129 130

/* SLAB cache for fs_struct structures (tsk->fs) */
131
struct kmem_cache *fs_cachep;
L
Linus Torvalds 已提交
132 133

/* SLAB cache for vm_area_struct structures */
134
struct kmem_cache *vm_area_cachep;
L
Linus Torvalds 已提交
135 136

/* SLAB cache for mm_struct structures (tsk->mm) */
137
static struct kmem_cache *mm_cachep;
L
Linus Torvalds 已提交
138 139 140

void free_task(struct task_struct *tsk)
{
P
Peter Zijlstra 已提交
141
	prop_local_destroy_single(&tsk->dirties);
R
Roman Zippel 已提交
142
	free_thread_info(tsk->stack);
I
Ingo Molnar 已提交
143
	rt_mutex_debug_task_free(tsk);
144
	ftrace_graph_exit_task(tsk);
L
Linus Torvalds 已提交
145 146 147 148
	free_task_struct(tsk);
}
EXPORT_SYMBOL(free_task);

149
void __put_task_struct(struct task_struct *tsk)
L
Linus Torvalds 已提交
150
{
E
Eugene Teo 已提交
151
	WARN_ON(!tsk->exit_state);
L
Linus Torvalds 已提交
152 153 154
	WARN_ON(atomic_read(&tsk->usage));
	WARN_ON(tsk == current);

155
	put_cred(tsk->real_cred);
156
	put_cred(tsk->cred);
157
	delayacct_tsk_free(tsk);
L
Linus Torvalds 已提交
158 159 160 161 162

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

S
Suresh Siddha 已提交
163 164 165 166 167 168 169
/*
 * 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
170

L
Linus Torvalds 已提交
171 172 173 174 175 176 177 178 179
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),
180
			ARCH_MIN_TASKALIGN, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
181 182
#endif

183 184 185
	/* do the arch specific task caches init */
	arch_task_cache_init();

L
Linus Torvalds 已提交
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204
	/*
	 * 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];
}

205 206 207 208 209 210 211
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
212 213 214 215
static struct task_struct *dup_task_struct(struct task_struct *orig)
{
	struct task_struct *tsk;
	struct thread_info *ti;
216 217
	unsigned long *stackend;

P
Peter Zijlstra 已提交
218
	int err;
L
Linus Torvalds 已提交
219 220 221 222 223 224 225 226 227 228 229 230 231

	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;
	}

232 233 234 235
 	err = arch_dup_task_struct(tsk, orig);
	if (err)
		goto out;

R
Roman Zippel 已提交
236
	tsk->stack = ti;
P
Peter Zijlstra 已提交
237 238

	err = prop_local_init_single(&tsk->dirties);
239 240
	if (err)
		goto out;
P
Peter Zijlstra 已提交
241

242
	setup_thread_stack(tsk, orig);
243 244
	stackend = end_of_stack(tsk);
	*stackend = STACK_END_MAGIC;	/* for overflow detection */
L
Linus Torvalds 已提交
245

246 247 248 249
#ifdef CONFIG_CC_STACKPROTECTOR
	tsk->stack_canary = get_random_int();
#endif

L
Linus Torvalds 已提交
250 251
	/* One for us, one for whoever does the "release_task()" (usually parent) */
	atomic_set(&tsk->usage,2);
252
	atomic_set(&tsk->fs_excl, 0);
253
#ifdef CONFIG_BLK_DEV_IO_TRACE
254
	tsk->btrace_seq = 0;
255
#endif
256
	tsk->splice_pipe = NULL;
L
Linus Torvalds 已提交
257
	return tsk;
258 259 260 261 262

out:
	free_thread_info(ti);
	free_task_struct(tsk);
	return NULL;
L
Linus Torvalds 已提交
263 264 265
}

#ifdef CONFIG_MMU
A
Alexey Dobriyan 已提交
266
static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
L
Linus Torvalds 已提交
267
{
H
Hugh Dickins 已提交
268
	struct vm_area_struct *mpnt, *tmp, **pprev;
L
Linus Torvalds 已提交
269 270 271 272 273 274
	struct rb_node **rb_link, *rb_parent;
	int retval;
	unsigned long charge;
	struct mempolicy *pol;

	down_write(&oldmm->mmap_sem);
275
	flush_cache_dup_mm(oldmm);
276 277 278 279
	/*
	 * Not linked in yet - no deadlock potential:
	 */
	down_write_nested(&mm->mmap_sem, SINGLE_DEPTH_NESTING);
280

L
Linus Torvalds 已提交
281 282 283 284
	mm->locked_vm = 0;
	mm->mmap = NULL;
	mm->mmap_cache = NULL;
	mm->free_area_cache = oldmm->mmap_base;
285
	mm->cached_hole_size = ~0UL;
L
Linus Torvalds 已提交
286
	mm->map_count = 0;
287
	cpumask_clear(mm_cpumask(mm));
L
Linus Torvalds 已提交
288 289 290 291 292
	mm->mm_rb = RB_ROOT;
	rb_link = &mm->mm_rb.rb_node;
	rb_parent = NULL;
	pprev = &mm->mmap;

H
Hugh Dickins 已提交
293
	for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) {
L
Linus Torvalds 已提交
294 295 296
		struct file *file;

		if (mpnt->vm_flags & VM_DONTCOPY) {
297 298
			long pages = vma_pages(mpnt);
			mm->total_vm -= pages;
299
			vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file,
300
								-pages);
L
Linus Torvalds 已提交
301 302 303 304 305 306 307 308 309
			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;
		}
310
		tmp = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
311 312 313
		if (!tmp)
			goto fail_nomem;
		*tmp = *mpnt;
314
		pol = mpol_dup(vma_policy(mpnt));
L
Linus Torvalds 已提交
315 316 317 318 319 320 321 322 323 324
		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) {
325
			struct inode *inode = file->f_path.dentry->d_inode;
H
Hugh Dickins 已提交
326 327
			struct address_space *mapping = file->f_mapping;

L
Linus Torvalds 已提交
328 329 330
			get_file(file);
			if (tmp->vm_flags & VM_DENYWRITE)
				atomic_dec(&inode->i_writecount);
H
Hugh Dickins 已提交
331 332 333
			spin_lock(&mapping->i_mmap_lock);
			if (tmp->vm_flags & VM_SHARED)
				mapping->i_mmap_writable++;
L
Linus Torvalds 已提交
334
			tmp->vm_truncate_count = mpnt->vm_truncate_count;
H
Hugh Dickins 已提交
335 336
			flush_dcache_mmap_lock(mapping);
			/* insert tmp into the share list, just after mpnt */
L
Linus Torvalds 已提交
337
			vma_prio_tree_add(tmp, mpnt);
H
Hugh Dickins 已提交
338 339
			flush_dcache_mmap_unlock(mapping);
			spin_unlock(&mapping->i_mmap_lock);
L
Linus Torvalds 已提交
340 341
		}

342 343 344 345 346 347 348 349
		/*
		 * 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
350
		/*
351
		 * Link in the new vma and copy the page table entries.
L
Linus Torvalds 已提交
352 353 354 355 356 357 358 359 360
		 */
		*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++;
361
		retval = copy_page_range(mm, oldmm, mpnt);
L
Linus Torvalds 已提交
362 363 364 365 366 367 368

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

		if (retval)
			goto out;
	}
369 370
	/* a new mm has just been created */
	arch_dup_mmap(oldmm, mm);
L
Linus Torvalds 已提交
371 372
	retval = 0;
out:
373
	up_write(&mm->mmap_sem);
H
Hugh Dickins 已提交
374
	flush_tlb_mm(oldmm);
L
Linus Torvalds 已提交
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
	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)
{
395
	pgd_free(mm, mm->pgd);
L
Linus Torvalds 已提交
396 397 398 399 400 401 402
}
#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
403
__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
L
Linus Torvalds 已提交
404

405
#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
L
Linus Torvalds 已提交
406 407
#define free_mm(mm)	(kmem_cache_free(mm_cachep, (mm)))

408 409 410 411 412 413 414 415 416 417 418 419
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
420 421
#include <linux/init_task.h>

422
static struct mm_struct * mm_init(struct mm_struct * mm, struct task_struct *p)
L
Linus Torvalds 已提交
423 424 425 426 427
{
	atomic_set(&mm->mm_users, 1);
	atomic_set(&mm->mm_count, 1);
	init_rwsem(&mm->mmap_sem);
	INIT_LIST_HEAD(&mm->mmlist);
428
	mm->flags = (current->mm) ? current->mm->flags : default_dump_filter;
429
	mm->core_state = NULL;
L
Linus Torvalds 已提交
430
	mm->nr_ptes = 0;
431
	set_mm_counter(mm, file_rss, 0);
H
Hugh Dickins 已提交
432
	set_mm_counter(mm, anon_rss, 0);
L
Linus Torvalds 已提交
433
	spin_lock_init(&mm->page_table_lock);
J
Jens Axboe 已提交
434 435
	spin_lock_init(&mm->ioctx_lock);
	INIT_HLIST_HEAD(&mm->ioctx_list);
L
Linus Torvalds 已提交
436
	mm->free_area_cache = TASK_UNMAPPED_BASE;
437
	mm->cached_hole_size = ~0UL;
438
	mm_init_owner(mm, p);
L
Linus Torvalds 已提交
439 440 441

	if (likely(!mm_alloc_pgd(mm))) {
		mm->def_flags = 0;
A
Andrea Arcangeli 已提交
442
		mmu_notifier_mm_init(mm);
L
Linus Torvalds 已提交
443 444
		return mm;
	}
445

L
Linus Torvalds 已提交
446 447 448 449 450 451 452 453 454 455 456 457 458 459
	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));
460
		mm = mm_init(mm, current);
L
Linus Torvalds 已提交
461 462 463 464 465 466 467 468 469
	}
	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.
 */
470
void __mmdrop(struct mm_struct *mm)
L
Linus Torvalds 已提交
471 472 473 474
{
	BUG_ON(mm == &init_mm);
	mm_free_pgd(mm);
	destroy_context(mm);
A
Andrea Arcangeli 已提交
475
	mmu_notifier_mm_destroy(mm);
L
Linus Torvalds 已提交
476 477
	free_mm(mm);
}
478
EXPORT_SYMBOL_GPL(__mmdrop);
L
Linus Torvalds 已提交
479 480 481 482 483 484

/*
 * Decrement the use count and release all resources for an mm.
 */
void mmput(struct mm_struct *mm)
{
A
Andrew Morton 已提交
485 486
	might_sleep();

L
Linus Torvalds 已提交
487 488 489
	if (atomic_dec_and_test(&mm->mm_users)) {
		exit_aio(mm);
		exit_mmap(mm);
M
Matt Helsley 已提交
490
		set_mm_exe_file(mm, NULL);
L
Linus Torvalds 已提交
491 492 493 494 495 496 497 498 499 500 501 502 503 504
		if (!list_empty(&mm->mmlist)) {
			spin_lock(&mmlist_lock);
			list_del(&mm->mmlist);
			spin_unlock(&mmlist_lock);
		}
		put_swap_token(mm);
		mmdrop(mm);
	}
}
EXPORT_SYMBOL_GPL(mmput);

/**
 * get_task_mm - acquire a reference to the task's mm
 *
505
 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
L
Linus Torvalds 已提交
506 507 508 509 510 511 512 513 514 515 516 517
 * 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) {
518
		if (task->flags & PF_KTHREAD)
L
Linus Torvalds 已提交
519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
			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;

545 546 547 548 549 550 551 552 553 554
	/* Get rid of any futexes when releasing the mm */
#ifdef CONFIG_FUTEX
	if (unlikely(tsk->robust_list))
		exit_robust_list(tsk);
#ifdef CONFIG_COMPAT
	if (unlikely(tsk->compat_robust_list))
		compat_exit_robust_list(tsk);
#endif
#endif

L
Linus Torvalds 已提交
555 556 557 558 559 560 561 562
	/* 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);
	}
563 564 565 566 567 568 569 570 571 572

	/*
	 * 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.
	 */
	if (tsk->clear_child_tid
	    && !(tsk->flags & PF_SIGNALED)
	    && atomic_read(&mm->mm_users) > 1) {
L
Linus Torvalds 已提交
573 574 575 576 577 578 579 580 581 582 583 584
		u32 __user * tidptr = tsk->clear_child_tid;
		tsk->clear_child_tid = NULL;

		/*
		 * We don't check the error code - if userspace has
		 * not set up a proper pointer then tough luck.
		 */
		put_user(0, tidptr);
		sys_futex(tidptr, FUTEX_WAKE, 1, NULL, NULL, 0);
	}
}

585 586 587 588
/*
 * Allocate a new mm structure and copy contents from the
 * mm structure of the passed in task structure.
 */
589
struct mm_struct *dup_mm(struct task_struct *tsk)
590 591 592 593 594 595 596 597 598 599 600 601 602
{
	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));

603 604 605 606
	/* Initializing for Swap token stuff */
	mm->token_priority = 0;
	mm->last_interval = 0;

607
	if (!mm_init(mm, tsk))
608 609 610 611 612
		goto fail_nomem;

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

M
Matt Helsley 已提交
613 614
	dup_mm_exe_file(oldmm, mm);

615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
	err = dup_mmap(mm, oldmm);
	if (err)
		goto free_pt;

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

	return mm;

free_pt:
	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
640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
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;

	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;
667
	mm = dup_mm(tsk);
L
Linus Torvalds 已提交
668 669 670 671
	if (!mm)
		goto fail_nomem;

good_mm:
672 673 674 675
	/* Initializing for Swap token stuff */
	mm->token_priority = 0;
	mm->last_interval = 0;

L
Linus Torvalds 已提交
676 677 678 679 680 681 682 683
	tsk->mm = mm;
	tsk->active_mm = mm;
	return 0;

fail_nomem:
	return retval;
}

A
Alexey Dobriyan 已提交
684
static struct fs_struct *__copy_fs_struct(struct fs_struct *old)
L
Linus Torvalds 已提交
685 686 687 688 689 690 691 692
{
	struct fs_struct *fs = kmem_cache_alloc(fs_cachep, GFP_KERNEL);
	/* We don't need to lock fs - think why ;-) */
	if (fs) {
		atomic_set(&fs->count, 1);
		rwlock_init(&fs->lock);
		fs->umask = old->umask;
		read_lock(&old->lock);
J
Jan Blunck 已提交
693 694 695 696
		fs->root = old->root;
		path_get(&old->root);
		fs->pwd = old->pwd;
		path_get(&old->pwd);
L
Linus Torvalds 已提交
697 698 699 700 701 702 703 704 705 706 707 708
		read_unlock(&old->lock);
	}
	return fs;
}

struct fs_struct *copy_fs_struct(struct fs_struct *old)
{
	return __copy_fs_struct(old);
}

EXPORT_SYMBOL_GPL(copy_fs_struct);

A
Alexey Dobriyan 已提交
709
static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
710 711 712 713 714 715 716 717 718 719 720
{
	if (clone_flags & CLONE_FS) {
		atomic_inc(&current->fs->count);
		return 0;
	}
	tsk->fs = __copy_fs_struct(current->fs);
	if (!tsk->fs)
		return -ENOMEM;
	return 0;
}

721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747
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;
}

748
static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
749 750 751 752 753 754
{
#ifdef CONFIG_BLOCK
	struct io_context *ioc = current->io_context;

	if (!ioc)
		return 0;
755 756 757 758 759 760 761 762
	/*
	 * 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)) {
763 764 765 766 767 768 769 770 771 772
		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
773
static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
774 775 776
{
	struct sighand_struct *sig;

Z
Zhaolei 已提交
777
	if (clone_flags & CLONE_SIGHAND) {
L
Linus Torvalds 已提交
778 779 780 781
		atomic_inc(&current->sighand->count);
		return 0;
	}
	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
I
Ingo Molnar 已提交
782
	rcu_assign_pointer(tsk->sighand, sig);
L
Linus Torvalds 已提交
783 784 785 786 787 788 789
	if (!sig)
		return -ENOMEM;
	atomic_set(&sig->count, 1);
	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
	return 0;
}

790
void __cleanup_sighand(struct sighand_struct *sighand)
791 792 793 794 795
{
	if (atomic_dec_and_test(&sighand->count))
		kmem_cache_free(sighand_cachep, sighand);
}

796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821

/*
 * 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. */
	sig->it_virt_expires = cputime_zero;
	sig->it_virt_incr = cputime_zero;
	sig->it_prof_expires = cputime_zero;
	sig->it_prof_incr = cputime_zero;

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

	/* 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
822
static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
823 824 825 826
{
	struct signal_struct *sig;

	if (clone_flags & CLONE_THREAD) {
827 828 829
		atomic_inc(&current->signal->count);
		atomic_inc(&current->signal->live);
		return 0;
L
Linus Torvalds 已提交
830 831
	}
	sig = kmem_cache_alloc(signal_cachep, GFP_KERNEL);
832 833 834 835

	if (sig)
		posix_cpu_timers_init_group(sig);

L
Linus Torvalds 已提交
836 837 838 839 840 841 842 843 844 845 846
	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;
	sig->group_exit_code = 0;
	sig->group_exit_task = NULL;
	sig->group_stop_count = 0;
847
	sig->curr_target = tsk;
L
Linus Torvalds 已提交
848 849 850
	init_sigpending(&sig->shared_pending);
	INIT_LIST_HEAD(&sig->posix_timers);

851
	hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
852
	sig->it_real_incr.tv64 = 0;
L
Linus Torvalds 已提交
853 854 855
	sig->real_timer.function = it_real_fn;

	sig->leader = 0;	/* session leadership doesn't inherit */
856
	sig->tty_old_pgrp = NULL;
A
Alan Cox 已提交
857
	sig->tty = NULL;
L
Linus Torvalds 已提交
858

859
	sig->utime = sig->stime = sig->cutime = sig->cstime = cputime_zero;
860 861
	sig->gtime = cputime_zero;
	sig->cgtime = cputime_zero;
L
Linus Torvalds 已提交
862 863
	sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0;
	sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0;
864
	sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0;
865
	task_io_accounting_init(&sig->ioac);
866
	sig->sum_sched_runtime = 0;
867
	taskstats_tgid_init(sig);
L
Linus Torvalds 已提交
868 869 870 871 872

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

873
	acct_init_pacct(&sig->pacct);
L
Linus Torvalds 已提交
874

M
Miloslav Trmac 已提交
875 876
	tty_audit_fork(sig);

L
Linus Torvalds 已提交
877 878 879
	return 0;
}

880 881
void __cleanup_signal(struct signal_struct *sig)
{
882
	thread_group_cputime_free(sig);
A
Alan Cox 已提交
883
	tty_kref_put(sig->tty);
884 885 886
	kmem_cache_free(signal_cachep, sig);
}

A
Alexey Dobriyan 已提交
887
static void cleanup_signal(struct task_struct *tsk)
888 889 890 891 892 893 894 895 896
{
	struct signal_struct *sig = tsk->signal;

	atomic_dec(&sig->live);

	if (atomic_dec_and_test(&sig->count))
		__cleanup_signal(sig);
}

A
Alexey Dobriyan 已提交
897
static void copy_flags(unsigned long clone_flags, struct task_struct *p)
L
Linus Torvalds 已提交
898 899 900
{
	unsigned long new_flags = p->flags;

901
	new_flags &= ~PF_SUPERPRIV;
L
Linus Torvalds 已提交
902
	new_flags |= PF_FORKNOEXEC;
R
Roland McGrath 已提交
903
	new_flags |= PF_STARTING;
L
Linus Torvalds 已提交
904
	p->flags = new_flags;
905
	clear_freeze_flag(p);
L
Linus Torvalds 已提交
906 907
}

908
SYSCALL_DEFINE1(set_tid_address, int __user *, tidptr)
L
Linus Torvalds 已提交
909 910 911
{
	current->clear_child_tid = tidptr;

912
	return task_pid_vnr(current);
L
Linus Torvalds 已提交
913 914
}

A
Alexey Dobriyan 已提交
915
static void rt_mutex_init_task(struct task_struct *p)
I
Ingo Molnar 已提交
916 917
{
	spin_lock_init(&p->pi_lock);
918
#ifdef CONFIG_RT_MUTEXES
I
Ingo Molnar 已提交
919 920 921 922 923
	plist_head_init(&p->pi_waiters, &p->pi_lock);
	p->pi_blocked_on = NULL;
#endif
}

924 925 926 927 928 929 930
#ifdef CONFIG_MM_OWNER
void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
{
	mm->owner = p;
}
#endif /* CONFIG_MM_OWNER */

931 932 933 934 935 936 937 938 939 940 941 942 943
/*
 * 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
944 945 946 947 948 949 950 951
/*
 * 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.
 */
952 953 954 955 956
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
957 958
					struct pid *pid,
					int trace)
L
Linus Torvalds 已提交
959 960
{
	int retval;
961
	struct task_struct *p;
962
	int cgroup_callbacks_done = 0;
L
Linus Torvalds 已提交
963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990

	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);

	retval = security_task_create(clone_flags);
	if (retval)
		goto fork_out;

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

991 992
	rt_mutex_init_task(p);

I
Ingo Molnar 已提交
993
#ifdef CONFIG_PROVE_LOCKING
994 995 996
	DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
#endif
L
Linus Torvalds 已提交
997
	retval = -EAGAIN;
998
	if (atomic_read(&p->real_cred->user->processes) >=
L
Linus Torvalds 已提交
999 1000
			p->signal->rlim[RLIMIT_NPROC].rlim_cur) {
		if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
1001
		    p->real_cred->user != INIT_USER)
L
Linus Torvalds 已提交
1002 1003 1004
			goto bad_fork_free;
	}

1005 1006 1007
	retval = copy_creds(p, clone_flags);
	if (retval < 0)
		goto bad_fork_free;
L
Linus Torvalds 已提交
1008 1009 1010 1011 1012 1013

	/*
	 * 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.
	 */
1014
	retval = -EAGAIN;
L
Linus Torvalds 已提交
1015 1016 1017
	if (nr_threads >= max_threads)
		goto bad_fork_cleanup_count;

A
Al Viro 已提交
1018
	if (!try_module_get(task_thread_info(p)->exec_domain->module))
L
Linus Torvalds 已提交
1019 1020 1021 1022 1023 1024
		goto bad_fork_cleanup_count;

	if (p->binfmt && !try_module_get(p->binfmt->module))
		goto bad_fork_cleanup_put_domain;

	p->did_exec = 0;
1025
	delayacct_tsk_init(p);	/* Must remain after dup_task_struct() */
L
Linus Torvalds 已提交
1026 1027 1028
	copy_flags(clone_flags, p);
	INIT_LIST_HEAD(&p->children);
	INIT_LIST_HEAD(&p->sibling);
P
Paul E. McKenney 已提交
1029 1030 1031 1032
#ifdef CONFIG_PREEMPT_RCU
	p->rcu_read_lock_nesting = 0;
	p->rcu_flipctr_idx = 0;
#endif /* #ifdef CONFIG_PREEMPT_RCU */
L
Linus Torvalds 已提交
1033 1034 1035 1036 1037 1038 1039 1040
	p->vfork_done = NULL;
	spin_lock_init(&p->alloc_lock);

	clear_tsk_thread_flag(p, TIF_SIGPENDING);
	init_sigpending(&p->pending);

	p->utime = cputime_zero;
	p->stime = cputime_zero;
1041
	p->gtime = cputime_zero;
1042 1043
	p->utimescaled = cputime_zero;
	p->stimescaled = cputime_zero;
P
Peter Zijlstra 已提交
1044
	p->prev_utime = cputime_zero;
1045
	p->prev_stime = cputime_zero;
1046

1047 1048
	p->default_timer_slack_ns = current->timer_slack_ns;

1049 1050 1051 1052 1053
#ifdef CONFIG_DETECT_SOFTLOCKUP
	p->last_switch_count = 0;
	p->last_switch_timestamp = 0;
#endif

1054
	task_io_accounting_init(&p->ioac);
L
Linus Torvalds 已提交
1055 1056
	acct_clear_integrals(p);

1057
	posix_cpu_timers_init(p);
L
Linus Torvalds 已提交
1058 1059 1060

	p->lock_depth = -1;		/* -1 = no lock */
	do_posix_clock_monotonic_gettime(&p->start_time);
1061 1062
	p->real_start_time = p->start_time;
	monotonic_to_bootbased(&p->real_start_time);
L
Linus Torvalds 已提交
1063 1064
	p->io_context = NULL;
	p->audit_context = NULL;
1065
	cgroup_fork(p);
L
Linus Torvalds 已提交
1066
#ifdef CONFIG_NUMA
1067
	p->mempolicy = mpol_dup(p->mempolicy);
L
Linus Torvalds 已提交
1068 1069 1070
 	if (IS_ERR(p->mempolicy)) {
 		retval = PTR_ERR(p->mempolicy);
 		p->mempolicy = NULL;
1071
 		goto bad_fork_cleanup_cgroup;
L
Linus Torvalds 已提交
1072
 	}
1073
	mpol_fix_fork_child_flag(p);
L
Linus Torvalds 已提交
1074
#endif
1075 1076
#ifdef CONFIG_TRACE_IRQFLAGS
	p->irq_events = 0;
1077 1078 1079
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
	p->hardirqs_enabled = 1;
#else
1080
	p->hardirqs_enabled = 0;
1081
#endif
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093
	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
1094 1095 1096 1097 1098
#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
1099

1100 1101 1102
#ifdef CONFIG_DEBUG_MUTEXES
	p->blocked_on = NULL; /* not blocked yet */
#endif
1103
	if (unlikely(current->ptrace))
1104
		ptrace_fork(p, clone_flags);
1105

1106 1107 1108
	/* Perform scheduler related setup. Assign this task to a CPU. */
	sched_fork(p, clone_flags);

L
Linus Torvalds 已提交
1109
	if ((retval = audit_alloc(p)))
1110
		goto bad_fork_cleanup_policy;
L
Linus Torvalds 已提交
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
	/* 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
1124
	if ((retval = copy_namespaces(clone_flags, p)))
D
David Howells 已提交
1125
		goto bad_fork_cleanup_mm;
1126
	if ((retval = copy_io(clone_flags, p)))
1127
		goto bad_fork_cleanup_namespaces;
L
Linus Torvalds 已提交
1128 1129
	retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs);
	if (retval)
1130
		goto bad_fork_cleanup_io;
L
Linus Torvalds 已提交
1131

1132 1133
	if (pid != &init_struct_pid) {
		retval = -ENOMEM;
1134
		pid = alloc_pid(p->nsproxy->pid_ns);
1135
		if (!pid)
1136
			goto bad_fork_cleanup_io;
1137 1138

		if (clone_flags & CLONE_NEWPID) {
1139
			retval = pid_ns_prepare_proc(p->nsproxy->pid_ns);
1140 1141 1142
			if (retval < 0)
				goto bad_fork_free_pid;
		}
1143 1144
	}

1145 1146
	ftrace_graph_init_task(p);

1147 1148 1149 1150 1151
	p->pid = pid_nr(pid);
	p->tgid = p->pid;
	if (clone_flags & CLONE_THREAD)
		p->tgid = current->tgid;

1152 1153 1154
	if (current->nsproxy != p->nsproxy) {
		retval = ns_cgroup_clone(p, pid);
		if (retval)
1155
			goto bad_fork_free_graph;
1156 1157
	}

L
Linus Torvalds 已提交
1158 1159 1160 1161 1162
	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;
1163
#ifdef CONFIG_FUTEX
1164 1165 1166 1167
	p->robust_list = NULL;
#ifdef CONFIG_COMPAT
	p->compat_robust_list = NULL;
#endif
1168 1169
	INIT_LIST_HEAD(&p->pi_state_list);
	p->pi_state_cache = NULL;
1170
#endif
1171 1172 1173 1174 1175 1176
	/*
	 * 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
1177 1178 1179 1180 1181
	/*
	 * Syscall tracing should be turned off in the child regardless
	 * of CLONE_PTRACE.
	 */
	clear_tsk_thread_flag(p, TIF_SYSCALL_TRACE);
1182 1183 1184
#ifdef TIF_SYSCALL_EMU
	clear_tsk_thread_flag(p, TIF_SYSCALL_EMU);
#endif
A
Arjan van de Ven 已提交
1185
	clear_all_latency_tracing(p);
L
Linus Torvalds 已提交
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196

	/* 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
1197
	INIT_LIST_HEAD(&p->thread_group);
L
Linus Torvalds 已提交
1198

1199 1200 1201 1202 1203 1204
	/* 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
1205 1206 1207 1208
	/* Need tasklist lock for parent etc handling! */
	write_lock_irq(&tasklist_lock);

	/*
N
Nick Piggin 已提交
1209 1210 1211 1212 1213 1214 1215
	 * 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
1216 1217
	 */
	p->cpus_allowed = current->cpus_allowed;
P
Peter Zijlstra 已提交
1218
	p->rt.nr_cpus_allowed = current->rt.nr_cpus_allowed;
1219 1220
	if (unlikely(!cpu_isset(task_cpu(p), p->cpus_allowed) ||
			!cpu_online(task_cpu(p))))
N
Nick Piggin 已提交
1221
		set_task_cpu(p, smp_processor_id());
L
Linus Torvalds 已提交
1222 1223

	/* CLONE_PARENT re-uses the old parent */
1224
	if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) {
L
Linus Torvalds 已提交
1225
		p->real_parent = current->real_parent;
1226 1227
		p->parent_exec_id = current->parent_exec_id;
	} else {
L
Linus Torvalds 已提交
1228
		p->real_parent = current;
1229 1230
		p->parent_exec_id = current->self_exec_id;
	}
L
Linus Torvalds 已提交
1231

1232
	spin_lock(&current->sighand->siglock);
1233 1234 1235 1236 1237 1238 1239 1240 1241

	/*
	 * 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
1242
	recalc_sigpending();
1243 1244 1245 1246
	if (signal_pending(current)) {
		spin_unlock(&current->sighand->siglock);
		write_unlock_irq(&tasklist_lock);
		retval = -ERESTARTNOINTR;
1247
		goto bad_fork_free_graph;
1248 1249
	}

L
Linus Torvalds 已提交
1250 1251
	if (clone_flags & CLONE_THREAD) {
		p->group_leader = current->group_leader;
O
Oleg Nesterov 已提交
1252
		list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group);
L
Linus Torvalds 已提交
1253 1254
	}

1255
	if (likely(p->pid)) {
R
Roland McGrath 已提交
1256
		list_add_tail(&p->sibling, &p->real_parent->children);
R
Roland McGrath 已提交
1257
		tracehook_finish_clone(p, clone_flags, trace);
1258 1259

		if (thread_group_leader(p)) {
E
Eric W. Biederman 已提交
1260
			if (clone_flags & CLONE_NEWPID)
1261
				p->nsproxy->pid_ns->child_reaper = p;
1262

1263
			p->signal->leader_pid = pid;
A
Alan Cox 已提交
1264 1265
			tty_kref_put(p->signal->tty);
			p->signal->tty = tty_kref_get(current->signal->tty);
E
Eric W. Biederman 已提交
1266 1267 1268 1269
			set_task_pgrp(p, task_pgrp_nr(current));
			set_task_session(p, task_session_nr(current));
			attach_pid(p, PIDTYPE_PGID, task_pgrp(current));
			attach_pid(p, PIDTYPE_SID, task_session(current));
1270
			list_add_tail_rcu(&p->tasks, &init_task.tasks);
L
Linus Torvalds 已提交
1271
			__get_cpu_var(process_counts)++;
1272
		}
1273
		attach_pid(p, PIDTYPE_PID, pid);
1274
		nr_threads++;
L
Linus Torvalds 已提交
1275 1276 1277
	}

	total_forks++;
1278
	spin_unlock(&current->sighand->siglock);
L
Linus Torvalds 已提交
1279
	write_unlock_irq(&tasklist_lock);
1280
	proc_fork_connector(p);
1281
	cgroup_post_fork(p);
L
Linus Torvalds 已提交
1282 1283
	return p;

1284 1285
bad_fork_free_graph:
	ftrace_graph_exit_task(p);
1286 1287 1288
bad_fork_free_pid:
	if (pid != &init_struct_pid)
		free_pid(pid);
1289 1290
bad_fork_cleanup_io:
	put_io_context(p->io_context);
S
Serge E. Hallyn 已提交
1291
bad_fork_cleanup_namespaces:
1292
	exit_task_namespaces(p);
L
Linus Torvalds 已提交
1293 1294 1295 1296
bad_fork_cleanup_mm:
	if (p->mm)
		mmput(p->mm);
bad_fork_cleanup_signal:
1297
	cleanup_signal(p);
L
Linus Torvalds 已提交
1298
bad_fork_cleanup_sighand:
1299
	__cleanup_sighand(p->sighand);
L
Linus Torvalds 已提交
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
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:
#ifdef CONFIG_NUMA
1310
	mpol_put(p->mempolicy);
1311
bad_fork_cleanup_cgroup:
L
Linus Torvalds 已提交
1312
#endif
1313
	cgroup_exit(p, cgroup_callbacks_done);
1314
	delayacct_tsk_free(p);
L
Linus Torvalds 已提交
1315 1316 1317
	if (p->binfmt)
		module_put(p->binfmt->module);
bad_fork_cleanup_put_domain:
A
Al Viro 已提交
1318
	module_put(task_thread_info(p)->exec_domain->module);
L
Linus Torvalds 已提交
1319
bad_fork_cleanup_count:
D
David Howells 已提交
1320
	atomic_dec(&p->cred->user->processes);
1321
	put_cred(p->real_cred);
1322
	put_cred(p->cred);
L
Linus Torvalds 已提交
1323 1324
bad_fork_free:
	free_task(p);
1325 1326
fork_out:
	return ERR_PTR(retval);
L
Linus Torvalds 已提交
1327 1328
}

A
Adrian Bunk 已提交
1329
noinline struct pt_regs * __cpuinit __attribute__((weak)) idle_regs(struct pt_regs *regs)
L
Linus Torvalds 已提交
1330 1331 1332 1333 1334
{
	memset(regs, 0, sizeof(struct pt_regs));
	return regs;
}

1335
struct task_struct * __cpuinit fork_idle(int cpu)
L
Linus Torvalds 已提交
1336
{
1337
	struct task_struct *task;
L
Linus Torvalds 已提交
1338 1339
	struct pt_regs regs;

1340
	task = copy_process(CLONE_VM, 0, idle_regs(&regs), 0, NULL,
R
Roland McGrath 已提交
1341
			    &init_struct_pid, 0);
1342 1343
	if (!IS_ERR(task))
		init_idle(task, cpu);
1344

L
Linus Torvalds 已提交
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
	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;
1363
	long nr;
L
Linus Torvalds 已提交
1364

1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
	/*
	 * 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
		 */
1375 1376
		if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SETUID) ||
				!capable(CAP_SETGID))
1377 1378 1379
			return -EPERM;
	}

1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
	/*
	 * 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
1397 1398 1399 1400 1401
	/*
	 * 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
1402

1403
	p = copy_process(clone_flags, stack_start, regs, stack_size,
R
Roland McGrath 已提交
1404
			 child_tidptr, NULL, trace);
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411
	/*
	 * 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;

1412 1413
		trace_sched_process_fork(current, p);

1414
		nr = task_pid_vnr(p);
1415 1416 1417

		if (clone_flags & CLONE_PARENT_SETTID)
			put_user(nr, parent_tidptr);
1418

L
Linus Torvalds 已提交
1419 1420 1421 1422 1423
		if (clone_flags & CLONE_VFORK) {
			p->vfork_done = &vfork;
			init_completion(&vfork);
		}

1424
		audit_finish_fork(p);
R
Roland McGrath 已提交
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435
		tracehook_report_clone(trace, regs, clone_flags, nr, p);

		/*
		 * 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
1436 1437 1438 1439 1440
			/*
			 * We'll start up with an immediate SIGSTOP.
			 */
			sigaddset(&p->pending.signal, SIGSTOP);
			set_tsk_thread_flag(p, TIF_SIGPENDING);
1441
			__set_task_state(p, TASK_STOPPED);
R
Roland McGrath 已提交
1442 1443
		} else {
			wake_up_new_task(p, clone_flags);
L
Linus Torvalds 已提交
1444 1445
		}

R
Roland McGrath 已提交
1446 1447 1448
		tracehook_report_clone_complete(trace, regs,
						clone_flags, nr, p);

L
Linus Torvalds 已提交
1449
		if (clone_flags & CLONE_VFORK) {
R
Rafael J. Wysocki 已提交
1450
			freezer_do_not_count();
L
Linus Torvalds 已提交
1451
			wait_for_completion(&vfork);
R
Rafael J. Wysocki 已提交
1452
			freezer_count();
R
Roland McGrath 已提交
1453
			tracehook_report_vfork_done(p, nr);
L
Linus Torvalds 已提交
1454 1455
		}
	} else {
1456
		nr = PTR_ERR(p);
L
Linus Torvalds 已提交
1457
	}
1458
	return nr;
L
Linus Torvalds 已提交
1459 1460
}

1461 1462 1463 1464
#ifndef ARCH_MIN_MMSTRUCT_ALIGN
#define ARCH_MIN_MMSTRUCT_ALIGN 0
#endif

1465
static void sighand_ctor(void *data)
1466 1467 1468
{
	struct sighand_struct *sighand = data;

C
Christoph Lameter 已提交
1469
	spin_lock_init(&sighand->siglock);
D
Davide Libenzi 已提交
1470
	init_waitqueue_head(&sighand->signalfd_wqh);
1471 1472
}

L
Linus Torvalds 已提交
1473 1474 1475 1476
void __init proc_caches_init(void)
{
	sighand_cachep = kmem_cache_create("sighand_cache",
			sizeof(struct sighand_struct), 0,
1477
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_DESTROY_BY_RCU,
1478
			sighand_ctor);
L
Linus Torvalds 已提交
1479 1480
	signal_cachep = kmem_cache_create("signal_cache",
			sizeof(struct signal_struct), 0,
1481 1482
			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
	files_cachep = kmem_cache_create("files_cache",
L
Linus Torvalds 已提交
1483
			sizeof(struct files_struct), 0,
1484 1485
			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
	fs_cachep = kmem_cache_create("fs_cache",
L
Linus Torvalds 已提交
1486
			sizeof(struct fs_struct), 0,
1487
			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
1488
	mm_cachep = kmem_cache_create("mm_struct",
1489
			sizeof(struct mm_struct), ARCH_MIN_MMSTRUCT_ALIGN,
1490
			SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
1491
	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC);
1492
	mmap_init();
L
Linus Torvalds 已提交
1493
}
1494 1495 1496 1497 1498

/*
 * 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
1499
static void check_unshare_flags(unsigned long *flags_ptr)
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540
{
	/*
	 * 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;
}

/*
1541
 * Unshare the filesystem structure if it is being shared
1542 1543 1544 1545 1546 1547
 */
static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp)
{
	struct fs_struct *fs = current->fs;

	if ((unshare_flags & CLONE_FS) &&
1548 1549 1550 1551 1552
	    (fs && atomic_read(&fs->count) > 1)) {
		*new_fsp = __copy_fs_struct(current->fs);
		if (!*new_fsp)
			return -ENOMEM;
	}
1553 1554 1555 1556 1557

	return 0;
}

/*
1558
 * Unsharing of sighand is not supported yet
1559 1560 1561 1562 1563
 */
static int unshare_sighand(unsigned long unshare_flags, struct sighand_struct **new_sighp)
{
	struct sighand_struct *sigh = current->sighand;

1564
	if ((unshare_flags & CLONE_SIGHAND) && atomic_read(&sigh->count) > 1)
1565 1566 1567 1568 1569 1570
		return -EINVAL;
	else
		return 0;
}

/*
1571
 * Unshare vm if it is being shared
1572 1573 1574 1575 1576 1577
 */
static int unshare_vm(unsigned long unshare_flags, struct mm_struct **new_mmp)
{
	struct mm_struct *mm = current->mm;

	if ((unshare_flags & CLONE_VM) &&
1578
	    (mm && atomic_read(&mm->mm_users) > 1)) {
1579
		return -EINVAL;
1580
	}
1581 1582 1583 1584 1585

	return 0;
}

/*
1586
 * Unshare file descriptor table if it is being shared
1587 1588 1589 1590
 */
static int unshare_fd(unsigned long unshare_flags, struct files_struct **new_fdp)
{
	struct files_struct *fd = current->files;
1591
	int error = 0;
1592 1593

	if ((unshare_flags & CLONE_FILES) &&
1594 1595 1596 1597 1598
	    (fd && atomic_read(&fd->count) > 1)) {
		*new_fdp = dup_fd(fd, &error);
		if (!*new_fdp)
			return error;
	}
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610

	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.
 */
1611
SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags)
1612 1613 1614
{
	int err = 0;
	struct fs_struct *fs, *new_fs = NULL;
1615
	struct sighand_struct *new_sigh = NULL;
1616 1617
	struct mm_struct *mm, *new_mm = NULL, *active_mm = NULL;
	struct files_struct *fd, *new_fd = NULL;
1618
	struct nsproxy *new_nsproxy = NULL;
1619
	int do_sysvsem = 0;
1620 1621 1622

	check_unshare_flags(&unshare_flags);

1623 1624 1625
	/* 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
1626
				CLONE_VM|CLONE_FILES|CLONE_SYSVSEM|
1627
				CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET))
1628 1629
		goto bad_unshare_out;

1630 1631 1632 1633 1634 1635
	/*
	 * 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))
1636
		do_sysvsem = 1;
1637 1638 1639 1640 1641
	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)))
1642
		goto bad_unshare_cleanup_fs;
1643 1644 1645 1646
	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;
1647 1648
	if ((err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy,
			new_fs)))
1649
		goto bad_unshare_cleanup_fd;
1650

1651 1652 1653 1654 1655 1656 1657
	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
1658

1659
		if (new_nsproxy) {
1660 1661
			switch_task_namespaces(current, new_nsproxy);
			new_nsproxy = NULL;
1662
		}
1663

1664 1665
		task_lock(current);

1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689
		if (new_fs) {
			fs = current->fs;
			current->fs = new_fs;
			new_fs = fs;
		}

		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);
	}

1690
	if (new_nsproxy)
1691
		put_nsproxy(new_nsproxy);
1692

1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
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)
		put_fs_struct(new_fs);

bad_unshare_cleanup_thread:
bad_unshare_out:
	return err;
}
1714 1715 1716 1717 1718 1719 1720 1721 1722 1723

/*
 *	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;
1724
	struct files_struct *copy = NULL;
1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737
	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;
}