fork.c 57.3 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14
/*
 *  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>
15
#include <linux/sched/autogroup.h>
16
#include <linux/sched/mm.h>
17
#include <linux/sched/coredump.h>
18
#include <linux/sched/user.h>
19
#include <linux/sched/numa_balancing.h>
20
#include <linux/sched/stat.h>
21
#include <linux/sched/task.h>
22
#include <linux/sched/task_stack.h>
23
#include <linux/sched/cputime.h>
24
#include <linux/rtmutex.h>
Linus Torvalds's avatar
Linus Torvalds committed
25 26 27 28 29 30 31 32 33
#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>
Al Viro's avatar
Al Viro committed
34
#include <linux/fdtable.h>
35
#include <linux/iocontext.h>
Linus Torvalds's avatar
Linus Torvalds committed
36 37 38
#include <linux/key.h>
#include <linux/binfmts.h>
#include <linux/mman.h>
Andrea Arcangeli's avatar
Andrea Arcangeli committed
39
#include <linux/mmu_notifier.h>
Linus Torvalds's avatar
Linus Torvalds committed
40
#include <linux/fs.h>
41 42
#include <linux/mm.h>
#include <linux/vmacache.h>
43
#include <linux/nsproxy.h>
44
#include <linux/capability.h>
Linus Torvalds's avatar
Linus Torvalds committed
45
#include <linux/cpu.h>
46
#include <linux/cgroup.h>
Linus Torvalds's avatar
Linus Torvalds committed
47
#include <linux/security.h>
48
#include <linux/hugetlb.h>
49
#include <linux/seccomp.h>
Linus Torvalds's avatar
Linus Torvalds committed
50 51 52 53
#include <linux/swap.h>
#include <linux/syscalls.h>
#include <linux/jiffies.h>
#include <linux/futex.h>
54
#include <linux/compat.h>
55
#include <linux/kthread.h>
56
#include <linux/task_io_accounting_ops.h>
57
#include <linux/rcupdate.h>
Linus Torvalds's avatar
Linus Torvalds committed
58 59 60
#include <linux/ptrace.h>
#include <linux/mount.h>
#include <linux/audit.h>
61
#include <linux/memcontrol.h>
62
#include <linux/ftrace.h>
63
#include <linux/proc_fs.h>
Linus Torvalds's avatar
Linus Torvalds committed
64 65
#include <linux/profile.h>
#include <linux/rmap.h>
66
#include <linux/ksm.h>
Linus Torvalds's avatar
Linus Torvalds committed
67
#include <linux/acct.h>
68
#include <linux/userfaultfd_k.h>
69
#include <linux/tsacct_kern.h>
70
#include <linux/cn_proc.h>
71
#include <linux/freezer.h>
72
#include <linux/delayacct.h>
73
#include <linux/taskstats_kern.h>
74
#include <linux/random.h>
75
#include <linux/tty.h>
76
#include <linux/blkdev.h>
77
#include <linux/fs_struct.h>
78
#include <linux/magic.h>
79
#include <linux/perf_event.h>
80
#include <linux/posix-timers.h>
81
#include <linux/user-return-notifier.h>
82
#include <linux/oom.h>
Andrea Arcangeli's avatar
Andrea Arcangeli committed
83
#include <linux/khugepaged.h>
84
#include <linux/signalfd.h>
85
#include <linux/uprobes.h>
86
#include <linux/aio.h>
87
#include <linux/compiler.h>
88
#include <linux/sysctl.h>
89
#include <linux/kcov.h>
Linus Torvalds's avatar
Linus Torvalds committed
90 91 92

#include <asm/pgtable.h>
#include <asm/pgalloc.h>
93
#include <linux/uaccess.h>
Linus Torvalds's avatar
Linus Torvalds committed
94 95 96 97
#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

98 99
#include <trace/events/sched.h>

100 101 102
#define CREATE_TRACE_POINTS
#include <trace/events/task.h>

103 104 105 106 107 108 109 110 111 112
/*
 * Minimum number of threads to boot the kernel
 */
#define MIN_THREADS 20

/*
 * Maximum number of threads
 */
#define MAX_THREADS FUTEX_TID_MASK

Linus Torvalds's avatar
Linus Torvalds committed
113 114 115 116
/*
 * Protected counters by write_lock_irq(&tasklist_lock)
 */
unsigned long total_forks;	/* Handle normal Linux uptimes. */
117
int nr_threads;			/* The idle threads do not count.. */
Linus Torvalds's avatar
Linus Torvalds committed
118 119 120 121 122

int max_threads;		/* tunable limit on nr_threads */

DEFINE_PER_CPU(unsigned long, process_counts) = 0;

123
__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
124 125 126 127 128 129 130 131

#ifdef CONFIG_PROVE_RCU
int lockdep_tasklist_lock_is_held(void)
{
	return lockdep_is_held(&tasklist_lock);
}
EXPORT_SYMBOL_GPL(lockdep_tasklist_lock_is_held);
#endif /* #ifdef CONFIG_PROVE_RCU */
Linus Torvalds's avatar
Linus Torvalds committed
132 133 134 135 136 137

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

138
	for_each_possible_cpu(cpu)
Linus Torvalds's avatar
Linus Torvalds committed
139 140 141 142 143
		total += per_cpu(process_counts, cpu);

	return total;
}

144 145 146 147
void __weak arch_release_task_struct(struct task_struct *tsk)
{
}

148
#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
149
static struct kmem_cache *task_struct_cachep;
150 151 152 153 154 155 156 157 158 159

static inline struct task_struct *alloc_task_struct_node(int node)
{
	return kmem_cache_alloc_node(task_struct_cachep, GFP_KERNEL, node);
}

static inline void free_task_struct(struct task_struct *tsk)
{
	kmem_cache_free(task_struct_cachep, tsk);
}
Linus Torvalds's avatar
Linus Torvalds committed
160 161
#endif

162
void __weak arch_release_thread_stack(unsigned long *stack)
163 164 165
{
}

166
#ifndef CONFIG_ARCH_THREAD_STACK_ALLOCATOR
167

168 169 170 171
/*
 * Allocate pages if THREAD_SIZE is >= PAGE_SIZE, otherwise use a
 * kmemcache based allocator.
 */
172
# if THREAD_SIZE >= PAGE_SIZE || defined(CONFIG_VMAP_STACK)
173 174 175 176 177 178 179 180 181 182

#ifdef CONFIG_VMAP_STACK
/*
 * vmalloc() is a bit slow, and calling vfree() enough times will force a TLB
 * flush.  Try to minimize the number of calls by caching stacks.
 */
#define NR_CACHED_STACKS 2
static DEFINE_PER_CPU(struct vm_struct *, cached_stacks[NR_CACHED_STACKS]);
#endif

183
static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node)
184
{
185
#ifdef CONFIG_VMAP_STACK
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
	void *stack;
	int i;

	local_irq_disable();
	for (i = 0; i < NR_CACHED_STACKS; i++) {
		struct vm_struct *s = this_cpu_read(cached_stacks[i]);

		if (!s)
			continue;
		this_cpu_write(cached_stacks[i], NULL);

		tsk->stack_vm_area = s;
		local_irq_enable();
		return s->addr;
	}
	local_irq_enable();

	stack = __vmalloc_node_range(THREAD_SIZE, THREAD_SIZE,
				     VMALLOC_START, VMALLOC_END,
				     THREADINFO_GFP | __GFP_HIGHMEM,
				     PAGE_KERNEL,
				     0, node, __builtin_return_address(0));
208 209 210 211 212 213 214 215 216 217

	/*
	 * We can't call find_vm_area() in interrupt context, and
	 * free_thread_stack() can be called in interrupt context,
	 * so cache the vm_struct.
	 */
	if (stack)
		tsk->stack_vm_area = find_vm_area(stack);
	return stack;
#else
218 219
	struct page *page = alloc_pages_node(node, THREADINFO_GFP,
					     THREAD_SIZE_ORDER);
220 221

	return page ? page_address(page) : NULL;
222
#endif
223 224
}

225
static inline void free_thread_stack(struct task_struct *tsk)
226
{
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242
#ifdef CONFIG_VMAP_STACK
	if (task_stack_vm_area(tsk)) {
		unsigned long flags;
		int i;

		local_irq_save(flags);
		for (i = 0; i < NR_CACHED_STACKS; i++) {
			if (this_cpu_read(cached_stacks[i]))
				continue;

			this_cpu_write(cached_stacks[i], tsk->stack_vm_area);
			local_irq_restore(flags);
			return;
		}
		local_irq_restore(flags);

243
		vfree_atomic(tsk->stack);
244 245 246 247 248
		return;
	}
#endif

	__free_pages(virt_to_page(tsk->stack), THREAD_SIZE_ORDER);
249
}
250
# else
251
static struct kmem_cache *thread_stack_cache;
252

253
static unsigned long *alloc_thread_stack_node(struct task_struct *tsk,
254 255
						  int node)
{
256
	return kmem_cache_alloc_node(thread_stack_cache, THREADINFO_GFP, node);
257 258
}

259
static void free_thread_stack(struct task_struct *tsk)
260
{
261
	kmem_cache_free(thread_stack_cache, tsk->stack);
262 263
}

264
void thread_stack_cache_init(void)
265
{
266
	thread_stack_cache = kmem_cache_create("thread_stack", THREAD_SIZE,
267
					      THREAD_SIZE, 0, NULL);
268
	BUG_ON(thread_stack_cache == NULL);
269 270
}
# endif
271 272
#endif

Linus Torvalds's avatar
Linus Torvalds committed
273
/* SLAB cache for signal_struct structures (tsk->signal) */
274
static struct kmem_cache *signal_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
275 276

/* SLAB cache for sighand_struct structures (tsk->sighand) */
277
struct kmem_cache *sighand_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
278 279

/* SLAB cache for files_struct structures (tsk->files) */
280
struct kmem_cache *files_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
281 282

/* SLAB cache for fs_struct structures (tsk->fs) */
283
struct kmem_cache *fs_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
284 285

/* SLAB cache for vm_area_struct structures */
286
struct kmem_cache *vm_area_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
287 288

/* SLAB cache for mm_struct structures (tsk->mm) */
289
static struct kmem_cache *mm_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
290

291
static void account_kernel_stack(struct task_struct *tsk, int account)
292
{
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
	void *stack = task_stack_page(tsk);
	struct vm_struct *vm = task_stack_vm_area(tsk);

	BUILD_BUG_ON(IS_ENABLED(CONFIG_VMAP_STACK) && PAGE_SIZE % 1024 != 0);

	if (vm) {
		int i;

		BUG_ON(vm->nr_pages != THREAD_SIZE / PAGE_SIZE);

		for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++) {
			mod_zone_page_state(page_zone(vm->pages[i]),
					    NR_KERNEL_STACK_KB,
					    PAGE_SIZE / 1024 * account);
		}

		/* All stack pages belong to the same memcg. */
		memcg_kmem_update_page_stat(vm->pages[0], MEMCG_KERNEL_STACK_KB,
					    account * (THREAD_SIZE / 1024));
	} else {
		/*
		 * All stack pages are in the same zone and belong to the
		 * same memcg.
		 */
		struct page *first_page = virt_to_page(stack);

		mod_zone_page_state(page_zone(first_page), NR_KERNEL_STACK_KB,
				    THREAD_SIZE / 1024 * account);

		memcg_kmem_update_page_stat(first_page, MEMCG_KERNEL_STACK_KB,
					    account * (THREAD_SIZE / 1024));
	}
325 326
}

327
static void release_task_stack(struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
328
{
329 330 331
	if (WARN_ON(tsk->state != TASK_DEAD))
		return;  /* Better to leak the stack than to free prematurely */

332
	account_kernel_stack(tsk, -1);
333
	arch_release_thread_stack(tsk->stack);
334
	free_thread_stack(tsk);
335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
	tsk->stack = NULL;
#ifdef CONFIG_VMAP_STACK
	tsk->stack_vm_area = NULL;
#endif
}

#ifdef CONFIG_THREAD_INFO_IN_TASK
void put_task_stack(struct task_struct *tsk)
{
	if (atomic_dec_and_test(&tsk->stack_refcount))
		release_task_stack(tsk);
}
#endif

void free_task(struct task_struct *tsk)
{
#ifndef CONFIG_THREAD_INFO_IN_TASK
	/*
	 * The task is finally done with both the stack and thread_info,
	 * so free both.
	 */
	release_task_stack(tsk);
#else
	/*
	 * If the task had a separate stack allocation, it should be gone
	 * by now.
	 */
	WARN_ON_ONCE(atomic_read(&tsk->stack_refcount) != 0);
#endif
364
	rt_mutex_debug_task_free(tsk);
365
	ftrace_graph_exit_task(tsk);
366
	put_seccomp_filter(tsk);
367
	arch_release_task_struct(tsk);
368 369
	if (tsk->flags & PF_KTHREAD)
		free_kthread_struct(tsk);
Linus Torvalds's avatar
Linus Torvalds committed
370 371 372 373
	free_task_struct(tsk);
}
EXPORT_SYMBOL(free_task);

374 375
static inline void free_signal_struct(struct signal_struct *sig)
{
376
	taskstats_tgid_free(sig);
377
	sched_autogroup_exit(sig);
378 379 380 381
	/*
	 * __mmdrop is not safe to call from softirq context on x86 due to
	 * pgd_dtor so postpone it to the async context
	 */
382
	if (sig->oom_mm)
383
		mmdrop_async(sig->oom_mm);
384 385 386 387 388
	kmem_cache_free(signal_cachep, sig);
}

static inline void put_signal_struct(struct signal_struct *sig)
{
389
	if (atomic_dec_and_test(&sig->sigcnt))
390 391 392
		free_signal_struct(sig);
}

393
void __put_task_struct(struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
394
{
Eugene Teo's avatar
Eugene Teo committed
395
	WARN_ON(!tsk->exit_state);
Linus Torvalds's avatar
Linus Torvalds committed
396 397 398
	WARN_ON(atomic_read(&tsk->usage));
	WARN_ON(tsk == current);

399
	cgroup_free(tsk);
400
	task_numa_free(tsk);
401
	security_task_free(tsk);
402
	exit_creds(tsk);
403
	delayacct_tsk_free(tsk);
404
	put_signal_struct(tsk->signal);
Linus Torvalds's avatar
Linus Torvalds committed
405 406 407 408

	if (!profile_handoff_task(tsk))
		free_task(tsk);
}
409
EXPORT_SYMBOL_GPL(__put_task_struct);
Linus Torvalds's avatar
Linus Torvalds committed
410

411
void __init __weak arch_task_cache_init(void) { }
412

413 414 415
/*
 * set_max_threads
 */
416
static void set_max_threads(unsigned int max_threads_suggested)
417
{
418
	u64 threads;
419 420

	/*
421 422
	 * The number of threads shall be limited such that the thread
	 * structures may only consume a small part of the available memory.
423
	 */
424 425 426 427 428 429
	if (fls64(totalram_pages) + fls64(PAGE_SIZE) > 64)
		threads = MAX_THREADS;
	else
		threads = div64_u64((u64) totalram_pages * (u64) PAGE_SIZE,
				    (u64) THREAD_SIZE * 8UL);

430 431 432
	if (threads > max_threads_suggested)
		threads = max_threads_suggested;

433
	max_threads = clamp_t(u64, threads, MIN_THREADS, MAX_THREADS);
434 435
}

436 437 438 439
#ifdef CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT
/* Initialized by the architecture: */
int arch_task_struct_size __read_mostly;
#endif
440

441
void __init fork_init(void)
Linus Torvalds's avatar
Linus Torvalds committed
442
{
443
	int i;
444
#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
Linus Torvalds's avatar
Linus Torvalds committed
445
#ifndef ARCH_MIN_TASKALIGN
446
#define ARCH_MIN_TASKALIGN	0
Linus Torvalds's avatar
Linus Torvalds committed
447
#endif
448
	int align = max_t(int, L1_CACHE_BYTES, ARCH_MIN_TASKALIGN);
449

Linus Torvalds's avatar
Linus Torvalds committed
450
	/* create a slab on which task_structs can be allocated */
451
	task_struct_cachep = kmem_cache_create("task_struct",
452
			arch_task_struct_size, align,
453
			SLAB_PANIC|SLAB_NOTRACK|SLAB_ACCOUNT, NULL);
Linus Torvalds's avatar
Linus Torvalds committed
454 455
#endif

456 457 458
	/* do the arch specific task caches init */
	arch_task_cache_init();

459
	set_max_threads(MAX_THREADS);
Linus Torvalds's avatar
Linus Torvalds committed
460 461 462 463 464

	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];
465

466 467 468
	for (i = 0; i < UCOUNT_COUNTS; i++) {
		init_user_ns.ucount_max[i] = max_threads/2;
	}
Linus Torvalds's avatar
Linus Torvalds committed
469 470
}

471
int __weak arch_dup_task_struct(struct task_struct *dst,
472 473 474 475 476 477
					       struct task_struct *src)
{
	*dst = *src;
	return 0;
}

478 479 480 481 482 483 484 485
void set_task_stack_end_magic(struct task_struct *tsk)
{
	unsigned long *stackend;

	stackend = end_of_stack(tsk);
	*stackend = STACK_END_MAGIC;	/* for overflow detection */
}

486
static struct task_struct *dup_task_struct(struct task_struct *orig, int node)
Linus Torvalds's avatar
Linus Torvalds committed
487 488
{
	struct task_struct *tsk;
489
	unsigned long *stack;
490
	struct vm_struct *stack_vm_area;
491
	int err;
Linus Torvalds's avatar
Linus Torvalds committed
492

493 494
	if (node == NUMA_NO_NODE)
		node = tsk_fork_get_node(orig);
495
	tsk = alloc_task_struct_node(node);
Linus Torvalds's avatar
Linus Torvalds committed
496 497 498
	if (!tsk)
		return NULL;

499 500
	stack = alloc_thread_stack_node(tsk, node);
	if (!stack)
501
		goto free_tsk;
Linus Torvalds's avatar
Linus Torvalds committed
502

503 504
	stack_vm_area = task_stack_vm_area(tsk);

505
	err = arch_dup_task_struct(tsk, orig);
506 507 508 509 510 511 512 513 514 515

	/*
	 * arch_dup_task_struct() clobbers the stack-related fields.  Make
	 * sure they're properly initialized before using any stack-related
	 * functions again.
	 */
	tsk->stack = stack;
#ifdef CONFIG_VMAP_STACK
	tsk->stack_vm_area = stack_vm_area;
#endif
516 517 518
#ifdef CONFIG_THREAD_INFO_IN_TASK
	atomic_set(&tsk->stack_refcount, 1);
#endif
519

520
	if (err)
521
		goto free_stack;
522

523 524 525 526 527 528 529 530 531
#ifdef CONFIG_SECCOMP
	/*
	 * We must handle setting up seccomp filters once we're under
	 * the sighand lock in case orig has changed between now and
	 * then. Until then, filter must be NULL to avoid messing up
	 * the usage counts on the error path calling free_task.
	 */
	tsk->seccomp.filter = NULL;
#endif
532 533

	setup_thread_stack(tsk, orig);
534
	clear_user_return_notifier(tsk);
535
	clear_tsk_need_resched(tsk);
536
	set_task_stack_end_magic(tsk);
Linus Torvalds's avatar
Linus Torvalds committed
537

538 539 540 541
#ifdef CONFIG_CC_STACKPROTECTOR
	tsk->stack_canary = get_random_int();
#endif

542 543 544 545 546
	/*
	 * One for us, one for whoever does the "release_task()" (usually
	 * parent)
	 */
	atomic_set(&tsk->usage, 2);
547
#ifdef CONFIG_BLK_DEV_IO_TRACE
548
	tsk->btrace_seq = 0;
549
#endif
550
	tsk->splice_pipe = NULL;
551
	tsk->task_frag.page = NULL;
552
	tsk->wake_q.next = NULL;
553

554
	account_kernel_stack(tsk, 1);
555

556 557
	kcov_task_init(tsk);

Linus Torvalds's avatar
Linus Torvalds committed
558
	return tsk;
559

560
free_stack:
561
	free_thread_stack(tsk);
562
free_tsk:
563 564
	free_task_struct(tsk);
	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
565 566 567
}

#ifdef CONFIG_MMU
568 569
static __latent_entropy int dup_mmap(struct mm_struct *mm,
					struct mm_struct *oldmm)
Linus Torvalds's avatar
Linus Torvalds committed
570
{
571
	struct vm_area_struct *mpnt, *tmp, *prev, **pprev;
Linus Torvalds's avatar
Linus Torvalds committed
572 573 574
	struct rb_node **rb_link, *rb_parent;
	int retval;
	unsigned long charge;
575
	LIST_HEAD(uf);
Linus Torvalds's avatar
Linus Torvalds committed
576

577
	uprobe_start_dup_mmap();
578 579 580 581
	if (down_write_killable(&oldmm->mmap_sem)) {
		retval = -EINTR;
		goto fail_uprobe_end;
	}
582
	flush_cache_dup_mm(oldmm);
583
	uprobe_dup_mmap(oldmm, mm);
584 585 586 587
	/*
	 * Not linked in yet - no deadlock potential:
	 */
	down_write_nested(&mm->mmap_sem, SINGLE_DEPTH_NESTING);
588

589 590 591
	/* No ordering required: file already has been exposed. */
	RCU_INIT_POINTER(mm->exe_file, get_mm_exe_file(oldmm));

592
	mm->total_vm = oldmm->total_vm;
593
	mm->data_vm = oldmm->data_vm;
594 595 596
	mm->exec_vm = oldmm->exec_vm;
	mm->stack_vm = oldmm->stack_vm;

Linus Torvalds's avatar
Linus Torvalds committed
597 598 599
	rb_link = &mm->mm_rb.rb_node;
	rb_parent = NULL;
	pprev = &mm->mmap;
600
	retval = ksm_fork(mm, oldmm);
Andrea Arcangeli's avatar
Andrea Arcangeli committed
601 602 603
	if (retval)
		goto out;
	retval = khugepaged_fork(mm, oldmm);
604 605
	if (retval)
		goto out;
Linus Torvalds's avatar
Linus Torvalds committed
606

607
	prev = NULL;
608
	for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) {
Linus Torvalds's avatar
Linus Torvalds committed
609 610 611
		struct file *file;

		if (mpnt->vm_flags & VM_DONTCOPY) {
612
			vm_stat_account(mm, mpnt->vm_flags, -vma_pages(mpnt));
Linus Torvalds's avatar
Linus Torvalds committed
613 614 615 616
			continue;
		}
		charge = 0;
		if (mpnt->vm_flags & VM_ACCOUNT) {
617 618
			unsigned long len = vma_pages(mpnt);

619
			if (security_vm_enough_memory_mm(oldmm, len)) /* sic */
Linus Torvalds's avatar
Linus Torvalds committed
620 621 622
				goto fail_nomem;
			charge = len;
		}
623
		tmp = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
Linus Torvalds's avatar
Linus Torvalds committed
624 625 626
		if (!tmp)
			goto fail_nomem;
		*tmp = *mpnt;
627
		INIT_LIST_HEAD(&tmp->anon_vma_chain);
628 629
		retval = vma_dup_policy(mpnt, tmp);
		if (retval)
Linus Torvalds's avatar
Linus Torvalds committed
630
			goto fail_nomem_policy;
631
		tmp->vm_mm = mm;
632 633 634
		retval = dup_userfaultfd(tmp, &uf);
		if (retval)
			goto fail_nomem_anon_vma_fork;
635 636
		if (anon_vma_fork(tmp, mpnt))
			goto fail_nomem_anon_vma_fork;
637
		tmp->vm_flags &= ~(VM_LOCKED | VM_LOCKONFAULT);
638
		tmp->vm_next = tmp->vm_prev = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
639 640
		file = tmp->vm_file;
		if (file) {
Al Viro's avatar
Al Viro committed
641
			struct inode *inode = file_inode(file);
642 643
			struct address_space *mapping = file->f_mapping;

Linus Torvalds's avatar
Linus Torvalds committed
644 645 646
			get_file(file);
			if (tmp->vm_flags & VM_DENYWRITE)
				atomic_dec(&inode->i_writecount);
647
			i_mmap_lock_write(mapping);
648
			if (tmp->vm_flags & VM_SHARED)
649
				atomic_inc(&mapping->i_mmap_writable);
650 651
			flush_dcache_mmap_lock(mapping);
			/* insert tmp into the share list, just after mpnt */
652 653
			vma_interval_tree_insert_after(tmp, mpnt,
					&mapping->i_mmap);
654
			flush_dcache_mmap_unlock(mapping);
655
			i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
656 657
		}

658 659 660 661 662 663 664 665
		/*
		 * 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);

Linus Torvalds's avatar
Linus Torvalds committed
666
		/*
667
		 * Link in the new vma and copy the page table entries.
Linus Torvalds's avatar
Linus Torvalds committed
668 669 670
		 */
		*pprev = tmp;
		pprev = &tmp->vm_next;
671 672
		tmp->vm_prev = prev;
		prev = tmp;
Linus Torvalds's avatar
Linus Torvalds committed
673 674 675 676 677 678

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

		mm->map_count++;
679
		retval = copy_page_range(mm, oldmm, mpnt);
Linus Torvalds's avatar
Linus Torvalds committed
680 681 682 683 684 685 686

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

		if (retval)
			goto out;
	}
687 688
	/* a new mm has just been created */
	arch_dup_mmap(oldmm, mm);
Linus Torvalds's avatar
Linus Torvalds committed
689 690
	retval = 0;
out:
691
	up_write(&mm->mmap_sem);
692
	flush_tlb_mm(oldmm);
Linus Torvalds's avatar
Linus Torvalds committed
693
	up_write(&oldmm->mmap_sem);
694
	dup_userfaultfd_complete(&uf);
695
fail_uprobe_end:
696
	uprobe_end_dup_mmap();
Linus Torvalds's avatar
Linus Torvalds committed
697
	return retval;
698
fail_nomem_anon_vma_fork:
699
	mpol_put(vma_policy(tmp));
Linus Torvalds's avatar
Linus Torvalds committed
700 701 702 703 704 705 706 707
fail_nomem_policy:
	kmem_cache_free(vm_area_cachep, tmp);
fail_nomem:
	retval = -ENOMEM;
	vm_unacct_memory(charge);
	goto out;
}

708
static inline int mm_alloc_pgd(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
709 710 711 712 713 714 715
{
	mm->pgd = pgd_alloc(mm);
	if (unlikely(!mm->pgd))
		return -ENOMEM;
	return 0;
}

716
static inline void mm_free_pgd(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
717
{
718
	pgd_free(mm, mm->pgd);
Linus Torvalds's avatar
Linus Torvalds committed
719 720
}
#else
721 722 723 724 725 726 727
static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
{
	down_write(&oldmm->mmap_sem);
	RCU_INIT_POINTER(mm->exe_file, get_mm_exe_file(oldmm));
	up_write(&oldmm->mmap_sem);
	return 0;
}
Linus Torvalds's avatar
Linus Torvalds committed
728 729 730 731
#define mm_alloc_pgd(mm)	(0)
#define mm_free_pgd(mm)
#endif /* CONFIG_MMU */

Daniel Walker's avatar
Daniel Walker committed
732
__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
Linus Torvalds's avatar
Linus Torvalds committed
733

734
#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
Linus Torvalds's avatar
Linus Torvalds committed
735 736
#define free_mm(mm)	(kmem_cache_free(mm_cachep, (mm)))

737 738 739 740 741 742 743 744 745 746 747 748
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);

Linus Torvalds's avatar
Linus Torvalds committed
749 750
#include <linux/init_task.h>

751 752 753 754
static void mm_init_aio(struct mm_struct *mm)
{
#ifdef CONFIG_AIO
	spin_lock_init(&mm->ioctx_lock);
755
	mm->ioctx_table = NULL;
756 757 758
#endif
}

759 760 761 762 763 764 765
static void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
{
#ifdef CONFIG_MEMCG
	mm->owner = p;
#endif
}

766 767
static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
	struct user_namespace *user_ns)
Linus Torvalds's avatar
Linus Torvalds committed
768
{
769 770 771
	mm->mmap = NULL;
	mm->mm_rb = RB_ROOT;
	mm->vmacache_seqnum = 0;
Linus Torvalds's avatar
Linus Torvalds committed
772 773 774 775
	atomic_set(&mm->mm_users, 1);
	atomic_set(&mm->mm_count, 1);
	init_rwsem(&mm->mmap_sem);
	INIT_LIST_HEAD(&mm->mmlist);
776
	mm->core_state = NULL;
777
	atomic_long_set(&mm->nr_ptes, 0);
778
	mm_nr_pmds_init(mm);
779 780
	mm->map_count = 0;
	mm->locked_vm = 0;
781
	mm->pinned_vm = 0;
782
	memset(&mm->rss_stat, 0, sizeof(mm->rss_stat));
Linus Torvalds's avatar
Linus Torvalds committed
783
	spin_lock_init(&mm->page_table_lock);
784
	mm_init_cpumask(mm);
785
	mm_init_aio(mm);
786
	mm_init_owner(mm, p);
787
	mmu_notifier_mm_init(mm);
788
	clear_tlb_flush_pending(mm);
789 790 791
#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
	mm->pmd_huge_pte = NULL;
#endif
Linus Torvalds's avatar
Linus Torvalds committed
792

793 794 795 796 797
	if (current->mm) {
		mm->flags = current->mm->flags & MMF_INIT_MASK;
		mm->def_flags = current->mm->def_flags & VM_INIT_DEF_MASK;
	} else {
		mm->flags = default_dump_filter;
Linus Torvalds's avatar
Linus Torvalds committed
798
		mm->def_flags = 0;
799 800
	}

801 802 803 804 805
	if (mm_alloc_pgd(mm))
		goto fail_nopgd;

	if (init_new_context(p, mm))
		goto fail_nocontext;
806

807
	mm->user_ns = get_user_ns(user_ns);
808 809 810 811 812
	return mm;

fail_nocontext:
	mm_free_pgd(mm);
fail_nopgd:
Linus Torvalds's avatar
Linus Torvalds committed
813 814 815 816
	free_mm(mm);
	return NULL;
}

817 818 819 820 821 822 823 824 825 826 827
static void check_mm(struct mm_struct *mm)
{
	int i;

	for (i = 0; i < NR_MM_COUNTERS; i++) {
		long x = atomic_long_read(&mm->rss_stat.count[i]);

		if (unlikely(x))
			printk(KERN_ALERT "BUG: Bad rss-counter state "
					  "mm:%p idx:%d val:%ld\n", mm, i, x);
	}
828 829 830 831 832 833 834 835

	if (atomic_long_read(&mm->nr_ptes))
		pr_alert("BUG: non-zero nr_ptes on freeing mm: %ld\n",
				atomic_long_read(&mm->nr_ptes));
	if (mm_nr_pmds(mm))
		pr_alert("BUG: non-zero nr_pmds on freeing mm: %ld\n",
				mm_nr_pmds(mm));

836
#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
837
	VM_BUG_ON_MM(mm->pmd_huge_pte, mm);
838 839 840
#endif
}

Linus Torvalds's avatar
Linus Torvalds committed
841 842 843
/*
 * Allocate and initialize an mm_struct.
 */
844
struct mm_struct *mm_alloc(void)
Linus Torvalds's avatar
Linus Torvalds committed
845
{
846
	struct mm_struct *mm;
Linus Torvalds's avatar
Linus Torvalds committed
847 848

	mm = allocate_mm();
849 850 851 852
	if (!mm)
		return NULL;

	memset(mm, 0, sizeof(*mm));
853
	return mm_init(mm, current, current_user_ns());
Linus Torvalds's avatar
Linus Torvalds committed
854 855 856 857 858 859 860
}

/*
 * 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.
 */
861
void __mmdrop(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
862 863 864 865
{
	BUG_ON(mm == &init_mm);
	mm_free_pgd(mm);
	destroy_context(mm);
Andrea Arcangeli's avatar
Andrea Arcangeli committed
866
	mmu_notifier_mm_destroy(mm);
867
	check_mm(mm);
868
	put_user_ns(mm->user_ns);
Linus Torvalds's avatar
Linus Torvalds committed
869 870
	free_mm(mm);
}
871
EXPORT_SYMBOL_GPL(__mmdrop);
Linus Torvalds's avatar
Linus Torvalds committed
872

873 874 875 876 877 878 879 880 881
static inline void __mmput(struct mm_struct *mm)
{
	VM_BUG_ON(atomic_read(&mm->mm_users));

	uprobe_clear_state(mm);
	exit_aio(mm);
	ksm_exit(mm);
	khugepaged_exit(mm); /* must run before exit_mmap */
	exit_mmap(mm);
882
	mm_put_huge_zero_page(mm);
883 884 885 886 887 888 889 890
	set_mm_exe_file(mm, NULL);
	if (!list_empty(&mm->mmlist)) {
		spin_lock(&mmlist_lock);
		list_del(&mm->mmlist);
		spin_unlock(&mmlist_lock);
	}
	if (mm->binfmt)
		module_put(mm->binfmt->module);
891
	set_bit(MMF_OOM_SKIP, &mm->flags);
892 893 894
	mmdrop(mm);
}

Linus Torvalds's avatar
Linus Torvalds committed
895 896 897 898 899
/*
 * Decrement the use count and release all resources for an mm.
 */
void mmput(struct mm_struct *mm)
{
900 901
	might_sleep();

902 903 904 905 906
	if (atomic_dec_and_test(&mm->mm_users))
		__mmput(mm);
}
EXPORT_SYMBOL_GPL(mmput);

907
#ifdef CONFIG_MMU
908 909 910 911 912 913 914 915
static void mmput_async_fn(struct work_struct *work)
{
	struct mm_struct *mm = container_of(work, struct mm_struct, async_put_work);
	__mmput(mm);
}

void mmput_async(struct mm_struct *mm)
{
Linus Torvalds's avatar
Linus Torvalds committed
916
	if (atomic_dec_and_test(&mm->mm_users)) {
917 918
		INIT_WORK(&mm->async_put_work, mmput_async_fn);
		schedule_work(&mm->async_put_work);
Linus Torvalds's avatar
Linus Torvalds committed
919 920
	}
}
921
#endif
Linus Torvalds's avatar
Linus Torvalds committed
922

923 924 925 926 927
/**
 * set_mm_exe_file - change a reference to the mm's executable file
 *
 * This changes mm's executable file (shown as symlink /proc/[pid]/exe).
 *
928 929 930 931 932
 * Main users are mmput() and sys_execve(). Callers prevent concurrent
 * invocations: in mmput() nobody alive left, in execve task is single
 * threaded. sys_prctl(PR_SET_MM_MAP/EXE_FILE) also needs to set the
 * mm->exe_file, but does so without using set_mm_exe_file() in order
 * to do avoid the need for any locks.
933
 */
934 935
void set_mm_exe_file(struct mm_struct *mm, struct file *new_exe_file)
{
936 937 938 939 940 941 942 943
	struct file *old_exe_file;

	/*
	 * It is safe to dereference the exe_file without RCU as
	 * this function is only called if nobody else can access
	 * this mm -- see comment above for justification.
	 */
	old_exe_file = rcu_dereference_raw(mm->exe_file);
944

945 946
	if (new_exe_file)
		get_file(new_exe_file);
947 948 949
	rcu_assign_pointer(mm->exe_file, new_exe_file);
	if (old_exe_file)
		fput(old_exe_file);
950 951
}

952 953 954 955 956 957
/**
 * get_mm_exe_file - acquire a reference to the mm's executable file
 *
 * Returns %NULL if mm has no associated executable file.
 * User must release file via fput().
 */
958 959 960 961
struct file *get_mm_exe_file(struct mm_struct *mm)
{
	struct file *exe_file;

962 963 964 965 966
	rcu_read_lock();
	exe_file = rcu_dereference(mm->exe_file);
	if (exe_file && !get_file_rcu(exe_file))
		exe_file = NULL;
	rcu_read_unlock();
967 968
	return exe_file;
}
969
EXPORT_SYMBOL(get_mm_exe_file);
970

971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992
/**
 * get_task_exe_file - acquire a reference to the task's executable file
 *
 * Returns %NULL if task's mm (if any) has no associated executable file or
 * this is a kernel thread with borrowed mm (see the comment above get_task_mm).
 * User must release file via fput().
 */
struct file *get_task_exe_file(struct task_struct *task)
{
	struct file *exe_file = NULL;
	struct mm_struct *mm;

	task_lock(task);
	mm = task->mm;
	if (mm) {
		if (!(task->flags & PF_KTHREAD))
			exe_file = get_mm_exe_file(mm);
	}
	task_unlock(task);
	return exe_file;
}
EXPORT_SYMBOL(get_task_exe_file);
993

Linus Torvalds's avatar
Linus Torvalds committed
994 995 996
/**
 * get_task_mm - acquire a reference to the task's mm
 *
997
 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
Linus Torvalds's avatar
Linus Torvalds committed
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
 * 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) {
1010
		if (task->flags & PF_KTHREAD)
Linus Torvalds's avatar
Linus Torvalds committed
1011 1012
			mm = NULL;
		else
1013
			mmget(mm);
Linus Torvalds's avatar
Linus Torvalds committed
1014 1015 1016 1017 1018 1019
	}
	task_unlock(task);
	return mm;
}
EXPORT_SYMBOL_GPL(get_task_mm);

1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
struct mm_struct *mm_access(struct task_struct *task, unsigned int mode)
{
	struct mm_struct *mm;
	int err;

	err =  mutex_lock_killable(&task->signal->cred_guard_mutex);
	if (err)
		return ERR_PTR(err);

	mm = get_task_mm(task);
	if (mm && mm != current->mm &&
			!ptrace_may_access(task, mode)) {
		mmput(mm);
		mm = ERR_PTR(-EACCES);
	}
	mutex_unlock(&task->signal->cred_guard_mutex);

	return mm;
}

1040
static void complete_vfork_done(struct task_struct *tsk)
1041
{
Oleg Nesterov's avatar
Oleg Nesterov committed
1042
	struct completion *vfork;
1043

Oleg Nesterov's avatar
Oleg Nesterov committed
1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
	task_lock(tsk);
	vfork = tsk->vfork_done;
	if (likely(vfork)) {
		tsk->vfork_done = NULL;
		complete(vfork);
	}
	task_unlock(tsk);
}

static int wait_for_vfork_done(struct task_struct *child,
				struct completion *vfork)
{
	int killed;

	freezer_do_not_count();
	killed = wait_for_completion_killable(vfork);
	freezer_count();

	if (killed) {
		task_lock(child);
		child->vfork_done = NULL;
		task_unlock(child);
	}

	put_task_struct(child);
	return killed;
1070 1071
}

Linus Torvalds's avatar
Linus Torvalds committed
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
/* 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)
{
1087 1088
	/* Get rid of any futexes when releasing the mm */
#ifdef CONFIG_FUTEX
1089
	if (unlikely(tsk->robust_list)) {
1090
		exit_robust_list(tsk);
1091 1092
		tsk->robust_list = NULL;
	}
1093
#ifdef CONFIG_COMPAT
1094
	if (unlikely(tsk->compat_robust_list)) {
1095
		compat_exit_robust_list(tsk);
1096 1097
		tsk->compat_robust_list = NULL;
	}
1098
#endif
1099 1100
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
1101 1102
#endif

1103 1104
	uprobe_free_utask(tsk);

Linus Torvalds's avatar
Linus Torvalds committed
1105 1106 1107
	/* Get rid of any cached register state */
	deactivate_mm(tsk, mm);

1108
	/*
1109 1110 1111
	 * Signal userspace if we're not exiting with a core dump
	 * because we want to leave the value intact for debugging
	 * purposes.
1112
	 */
1113
	if (tsk->clear_child_tid) {
1114
		if (!(tsk->signal->flags & SIGNAL_GROUP_COREDUMP) &&
1115 1116 1117 1118 1119 1120 1121 1122 1123
		    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);
		}
Linus Torvalds's avatar
Linus Torvalds committed
1124 1125
		tsk->clear_child_tid = NULL;
	}
1126 1127 1128 1129 1130 1131 1132

	/*
	 * All done, finally we can wake up parent and return this mm to him.
	 * Also kthread_stop() uses this completion for synchronization.
	 */
	if (tsk->vfork_done)
		complete_vfork_done(tsk);
Linus Torvalds's avatar
Linus Torvalds committed
1133 1134
}

1135 1136 1137 1138
/*
 * Allocate a new mm structure and copy contents from the
 * mm structure of the passed in task structure.
 */
1139
static struct mm_struct *dup_mm(struct task_struct *tsk)
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
{
	struct mm_struct *mm, *oldmm = current->mm;
	int err;

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

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

1150
	if (!mm_init(mm, tsk, mm->user_ns))
1151 1152 1153 1154 1155 1156 1157 1158 1159
		goto fail_nomem;

	err = dup_mmap(mm, oldmm);
	if (err)
		goto free_pt;

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

1160 1161 1162
	if (mm->binfmt && !try_module_get(mm->binfmt->module))
		goto free_pt;

1163 1164 1165
	return mm;

free_pt:
1166 1167
	/* don't put binfmt in mmput, we haven't got module yet */
	mm->binfmt = NULL;
1168 1169 1170 1171 1172 1173
	mmput(mm);

fail_nomem:
	return NULL;
}

1174
static int copy_mm(unsigned long clone_flags, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
1175
{
1176
	struct mm_struct *mm, *oldmm;
Linus Torvalds's avatar
Linus Torvalds committed
1177 1178 1179 1180
	int retval;

	tsk->min_flt = tsk->maj_flt = 0;
	tsk->nvcsw = tsk->nivcsw = 0;
1181 1182 1183
#ifdef CONFIG_DETECT_HUNG_TASK
	tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw;
#endif
Linus Torvalds's avatar
Linus Torvalds committed
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196

	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;

1197 1198 1199
	/* initialize the new vmacache entries */
	vmacache_flush(tsk);

Linus Torvalds's avatar
Linus Torvalds committed
1200
	if (clone_flags & CLONE_VM) {
1201
		mmget(oldmm);
Linus Torvalds's avatar
Linus Torvalds committed
1202 1203 1204 1205 1206
		mm = oldmm;
		goto good_mm;
	}

	retval = -ENOMEM;
1207
	mm = dup_mm(tsk);
Linus Torvalds's avatar
Linus Torvalds committed
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219
	if (!mm)
		goto fail_nomem;

good_mm:
	tsk->mm = mm;
	tsk->active_mm = mm;
	return 0;

fail_nomem:
	return retval;
}

Alexey Dobriyan's avatar
Alexey Dobriyan committed
1220
static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
1221
{
1222
	struct fs_struct *fs = current->fs;
Linus Torvalds's avatar
Linus Torvalds committed
1223
	if (clone_flags & CLONE_FS) {
1224
		/* tsk->fs is already what we want */
1225
		spin_lock(&fs->lock);
1226
		if (fs->in_exec) {
1227
			spin_unlock(&fs->lock);
1228 1229 1230
			return -EAGAIN;
		}
		fs->users++;
1231
		spin_unlock(&fs->lock);
Linus Torvalds's avatar
Linus Torvalds committed
1232 1233
		return 0;
	}
1234
	tsk->fs = copy_fs_struct(fs);
Linus Torvalds's avatar
Linus Torvalds committed
1235 1236 1237 1238 1239
	if (!tsk->fs)
		return -ENOMEM;
	return 0;
}

1240
static int copy_files(unsigned long clone_flags, struct task_struct *tsk)
1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266
{
	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;
}

1267
static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
1268 1269 1270
{
#ifdef CONFIG_BLOCK
	struct io_context *ioc = current->io_context;
1271
	struct io_context *new_ioc;
1272 1273 1274

	if (!ioc)
		return 0;
1275 1276 1277 1278
	/*
	 * Share io context with parent, if CLONE_IO is set
	 */
	if (clone_flags & CLONE_IO) {
1279 1280
		ioc_task_link(ioc);
		tsk->io_context = ioc;
1281
	} else if (ioprio_valid(ioc->ioprio)) {
1282 1283
		new_ioc = get_task_io_context(tsk, GFP_KERNEL, NUMA_NO_NODE);
		if (unlikely(!new_ioc))
1284 1285
			return -ENOMEM;

1286
		new_ioc->ioprio = ioc->ioprio;
1287
		put_io_context(new_ioc);
1288 1289 1290 1291 1292
	}
#endif
	return 0;
}

Alexey Dobriyan's avatar
Alexey Dobriyan committed
1293
static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
1294 1295 1296
{
	struct sighand_struct *sig;

1297
	if (clone_flags & CLONE_SIGHAND) {
Linus Torvalds's avatar
Linus Torvalds committed
1298 1299 1300 1301
		atomic_inc(&current->sighand->count);
		return 0;
	}
	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
1302
	rcu_assign_pointer(tsk->sighand, sig);
Linus Torvalds's avatar
Linus Torvalds committed
1303 1304
	if (!sig)
		return -ENOMEM;
1305

Linus Torvalds's avatar
Linus Torvalds committed
1306 1307 1308 1309 1310
	atomic_set(&sig->count, 1);
	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
	return 0;
}

1311
void __cleanup_sighand(struct sighand_struct *sighand)
1312
{
1313 1314
	if (atomic_dec_and_test(&sighand->count)) {
		signalfd_cleanup(sighand);
1315 1316 1317 1318
		/*
		 * sighand_cachep is SLAB_DESTROY_BY_RCU so we can free it
		 * without an RCU grace period, see __lock_task_sighand().
		 */
1319
		kmem_cache_free(sighand_cachep, sighand);
1320
	}
1321 1322
}

1323
#ifdef CONFIG_POSIX_TIMERS
1324 1325 1326 1327 1328
/*
 * Initialize POSIX timer handling for a thread group.
 */
static void posix_cpu_timers_init_group(struct signal_struct *sig)
{
1329 1330
	unsigned long cpu_limit;

1331
	cpu_limit = READ_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur);
1332
	if (cpu_limit != RLIM_INFINITY) {
1333
		sig->cputime_expires.prof_exp = cpu_limit * NSEC_PER_SEC;
1334
		sig->cputimer.running = true;
1335 1336
	}

1337 1338 1339 1340 1341
	/* The timer lists. */
	INIT_LIST_HEAD(&sig->cpu_timers[0]);
	INIT_LIST_HEAD(&sig->cpu_timers[1]);
	INIT_LIST_HEAD(&sig->cpu_timers[2]);
}
1342 1343 1344
#else
static inline void posix_cpu_timers_init_group(struct signal_struct *sig) { }
#endif
1345

Alexey Dobriyan's avatar
Alexey Dobriyan committed
1346
static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)