fork.c 60.1 KB
Newer Older
L
Linus Torvalds 已提交
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>
L
Linus Torvalds 已提交
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>
A
Al Viro 已提交
34
#include <linux/fdtable.h>
35
#include <linux/iocontext.h>
L
Linus Torvalds 已提交
36 37 38
#include <linux/key.h>
#include <linux/binfmts.h>
#include <linux/mman.h>
A
Andrea Arcangeli 已提交
39
#include <linux/mmu_notifier.h>
40
#include <linux/hmm.h>
L
Linus Torvalds 已提交
41
#include <linux/fs.h>
D
Davidlohr Bueso 已提交
42 43
#include <linux/mm.h>
#include <linux/vmacache.h>
S
Serge E. Hallyn 已提交
44
#include <linux/nsproxy.h>
45
#include <linux/capability.h>
L
Linus Torvalds 已提交
46
#include <linux/cpu.h>
47
#include <linux/cgroup.h>
L
Linus Torvalds 已提交
48
#include <linux/security.h>
49
#include <linux/hugetlb.h>
50
#include <linux/seccomp.h>
L
Linus Torvalds 已提交
51 52 53 54
#include <linux/swap.h>
#include <linux/syscalls.h>
#include <linux/jiffies.h>
#include <linux/futex.h>
55
#include <linux/compat.h>
56
#include <linux/kthread.h>
57
#include <linux/task_io_accounting_ops.h>
58
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
59 60 61
#include <linux/ptrace.h>
#include <linux/mount.h>
#include <linux/audit.h>
62
#include <linux/memcontrol.h>
63
#include <linux/ftrace.h>
64
#include <linux/proc_fs.h>
L
Linus Torvalds 已提交
65 66
#include <linux/profile.h>
#include <linux/rmap.h>
H
Hugh Dickins 已提交
67
#include <linux/ksm.h>
L
Linus Torvalds 已提交
68
#include <linux/acct.h>
69
#include <linux/userfaultfd_k.h>
70
#include <linux/tsacct_kern.h>
M
Matt Helsley 已提交
71
#include <linux/cn_proc.h>
R
Rafael J. Wysocki 已提交
72
#include <linux/freezer.h>
73
#include <linux/delayacct.h>
74
#include <linux/taskstats_kern.h>
75
#include <linux/random.h>
M
Miloslav Trmac 已提交
76
#include <linux/tty.h>
77
#include <linux/blkdev.h>
78
#include <linux/fs_struct.h>
79
#include <linux/magic.h>
80
#include <linux/sched/mm.h>
81
#include <linux/perf_event.h>
82
#include <linux/posix-timers.h>
83
#include <linux/user-return-notifier.h>
Y
Ying Han 已提交
84
#include <linux/oom.h>
A
Andrea Arcangeli 已提交
85
#include <linux/khugepaged.h>
86
#include <linux/signalfd.h>
87
#include <linux/uprobes.h>
88
#include <linux/aio.h>
89
#include <linux/compiler.h>
90
#include <linux/sysctl.h>
D
Dmitry Vyukov 已提交
91
#include <linux/kcov.h>
92
#include <linux/livepatch.h>
93
#include <linux/thread_info.h>
L
Linus Torvalds 已提交
94 95 96

#include <asm/pgtable.h>
#include <asm/pgalloc.h>
97
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
98 99 100 101
#include <asm/mmu_context.h>
#include <asm/cacheflush.h>
#include <asm/tlbflush.h>

102 103
#include <trace/events/sched.h>

104 105 106
#define CREATE_TRACE_POINTS
#include <trace/events/task.h>

107 108 109 110 111 112 113 114 115 116
/*
 * Minimum number of threads to boot the kernel
 */
#define MIN_THREADS 20

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

L
Linus Torvalds 已提交
117 118 119 120
/*
 * Protected counters by write_lock_irq(&tasklist_lock)
 */
unsigned long total_forks;	/* Handle normal Linux uptimes. */
121
int nr_threads;			/* The idle threads do not count.. */
L
Linus Torvalds 已提交
122 123 124 125 126

int max_threads;		/* tunable limit on nr_threads */

DEFINE_PER_CPU(unsigned long, process_counts) = 0;

127
__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
128 129 130 131 132 133 134 135

#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 */
L
Linus Torvalds 已提交
136 137 138 139 140 141

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

142
	for_each_possible_cpu(cpu)
L
Linus Torvalds 已提交
143 144 145 146 147
		total += per_cpu(process_counts, cpu);

	return total;
}

148 149 150 151
void __weak arch_release_task_struct(struct task_struct *tsk)
{
}

152
#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
153
static struct kmem_cache *task_struct_cachep;
154 155 156 157 158 159 160 161 162 163

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);
}
L
Linus Torvalds 已提交
164 165
#endif

166
void __weak arch_release_thread_stack(unsigned long *stack)
167 168 169
{
}

170
#ifndef CONFIG_ARCH_THREAD_STACK_ALLOCATOR
171

172 173 174 175
/*
 * Allocate pages if THREAD_SIZE is >= PAGE_SIZE, otherwise use a
 * kmemcache based allocator.
 */
176
# if THREAD_SIZE >= PAGE_SIZE || defined(CONFIG_VMAP_STACK)
177 178 179 180 181 182 183 184

#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]);
185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202

static int free_vm_stack_cache(unsigned int cpu)
{
	struct vm_struct **cached_vm_stacks = per_cpu_ptr(cached_stacks, cpu);
	int i;

	for (i = 0; i < NR_CACHED_STACKS; i++) {
		struct vm_struct *vm_stack = cached_vm_stacks[i];

		if (!vm_stack)
			continue;

		vfree(vm_stack->addr);
		cached_vm_stacks[i] = NULL;
	}

	return 0;
}
203 204
#endif

205
static unsigned long *alloc_thread_stack_node(struct task_struct *tsk, int node)
206
{
207
#ifdef CONFIG_VMAP_STACK
208 209 210 211
	void *stack;
	int i;

	for (i = 0; i < NR_CACHED_STACKS; i++) {
212 213 214
		struct vm_struct *s;

		s = this_cpu_xchg(cached_stacks[i], NULL);
215 216 217 218

		if (!s)
			continue;

219 220 221 222
#ifdef CONFIG_DEBUG_KMEMLEAK
		/* Clear stale pointers from reused stack. */
		memset(s->addr, 0, THREAD_SIZE);
#endif
223 224 225 226
		tsk->stack_vm_area = s;
		return s->addr;
	}

227
	stack = __vmalloc_node_range(THREAD_SIZE, THREAD_ALIGN,
228
				     VMALLOC_START, VMALLOC_END,
229
				     THREADINFO_GFP,
230 231
				     PAGE_KERNEL,
				     0, node, __builtin_return_address(0));
232 233 234 235 236 237 238 239 240 241

	/*
	 * 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
242 243
	struct page *page = alloc_pages_node(node, THREADINFO_GFP,
					     THREAD_SIZE_ORDER);
244 245

	return page ? page_address(page) : NULL;
246
#endif
247 248
}

249
static inline void free_thread_stack(struct task_struct *tsk)
250
{
251 252 253 254 255
#ifdef CONFIG_VMAP_STACK
	if (task_stack_vm_area(tsk)) {
		int i;

		for (i = 0; i < NR_CACHED_STACKS; i++) {
256 257
			if (this_cpu_cmpxchg(cached_stacks[i],
					NULL, tsk->stack_vm_area) != NULL)
258 259 260 261 262
				continue;

			return;
		}

263
		vfree_atomic(tsk->stack);
264 265 266 267 268
		return;
	}
#endif

	__free_pages(virt_to_page(tsk->stack), THREAD_SIZE_ORDER);
269
}
270
# else
271
static struct kmem_cache *thread_stack_cache;
272

273
static unsigned long *alloc_thread_stack_node(struct task_struct *tsk,
274 275
						  int node)
{
276
	return kmem_cache_alloc_node(thread_stack_cache, THREADINFO_GFP, node);
277 278
}

279
static void free_thread_stack(struct task_struct *tsk)
280
{
281
	kmem_cache_free(thread_stack_cache, tsk->stack);
282 283
}

284
void thread_stack_cache_init(void)
285
{
286 287 288
	thread_stack_cache = kmem_cache_create_usercopy("thread_stack",
					THREAD_SIZE, THREAD_SIZE, 0, 0,
					THREAD_SIZE, NULL);
289
	BUG_ON(thread_stack_cache == NULL);
290 291
}
# endif
292 293
#endif

L
Linus Torvalds 已提交
294
/* SLAB cache for signal_struct structures (tsk->signal) */
295
static struct kmem_cache *signal_cachep;
L
Linus Torvalds 已提交
296 297

/* SLAB cache for sighand_struct structures (tsk->sighand) */
298
struct kmem_cache *sighand_cachep;
L
Linus Torvalds 已提交
299 300

/* SLAB cache for files_struct structures (tsk->files) */
301
struct kmem_cache *files_cachep;
L
Linus Torvalds 已提交
302 303

/* SLAB cache for fs_struct structures (tsk->fs) */
304
struct kmem_cache *fs_cachep;
L
Linus Torvalds 已提交
305 306

/* SLAB cache for vm_area_struct structures */
307
struct kmem_cache *vm_area_cachep;
L
Linus Torvalds 已提交
308 309

/* SLAB cache for mm_struct structures (tsk->mm) */
310
static struct kmem_cache *mm_cachep;
L
Linus Torvalds 已提交
311

312
static void account_kernel_stack(struct task_struct *tsk, int account)
313
{
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330
	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. */
331 332
		mod_memcg_page_state(vm->pages[0], MEMCG_KERNEL_STACK_KB,
				     account * (THREAD_SIZE / 1024));
333 334 335 336 337 338 339 340 341 342
	} 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);

343 344
		mod_memcg_page_state(first_page, MEMCG_KERNEL_STACK_KB,
				     account * (THREAD_SIZE / 1024));
345
	}
346 347
}

348
static void release_task_stack(struct task_struct *tsk)
L
Linus Torvalds 已提交
349
{
350 351 352
	if (WARN_ON(tsk->state != TASK_DEAD))
		return;  /* Better to leak the stack than to free prematurely */

353
	account_kernel_stack(tsk, -1);
354
	arch_release_thread_stack(tsk->stack);
355
	free_thread_stack(tsk);
356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384
	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
I
Ingo Molnar 已提交
385
	rt_mutex_debug_task_free(tsk);
386
	ftrace_graph_exit_task(tsk);
387
	put_seccomp_filter(tsk);
388
	arch_release_task_struct(tsk);
389 390
	if (tsk->flags & PF_KTHREAD)
		free_kthread_struct(tsk);
L
Linus Torvalds 已提交
391 392 393 394
	free_task_struct(tsk);
}
EXPORT_SYMBOL(free_task);

395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
#ifdef CONFIG_MMU
static __latent_entropy int dup_mmap(struct mm_struct *mm,
					struct mm_struct *oldmm)
{
	struct vm_area_struct *mpnt, *tmp, *prev, **pprev;
	struct rb_node **rb_link, *rb_parent;
	int retval;
	unsigned long charge;
	LIST_HEAD(uf);

	uprobe_start_dup_mmap();
	if (down_write_killable(&oldmm->mmap_sem)) {
		retval = -EINTR;
		goto fail_uprobe_end;
	}
	flush_cache_dup_mm(oldmm);
	uprobe_dup_mmap(oldmm, mm);
	/*
	 * Not linked in yet - no deadlock potential:
	 */
	down_write_nested(&mm->mmap_sem, SINGLE_DEPTH_NESTING);

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

	mm->total_vm = oldmm->total_vm;
	mm->data_vm = oldmm->data_vm;
	mm->exec_vm = oldmm->exec_vm;
	mm->stack_vm = oldmm->stack_vm;

	rb_link = &mm->mm_rb.rb_node;
	rb_parent = NULL;
	pprev = &mm->mmap;
	retval = ksm_fork(mm, oldmm);
	if (retval)
		goto out;
	retval = khugepaged_fork(mm, oldmm);
	if (retval)
		goto out;

	prev = NULL;
	for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) {
		struct file *file;

		if (mpnt->vm_flags & VM_DONTCOPY) {
			vm_stat_account(mm, mpnt->vm_flags, -vma_pages(mpnt));
			continue;
		}
		charge = 0;
		if (mpnt->vm_flags & VM_ACCOUNT) {
			unsigned long len = vma_pages(mpnt);

			if (security_vm_enough_memory_mm(oldmm, len)) /* sic */
				goto fail_nomem;
			charge = len;
		}
		tmp = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
		if (!tmp)
			goto fail_nomem;
		*tmp = *mpnt;
		INIT_LIST_HEAD(&tmp->anon_vma_chain);
		retval = vma_dup_policy(mpnt, tmp);
		if (retval)
			goto fail_nomem_policy;
		tmp->vm_mm = mm;
		retval = dup_userfaultfd(tmp, &uf);
		if (retval)
			goto fail_nomem_anon_vma_fork;
		if (tmp->vm_flags & VM_WIPEONFORK) {
			/* VM_WIPEONFORK gets a clean slate in the child. */
			tmp->anon_vma = NULL;
			if (anon_vma_prepare(tmp))
				goto fail_nomem_anon_vma_fork;
		} else if (anon_vma_fork(tmp, mpnt))
			goto fail_nomem_anon_vma_fork;
		tmp->vm_flags &= ~(VM_LOCKED | VM_LOCKONFAULT);
		tmp->vm_next = tmp->vm_prev = NULL;
		file = tmp->vm_file;
		if (file) {
			struct inode *inode = file_inode(file);
			struct address_space *mapping = file->f_mapping;

			get_file(file);
			if (tmp->vm_flags & VM_DENYWRITE)
				atomic_dec(&inode->i_writecount);
			i_mmap_lock_write(mapping);
			if (tmp->vm_flags & VM_SHARED)
				atomic_inc(&mapping->i_mmap_writable);
			flush_dcache_mmap_lock(mapping);
			/* insert tmp into the share list, just after mpnt */
			vma_interval_tree_insert_after(tmp, mpnt,
					&mapping->i_mmap);
			flush_dcache_mmap_unlock(mapping);
			i_mmap_unlock_write(mapping);
		}

		/*
		 * 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);

		/*
		 * Link in the new vma and copy the page table entries.
		 */
		*pprev = tmp;
		pprev = &tmp->vm_next;
		tmp->vm_prev = prev;
		prev = tmp;

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

		mm->map_count++;
		if (!(tmp->vm_flags & VM_WIPEONFORK))
			retval = copy_page_range(mm, oldmm, mpnt);

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

		if (retval)
			goto out;
	}
	/* a new mm has just been created */
	arch_dup_mmap(oldmm, mm);
	retval = 0;
out:
	up_write(&mm->mmap_sem);
	flush_tlb_mm(oldmm);
	up_write(&oldmm->mmap_sem);
	dup_userfaultfd_complete(&uf);
fail_uprobe_end:
	uprobe_end_dup_mmap();
	return retval;
fail_nomem_anon_vma_fork:
	mpol_put(vma_policy(tmp));
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)
{
	pgd_free(mm, mm->pgd);
}
#else
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;
}
#define mm_alloc_pgd(mm)	(0)
#define mm_free_pgd(mm)
#endif /* CONFIG_MMU */

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);
	}

	if (mm_pgtables_bytes(mm))
		pr_alert("BUG: non-zero pgtables_bytes on freeing mm: %ld\n",
				mm_pgtables_bytes(mm));

#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
	VM_BUG_ON_MM(mm->pmd_huge_pte, mm);
#endif
}

#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
#define free_mm(mm)	(kmem_cache_free(mm_cachep, (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.
 */
static void __mmdrop(struct mm_struct *mm)
{
	BUG_ON(mm == &init_mm);
	mm_free_pgd(mm);
	destroy_context(mm);
	hmm_mm_destroy(mm);
	mmu_notifier_mm_destroy(mm);
	check_mm(mm);
	put_user_ns(mm->user_ns);
	free_mm(mm);
}

void mmdrop(struct mm_struct *mm)
{
	if (unlikely(atomic_dec_and_test(&mm->mm_count)))
		__mmdrop(mm);
}
EXPORT_SYMBOL_GPL(mmdrop);

static void mmdrop_async_fn(struct work_struct *work)
{
	struct mm_struct *mm;

	mm = container_of(work, struct mm_struct, async_put_work);
	__mmdrop(mm);
}

static void mmdrop_async(struct mm_struct *mm)
{
	if (unlikely(atomic_dec_and_test(&mm->mm_count))) {
		INIT_WORK(&mm->async_put_work, mmdrop_async_fn);
		schedule_work(&mm->async_put_work);
	}
}

630 631
static inline void free_signal_struct(struct signal_struct *sig)
{
632
	taskstats_tgid_free(sig);
633
	sched_autogroup_exit(sig);
634 635 636 637
	/*
	 * __mmdrop is not safe to call from softirq context on x86 due to
	 * pgd_dtor so postpone it to the async context
	 */
638
	if (sig->oom_mm)
639
		mmdrop_async(sig->oom_mm);
640 641 642 643 644
	kmem_cache_free(signal_cachep, sig);
}

static inline void put_signal_struct(struct signal_struct *sig)
{
645
	if (atomic_dec_and_test(&sig->sigcnt))
646 647 648
		free_signal_struct(sig);
}

649
void __put_task_struct(struct task_struct *tsk)
L
Linus Torvalds 已提交
650
{
E
Eugene Teo 已提交
651
	WARN_ON(!tsk->exit_state);
L
Linus Torvalds 已提交
652 653 654
	WARN_ON(atomic_read(&tsk->usage));
	WARN_ON(tsk == current);

655
	cgroup_free(tsk);
656
	task_numa_free(tsk);
657
	security_task_free(tsk);
658
	exit_creds(tsk);
659
	delayacct_tsk_free(tsk);
660
	put_signal_struct(tsk->signal);
L
Linus Torvalds 已提交
661 662 663 664

	if (!profile_handoff_task(tsk))
		free_task(tsk);
}
665
EXPORT_SYMBOL_GPL(__put_task_struct);
L
Linus Torvalds 已提交
666

T
Thomas Gleixner 已提交
667
void __init __weak arch_task_cache_init(void) { }
668

669 670 671
/*
 * set_max_threads
 */
672
static void set_max_threads(unsigned int max_threads_suggested)
673
{
674
	u64 threads;
675 676

	/*
677 678
	 * The number of threads shall be limited such that the thread
	 * structures may only consume a small part of the available memory.
679
	 */
680 681 682 683 684 685
	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);

686 687 688
	if (threads > max_threads_suggested)
		threads = max_threads_suggested;

689
	max_threads = clamp_t(u64, threads, MIN_THREADS, MAX_THREADS);
690 691
}

692 693 694 695
#ifdef CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT
/* Initialized by the architecture: */
int arch_task_struct_size __read_mostly;
#endif
696

697 698 699 700 701 702 703 704 705 706 707 708 709 710 711
static void task_struct_whitelist(unsigned long *offset, unsigned long *size)
{
	/* Fetch thread_struct whitelist for the architecture. */
	arch_thread_struct_whitelist(offset, size);

	/*
	 * Handle zero-sized whitelist or empty thread_struct, otherwise
	 * adjust offset to position of thread_struct in task_struct.
	 */
	if (unlikely(*size == 0))
		*offset = 0;
	else
		*offset += offsetof(struct task_struct, thread);
}

712
void __init fork_init(void)
L
Linus Torvalds 已提交
713
{
714
	int i;
715
#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
L
Linus Torvalds 已提交
716
#ifndef ARCH_MIN_TASKALIGN
717
#define ARCH_MIN_TASKALIGN	0
L
Linus Torvalds 已提交
718
#endif
P
Peter Zijlstra 已提交
719
	int align = max_t(int, L1_CACHE_BYTES, ARCH_MIN_TASKALIGN);
720
	unsigned long useroffset, usersize;
721

L
Linus Torvalds 已提交
722
	/* create a slab on which task_structs can be allocated */
723 724
	task_struct_whitelist(&useroffset, &usersize);
	task_struct_cachep = kmem_cache_create_usercopy("task_struct",
725
			arch_task_struct_size, align,
726 727
			SLAB_PANIC|SLAB_ACCOUNT,
			useroffset, usersize, NULL);
L
Linus Torvalds 已提交
728 729
#endif

730 731 732
	/* do the arch specific task caches init */
	arch_task_cache_init();

733
	set_max_threads(MAX_THREADS);
L
Linus Torvalds 已提交
734 735 736 737 738

	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];
739

740 741 742
	for (i = 0; i < UCOUNT_COUNTS; i++) {
		init_user_ns.ucount_max[i] = max_threads/2;
	}
743 744 745 746 747

#ifdef CONFIG_VMAP_STACK
	cpuhp_setup_state(CPUHP_BP_PREPARE_DYN, "fork:vm_stack_cache",
			  NULL, free_vm_stack_cache);
#endif
748 749

	lockdep_init_task(&init_task);
L
Linus Torvalds 已提交
750 751
}

752
int __weak arch_dup_task_struct(struct task_struct *dst,
753 754 755 756 757 758
					       struct task_struct *src)
{
	*dst = *src;
	return 0;
}

759 760 761 762 763 764 765 766
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 */
}

767
static struct task_struct *dup_task_struct(struct task_struct *orig, int node)
L
Linus Torvalds 已提交
768 769
{
	struct task_struct *tsk;
770
	unsigned long *stack;
771
	struct vm_struct *stack_vm_area;
P
Peter Zijlstra 已提交
772
	int err;
L
Linus Torvalds 已提交
773

774 775
	if (node == NUMA_NO_NODE)
		node = tsk_fork_get_node(orig);
776
	tsk = alloc_task_struct_node(node);
L
Linus Torvalds 已提交
777 778 779
	if (!tsk)
		return NULL;

780 781
	stack = alloc_thread_stack_node(tsk, node);
	if (!stack)
782
		goto free_tsk;
L
Linus Torvalds 已提交
783

784 785
	stack_vm_area = task_stack_vm_area(tsk);

786
	err = arch_dup_task_struct(tsk, orig);
787 788 789 790 791 792 793 794 795 796

	/*
	 * 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
797 798 799
#ifdef CONFIG_THREAD_INFO_IN_TASK
	atomic_set(&tsk->stack_refcount, 1);
#endif
800

801
	if (err)
802
		goto free_stack;
803

K
Kees Cook 已提交
804 805 806 807 808 809 810 811 812
#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
813 814

	setup_thread_stack(tsk, orig);
815
	clear_user_return_notifier(tsk);
816
	clear_tsk_need_resched(tsk);
817
	set_task_stack_end_magic(tsk);
L
Linus Torvalds 已提交
818

819
#ifdef CONFIG_CC_STACKPROTECTOR
820
	tsk->stack_canary = get_random_canary();
821 822
#endif

823 824 825 826 827
	/*
	 * One for us, one for whoever does the "release_task()" (usually
	 * parent)
	 */
	atomic_set(&tsk->usage, 2);
828
#ifdef CONFIG_BLK_DEV_IO_TRACE
829
	tsk->btrace_seq = 0;
830
#endif
831
	tsk->splice_pipe = NULL;
832
	tsk->task_frag.page = NULL;
833
	tsk->wake_q.next = NULL;
834

835
	account_kernel_stack(tsk, 1);
836

D
Dmitry Vyukov 已提交
837 838
	kcov_task_init(tsk);

839 840 841 842
#ifdef CONFIG_FAULT_INJECTION
	tsk->fail_nth = 0;
#endif

L
Linus Torvalds 已提交
843
	return tsk;
844

845
free_stack:
846
	free_thread_stack(tsk);
847
free_tsk:
848 849
	free_task_struct(tsk);
	return NULL;
L
Linus Torvalds 已提交
850 851
}

D
Daniel Walker 已提交
852
__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
L
Linus Torvalds 已提交
853

854 855 856 857 858 859 860 861 862 863 864 865
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
866 867
#include <linux/init_task.h>

A
Alexey Dobriyan 已提交
868 869 870 871
static void mm_init_aio(struct mm_struct *mm)
{
#ifdef CONFIG_AIO
	spin_lock_init(&mm->ioctx_lock);
872
	mm->ioctx_table = NULL;
A
Alexey Dobriyan 已提交
873 874 875
#endif
}

876 877 878 879 880 881 882
static void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
{
#ifdef CONFIG_MEMCG
	mm->owner = p;
#endif
}

883 884 885 886 887 888 889
static void mm_init_uprobes_state(struct mm_struct *mm)
{
#ifdef CONFIG_UPROBES
	mm->uprobes_state.xol_area = NULL;
#endif
}

890 891
static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
	struct user_namespace *user_ns)
L
Linus Torvalds 已提交
892
{
893 894 895
	mm->mmap = NULL;
	mm->mm_rb = RB_ROOT;
	mm->vmacache_seqnum = 0;
L
Linus Torvalds 已提交
896 897 898 899
	atomic_set(&mm->mm_users, 1);
	atomic_set(&mm->mm_count, 1);
	init_rwsem(&mm->mmap_sem);
	INIT_LIST_HEAD(&mm->mmlist);
900
	mm->core_state = NULL;
901
	mm_pgtables_bytes_init(mm);
902 903
	mm->map_count = 0;
	mm->locked_vm = 0;
V
Vladimir Davydov 已提交
904
	mm->pinned_vm = 0;
K
KAMEZAWA Hiroyuki 已提交
905
	memset(&mm->rss_stat, 0, sizeof(mm->rss_stat));
L
Linus Torvalds 已提交
906
	spin_lock_init(&mm->page_table_lock);
907
	mm_init_cpumask(mm);
A
Alexey Dobriyan 已提交
908
	mm_init_aio(mm);
909
	mm_init_owner(mm, p);
910
	RCU_INIT_POINTER(mm->exe_file, NULL);
911
	mmu_notifier_mm_init(mm);
912
	hmm_mm_init(mm);
913
	init_tlb_flush_pending(mm);
914 915 916
#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
	mm->pmd_huge_pte = NULL;
#endif
917
	mm_init_uprobes_state(mm);
L
Linus Torvalds 已提交
918

919 920 921 922 923
	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;
L
Linus Torvalds 已提交
924
		mm->def_flags = 0;
925 926
	}

927 928 929 930 931
	if (mm_alloc_pgd(mm))
		goto fail_nopgd;

	if (init_new_context(p, mm))
		goto fail_nocontext;
932

933
	mm->user_ns = get_user_ns(user_ns);
934 935 936 937 938
	return mm;

fail_nocontext:
	mm_free_pgd(mm);
fail_nopgd:
L
Linus Torvalds 已提交
939 940 941 942 943 944 945
	free_mm(mm);
	return NULL;
}

/*
 * Allocate and initialize an mm_struct.
 */
946
struct mm_struct *mm_alloc(void)
L
Linus Torvalds 已提交
947
{
948
	struct mm_struct *mm;
L
Linus Torvalds 已提交
949 950

	mm = allocate_mm();
951 952 953 954
	if (!mm)
		return NULL;

	memset(mm, 0, sizeof(*mm));
955
	return mm_init(mm, current, current_user_ns());
L
Linus Torvalds 已提交
956 957
}

958 959 960 961 962 963 964 965 966
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);
967
	mm_put_huge_zero_page(mm);
968 969 970 971 972 973 974 975 976 977 978
	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);
	mmdrop(mm);
}

L
Linus Torvalds 已提交
979 980 981 982 983
/*
 * Decrement the use count and release all resources for an mm.
 */
void mmput(struct mm_struct *mm)
{
A
Andrew Morton 已提交
984 985
	might_sleep();

986 987 988 989 990
	if (atomic_dec_and_test(&mm->mm_users))
		__mmput(mm);
}
EXPORT_SYMBOL_GPL(mmput);

991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008
#ifdef CONFIG_MMU
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)
{
	if (atomic_dec_and_test(&mm->mm_users)) {
		INIT_WORK(&mm->async_put_work, mmput_async_fn);
		schedule_work(&mm->async_put_work);
	}
}
#endif

1009 1010 1011 1012 1013
/**
 * 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).
 *
1014 1015 1016 1017 1018
 * 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.
1019
 */
1020 1021
void set_mm_exe_file(struct mm_struct *mm, struct file *new_exe_file)
{
1022 1023 1024 1025 1026 1027 1028 1029
	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);
1030

1031 1032
	if (new_exe_file)
		get_file(new_exe_file);
1033 1034 1035
	rcu_assign_pointer(mm->exe_file, new_exe_file);
	if (old_exe_file)
		fput(old_exe_file);
1036 1037
}

1038 1039 1040 1041 1042 1043
/**
 * 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().
 */
1044 1045 1046 1047
struct file *get_mm_exe_file(struct mm_struct *mm)
{
	struct file *exe_file;

1048 1049 1050 1051 1052
	rcu_read_lock();
	exe_file = rcu_dereference(mm->exe_file);
	if (exe_file && !get_file_rcu(exe_file))
		exe_file = NULL;
	rcu_read_unlock();
1053 1054
	return exe_file;
}
1055
EXPORT_SYMBOL(get_mm_exe_file);
1056

M
Mateusz Guzik 已提交
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
/**
 * 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);
1079

L
Linus Torvalds 已提交
1080 1081 1082
/**
 * get_task_mm - acquire a reference to the task's mm
 *
1083
 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
L
Linus Torvalds 已提交
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
 * 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) {
1096
		if (task->flags & PF_KTHREAD)
L
Linus Torvalds 已提交
1097 1098
			mm = NULL;
		else
V
Vegard Nossum 已提交
1099
			mmget(mm);
L
Linus Torvalds 已提交
1100 1101 1102 1103 1104 1105
	}
	task_unlock(task);
	return mm;
}
EXPORT_SYMBOL_GPL(get_task_mm);

1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
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;
}

1126
static void complete_vfork_done(struct task_struct *tsk)
1127
{
O
Oleg Nesterov 已提交
1128
	struct completion *vfork;
1129

O
Oleg Nesterov 已提交
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
	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;
1156 1157
}

L
Linus Torvalds 已提交
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172
/* 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)
{
1173 1174
	/* Get rid of any futexes when releasing the mm */
#ifdef CONFIG_FUTEX
1175
	if (unlikely(tsk->robust_list)) {
1176
		exit_robust_list(tsk);
1177 1178
		tsk->robust_list = NULL;
	}
1179
#ifdef CONFIG_COMPAT
1180
	if (unlikely(tsk->compat_robust_list)) {
1181
		compat_exit_robust_list(tsk);
1182 1183
		tsk->compat_robust_list = NULL;
	}
1184
#endif
1185 1186
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
1187 1188
#endif

1189 1190
	uprobe_free_utask(tsk);

L
Linus Torvalds 已提交
1191 1192 1193
	/* Get rid of any cached register state */
	deactivate_mm(tsk, mm);

1194
	/*
1195 1196 1197
	 * Signal userspace if we're not exiting with a core dump
	 * because we want to leave the value intact for debugging
	 * purposes.
1198
	 */
1199
	if (tsk->clear_child_tid) {
1200
		if (!(tsk->signal->flags & SIGNAL_GROUP_COREDUMP) &&
1201 1202 1203 1204 1205 1206 1207 1208 1209
		    atomic_read(&mm->mm_users) > 1) {
			/*
			 * We don't check the error code - if userspace has
			 * not set up a proper pointer then tough luck.
			 */
			put_user(0, tsk->clear_child_tid);
			sys_futex(tsk->clear_child_tid, FUTEX_WAKE,
					1, NULL, NULL, 0);
		}
L
Linus Torvalds 已提交
1210 1211
		tsk->clear_child_tid = NULL;
	}
1212 1213 1214 1215 1216 1217 1218

	/*
	 * 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);
L
Linus Torvalds 已提交
1219 1220
}

1221 1222 1223 1224
/*
 * Allocate a new mm structure and copy contents from the
 * mm structure of the passed in task structure.
 */
1225
static struct mm_struct *dup_mm(struct task_struct *tsk)
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235
{
	struct mm_struct *mm, *oldmm = current->mm;
	int err;

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

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

1236
	if (!mm_init(mm, tsk, mm->user_ns))
1237 1238 1239 1240 1241 1242 1243 1244 1245
		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;

1246 1247 1248
	if (mm->binfmt && !try_module_get(mm->binfmt->module))
		goto free_pt;

1249 1250 1251
	return mm;

free_pt:
1252 1253
	/* don't put binfmt in mmput, we haven't got module yet */
	mm->binfmt = NULL;
1254 1255 1256 1257 1258 1259
	mmput(mm);

fail_nomem:
	return NULL;
}

1260
static int copy_mm(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
1261
{
1262
	struct mm_struct *mm, *oldmm;
L
Linus Torvalds 已提交
1263 1264 1265 1266
	int retval;

	tsk->min_flt = tsk->maj_flt = 0;
	tsk->nvcsw = tsk->nivcsw = 0;
1267 1268 1269
#ifdef CONFIG_DETECT_HUNG_TASK
	tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw;
#endif
L
Linus Torvalds 已提交
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282

	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;

D
Davidlohr Bueso 已提交
1283 1284 1285
	/* initialize the new vmacache entries */
	vmacache_flush(tsk);

L
Linus Torvalds 已提交
1286
	if (clone_flags & CLONE_VM) {
V
Vegard Nossum 已提交
1287
		mmget(oldmm);
L
Linus Torvalds 已提交
1288 1289 1290 1291 1292
		mm = oldmm;
		goto good_mm;
	}

	retval = -ENOMEM;
1293
	mm = dup_mm(tsk);
L
Linus Torvalds 已提交
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305
	if (!mm)
		goto fail_nomem;

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

fail_nomem:
	return retval;
}

A
Alexey Dobriyan 已提交
1306
static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
1307
{
A
Al Viro 已提交
1308
	struct fs_struct *fs = current->fs;
L
Linus Torvalds 已提交
1309
	if (clone_flags & CLONE_FS) {
A
Al Viro 已提交
1310
		/* tsk->fs is already what we want */
N
Nick Piggin 已提交
1311
		spin_lock(&fs->lock);
A
Al Viro 已提交
1312
		if (fs->in_exec) {
N
Nick Piggin 已提交
1313
			spin_unlock(&fs->lock);
A
Al Viro 已提交
1314 1315 1316
			return -EAGAIN;
		}
		fs->users++;
N
Nick Piggin 已提交
1317
		spin_unlock(&fs->lock);
L
Linus Torvalds 已提交
1318 1319
		return 0;
	}
A
Al Viro 已提交
1320
	tsk->fs = copy_fs_struct(fs);
L
Linus Torvalds 已提交
1321 1322 1323 1324 1325
	if (!tsk->fs)
		return -ENOMEM;
	return 0;
}

1326
static int copy_files(unsigned long clone_flags, struct task_struct *tsk)
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
{
	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;
}

1353
static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
1354 1355 1356
{
#ifdef CONFIG_BLOCK
	struct io_context *ioc = current->io_context;
1357
	struct io_context *new_ioc;
1358 1359 1360

	if (!ioc)
		return 0;
1361 1362 1363 1364
	/*
	 * Share io context with parent, if CLONE_IO is set
	 */
	if (clone_flags & CLONE_IO) {
T
Tejun Heo 已提交
1365 1366
		ioc_task_link(ioc);
		tsk->io_context = ioc;
1367
	} else if (ioprio_valid(ioc->ioprio)) {
1368 1369
		new_ioc = get_task_io_context(tsk, GFP_KERNEL, NUMA_NO_NODE);
		if (unlikely(!new_ioc))
1370 1371
			return -ENOMEM;

1372
		new_ioc->ioprio = ioc->ioprio;
1373
		put_io_context(new_ioc);
1374 1375 1376 1377 1378
	}
#endif
	return 0;
}

A
Alexey Dobriyan 已提交
1379
static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
1380 1381 1382
{
	struct sighand_struct *sig;

Z
Zhaolei 已提交
1383
	if (clone_flags & CLONE_SIGHAND) {
L
Linus Torvalds 已提交
1384 1385 1386 1387
		atomic_inc(&current->sighand->count);
		return 0;
	}
	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
I
Ingo Molnar 已提交
1388
	rcu_assign_pointer(tsk->sighand, sig);
L
Linus Torvalds 已提交
1389 1390
	if (!sig)
		return -ENOMEM;
1391

L
Linus Torvalds 已提交
1392 1393 1394 1395 1396
	atomic_set(&sig->count, 1);
	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
	return 0;
}

1397
void __cleanup_sighand(struct sighand_struct *sighand)
1398
{
1399 1400
	if (atomic_dec_and_test(&sighand->count)) {
		signalfd_cleanup(sighand);
1401
		/*
1402
		 * sighand_cachep is SLAB_TYPESAFE_BY_RCU so we can free it
1403 1404
		 * without an RCU grace period, see __lock_task_sighand().
		 */
1405
		kmem_cache_free(sighand_cachep, sighand);
1406
	}
1407 1408
}

1409
#ifdef CONFIG_POSIX_TIMERS
1410 1411 1412 1413 1414
/*
 * Initialize POSIX timer handling for a thread group.
 */
static void posix_cpu_timers_init_group(struct signal_struct *sig)
{
1415 1416
	unsigned long cpu_limit;

1417
	cpu_limit = READ_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur);
1418
	if (cpu_limit != RLIM_INFINITY) {
1419
		sig->cputime_expires.prof_exp = cpu_limit * NSEC_PER_SEC;
1420
		sig->cputimer.running = true;
1421 1422
	}

1423 1424 1425 1426 1427
	/* The timer lists. */
	INIT_LIST_HEAD(&sig->cpu_timers[0]);
	INIT_LIST_HEAD(&sig->cpu_timers[1]);
	INIT_LIST_HEAD(&sig->cpu_timers[2]);
}
1428 1429 1430
#else
static inline void posix_cpu_timers_init_group(struct signal_struct *sig) { }
#endif
1431

A
Alexey Dobriyan 已提交
1432
static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
L
Linus Torvalds 已提交
1433 1434 1435
{
	struct signal_struct *sig;

1436
	if (clone_flags & CLONE_THREAD)
1437 1438
		return 0;

1439
	sig = kmem_cache_zalloc(signal_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
1440 1441 1442 1443
	tsk->signal = sig;
	if (!sig)
		return -ENOMEM;

1444
	sig->nr_threads = 1;
L
Linus Torvalds 已提交
1445
	atomic_set(&sig->live, 1);
1446
	atomic_set(&sig->sigcnt, 1);
1447 1448 1449 1450 1451

	/* list_add(thread_node, thread_head) without INIT_LIST_HEAD() */
	sig->thread_head = (struct list_head)LIST_HEAD_INIT(tsk->thread_node);
	tsk->thread_node = (struct list_head)LIST_HEAD_INIT(sig->thread_head);

L
Linus Torvalds 已提交
1452
	init_waitqueue_head(&sig->wait_chldexit);
1453
	sig->curr_target = tsk;
L
Linus Torvalds 已提交
1454
	init_sigpending(&sig->shared_pending);
1455
	seqlock_init(&sig->stats_lock);
1456
	prev_cputime_init(&sig->prev_cputime);
L
Linus Torvalds 已提交
1457

1458
#ifdef CONFIG_POSIX_TIMERS
1459
	INIT_LIST_HEAD(&sig->posix_timers);
1460
	hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
L
Linus Torvalds 已提交
1461
	sig->real_timer.function = it_real_fn;
1462
#endif
L
Linus Torvalds 已提交
1463 1464 1465 1466 1467

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

1468 1469
	posix_cpu_timers_init_group(sig);

M
Miloslav Trmac 已提交
1470
	tty_audit_fork(sig);
1471
	sched_autogroup_fork(sig);
M
Miloslav Trmac 已提交
1472

D
David Rientjes 已提交
1473
	sig->oom_score_adj = current->signal->oom_score_adj;
1474
	sig->oom_score_adj_min = current->signal->oom_score_adj_min;
1475

1476 1477
	mutex_init(&sig->cred_guard_mutex);

L
Linus Torvalds 已提交
1478 1479 1480
	return 0;
}

K
Kees Cook 已提交
1481 1482 1483 1484 1485 1486 1487 1488 1489
static void copy_seccomp(struct task_struct *p)
{
#ifdef CONFIG_SECCOMP
	/*
	 * Must be called with sighand->lock held, which is common to
	 * all threads in the group. Holding cred_guard_mutex is not
	 * needed because this new task is not yet running and cannot
	 * be racing exec.
	 */
1490
	assert_spin_locked(&current->sighand->siglock);
K
Kees Cook 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513

	/* Ref-count the new filter user, and assign it. */
	get_seccomp_filter(current);
	p->seccomp = current->seccomp;

	/*
	 * Explicitly enable no_new_privs here in case it got set
	 * between the task_struct being duplicated and holding the
	 * sighand lock. The seccomp state and nnp must be in sync.
	 */
	if (task_no_new_privs(current))
		task_set_no_new_privs(p);

	/*
	 * If the parent gained a seccomp mode after copying thread
	 * flags and between before we held the sighand lock, we have
	 * to manually enable the seccomp thread flag here.
	 */
	if (p->seccomp.mode != SECCOMP_MODE_DISABLED)
		set_tsk_thread_flag(p, TIF_SECCOMP);
#endif
}

1514
SYSCALL_DEFINE1(set_tid_address, int __user *, tidptr)
L
Linus Torvalds 已提交
1515 1516 1517
{
	current->clear_child_tid = tidptr;

1518
	return task_pid_vnr(current);
L
Linus Torvalds 已提交
1519 1520
}

A
Alexey Dobriyan 已提交
1521
static void rt_mutex_init_task(struct task_struct *p)
I
Ingo Molnar 已提交
1522
{
1523
	raw_spin_lock_init(&p->pi_lock);
1524
#ifdef CONFIG_RT_MUTEXES
1525
	p->pi_waiters = RB_ROOT_CACHED;
1526
	p->pi_top_task = NULL;
I
Ingo Molnar 已提交
1527 1528 1529 1530
	p->pi_blocked_on = NULL;
#endif
}

1531
#ifdef CONFIG_POSIX_TIMERS
1532 1533 1534 1535 1536
/*
 * Initialize POSIX timer handling for a single task.
 */
static void posix_cpu_timers_init(struct task_struct *tsk)
{
1537 1538
	tsk->cputime_expires.prof_exp = 0;
	tsk->cputime_expires.virt_exp = 0;
1539 1540 1541 1542 1543
	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]);
}
1544 1545 1546
#else
static inline void posix_cpu_timers_init(struct task_struct *tsk) { }
#endif
1547

1548 1549 1550 1551 1552 1553
static inline void
init_task_pid(struct task_struct *task, enum pid_type type, struct pid *pid)
{
	 task->pids[type].pid = pid;
}

1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
static inline void rcu_copy_process(struct task_struct *p)
{
#ifdef CONFIG_PREEMPT_RCU
	p->rcu_read_lock_nesting = 0;
	p->rcu_read_unlock_special.s = 0;
	p->rcu_blocked_node = NULL;
	INIT_LIST_HEAD(&p->rcu_node_entry);
#endif /* #ifdef CONFIG_PREEMPT_RCU */
#ifdef CONFIG_TASKS_RCU
	p->rcu_tasks_holdout = false;
	INIT_LIST_HEAD(&p->rcu_tasks_holdout_list);
	p->rcu_tasks_idle_cpu = -1;
#endif /* #ifdef CONFIG_TASKS_RCU */
}

L
Linus Torvalds 已提交
1569 1570 1571 1572 1573 1574 1575 1576
/*
 * 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.
 */
1577 1578
static __latent_entropy struct task_struct *copy_process(
					unsigned long clone_flags,
1579 1580 1581
					unsigned long stack_start,
					unsigned long stack_size,
					int __user *child_tidptr,
R
Roland McGrath 已提交
1582
					struct pid *pid,
1583
					int trace,
1584 1585
					unsigned long tls,
					int node)
L
Linus Torvalds 已提交
1586 1587
{
	int retval;
1588
	struct task_struct *p;
L
Linus Torvalds 已提交
1589 1590 1591 1592

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

1593 1594 1595
	if ((clone_flags & (CLONE_NEWUSER|CLONE_FS)) == (CLONE_NEWUSER|CLONE_FS))
		return ERR_PTR(-EINVAL);

L
Linus Torvalds 已提交
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610
	/*
	 * 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);

1611 1612 1613 1614 1615 1616 1617 1618 1619 1620
	/*
	 * Siblings of global init remain as zombies on exit since they are
	 * not reaped by their parent (swapper). To solve this and to avoid
	 * multi-rooted process trees, prevent global and container-inits
	 * from creating siblings.
	 */
	if ((clone_flags & CLONE_PARENT) &&
				current->signal->flags & SIGNAL_UNKILLABLE)
		return ERR_PTR(-EINVAL);

1621
	/*
1622
	 * If the new process will be in a different pid or user namespace
1623
	 * do not allow it to share a thread group with the forking task.
1624
	 */
1625
	if (clone_flags & CLONE_THREAD) {
1626 1627 1628 1629 1630
		if ((clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) ||
		    (task_active_pid_ns(current) !=
				current->nsproxy->pid_ns_for_children))
			return ERR_PTR(-EINVAL);
	}
1631

L
Linus Torvalds 已提交
1632
	retval = -ENOMEM;
1633
	p = dup_task_struct(current, node);
L
Linus Torvalds 已提交
1634 1635 1636
	if (!p)
		goto fork_out;

1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
	/*
	 * This _must_ happen before we call free_task(), i.e. before we jump
	 * to any of the bad_fork_* labels. This is to avoid freeing
	 * p->set_child_tid which is (ab)used as a kthread's data pointer for
	 * kernel threads (PF_KTHREAD).
	 */
	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;

1649 1650
	ftrace_graph_init_task(p);

1651 1652
	rt_mutex_init_task(p);

I
Ingo Molnar 已提交
1653
#ifdef CONFIG_PROVE_LOCKING
1654 1655 1656
	DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
#endif
L
Linus Torvalds 已提交
1657
	retval = -EAGAIN;
1658
	if (atomic_read(&p->real_cred->user->processes) >=
1659
			task_rlimit(p, RLIMIT_NPROC)) {
1660 1661
		if (p->real_cred->user != INIT_USER &&
		    !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN))
L
Linus Torvalds 已提交
1662 1663
			goto bad_fork_free;
	}
1664
	current->flags &= ~PF_NPROC_EXCEEDED;
L
Linus Torvalds 已提交
1665

1666 1667 1668
	retval = copy_creds(p, clone_flags);
	if (retval < 0)
		goto bad_fork_free;
L
Linus Torvalds 已提交
1669 1670 1671 1672 1673 1674

	/*
	 * 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.
	 */
1675
	retval = -EAGAIN;
L
Linus Torvalds 已提交
1676 1677 1678
	if (nr_threads >= max_threads)
		goto bad_fork_cleanup_count;

1679
	delayacct_tsk_init(p);	/* Must remain after dup_task_struct() */
1680
	p->flags &= ~(PF_SUPERPRIV | PF_WQ_WORKER | PF_IDLE);
1681
	p->flags |= PF_FORKNOEXEC;
L
Linus Torvalds 已提交
1682 1683
	INIT_LIST_HEAD(&p->children);
	INIT_LIST_HEAD(&p->sibling);
1684
	rcu_copy_process(p);
L
Linus Torvalds 已提交
1685 1686 1687 1688 1689
	p->vfork_done = NULL;
	spin_lock_init(&p->alloc_lock);

	init_sigpending(&p->pending);

1690
	p->utime = p->stime = p->gtime = 0;
1691
#ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME
1692
	p->utimescaled = p->stimescaled = 0;
1693
#endif
1694 1695
	prev_cputime_init(&p->prev_cputime);

1696
#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
1697 1698 1699
	seqcount_init(&p->vtime.seqcount);
	p->vtime.starttime = 0;
	p->vtime.state = VTIME_INACTIVE;
1700 1701
#endif

1702 1703 1704
#if defined(SPLIT_RSS_COUNTING)
	memset(&p->rss_stat, 0, sizeof(p->rss_stat));
#endif
1705

1706 1707
	p->default_timer_slack_ns = current->timer_slack_ns;

1708
	task_io_accounting_init(&p->ioac);
L
Linus Torvalds 已提交
1709 1710
	acct_clear_integrals(p);

1711
	posix_cpu_timers_init(p);
L
Linus Torvalds 已提交
1712

1713
	p->start_time = ktime_get_ns();
1714
	p->real_start_time = ktime_get_boot_ns();
L
Linus Torvalds 已提交
1715 1716
	p->io_context = NULL;
	p->audit_context = NULL;
1717
	cgroup_fork(p);
L
Linus Torvalds 已提交
1718
#ifdef CONFIG_NUMA
1719
	p->mempolicy = mpol_dup(p->mempolicy);
1720 1721 1722
	if (IS_ERR(p->mempolicy)) {
		retval = PTR_ERR(p->mempolicy);
		p->mempolicy = NULL;
1723
		goto bad_fork_cleanup_threadgroup_lock;
1724
	}
L
Linus Torvalds 已提交
1725
#endif
1726 1727 1728
#ifdef CONFIG_CPUSETS
	p->cpuset_mem_spread_rotor = NUMA_NO_NODE;
	p->cpuset_slab_spread_rotor = NUMA_NO_NODE;
1729
	seqcount_init(&p->mems_allowed_seq);
1730
#endif
1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745
#ifdef CONFIG_TRACE_IRQFLAGS
	p->irq_events = 0;
	p->hardirqs_enabled = 0;
	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
1746 1747 1748

	p->pagefault_disabled = 0;

I
Ingo Molnar 已提交
1749 1750 1751 1752
#ifdef CONFIG_LOCKDEP
	p->lockdep_depth = 0; /* no locks held yet */
	p->curr_chain_key = 0;
	p->lockdep_recursion = 0;
1753
	lockdep_init_task(p);
I
Ingo Molnar 已提交
1754
#endif
L
Linus Torvalds 已提交
1755

1756 1757 1758
#ifdef CONFIG_DEBUG_MUTEXES
	p->blocked_on = NULL; /* not blocked yet */
#endif
K
Kent Overstreet 已提交
1759 1760 1761 1762
#ifdef CONFIG_BCACHE
	p->sequential_io	= 0;
	p->sequential_io_avg	= 0;
#endif
1763

1764
	/* Perform scheduler related setup. Assign this task to a CPU. */
1765 1766 1767
	retval = sched_fork(clone_flags, p);
	if (retval)
		goto bad_fork_cleanup_policy;
1768

1769
	retval = perf_event_init_task(p);
1770 1771
	if (retval)
		goto bad_fork_cleanup_policy;
1772 1773
	retval = audit_alloc(p);
	if (retval)
P
Peter Zijlstra 已提交
1774
		goto bad_fork_cleanup_perf;
L
Linus Torvalds 已提交
1775
	/* copy all the process information */
1776
	shm_init_task(p);
1777
	retval = security_task_alloc(p, clone_flags);
1778
	if (retval)
L
Linus Torvalds 已提交
1779
		goto bad_fork_cleanup_audit;
1780 1781 1782
	retval = copy_semundo(clone_flags, p);
	if (retval)
		goto bad_fork_cleanup_security;
1783 1784
	retval = copy_files(clone_flags, p);
	if (retval)
L
Linus Torvalds 已提交
1785
		goto bad_fork_cleanup_semundo;
1786 1787
	retval = copy_fs(clone_flags, p);
	if (retval)
L
Linus Torvalds 已提交
1788
		goto bad_fork_cleanup_files;
1789 1790
	retval = copy_sighand(clone_flags, p);
	if (retval)
L
Linus Torvalds 已提交
1791
		goto bad_fork_cleanup_fs;
1792 1793
	retval = copy_signal(clone_flags, p);
	if (retval)
L
Linus Torvalds 已提交
1794
		goto bad_fork_cleanup_sighand;
1795 1796
	retval = copy_mm(clone_flags, p);
	if (retval)
L
Linus Torvalds 已提交
1797
		goto bad_fork_cleanup_signal;
1798 1799
	retval = copy_namespaces(clone_flags, p);
	if (retval)
D
David Howells 已提交
1800
		goto bad_fork_cleanup_mm;
1801 1802
	retval = copy_io(clone_flags, p);
	if (retval)
1803
		goto bad_fork_cleanup_namespaces;
1804
	retval = copy_thread_tls(clone_flags, stack_start, stack_size, p, tls);
L
Linus Torvalds 已提交
1805
	if (retval)
1806
		goto bad_fork_cleanup_io;
L
Linus Torvalds 已提交
1807

1808
	if (pid != &init_struct_pid) {
1809
		pid = alloc_pid(p->nsproxy->pid_ns_for_children);
1810 1811
		if (IS_ERR(pid)) {
			retval = PTR_ERR(pid);
1812
			goto bad_fork_cleanup_thread;
1813
		}
1814 1815
	}

1816 1817 1818
#ifdef CONFIG_BLOCK
	p->plug = NULL;
#endif
1819
#ifdef CONFIG_FUTEX
1820 1821 1822 1823
	p->robust_list = NULL;
#ifdef CONFIG_COMPAT
	p->compat_robust_list = NULL;
#endif
1824 1825
	INIT_LIST_HEAD(&p->pi_state_list);
	p->pi_state_cache = NULL;
1826
#endif
1827 1828 1829 1830
	/*
	 * sigaltstack should be cleared when sharing the same VM
	 */
	if ((clone_flags & (CLONE_VM|CLONE_VFORK)) == CLONE_VM)
1831
		sas_ss_reset(p);
1832

L
Linus Torvalds 已提交
1833
	/*
1834 1835
	 * Syscall tracing and stepping should be turned off in the
	 * child regardless of CLONE_PTRACE.
L
Linus Torvalds 已提交
1836
	 */
1837
	user_disable_single_step(p);
L
Linus Torvalds 已提交
1838
	clear_tsk_thread_flag(p, TIF_SYSCALL_TRACE);
1839 1840 1841
#ifdef TIF_SYSCALL_EMU
	clear_tsk_thread_flag(p, TIF_SYSCALL_EMU);
#endif
A
Arjan van de Ven 已提交
1842
	clear_all_latency_tracing(p);
L
Linus Torvalds 已提交
1843 1844

	/* ok, now we should be set up.. */
1845 1846
	p->pid = pid_nr(pid);
	if (clone_flags & CLONE_THREAD) {
1847
		p->exit_signal = -1;
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857
		p->group_leader = current->group_leader;
		p->tgid = current->tgid;
	} else {
		if (clone_flags & CLONE_PARENT)
			p->exit_signal = current->group_leader->exit_signal;
		else
			p->exit_signal = (clone_flags & CSIGNAL);
		p->group_leader = p;
		p->tgid = p->pid;
	}
1858

1859 1860
	p->nr_dirtied = 0;
	p->nr_dirtied_pause = 128 >> (PAGE_SHIFT - 10);
1861
	p->dirty_paused_when = 0;
1862

1863
	p->pdeath_signal = 0;
O
Oleg Nesterov 已提交
1864
	INIT_LIST_HEAD(&p->thread_group);
A
Al Viro 已提交
1865
	p->task_works = NULL;
L
Linus Torvalds 已提交
1866

1867
	cgroup_threadgroup_change_begin(current);
1868 1869 1870 1871 1872 1873
	/*
	 * Ensure that the cgroup subsystem policies allow the new process to be
	 * forked. It should be noted the the new process's css_set can be changed
	 * between here and cgroup_post_fork() if an organisation operation is in
	 * progress.
	 */
1874
	retval = cgroup_can_fork(p);
1875 1876 1877
	if (retval)
		goto bad_fork_free_pid;

1878 1879 1880 1881
	/*
	 * Make it visible to the rest of the system, but dont wake it up yet.
	 * Need tasklist lock for parent etc handling!
	 */
L
Linus Torvalds 已提交
1882 1883 1884
	write_lock_irq(&tasklist_lock);

	/* CLONE_PARENT re-uses the old parent */
1885
	if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) {
L
Linus Torvalds 已提交
1886
		p->real_parent = current->real_parent;
1887 1888
		p->parent_exec_id = current->parent_exec_id;
	} else {
L
Linus Torvalds 已提交
1889
		p->real_parent = current;
1890 1891
		p->parent_exec_id = current->self_exec_id;
	}
L
Linus Torvalds 已提交
1892

1893 1894
	klp_copy_process(p);

1895
	spin_lock(&current->sighand->siglock);
1896

K
Kees Cook 已提交
1897 1898 1899 1900 1901 1902
	/*
	 * Copy seccomp details explicitly here, in case they were changed
	 * before holding sighand lock.
	 */
	copy_seccomp(p);

1903 1904 1905 1906 1907 1908 1909
	/*
	 * 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).
1910
	*/
D
Daniel Walker 已提交
1911
	recalc_sigpending();
1912 1913
	if (signal_pending(current)) {
		retval = -ERESTARTNOINTR;
1914
		goto bad_fork_cancel_cgroup;
1915
	}
G
Gargi Sharma 已提交
1916
	if (unlikely(!(ns_of_pid(pid)->pid_allocated & PIDNS_ADDING))) {
1917 1918 1919
		retval = -ENOMEM;
		goto bad_fork_cancel_cgroup;
	}
1920

1921
	if (likely(p->pid)) {
T
Tejun Heo 已提交
1922
		ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace);
1923

1924
		init_task_pid(p, PIDTYPE_PID, pid);
1925
		if (thread_group_leader(p)) {
1926 1927 1928
			init_task_pid(p, PIDTYPE_PGID, task_pgrp(current));
			init_task_pid(p, PIDTYPE_SID, task_session(current));

1929
			if (is_child_reaper(pid)) {
1930
				ns_of_pid(pid)->child_reaper = p;
1931 1932
				p->signal->flags |= SIGNAL_UNKILLABLE;
			}
1933

1934
			p->signal->leader_pid = pid;
A
Alan Cox 已提交
1935
			p->signal->tty = tty_kref_get(current->signal->tty);
1936 1937 1938 1939 1940 1941 1942
			/*
			 * Inherit has_child_subreaper flag under the same
			 * tasklist_lock with adding child to the process tree
			 * for propagate_has_child_subreaper optimization.
			 */
			p->signal->has_child_subreaper = p->real_parent->signal->has_child_subreaper ||
							 p->real_parent->signal->is_child_subreaper;
1943
			list_add_tail(&p->sibling, &p->real_parent->children);
1944
			list_add_tail_rcu(&p->tasks, &init_task.tasks);
1945 1946
			attach_pid(p, PIDTYPE_PGID);
			attach_pid(p, PIDTYPE_SID);
1947
			__this_cpu_inc(process_counts);
1948 1949 1950 1951 1952 1953
		} else {
			current->signal->nr_threads++;
			atomic_inc(&current->signal->live);
			atomic_inc(&current->signal->sigcnt);
			list_add_tail_rcu(&p->thread_group,
					  &p->group_leader->thread_group);
1954 1955
			list_add_tail_rcu(&p->thread_node,
					  &p->signal->thread_head);
1956
		}
1957
		attach_pid(p, PIDTYPE_PID);
1958
		nr_threads++;
L
Linus Torvalds 已提交
1959 1960 1961
	}

	total_forks++;
1962
	spin_unlock(&current->sighand->siglock);
1963
	syscall_tracepoint_update(p);
L
Linus Torvalds 已提交
1964
	write_unlock_irq(&tasklist_lock);
1965

1966
	proc_fork_connector(p);
1967
	cgroup_post_fork(p);
1968
	cgroup_threadgroup_change_end(current);
1969
	perf_event_fork(p);
1970 1971

	trace_task_newtask(p, clone_flags);
1972
	uprobe_copy_process(p, clone_flags);
1973

L
Linus Torvalds 已提交
1974 1975
	return p;

1976
bad_fork_cancel_cgroup:
1977 1978
	spin_unlock(&current->sighand->siglock);
	write_unlock_irq(&tasklist_lock);
1979
	cgroup_cancel_fork(p);
1980
bad_fork_free_pid:
1981
	cgroup_threadgroup_change_end(current);
1982 1983
	if (pid != &init_struct_pid)
		free_pid(pid);
1984 1985
bad_fork_cleanup_thread:
	exit_thread(p);
1986
bad_fork_cleanup_io:
1987 1988
	if (p->io_context)
		exit_io_context(p);
S
Serge E. Hallyn 已提交
1989
bad_fork_cleanup_namespaces:
1990
	exit_task_namespaces(p);
L
Linus Torvalds 已提交
1991
bad_fork_cleanup_mm:
D
David Rientjes 已提交
1992
	if (p->mm)
L
Linus Torvalds 已提交
1993 1994
		mmput(p->mm);
bad_fork_cleanup_signal:
1995
	if (!(clone_flags & CLONE_THREAD))
1996
		free_signal_struct(p->signal);
L
Linus Torvalds 已提交
1997
bad_fork_cleanup_sighand:
1998
	__cleanup_sighand(p->sighand);
L
Linus Torvalds 已提交
1999 2000 2001 2002 2003 2004
bad_fork_cleanup_fs:
	exit_fs(p); /* blocking */
bad_fork_cleanup_files:
	exit_files(p); /* blocking */
bad_fork_cleanup_semundo:
	exit_sem(p);
2005 2006
bad_fork_cleanup_security:
	security_task_free(p);
L
Linus Torvalds 已提交
2007 2008
bad_fork_cleanup_audit:
	audit_free(p);
P
Peter Zijlstra 已提交
2009
bad_fork_cleanup_perf:
2010
	perf_event_free_task(p);
P
Peter Zijlstra 已提交
2011
bad_fork_cleanup_policy:
2012
	lockdep_free_task(p);
L
Linus Torvalds 已提交
2013
#ifdef CONFIG_NUMA
2014
	mpol_put(p->mempolicy);
2015
bad_fork_cleanup_threadgroup_lock:
L
Linus Torvalds 已提交
2016
#endif
2017
	delayacct_tsk_free(p);
L
Linus Torvalds 已提交
2018
bad_fork_cleanup_count:
D
David Howells 已提交
2019
	atomic_dec(&p->cred->user->processes);
2020
	exit_creds(p);
L
Linus Torvalds 已提交
2021
bad_fork_free:
2022
	p->state = TASK_DEAD;
2023
	put_task_stack(p);
L
Linus Torvalds 已提交
2024
	free_task(p);
2025 2026
fork_out:
	return ERR_PTR(retval);
L
Linus Torvalds 已提交
2027 2028
}

2029 2030 2031 2032 2033 2034 2035 2036 2037 2038
static inline void init_idle_pids(struct pid_link *links)
{
	enum pid_type type;

	for (type = PIDTYPE_PID; type < PIDTYPE_MAX; ++type) {
		INIT_HLIST_NODE(&links[type].node); /* not really needed */
		links[type].pid = &init_struct_pid;
	}
}

2039
struct task_struct *fork_idle(int cpu)
L
Linus Torvalds 已提交
2040
{
2041
	struct task_struct *task;
2042 2043
	task = copy_process(CLONE_VM, 0, 0, NULL, &init_struct_pid, 0, 0,
			    cpu_to_node(cpu));
2044 2045
	if (!IS_ERR(task)) {
		init_idle_pids(task->pids);
2046
		init_idle(task, cpu);
2047
	}
2048

L
Linus Torvalds 已提交
2049 2050 2051 2052 2053 2054 2055 2056 2057
	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.
 */
2058
long _do_fork(unsigned long clone_flags,
L
Linus Torvalds 已提交
2059 2060 2061
	      unsigned long stack_start,
	      unsigned long stack_size,
	      int __user *parent_tidptr,
2062 2063
	      int __user *child_tidptr,
	      unsigned long tls)
L
Linus Torvalds 已提交
2064 2065 2066
{
	struct task_struct *p;
	int trace = 0;
2067
	long nr;
L
Linus Torvalds 已提交
2068

R
Roland McGrath 已提交
2069
	/*
T
Tejun Heo 已提交
2070 2071 2072 2073
	 * Determine whether and which event to report to ptracer.  When
	 * called from kernel_thread or CLONE_UNTRACED is explicitly
	 * requested, no event is reported; otherwise, report if the event
	 * for the type of forking is enabled.
R
Roland McGrath 已提交
2074
	 */
2075
	if (!(clone_flags & CLONE_UNTRACED)) {
T
Tejun Heo 已提交
2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
		if (clone_flags & CLONE_VFORK)
			trace = PTRACE_EVENT_VFORK;
		else if ((clone_flags & CSIGNAL) != SIGCHLD)
			trace = PTRACE_EVENT_CLONE;
		else
			trace = PTRACE_EVENT_FORK;

		if (likely(!ptrace_event_enabled(current, trace)))
			trace = 0;
	}
L
Linus Torvalds 已提交
2086

A
Al Viro 已提交
2087
	p = copy_process(clone_flags, stack_start, stack_size,
2088
			 child_tidptr, NULL, trace, tls, NUMA_NO_NODE);
2089
	add_latent_entropy();
L
Linus Torvalds 已提交
2090 2091 2092 2093 2094 2095
	/*
	 * 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;
2096
		struct pid *pid;
L
Linus Torvalds 已提交
2097

2098 2099
		trace_sched_process_fork(current, p);

2100 2101
		pid = get_task_pid(p, PIDTYPE_PID);
		nr = pid_vnr(pid);
2102 2103 2104

		if (clone_flags & CLONE_PARENT_SETTID)
			put_user(nr, parent_tidptr);
2105

L
Linus Torvalds 已提交
2106 2107 2108
		if (clone_flags & CLONE_VFORK) {
			p->vfork_done = &vfork;
			init_completion(&vfork);
O
Oleg Nesterov 已提交
2109
			get_task_struct(p);
L
Linus Torvalds 已提交
2110 2111
		}

2112
		wake_up_new_task(p);
L
Linus Torvalds 已提交
2113

T
Tejun Heo 已提交
2114 2115
		/* forking complete and child started to run, tell ptracer */
		if (unlikely(trace))
2116
			ptrace_event_pid(trace, pid);
R
Roland McGrath 已提交
2117

L
Linus Torvalds 已提交
2118
		if (clone_flags & CLONE_VFORK) {
O
Oleg Nesterov 已提交
2119
			if (!wait_for_vfork_done(p, &vfork))
2120
				ptrace_event_pid(PTRACE_EVENT_VFORK_DONE, pid);
L
Linus Torvalds 已提交
2121
		}
2122 2123

		put_pid(pid);
L
Linus Torvalds 已提交
2124
	} else {
2125
		nr = PTR_ERR(p);
L
Linus Torvalds 已提交
2126
	}
2127
	return nr;
L
Linus Torvalds 已提交
2128 2129
}

2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
#ifndef CONFIG_HAVE_COPY_THREAD_TLS
/* For compatibility with architectures that call do_fork directly rather than
 * using the syscall entry points below. */
long do_fork(unsigned long clone_flags,
	      unsigned long stack_start,
	      unsigned long stack_size,
	      int __user *parent_tidptr,
	      int __user *child_tidptr)
{
	return _do_fork(clone_flags, stack_start, stack_size,
			parent_tidptr, child_tidptr, 0);
}
#endif

2144 2145 2146 2147 2148
/*
 * Create a kernel thread.
 */
pid_t kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
{
2149 2150
	return _do_fork(flags|CLONE_VM|CLONE_UNTRACED, (unsigned long)fn,
		(unsigned long)arg, NULL, NULL, 0);
2151 2152
}

A
Al Viro 已提交
2153 2154 2155 2156
#ifdef __ARCH_WANT_SYS_FORK
SYSCALL_DEFINE0(fork)
{
#ifdef CONFIG_MMU
2157
	return _do_fork(SIGCHLD, 0, 0, NULL, NULL, 0);
A
Al Viro 已提交
2158 2159
#else
	/* can not support in nommu mode */
2160
	return -EINVAL;
A
Al Viro 已提交
2161 2162 2163 2164 2165 2166 2167
#endif
}
#endif

#ifdef __ARCH_WANT_SYS_VFORK
SYSCALL_DEFINE0(vfork)
{
2168 2169
	return _do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, 0,
			0, NULL, NULL, 0);
A
Al Viro 已提交
2170 2171 2172 2173 2174 2175 2176
}
#endif

#ifdef __ARCH_WANT_SYS_CLONE
#ifdef CONFIG_CLONE_BACKWARDS
SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp,
		 int __user *, parent_tidptr,
2177
		 unsigned long, tls,
A
Al Viro 已提交
2178 2179 2180 2181 2182
		 int __user *, child_tidptr)
#elif defined(CONFIG_CLONE_BACKWARDS2)
SYSCALL_DEFINE5(clone, unsigned long, newsp, unsigned long, clone_flags,
		 int __user *, parent_tidptr,
		 int __user *, child_tidptr,
2183
		 unsigned long, tls)
M
Michal Simek 已提交
2184 2185 2186 2187 2188
#elif defined(CONFIG_CLONE_BACKWARDS3)
SYSCALL_DEFINE6(clone, unsigned long, clone_flags, unsigned long, newsp,
		int, stack_size,
		int __user *, parent_tidptr,
		int __user *, child_tidptr,
2189
		unsigned long, tls)
A
Al Viro 已提交
2190 2191 2192 2193
#else
SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp,
		 int __user *, parent_tidptr,
		 int __user *, child_tidptr,
2194
		 unsigned long, tls)
A
Al Viro 已提交
2195 2196
#endif
{
2197
	return _do_fork(clone_flags, newsp, 0, parent_tidptr, child_tidptr, tls);
A
Al Viro 已提交
2198 2199 2200
}
#endif

2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
void walk_process_tree(struct task_struct *top, proc_visitor visitor, void *data)
{
	struct task_struct *leader, *parent, *child;
	int res;

	read_lock(&tasklist_lock);
	leader = top = top->group_leader;
down:
	for_each_thread(leader, parent) {
		list_for_each_entry(child, &parent->children, sibling) {
			res = visitor(child, data);
			if (res) {
				if (res < 0)
					goto out;
				leader = child;
				goto down;
			}
up:
			;
		}
	}

	if (leader != top) {
		child = leader;
		parent = child->real_parent;
		leader = parent->group_leader;
		goto up;
	}
out:
	read_unlock(&tasklist_lock);
}

2233 2234 2235 2236
#ifndef ARCH_MIN_MMSTRUCT_ALIGN
#define ARCH_MIN_MMSTRUCT_ALIGN 0
#endif

2237
static void sighand_ctor(void *data)
2238 2239 2240
{
	struct sighand_struct *sighand = data;

C
Christoph Lameter 已提交
2241
	spin_lock_init(&sighand->siglock);
D
Davide Libenzi 已提交
2242
	init_waitqueue_head(&sighand->signalfd_wqh);
2243 2244
}

L
Linus Torvalds 已提交
2245 2246 2247 2248
void __init proc_caches_init(void)
{
	sighand_cachep = kmem_cache_create("sighand_cache",
			sizeof(struct sighand_struct), 0,
2249
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_TYPESAFE_BY_RCU|
2250
			SLAB_ACCOUNT, sighand_ctor);
L
Linus Torvalds 已提交
2251 2252
	signal_cachep = kmem_cache_create("signal_cache",
			sizeof(struct signal_struct), 0,
2253
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
2254
			NULL);
2255
	files_cachep = kmem_cache_create("files_cache",
L
Linus Torvalds 已提交
2256
			sizeof(struct files_struct), 0,
2257
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
2258
			NULL);
2259
	fs_cachep = kmem_cache_create("fs_cache",
L
Linus Torvalds 已提交
2260
			sizeof(struct fs_struct), 0,
2261
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
2262
			NULL);
2263 2264 2265 2266 2267 2268 2269
	/*
	 * FIXME! The "sizeof(struct mm_struct)" currently includes the
	 * whole struct cpumask for the OFFSTACK case. We could change
	 * this to *only* allocate as much of it as required by the
	 * maximum number of CPU's we can ever have.  The cpumask_allocation
	 * is at the end of the structure, exactly for that reason.
	 */
2270
	mm_cachep = kmem_cache_create_usercopy("mm_struct",
2271
			sizeof(struct mm_struct), ARCH_MIN_MMSTRUCT_ALIGN,
2272
			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
2273 2274
			offsetof(struct mm_struct, saved_auxv),
			sizeof_field(struct mm_struct, saved_auxv),
2275 2276
			NULL);
	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC|SLAB_ACCOUNT);
2277
	mmap_init();
2278
	nsproxy_cache_init();
L
Linus Torvalds 已提交
2279
}
2280 2281

/*
2282
 * Check constraints on flags passed to the unshare system call.
2283
 */
2284
static int check_unshare_flags(unsigned long unshare_flags)
2285
{
2286 2287
	if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND|
				CLONE_VM|CLONE_FILES|CLONE_SYSVSEM|
2288
				CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET|
2289
				CLONE_NEWUSER|CLONE_NEWPID|CLONE_NEWCGROUP))
2290
		return -EINVAL;
2291
	/*
2292 2293 2294 2295
	 * Not implemented, but pretend it works if there is nothing
	 * to unshare.  Note that unsharing the address space or the
	 * signal handlers also need to unshare the signal queues (aka
	 * CLONE_THREAD).
2296
	 */
2297
	if (unshare_flags & (CLONE_THREAD | CLONE_SIGHAND | CLONE_VM)) {
2298 2299 2300 2301 2302 2303 2304 2305 2306
		if (!thread_group_empty(current))
			return -EINVAL;
	}
	if (unshare_flags & (CLONE_SIGHAND | CLONE_VM)) {
		if (atomic_read(&current->sighand->count) > 1)
			return -EINVAL;
	}
	if (unshare_flags & CLONE_VM) {
		if (!current_is_single_threaded())
2307 2308
			return -EINVAL;
	}
2309 2310 2311 2312 2313

	return 0;
}

/*
2314
 * Unshare the filesystem structure if it is being shared
2315 2316 2317 2318 2319
 */
static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp)
{
	struct fs_struct *fs = current->fs;

A
Al Viro 已提交
2320 2321 2322 2323 2324 2325 2326 2327 2328 2329
	if (!(unshare_flags & CLONE_FS) || !fs)
		return 0;

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

	*new_fsp = copy_fs_struct(fs);
	if (!*new_fsp)
		return -ENOMEM;
2330 2331 2332 2333 2334

	return 0;
}

/*
2335
 * Unshare file descriptor table if it is being shared
2336 2337 2338 2339
 */
static int unshare_fd(unsigned long unshare_flags, struct files_struct **new_fdp)
{
	struct files_struct *fd = current->files;
2340
	int error = 0;
2341 2342

	if ((unshare_flags & CLONE_FILES) &&
2343 2344 2345 2346 2347
	    (fd && atomic_read(&fd->count) > 1)) {
		*new_fdp = dup_fd(fd, &error);
		if (!*new_fdp)
			return error;
	}
2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359

	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.
 */
2360
SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags)
2361 2362 2363
{
	struct fs_struct *fs, *new_fs = NULL;
	struct files_struct *fd, *new_fd = NULL;
2364
	struct cred *new_cred = NULL;
2365
	struct nsproxy *new_nsproxy = NULL;
2366
	int do_sysvsem = 0;
2367
	int err;
2368

2369
	/*
2370 2371
	 * If unsharing a user namespace must also unshare the thread group
	 * and unshare the filesystem root and working directories.
2372 2373
	 */
	if (unshare_flags & CLONE_NEWUSER)
2374
		unshare_flags |= CLONE_THREAD | CLONE_FS;
2375 2376 2377 2378 2379
	/*
	 * If unsharing vm, must also unshare signal handlers.
	 */
	if (unshare_flags & CLONE_VM)
		unshare_flags |= CLONE_SIGHAND;
2380 2381 2382 2383 2384
	/*
	 * If unsharing a signal handlers, must also unshare the signal queues.
	 */
	if (unshare_flags & CLONE_SIGHAND)
		unshare_flags |= CLONE_THREAD;
2385 2386 2387 2388 2389
	/*
	 * If unsharing namespace, must also unshare filesystem information.
	 */
	if (unshare_flags & CLONE_NEWNS)
		unshare_flags |= CLONE_FS;
2390 2391 2392 2393

	err = check_unshare_flags(unshare_flags);
	if (err)
		goto bad_unshare_out;
2394 2395 2396 2397 2398 2399
	/*
	 * 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))
2400
		do_sysvsem = 1;
2401 2402
	err = unshare_fs(unshare_flags, &new_fs);
	if (err)
2403
		goto bad_unshare_out;
2404 2405
	err = unshare_fd(unshare_flags, &new_fd);
	if (err)
2406
		goto bad_unshare_cleanup_fs;
2407
	err = unshare_userns(unshare_flags, &new_cred);
2408
	if (err)
2409
		goto bad_unshare_cleanup_fd;
2410 2411 2412 2413
	err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy,
					 new_cred, new_fs);
	if (err)
		goto bad_unshare_cleanup_cred;
2414

2415
	if (new_fs || new_fd || do_sysvsem || new_cred || new_nsproxy) {
2416 2417 2418 2419 2420 2421
		if (do_sysvsem) {
			/*
			 * CLONE_SYSVSEM is equivalent to sys_exit().
			 */
			exit_sem(current);
		}
2422 2423 2424 2425 2426
		if (unshare_flags & CLONE_NEWIPC) {
			/* Orphan segments in old ns (see sem above). */
			exit_shm(current);
			shm_init_task(current);
		}
S
Serge E. Hallyn 已提交
2427

A
Alan Cox 已提交
2428
		if (new_nsproxy)
2429
			switch_task_namespaces(current, new_nsproxy);
2430

2431 2432
		task_lock(current);

2433 2434
		if (new_fs) {
			fs = current->fs;
N
Nick Piggin 已提交
2435
			spin_lock(&fs->lock);
2436
			current->fs = new_fs;
A
Al Viro 已提交
2437 2438 2439 2440
			if (--fs->users)
				new_fs = NULL;
			else
				new_fs = fs;
N
Nick Piggin 已提交
2441
			spin_unlock(&fs->lock);
2442 2443 2444 2445 2446 2447 2448 2449 2450
		}

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

		task_unlock(current);
2451 2452 2453 2454 2455 2456

		if (new_cred) {
			/* Install the new user namespace */
			commit_creds(new_cred);
			new_cred = NULL;
		}
2457 2458
	}

2459 2460
	perf_event_namespaces(current);

2461 2462 2463
bad_unshare_cleanup_cred:
	if (new_cred)
		put_cred(new_cred);
2464 2465 2466 2467 2468 2469
bad_unshare_cleanup_fd:
	if (new_fd)
		put_files_struct(new_fd);

bad_unshare_cleanup_fs:
	if (new_fs)
A
Al Viro 已提交
2470
		free_fs_struct(new_fs);
2471 2472 2473 2474

bad_unshare_out:
	return err;
}
2475 2476 2477 2478 2479 2480 2481 2482 2483 2484

/*
 *	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;
2485
	struct files_struct *copy = NULL;
2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498
	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;
}
2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521

int sysctl_max_threads(struct ctl_table *table, int write,
		       void __user *buffer, size_t *lenp, loff_t *ppos)
{
	struct ctl_table t;
	int ret;
	int threads = max_threads;
	int min = MIN_THREADS;
	int max = MAX_THREADS;

	t = *table;
	t.data = &threads;
	t.extra1 = &min;
	t.extra2 = &max;

	ret = proc_dointvec_minmax(&t, write, buffer, lenp, ppos);
	if (ret || !write)
		return ret;

	set_max_threads(threads);

	return 0;
}