Linux Audio

Check our new training course

Loading...
   1// SPDX-License-Identifier: GPL-2.0-only
   2/*
   3 *  linux/kernel/fork.c
   4 *
   5 *  Copyright (C) 1991, 1992  Linus Torvalds
   6 */
   7
   8/*
   9 *  'fork.c' contains the help-routines for the 'fork' system call
  10 * (see also entry.S and others).
  11 * Fork is rather simple, once you get the hang of it, but the memory
  12 * management can be a bitch. See 'mm/memory.c': 'copy_page_range()'
  13 */
  14
  15#include <linux/anon_inodes.h>
  16#include <linux/slab.h>
  17#include <linux/sched/autogroup.h>
  18#include <linux/sched/mm.h>
  19#include <linux/sched/coredump.h>
  20#include <linux/sched/user.h>
  21#include <linux/sched/numa_balancing.h>
  22#include <linux/sched/stat.h>
  23#include <linux/sched/task.h>
  24#include <linux/sched/task_stack.h>
  25#include <linux/sched/cputime.h>
  26#include <linux/seq_file.h>
  27#include <linux/rtmutex.h>
  28#include <linux/init.h>
  29#include <linux/unistd.h>
  30#include <linux/module.h>
  31#include <linux/vmalloc.h>
  32#include <linux/completion.h>
  33#include <linux/personality.h>
  34#include <linux/mempolicy.h>
  35#include <linux/sem.h>
  36#include <linux/file.h>
  37#include <linux/fdtable.h>
  38#include <linux/iocontext.h>
  39#include <linux/key.h>
  40#include <linux/kmsan.h>
  41#include <linux/binfmts.h>
  42#include <linux/mman.h>
  43#include <linux/mmu_notifier.h>
  44#include <linux/fs.h>
  45#include <linux/mm.h>
  46#include <linux/mm_inline.h>
  47#include <linux/nsproxy.h>
  48#include <linux/capability.h>
  49#include <linux/cpu.h>
  50#include <linux/cgroup.h>
  51#include <linux/security.h>
  52#include <linux/hugetlb.h>
  53#include <linux/seccomp.h>
  54#include <linux/swap.h>
  55#include <linux/syscalls.h>
  56#include <linux/syscall_user_dispatch.h>
  57#include <linux/jiffies.h>
  58#include <linux/futex.h>
  59#include <linux/compat.h>
  60#include <linux/kthread.h>
  61#include <linux/task_io_accounting_ops.h>
  62#include <linux/rcupdate.h>
  63#include <linux/ptrace.h>
  64#include <linux/mount.h>
  65#include <linux/audit.h>
  66#include <linux/memcontrol.h>
  67#include <linux/ftrace.h>
  68#include <linux/proc_fs.h>
  69#include <linux/profile.h>
  70#include <linux/rmap.h>
  71#include <linux/ksm.h>
  72#include <linux/acct.h>
  73#include <linux/userfaultfd_k.h>
  74#include <linux/tsacct_kern.h>
  75#include <linux/cn_proc.h>
  76#include <linux/freezer.h>
  77#include <linux/delayacct.h>
  78#include <linux/taskstats_kern.h>
  79#include <linux/tty.h>
  80#include <linux/fs_struct.h>
  81#include <linux/magic.h>
  82#include <linux/perf_event.h>
  83#include <linux/posix-timers.h>
  84#include <linux/user-return-notifier.h>
  85#include <linux/oom.h>
  86#include <linux/khugepaged.h>
  87#include <linux/signalfd.h>
  88#include <linux/uprobes.h>
  89#include <linux/aio.h>
  90#include <linux/compiler.h>
  91#include <linux/sysctl.h>
  92#include <linux/kcov.h>
  93#include <linux/livepatch.h>
  94#include <linux/thread_info.h>
  95#include <linux/stackleak.h>
  96#include <linux/kasan.h>
  97#include <linux/scs.h>
  98#include <linux/io_uring.h>
  99#include <linux/bpf.h>
 100#include <linux/stackprotector.h>
 101#include <linux/user_events.h>
 102#include <linux/iommu.h>
 103#include <linux/rseq.h>
 104#include <uapi/linux/pidfd.h>
 105#include <linux/pidfs.h>
 106
 107#include <asm/pgalloc.h>
 108#include <linux/uaccess.h>
 109#include <asm/mmu_context.h>
 110#include <asm/cacheflush.h>
 111#include <asm/tlbflush.h>
 112
 113#include <trace/events/sched.h>
 114
 115#define CREATE_TRACE_POINTS
 116#include <trace/events/task.h>
 117
 118/*
 119 * Minimum number of threads to boot the kernel
 120 */
 121#define MIN_THREADS 20
 122
 123/*
 124 * Maximum number of threads
 125 */
 126#define MAX_THREADS FUTEX_TID_MASK
 127
 128/*
 129 * Protected counters by write_lock_irq(&tasklist_lock)
 130 */
 131unsigned long total_forks;	/* Handle normal Linux uptimes. */
 132int nr_threads;			/* The idle threads do not count.. */
 133
 134static int max_threads;		/* tunable limit on nr_threads */
 135
 136#define NAMED_ARRAY_INDEX(x)	[x] = __stringify(x)
 137
 138static const char * const resident_page_types[] = {
 139	NAMED_ARRAY_INDEX(MM_FILEPAGES),
 140	NAMED_ARRAY_INDEX(MM_ANONPAGES),
 141	NAMED_ARRAY_INDEX(MM_SWAPENTS),
 142	NAMED_ARRAY_INDEX(MM_SHMEMPAGES),
 143};
 144
 145DEFINE_PER_CPU(unsigned long, process_counts) = 0;
 146
 147__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
 148
 149#ifdef CONFIG_PROVE_RCU
 150int lockdep_tasklist_lock_is_held(void)
 151{
 152	return lockdep_is_held(&tasklist_lock);
 153}
 154EXPORT_SYMBOL_GPL(lockdep_tasklist_lock_is_held);
 155#endif /* #ifdef CONFIG_PROVE_RCU */
 156
 157int nr_processes(void)
 158{
 159	int cpu;
 160	int total = 0;
 161
 162	for_each_possible_cpu(cpu)
 163		total += per_cpu(process_counts, cpu);
 164
 165	return total;
 166}
 167
 168void __weak arch_release_task_struct(struct task_struct *tsk)
 169{
 170}
 171
 172static struct kmem_cache *task_struct_cachep;
 173
 174static inline struct task_struct *alloc_task_struct_node(int node)
 175{
 176	return kmem_cache_alloc_node(task_struct_cachep, GFP_KERNEL, node);
 177}
 178
 179static inline void free_task_struct(struct task_struct *tsk)
 180{
 181	kmem_cache_free(task_struct_cachep, tsk);
 182}
 183
 184/*
 185 * Allocate pages if THREAD_SIZE is >= PAGE_SIZE, otherwise use a
 186 * kmemcache based allocator.
 187 */
 188# if THREAD_SIZE >= PAGE_SIZE || defined(CONFIG_VMAP_STACK)
 189
 190#  ifdef CONFIG_VMAP_STACK
 191/*
 192 * vmalloc() is a bit slow, and calling vfree() enough times will force a TLB
 193 * flush.  Try to minimize the number of calls by caching stacks.
 194 */
 195#define NR_CACHED_STACKS 2
 196static DEFINE_PER_CPU(struct vm_struct *, cached_stacks[NR_CACHED_STACKS]);
 197
 198struct vm_stack {
 199	struct rcu_head rcu;
 200	struct vm_struct *stack_vm_area;
 201};
 202
 203static bool try_release_thread_stack_to_cache(struct vm_struct *vm)
 204{
 205	unsigned int i;
 206
 207	for (i = 0; i < NR_CACHED_STACKS; i++) {
 208		if (this_cpu_cmpxchg(cached_stacks[i], NULL, vm) != NULL)
 209			continue;
 210		return true;
 211	}
 212	return false;
 213}
 214
 215static void thread_stack_free_rcu(struct rcu_head *rh)
 216{
 217	struct vm_stack *vm_stack = container_of(rh, struct vm_stack, rcu);
 218
 219	if (try_release_thread_stack_to_cache(vm_stack->stack_vm_area))
 220		return;
 221
 222	vfree(vm_stack);
 223}
 224
 225static void thread_stack_delayed_free(struct task_struct *tsk)
 226{
 227	struct vm_stack *vm_stack = tsk->stack;
 228
 229	vm_stack->stack_vm_area = tsk->stack_vm_area;
 230	call_rcu(&vm_stack->rcu, thread_stack_free_rcu);
 231}
 232
 233static int free_vm_stack_cache(unsigned int cpu)
 234{
 235	struct vm_struct **cached_vm_stacks = per_cpu_ptr(cached_stacks, cpu);
 236	int i;
 237
 238	for (i = 0; i < NR_CACHED_STACKS; i++) {
 239		struct vm_struct *vm_stack = cached_vm_stacks[i];
 240
 241		if (!vm_stack)
 242			continue;
 243
 244		vfree(vm_stack->addr);
 245		cached_vm_stacks[i] = NULL;
 246	}
 247
 248	return 0;
 249}
 250
 251static int memcg_charge_kernel_stack(struct vm_struct *vm)
 252{
 253	int i;
 254	int ret;
 255	int nr_charged = 0;
 256
 257	BUG_ON(vm->nr_pages != THREAD_SIZE / PAGE_SIZE);
 258
 259	for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++) {
 260		ret = memcg_kmem_charge_page(vm->pages[i], GFP_KERNEL, 0);
 261		if (ret)
 262			goto err;
 263		nr_charged++;
 264	}
 265	return 0;
 266err:
 267	for (i = 0; i < nr_charged; i++)
 268		memcg_kmem_uncharge_page(vm->pages[i], 0);
 269	return ret;
 270}
 271
 272static int alloc_thread_stack_node(struct task_struct *tsk, int node)
 273{
 274	struct vm_struct *vm;
 275	void *stack;
 276	int i;
 277
 278	for (i = 0; i < NR_CACHED_STACKS; i++) {
 279		struct vm_struct *s;
 280
 281		s = this_cpu_xchg(cached_stacks[i], NULL);
 282
 283		if (!s)
 284			continue;
 285
 286		/* Reset stack metadata. */
 287		kasan_unpoison_range(s->addr, THREAD_SIZE);
 288
 289		stack = kasan_reset_tag(s->addr);
 290
 291		/* Clear stale pointers from reused stack. */
 292		memset(stack, 0, THREAD_SIZE);
 293
 294		if (memcg_charge_kernel_stack(s)) {
 295			vfree(s->addr);
 296			return -ENOMEM;
 297		}
 298
 299		tsk->stack_vm_area = s;
 300		tsk->stack = stack;
 301		return 0;
 302	}
 303
 304	/*
 305	 * Allocated stacks are cached and later reused by new threads,
 306	 * so memcg accounting is performed manually on assigning/releasing
 307	 * stacks to tasks. Drop __GFP_ACCOUNT.
 308	 */
 309	stack = __vmalloc_node_range(THREAD_SIZE, THREAD_ALIGN,
 310				     VMALLOC_START, VMALLOC_END,
 311				     THREADINFO_GFP & ~__GFP_ACCOUNT,
 312				     PAGE_KERNEL,
 313				     0, node, __builtin_return_address(0));
 314	if (!stack)
 315		return -ENOMEM;
 316
 317	vm = find_vm_area(stack);
 318	if (memcg_charge_kernel_stack(vm)) {
 319		vfree(stack);
 320		return -ENOMEM;
 321	}
 322	/*
 323	 * We can't call find_vm_area() in interrupt context, and
 324	 * free_thread_stack() can be called in interrupt context,
 325	 * so cache the vm_struct.
 326	 */
 327	tsk->stack_vm_area = vm;
 328	stack = kasan_reset_tag(stack);
 329	tsk->stack = stack;
 330	return 0;
 331}
 332
 333static void free_thread_stack(struct task_struct *tsk)
 334{
 335	if (!try_release_thread_stack_to_cache(tsk->stack_vm_area))
 336		thread_stack_delayed_free(tsk);
 337
 338	tsk->stack = NULL;
 339	tsk->stack_vm_area = NULL;
 340}
 341
 342#  else /* !CONFIG_VMAP_STACK */
 343
 344static void thread_stack_free_rcu(struct rcu_head *rh)
 345{
 346	__free_pages(virt_to_page(rh), THREAD_SIZE_ORDER);
 347}
 348
 349static void thread_stack_delayed_free(struct task_struct *tsk)
 350{
 351	struct rcu_head *rh = tsk->stack;
 352
 353	call_rcu(rh, thread_stack_free_rcu);
 354}
 355
 356static int alloc_thread_stack_node(struct task_struct *tsk, int node)
 357{
 358	struct page *page = alloc_pages_node(node, THREADINFO_GFP,
 359					     THREAD_SIZE_ORDER);
 360
 361	if (likely(page)) {
 362		tsk->stack = kasan_reset_tag(page_address(page));
 363		return 0;
 364	}
 365	return -ENOMEM;
 366}
 367
 368static void free_thread_stack(struct task_struct *tsk)
 369{
 370	thread_stack_delayed_free(tsk);
 371	tsk->stack = NULL;
 372}
 373
 374#  endif /* CONFIG_VMAP_STACK */
 375# else /* !(THREAD_SIZE >= PAGE_SIZE || defined(CONFIG_VMAP_STACK)) */
 376
 377static struct kmem_cache *thread_stack_cache;
 378
 379static void thread_stack_free_rcu(struct rcu_head *rh)
 380{
 381	kmem_cache_free(thread_stack_cache, rh);
 382}
 383
 384static void thread_stack_delayed_free(struct task_struct *tsk)
 385{
 386	struct rcu_head *rh = tsk->stack;
 387
 388	call_rcu(rh, thread_stack_free_rcu);
 389}
 390
 391static int alloc_thread_stack_node(struct task_struct *tsk, int node)
 392{
 393	unsigned long *stack;
 394	stack = kmem_cache_alloc_node(thread_stack_cache, THREADINFO_GFP, node);
 395	stack = kasan_reset_tag(stack);
 396	tsk->stack = stack;
 397	return stack ? 0 : -ENOMEM;
 398}
 399
 400static void free_thread_stack(struct task_struct *tsk)
 401{
 402	thread_stack_delayed_free(tsk);
 403	tsk->stack = NULL;
 404}
 405
 406void thread_stack_cache_init(void)
 407{
 408	thread_stack_cache = kmem_cache_create_usercopy("thread_stack",
 409					THREAD_SIZE, THREAD_SIZE, 0, 0,
 410					THREAD_SIZE, NULL);
 411	BUG_ON(thread_stack_cache == NULL);
 412}
 413
 414# endif /* THREAD_SIZE >= PAGE_SIZE || defined(CONFIG_VMAP_STACK) */
 415
 416/* SLAB cache for signal_struct structures (tsk->signal) */
 417static struct kmem_cache *signal_cachep;
 418
 419/* SLAB cache for sighand_struct structures (tsk->sighand) */
 420struct kmem_cache *sighand_cachep;
 421
 422/* SLAB cache for files_struct structures (tsk->files) */
 423struct kmem_cache *files_cachep;
 424
 425/* SLAB cache for fs_struct structures (tsk->fs) */
 426struct kmem_cache *fs_cachep;
 427
 428/* SLAB cache for vm_area_struct structures */
 429static struct kmem_cache *vm_area_cachep;
 430
 431/* SLAB cache for mm_struct structures (tsk->mm) */
 432static struct kmem_cache *mm_cachep;
 433
 434#ifdef CONFIG_PER_VMA_LOCK
 435
 436/* SLAB cache for vm_area_struct.lock */
 437static struct kmem_cache *vma_lock_cachep;
 438
 439static bool vma_lock_alloc(struct vm_area_struct *vma)
 440{
 441	vma->vm_lock = kmem_cache_alloc(vma_lock_cachep, GFP_KERNEL);
 442	if (!vma->vm_lock)
 443		return false;
 444
 445	init_rwsem(&vma->vm_lock->lock);
 446	vma->vm_lock_seq = -1;
 447
 448	return true;
 449}
 450
 451static inline void vma_lock_free(struct vm_area_struct *vma)
 452{
 453	kmem_cache_free(vma_lock_cachep, vma->vm_lock);
 454}
 455
 456#else /* CONFIG_PER_VMA_LOCK */
 457
 458static inline bool vma_lock_alloc(struct vm_area_struct *vma) { return true; }
 459static inline void vma_lock_free(struct vm_area_struct *vma) {}
 460
 461#endif /* CONFIG_PER_VMA_LOCK */
 462
 463struct vm_area_struct *vm_area_alloc(struct mm_struct *mm)
 464{
 465	struct vm_area_struct *vma;
 466
 467	vma = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
 468	if (!vma)
 469		return NULL;
 470
 471	vma_init(vma, mm);
 472	if (!vma_lock_alloc(vma)) {
 473		kmem_cache_free(vm_area_cachep, vma);
 474		return NULL;
 475	}
 476
 477	return vma;
 478}
 479
 480struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig)
 481{
 482	struct vm_area_struct *new = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
 483
 484	if (!new)
 485		return NULL;
 486
 487	ASSERT_EXCLUSIVE_WRITER(orig->vm_flags);
 488	ASSERT_EXCLUSIVE_WRITER(orig->vm_file);
 489	/*
 490	 * orig->shared.rb may be modified concurrently, but the clone
 491	 * will be reinitialized.
 492	 */
 493	data_race(memcpy(new, orig, sizeof(*new)));
 494	if (!vma_lock_alloc(new)) {
 495		kmem_cache_free(vm_area_cachep, new);
 496		return NULL;
 497	}
 498	INIT_LIST_HEAD(&new->anon_vma_chain);
 499	vma_numab_state_init(new);
 500	dup_anon_vma_name(orig, new);
 501
 502	return new;
 503}
 504
 505void __vm_area_free(struct vm_area_struct *vma)
 506{
 507	vma_numab_state_free(vma);
 508	free_anon_vma_name(vma);
 509	vma_lock_free(vma);
 510	kmem_cache_free(vm_area_cachep, vma);
 511}
 512
 513#ifdef CONFIG_PER_VMA_LOCK
 514static void vm_area_free_rcu_cb(struct rcu_head *head)
 515{
 516	struct vm_area_struct *vma = container_of(head, struct vm_area_struct,
 517						  vm_rcu);
 518
 519	/* The vma should not be locked while being destroyed. */
 520	VM_BUG_ON_VMA(rwsem_is_locked(&vma->vm_lock->lock), vma);
 521	__vm_area_free(vma);
 522}
 523#endif
 524
 525void vm_area_free(struct vm_area_struct *vma)
 526{
 527#ifdef CONFIG_PER_VMA_LOCK
 528	call_rcu(&vma->vm_rcu, vm_area_free_rcu_cb);
 529#else
 530	__vm_area_free(vma);
 531#endif
 532}
 533
 534static void account_kernel_stack(struct task_struct *tsk, int account)
 535{
 536	if (IS_ENABLED(CONFIG_VMAP_STACK)) {
 537		struct vm_struct *vm = task_stack_vm_area(tsk);
 538		int i;
 539
 540		for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++)
 541			mod_lruvec_page_state(vm->pages[i], NR_KERNEL_STACK_KB,
 542					      account * (PAGE_SIZE / 1024));
 543	} else {
 544		void *stack = task_stack_page(tsk);
 545
 546		/* All stack pages are in the same node. */
 547		mod_lruvec_kmem_state(stack, NR_KERNEL_STACK_KB,
 548				      account * (THREAD_SIZE / 1024));
 549	}
 550}
 551
 552void exit_task_stack_account(struct task_struct *tsk)
 553{
 554	account_kernel_stack(tsk, -1);
 555
 556	if (IS_ENABLED(CONFIG_VMAP_STACK)) {
 557		struct vm_struct *vm;
 558		int i;
 559
 560		vm = task_stack_vm_area(tsk);
 561		for (i = 0; i < THREAD_SIZE / PAGE_SIZE; i++)
 562			memcg_kmem_uncharge_page(vm->pages[i], 0);
 563	}
 564}
 565
 566static void release_task_stack(struct task_struct *tsk)
 567{
 568	if (WARN_ON(READ_ONCE(tsk->__state) != TASK_DEAD))
 569		return;  /* Better to leak the stack than to free prematurely */
 570
 571	free_thread_stack(tsk);
 572}
 573
 574#ifdef CONFIG_THREAD_INFO_IN_TASK
 575void put_task_stack(struct task_struct *tsk)
 576{
 577	if (refcount_dec_and_test(&tsk->stack_refcount))
 578		release_task_stack(tsk);
 579}
 580#endif
 581
 582void free_task(struct task_struct *tsk)
 583{
 584#ifdef CONFIG_SECCOMP
 585	WARN_ON_ONCE(tsk->seccomp.filter);
 586#endif
 587	release_user_cpus_ptr(tsk);
 588	scs_release(tsk);
 589
 590#ifndef CONFIG_THREAD_INFO_IN_TASK
 591	/*
 592	 * The task is finally done with both the stack and thread_info,
 593	 * so free both.
 594	 */
 595	release_task_stack(tsk);
 596#else
 597	/*
 598	 * If the task had a separate stack allocation, it should be gone
 599	 * by now.
 600	 */
 601	WARN_ON_ONCE(refcount_read(&tsk->stack_refcount) != 0);
 602#endif
 603	rt_mutex_debug_task_free(tsk);
 604	ftrace_graph_exit_task(tsk);
 605	arch_release_task_struct(tsk);
 606	if (tsk->flags & PF_KTHREAD)
 607		free_kthread_struct(tsk);
 608	bpf_task_storage_free(tsk);
 609	free_task_struct(tsk);
 610}
 611EXPORT_SYMBOL(free_task);
 612
 613static void dup_mm_exe_file(struct mm_struct *mm, struct mm_struct *oldmm)
 614{
 615	struct file *exe_file;
 616
 617	exe_file = get_mm_exe_file(oldmm);
 618	RCU_INIT_POINTER(mm->exe_file, exe_file);
 619	/*
 620	 * We depend on the oldmm having properly denied write access to the
 621	 * exe_file already.
 622	 */
 623	if (exe_file && deny_write_access(exe_file))
 624		pr_warn_once("deny_write_access() failed in %s\n", __func__);
 625}
 626
 627#ifdef CONFIG_MMU
 628static __latent_entropy int dup_mmap(struct mm_struct *mm,
 629					struct mm_struct *oldmm)
 630{
 631	struct vm_area_struct *mpnt, *tmp;
 632	int retval;
 633	unsigned long charge = 0;
 634	LIST_HEAD(uf);
 635	VMA_ITERATOR(vmi, mm, 0);
 636
 637	uprobe_start_dup_mmap();
 638	if (mmap_write_lock_killable(oldmm)) {
 639		retval = -EINTR;
 640		goto fail_uprobe_end;
 641	}
 642	flush_cache_dup_mm(oldmm);
 643	uprobe_dup_mmap(oldmm, mm);
 644	/*
 645	 * Not linked in yet - no deadlock potential:
 646	 */
 647	mmap_write_lock_nested(mm, SINGLE_DEPTH_NESTING);
 648
 649	/* No ordering required: file already has been exposed. */
 650	dup_mm_exe_file(mm, oldmm);
 651
 652	mm->total_vm = oldmm->total_vm;
 653	mm->data_vm = oldmm->data_vm;
 654	mm->exec_vm = oldmm->exec_vm;
 655	mm->stack_vm = oldmm->stack_vm;
 656
 657	retval = ksm_fork(mm, oldmm);
 658	if (retval)
 659		goto out;
 660	khugepaged_fork(mm, oldmm);
 661
 662	/* Use __mt_dup() to efficiently build an identical maple tree. */
 663	retval = __mt_dup(&oldmm->mm_mt, &mm->mm_mt, GFP_KERNEL);
 664	if (unlikely(retval))
 665		goto out;
 666
 667	mt_clear_in_rcu(vmi.mas.tree);
 668	for_each_vma(vmi, mpnt) {
 669		struct file *file;
 670
 671		vma_start_write(mpnt);
 672		if (mpnt->vm_flags & VM_DONTCOPY) {
 673			retval = vma_iter_clear_gfp(&vmi, mpnt->vm_start,
 674						    mpnt->vm_end, GFP_KERNEL);
 675			if (retval)
 676				goto loop_out;
 677
 678			vm_stat_account(mm, mpnt->vm_flags, -vma_pages(mpnt));
 679			continue;
 680		}
 681		charge = 0;
 682		/*
 683		 * Don't duplicate many vmas if we've been oom-killed (for
 684		 * example)
 685		 */
 686		if (fatal_signal_pending(current)) {
 687			retval = -EINTR;
 688			goto loop_out;
 689		}
 690		if (mpnt->vm_flags & VM_ACCOUNT) {
 691			unsigned long len = vma_pages(mpnt);
 692
 693			if (security_vm_enough_memory_mm(oldmm, len)) /* sic */
 694				goto fail_nomem;
 695			charge = len;
 696		}
 697		tmp = vm_area_dup(mpnt);
 698		if (!tmp)
 699			goto fail_nomem;
 700		retval = vma_dup_policy(mpnt, tmp);
 701		if (retval)
 702			goto fail_nomem_policy;
 703		tmp->vm_mm = mm;
 704		retval = dup_userfaultfd(tmp, &uf);
 705		if (retval)
 706			goto fail_nomem_anon_vma_fork;
 707		if (tmp->vm_flags & VM_WIPEONFORK) {
 708			/*
 709			 * VM_WIPEONFORK gets a clean slate in the child.
 710			 * Don't prepare anon_vma until fault since we don't
 711			 * copy page for current vma.
 712			 */
 713			tmp->anon_vma = NULL;
 714		} else if (anon_vma_fork(tmp, mpnt))
 715			goto fail_nomem_anon_vma_fork;
 716		vm_flags_clear(tmp, VM_LOCKED_MASK);
 717		/*
 718		 * Copy/update hugetlb private vma information.
 719		 */
 720		if (is_vm_hugetlb_page(tmp))
 721			hugetlb_dup_vma_private(tmp);
 722
 723		/*
 724		 * Link the vma into the MT. After using __mt_dup(), memory
 725		 * allocation is not necessary here, so it cannot fail.
 726		 */
 727		vma_iter_bulk_store(&vmi, tmp);
 728
 729		mm->map_count++;
 730
 731		if (tmp->vm_ops && tmp->vm_ops->open)
 732			tmp->vm_ops->open(tmp);
 733
 734		file = tmp->vm_file;
 735		if (file) {
 736			struct address_space *mapping = file->f_mapping;
 737
 738			get_file(file);
 739			i_mmap_lock_write(mapping);
 740			if (vma_is_shared_maywrite(tmp))
 741				mapping_allow_writable(mapping);
 742			flush_dcache_mmap_lock(mapping);
 743			/* insert tmp into the share list, just after mpnt */
 744			vma_interval_tree_insert_after(tmp, mpnt,
 745					&mapping->i_mmap);
 746			flush_dcache_mmap_unlock(mapping);
 747			i_mmap_unlock_write(mapping);
 748		}
 749
 750		if (!(tmp->vm_flags & VM_WIPEONFORK))
 751			retval = copy_page_range(tmp, mpnt);
 752
 753		if (retval) {
 754			mpnt = vma_next(&vmi);
 755			goto loop_out;
 756		}
 757	}
 758	/* a new mm has just been created */
 759	retval = arch_dup_mmap(oldmm, mm);
 760loop_out:
 761	vma_iter_free(&vmi);
 762	if (!retval) {
 763		mt_set_in_rcu(vmi.mas.tree);
 764	} else if (mpnt) {
 765		/*
 766		 * The entire maple tree has already been duplicated. If the
 767		 * mmap duplication fails, mark the failure point with
 768		 * XA_ZERO_ENTRY. In exit_mmap(), if this marker is encountered,
 769		 * stop releasing VMAs that have not been duplicated after this
 770		 * point.
 771		 */
 772		mas_set_range(&vmi.mas, mpnt->vm_start, mpnt->vm_end - 1);
 773		mas_store(&vmi.mas, XA_ZERO_ENTRY);
 774	}
 775out:
 776	mmap_write_unlock(mm);
 777	flush_tlb_mm(oldmm);
 778	mmap_write_unlock(oldmm);
 779	dup_userfaultfd_complete(&uf);
 780fail_uprobe_end:
 781	uprobe_end_dup_mmap();
 782	return retval;
 783
 784fail_nomem_anon_vma_fork:
 785	mpol_put(vma_policy(tmp));
 786fail_nomem_policy:
 787	vm_area_free(tmp);
 788fail_nomem:
 789	retval = -ENOMEM;
 790	vm_unacct_memory(charge);
 791	goto loop_out;
 792}
 793
 794static inline int mm_alloc_pgd(struct mm_struct *mm)
 795{
 796	mm->pgd = pgd_alloc(mm);
 797	if (unlikely(!mm->pgd))
 798		return -ENOMEM;
 799	return 0;
 800}
 801
 802static inline void mm_free_pgd(struct mm_struct *mm)
 803{
 804	pgd_free(mm, mm->pgd);
 805}
 806#else
 807static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
 808{
 809	mmap_write_lock(oldmm);
 810	dup_mm_exe_file(mm, oldmm);
 811	mmap_write_unlock(oldmm);
 812	return 0;
 813}
 814#define mm_alloc_pgd(mm)	(0)
 815#define mm_free_pgd(mm)
 816#endif /* CONFIG_MMU */
 817
 818static void check_mm(struct mm_struct *mm)
 819{
 820	int i;
 821
 822	BUILD_BUG_ON_MSG(ARRAY_SIZE(resident_page_types) != NR_MM_COUNTERS,
 823			 "Please make sure 'struct resident_page_types[]' is updated as well");
 824
 825	for (i = 0; i < NR_MM_COUNTERS; i++) {
 826		long x = percpu_counter_sum(&mm->rss_stat[i]);
 827
 828		if (unlikely(x))
 829			pr_alert("BUG: Bad rss-counter state mm:%p type:%s val:%ld\n",
 830				 mm, resident_page_types[i], x);
 831	}
 832
 833	if (mm_pgtables_bytes(mm))
 834		pr_alert("BUG: non-zero pgtables_bytes on freeing mm: %ld\n",
 835				mm_pgtables_bytes(mm));
 836
 837#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
 838	VM_BUG_ON_MM(mm->pmd_huge_pte, mm);
 839#endif
 840}
 841
 842#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
 843#define free_mm(mm)	(kmem_cache_free(mm_cachep, (mm)))
 844
 845static void do_check_lazy_tlb(void *arg)
 846{
 847	struct mm_struct *mm = arg;
 848
 849	WARN_ON_ONCE(current->active_mm == mm);
 850}
 851
 852static void do_shoot_lazy_tlb(void *arg)
 853{
 854	struct mm_struct *mm = arg;
 855
 856	if (current->active_mm == mm) {
 857		WARN_ON_ONCE(current->mm);
 858		current->active_mm = &init_mm;
 859		switch_mm(mm, &init_mm, current);
 860	}
 861}
 862
 863static void cleanup_lazy_tlbs(struct mm_struct *mm)
 864{
 865	if (!IS_ENABLED(CONFIG_MMU_LAZY_TLB_SHOOTDOWN)) {
 866		/*
 867		 * In this case, lazy tlb mms are refounted and would not reach
 868		 * __mmdrop until all CPUs have switched away and mmdrop()ed.
 869		 */
 870		return;
 871	}
 872
 873	/*
 874	 * Lazy mm shootdown does not refcount "lazy tlb mm" usage, rather it
 875	 * requires lazy mm users to switch to another mm when the refcount
 876	 * drops to zero, before the mm is freed. This requires IPIs here to
 877	 * switch kernel threads to init_mm.
 878	 *
 879	 * archs that use IPIs to flush TLBs can piggy-back that lazy tlb mm
 880	 * switch with the final userspace teardown TLB flush which leaves the
 881	 * mm lazy on this CPU but no others, reducing the need for additional
 882	 * IPIs here. There are cases where a final IPI is still required here,
 883	 * such as the final mmdrop being performed on a different CPU than the
 884	 * one exiting, or kernel threads using the mm when userspace exits.
 885	 *
 886	 * IPI overheads have not found to be expensive, but they could be
 887	 * reduced in a number of possible ways, for example (roughly
 888	 * increasing order of complexity):
 889	 * - The last lazy reference created by exit_mm() could instead switch
 890	 *   to init_mm, however it's probable this will run on the same CPU
 891	 *   immediately afterwards, so this may not reduce IPIs much.
 892	 * - A batch of mms requiring IPIs could be gathered and freed at once.
 893	 * - CPUs store active_mm where it can be remotely checked without a
 894	 *   lock, to filter out false-positives in the cpumask.
 895	 * - After mm_users or mm_count reaches zero, switching away from the
 896	 *   mm could clear mm_cpumask to reduce some IPIs, perhaps together
 897	 *   with some batching or delaying of the final IPIs.
 898	 * - A delayed freeing and RCU-like quiescing sequence based on mm
 899	 *   switching to avoid IPIs completely.
 900	 */
 901	on_each_cpu_mask(mm_cpumask(mm), do_shoot_lazy_tlb, (void *)mm, 1);
 902	if (IS_ENABLED(CONFIG_DEBUG_VM_SHOOT_LAZIES))
 903		on_each_cpu(do_check_lazy_tlb, (void *)mm, 1);
 904}
 905
 906/*
 907 * Called when the last reference to the mm
 908 * is dropped: either by a lazy thread or by
 909 * mmput. Free the page directory and the mm.
 910 */
 911void __mmdrop(struct mm_struct *mm)
 912{
 913	BUG_ON(mm == &init_mm);
 914	WARN_ON_ONCE(mm == current->mm);
 915
 916	/* Ensure no CPUs are using this as their lazy tlb mm */
 917	cleanup_lazy_tlbs(mm);
 918
 919	WARN_ON_ONCE(mm == current->active_mm);
 920	mm_free_pgd(mm);
 921	destroy_context(mm);
 922	mmu_notifier_subscriptions_destroy(mm);
 923	check_mm(mm);
 924	put_user_ns(mm->user_ns);
 925	mm_pasid_drop(mm);
 926	mm_destroy_cid(mm);
 927	percpu_counter_destroy_many(mm->rss_stat, NR_MM_COUNTERS);
 928
 929	free_mm(mm);
 930}
 931EXPORT_SYMBOL_GPL(__mmdrop);
 932
 933static void mmdrop_async_fn(struct work_struct *work)
 934{
 935	struct mm_struct *mm;
 936
 937	mm = container_of(work, struct mm_struct, async_put_work);
 938	__mmdrop(mm);
 939}
 940
 941static void mmdrop_async(struct mm_struct *mm)
 942{
 943	if (unlikely(atomic_dec_and_test(&mm->mm_count))) {
 944		INIT_WORK(&mm->async_put_work, mmdrop_async_fn);
 945		schedule_work(&mm->async_put_work);
 946	}
 947}
 948
 949static inline void free_signal_struct(struct signal_struct *sig)
 950{
 951	taskstats_tgid_free(sig);
 952	sched_autogroup_exit(sig);
 953	/*
 954	 * __mmdrop is not safe to call from softirq context on x86 due to
 955	 * pgd_dtor so postpone it to the async context
 956	 */
 957	if (sig->oom_mm)
 958		mmdrop_async(sig->oom_mm);
 959	kmem_cache_free(signal_cachep, sig);
 960}
 961
 962static inline void put_signal_struct(struct signal_struct *sig)
 963{
 964	if (refcount_dec_and_test(&sig->sigcnt))
 965		free_signal_struct(sig);
 966}
 967
 968void __put_task_struct(struct task_struct *tsk)
 969{
 970	WARN_ON(!tsk->exit_state);
 971	WARN_ON(refcount_read(&tsk->usage));
 972	WARN_ON(tsk == current);
 973
 974	io_uring_free(tsk);
 975	cgroup_free(tsk);
 976	task_numa_free(tsk, true);
 977	security_task_free(tsk);
 978	exit_creds(tsk);
 979	delayacct_tsk_free(tsk);
 980	put_signal_struct(tsk->signal);
 981	sched_core_free(tsk);
 982	free_task(tsk);
 983}
 984EXPORT_SYMBOL_GPL(__put_task_struct);
 985
 986void __put_task_struct_rcu_cb(struct rcu_head *rhp)
 987{
 988	struct task_struct *task = container_of(rhp, struct task_struct, rcu);
 989
 990	__put_task_struct(task);
 991}
 992EXPORT_SYMBOL_GPL(__put_task_struct_rcu_cb);
 993
 994void __init __weak arch_task_cache_init(void) { }
 995
 996/*
 997 * set_max_threads
 998 */
 999static void set_max_threads(unsigned int max_threads_suggested)
1000{
1001	u64 threads;
1002	unsigned long nr_pages = totalram_pages();
1003
1004	/*
1005	 * The number of threads shall be limited such that the thread
1006	 * structures may only consume a small part of the available memory.
1007	 */
1008	if (fls64(nr_pages) + fls64(PAGE_SIZE) > 64)
1009		threads = MAX_THREADS;
1010	else
1011		threads = div64_u64((u64) nr_pages * (u64) PAGE_SIZE,
1012				    (u64) THREAD_SIZE * 8UL);
1013
1014	if (threads > max_threads_suggested)
1015		threads = max_threads_suggested;
1016
1017	max_threads = clamp_t(u64, threads, MIN_THREADS, MAX_THREADS);
1018}
1019
1020#ifdef CONFIG_ARCH_WANTS_DYNAMIC_TASK_STRUCT
1021/* Initialized by the architecture: */
1022int arch_task_struct_size __read_mostly;
1023#endif
1024
1025static void task_struct_whitelist(unsigned long *offset, unsigned long *size)
1026{
1027	/* Fetch thread_struct whitelist for the architecture. */
1028	arch_thread_struct_whitelist(offset, size);
1029
1030	/*
1031	 * Handle zero-sized whitelist or empty thread_struct, otherwise
1032	 * adjust offset to position of thread_struct in task_struct.
1033	 */
1034	if (unlikely(*size == 0))
1035		*offset = 0;
1036	else
1037		*offset += offsetof(struct task_struct, thread);
1038}
1039
1040void __init fork_init(void)
1041{
1042	int i;
1043#ifndef ARCH_MIN_TASKALIGN
1044#define ARCH_MIN_TASKALIGN	0
1045#endif
1046	int align = max_t(int, L1_CACHE_BYTES, ARCH_MIN_TASKALIGN);
1047	unsigned long useroffset, usersize;
1048
1049	/* create a slab on which task_structs can be allocated */
1050	task_struct_whitelist(&useroffset, &usersize);
1051	task_struct_cachep = kmem_cache_create_usercopy("task_struct",
1052			arch_task_struct_size, align,
1053			SLAB_PANIC|SLAB_ACCOUNT,
1054			useroffset, usersize, NULL);
1055
1056	/* do the arch specific task caches init */
1057	arch_task_cache_init();
1058
1059	set_max_threads(MAX_THREADS);
1060
1061	init_task.signal->rlim[RLIMIT_NPROC].rlim_cur = max_threads/2;
1062	init_task.signal->rlim[RLIMIT_NPROC].rlim_max = max_threads/2;
1063	init_task.signal->rlim[RLIMIT_SIGPENDING] =
1064		init_task.signal->rlim[RLIMIT_NPROC];
1065
1066	for (i = 0; i < UCOUNT_COUNTS; i++)
1067		init_user_ns.ucount_max[i] = max_threads/2;
1068
1069	set_userns_rlimit_max(&init_user_ns, UCOUNT_RLIMIT_NPROC,      RLIM_INFINITY);
1070	set_userns_rlimit_max(&init_user_ns, UCOUNT_RLIMIT_MSGQUEUE,   RLIM_INFINITY);
1071	set_userns_rlimit_max(&init_user_ns, UCOUNT_RLIMIT_SIGPENDING, RLIM_INFINITY);
1072	set_userns_rlimit_max(&init_user_ns, UCOUNT_RLIMIT_MEMLOCK,    RLIM_INFINITY);
1073
1074#ifdef CONFIG_VMAP_STACK
1075	cpuhp_setup_state(CPUHP_BP_PREPARE_DYN, "fork:vm_stack_cache",
1076			  NULL, free_vm_stack_cache);
1077#endif
1078
1079	scs_init();
1080
1081	lockdep_init_task(&init_task);
1082	uprobes_init();
1083}
1084
1085int __weak arch_dup_task_struct(struct task_struct *dst,
1086					       struct task_struct *src)
1087{
1088	*dst = *src;
1089	return 0;
1090}
1091
1092void set_task_stack_end_magic(struct task_struct *tsk)
1093{
1094	unsigned long *stackend;
1095
1096	stackend = end_of_stack(tsk);
1097	*stackend = STACK_END_MAGIC;	/* for overflow detection */
1098}
1099
1100static struct task_struct *dup_task_struct(struct task_struct *orig, int node)
1101{
1102	struct task_struct *tsk;
1103	int err;
1104
1105	if (node == NUMA_NO_NODE)
1106		node = tsk_fork_get_node(orig);
1107	tsk = alloc_task_struct_node(node);
1108	if (!tsk)
1109		return NULL;
1110
1111	err = arch_dup_task_struct(tsk, orig);
1112	if (err)
1113		goto free_tsk;
1114
1115	err = alloc_thread_stack_node(tsk, node);
1116	if (err)
1117		goto free_tsk;
1118
1119#ifdef CONFIG_THREAD_INFO_IN_TASK
1120	refcount_set(&tsk->stack_refcount, 1);
1121#endif
1122	account_kernel_stack(tsk, 1);
1123
1124	err = scs_prepare(tsk, node);
1125	if (err)
1126		goto free_stack;
1127
1128#ifdef CONFIG_SECCOMP
1129	/*
1130	 * We must handle setting up seccomp filters once we're under
1131	 * the sighand lock in case orig has changed between now and
1132	 * then. Until then, filter must be NULL to avoid messing up
1133	 * the usage counts on the error path calling free_task.
1134	 */
1135	tsk->seccomp.filter = NULL;
1136#endif
1137
1138	setup_thread_stack(tsk, orig);
1139	clear_user_return_notifier(tsk);
1140	clear_tsk_need_resched(tsk);
1141	set_task_stack_end_magic(tsk);
1142	clear_syscall_work_syscall_user_dispatch(tsk);
1143
1144#ifdef CONFIG_STACKPROTECTOR
1145	tsk->stack_canary = get_random_canary();
1146#endif
1147	if (orig->cpus_ptr == &orig->cpus_mask)
1148		tsk->cpus_ptr = &tsk->cpus_mask;
1149	dup_user_cpus_ptr(tsk, orig, node);
1150
1151	/*
1152	 * One for the user space visible state that goes away when reaped.
1153	 * One for the scheduler.
1154	 */
1155	refcount_set(&tsk->rcu_users, 2);
1156	/* One for the rcu users */
1157	refcount_set(&tsk->usage, 1);
1158#ifdef CONFIG_BLK_DEV_IO_TRACE
1159	tsk->btrace_seq = 0;
1160#endif
1161	tsk->splice_pipe = NULL;
1162	tsk->task_frag.page = NULL;
1163	tsk->wake_q.next = NULL;
1164	tsk->worker_private = NULL;
1165
1166	kcov_task_init(tsk);
1167	kmsan_task_create(tsk);
1168	kmap_local_fork(tsk);
1169
1170#ifdef CONFIG_FAULT_INJECTION
1171	tsk->fail_nth = 0;
1172#endif
1173
1174#ifdef CONFIG_BLK_CGROUP
1175	tsk->throttle_disk = NULL;
1176	tsk->use_memdelay = 0;
1177#endif
1178
1179#ifdef CONFIG_ARCH_HAS_CPU_PASID
1180	tsk->pasid_activated = 0;
1181#endif
1182
1183#ifdef CONFIG_MEMCG
1184	tsk->active_memcg = NULL;
1185#endif
1186
1187#ifdef CONFIG_CPU_SUP_INTEL
1188	tsk->reported_split_lock = 0;
1189#endif
1190
1191#ifdef CONFIG_SCHED_MM_CID
1192	tsk->mm_cid = -1;
1193	tsk->last_mm_cid = -1;
1194	tsk->mm_cid_active = 0;
1195	tsk->migrate_from_cpu = -1;
1196#endif
1197	return tsk;
1198
1199free_stack:
1200	exit_task_stack_account(tsk);
1201	free_thread_stack(tsk);
1202free_tsk:
1203	free_task_struct(tsk);
1204	return NULL;
1205}
1206
1207__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
1208
1209static unsigned long default_dump_filter = MMF_DUMP_FILTER_DEFAULT;
1210
1211static int __init coredump_filter_setup(char *s)
1212{
1213	default_dump_filter =
1214		(simple_strtoul(s, NULL, 0) << MMF_DUMP_FILTER_SHIFT) &
1215		MMF_DUMP_FILTER_MASK;
1216	return 1;
1217}
1218
1219__setup("coredump_filter=", coredump_filter_setup);
1220
1221#include <linux/init_task.h>
1222
1223static void mm_init_aio(struct mm_struct *mm)
1224{
1225#ifdef CONFIG_AIO
1226	spin_lock_init(&mm->ioctx_lock);
1227	mm->ioctx_table = NULL;
1228#endif
1229}
1230
1231static __always_inline void mm_clear_owner(struct mm_struct *mm,
1232					   struct task_struct *p)
1233{
1234#ifdef CONFIG_MEMCG
1235	if (mm->owner == p)
1236		WRITE_ONCE(mm->owner, NULL);
1237#endif
1238}
1239
1240static void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
1241{
1242#ifdef CONFIG_MEMCG
1243	mm->owner = p;
1244#endif
1245}
1246
1247static void mm_init_uprobes_state(struct mm_struct *mm)
1248{
1249#ifdef CONFIG_UPROBES
1250	mm->uprobes_state.xol_area = NULL;
1251#endif
1252}
1253
1254static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
1255	struct user_namespace *user_ns)
1256{
1257	mt_init_flags(&mm->mm_mt, MM_MT_FLAGS);
1258	mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock);
1259	atomic_set(&mm->mm_users, 1);
1260	atomic_set(&mm->mm_count, 1);
1261	seqcount_init(&mm->write_protect_seq);
1262	mmap_init_lock(mm);
1263	INIT_LIST_HEAD(&mm->mmlist);
1264#ifdef CONFIG_PER_VMA_LOCK
1265	mm->mm_lock_seq = 0;
1266#endif
1267	mm_pgtables_bytes_init(mm);
1268	mm->map_count = 0;
1269	mm->locked_vm = 0;
1270	atomic64_set(&mm->pinned_vm, 0);
1271	memset(&mm->rss_stat, 0, sizeof(mm->rss_stat));
1272	spin_lock_init(&mm->page_table_lock);
1273	spin_lock_init(&mm->arg_lock);
1274	mm_init_cpumask(mm);
1275	mm_init_aio(mm);
1276	mm_init_owner(mm, p);
1277	mm_pasid_init(mm);
1278	RCU_INIT_POINTER(mm->exe_file, NULL);
1279	mmu_notifier_subscriptions_init(mm);
1280	init_tlb_flush_pending(mm);
1281#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS
1282	mm->pmd_huge_pte = NULL;
1283#endif
1284	mm_init_uprobes_state(mm);
1285	hugetlb_count_init(mm);
1286
1287	if (current->mm) {
1288		mm->flags = mmf_init_flags(current->mm->flags);
1289		mm->def_flags = current->mm->def_flags & VM_INIT_DEF_MASK;
1290	} else {
1291		mm->flags = default_dump_filter;
1292		mm->def_flags = 0;
1293	}
1294
1295	if (mm_alloc_pgd(mm))
1296		goto fail_nopgd;
1297
1298	if (init_new_context(p, mm))
1299		goto fail_nocontext;
1300
1301	if (mm_alloc_cid(mm))
1302		goto fail_cid;
1303
1304	if (percpu_counter_init_many(mm->rss_stat, 0, GFP_KERNEL_ACCOUNT,
1305				     NR_MM_COUNTERS))
1306		goto fail_pcpu;
1307
1308	mm->user_ns = get_user_ns(user_ns);
1309	lru_gen_init_mm(mm);
1310	return mm;
1311
1312fail_pcpu:
1313	mm_destroy_cid(mm);
1314fail_cid:
1315	destroy_context(mm);
1316fail_nocontext:
1317	mm_free_pgd(mm);
1318fail_nopgd:
1319	free_mm(mm);
1320	return NULL;
1321}
1322
1323/*
1324 * Allocate and initialize an mm_struct.
1325 */
1326struct mm_struct *mm_alloc(void)
1327{
1328	struct mm_struct *mm;
1329
1330	mm = allocate_mm();
1331	if (!mm)
1332		return NULL;
1333
1334	memset(mm, 0, sizeof(*mm));
1335	return mm_init(mm, current, current_user_ns());
1336}
1337
1338static inline void __mmput(struct mm_struct *mm)
1339{
1340	VM_BUG_ON(atomic_read(&mm->mm_users));
1341
1342	uprobe_clear_state(mm);
1343	exit_aio(mm);
1344	ksm_exit(mm);
1345	khugepaged_exit(mm); /* must run before exit_mmap */
1346	exit_mmap(mm);
1347	mm_put_huge_zero_page(mm);
1348	set_mm_exe_file(mm, NULL);
1349	if (!list_empty(&mm->mmlist)) {
1350		spin_lock(&mmlist_lock);
1351		list_del(&mm->mmlist);
1352		spin_unlock(&mmlist_lock);
1353	}
1354	if (mm->binfmt)
1355		module_put(mm->binfmt->module);
1356	lru_gen_del_mm(mm);
1357	mmdrop(mm);
1358}
1359
1360/*
1361 * Decrement the use count and release all resources for an mm.
1362 */
1363void mmput(struct mm_struct *mm)
1364{
1365	might_sleep();
1366
1367	if (atomic_dec_and_test(&mm->mm_users))
1368		__mmput(mm);
1369}
1370EXPORT_SYMBOL_GPL(mmput);
1371
1372#ifdef CONFIG_MMU
1373static void mmput_async_fn(struct work_struct *work)
1374{
1375	struct mm_struct *mm = container_of(work, struct mm_struct,
1376					    async_put_work);
1377
1378	__mmput(mm);
1379}
1380
1381void mmput_async(struct mm_struct *mm)
1382{
1383	if (atomic_dec_and_test(&mm->mm_users)) {
1384		INIT_WORK(&mm->async_put_work, mmput_async_fn);
1385		schedule_work(&mm->async_put_work);
1386	}
1387}
1388EXPORT_SYMBOL_GPL(mmput_async);
1389#endif
1390
1391/**
1392 * set_mm_exe_file - change a reference to the mm's executable file
1393 * @mm: The mm to change.
1394 * @new_exe_file: The new file to use.
1395 *
1396 * This changes mm's executable file (shown as symlink /proc/[pid]/exe).
1397 *
1398 * Main users are mmput() and sys_execve(). Callers prevent concurrent
1399 * invocations: in mmput() nobody alive left, in execve it happens before
1400 * the new mm is made visible to anyone.
1401 *
1402 * Can only fail if new_exe_file != NULL.
1403 */
1404int set_mm_exe_file(struct mm_struct *mm, struct file *new_exe_file)
1405{
1406	struct file *old_exe_file;
1407
1408	/*
1409	 * It is safe to dereference the exe_file without RCU as
1410	 * this function is only called if nobody else can access
1411	 * this mm -- see comment above for justification.
1412	 */
1413	old_exe_file = rcu_dereference_raw(mm->exe_file);
1414
1415	if (new_exe_file) {
1416		/*
1417		 * We expect the caller (i.e., sys_execve) to already denied
1418		 * write access, so this is unlikely to fail.
1419		 */
1420		if (unlikely(deny_write_access(new_exe_file)))
1421			return -EACCES;
1422		get_file(new_exe_file);
1423	}
1424	rcu_assign_pointer(mm->exe_file, new_exe_file);
1425	if (old_exe_file) {
1426		allow_write_access(old_exe_file);
1427		fput(old_exe_file);
1428	}
1429	return 0;
1430}
1431
1432/**
1433 * replace_mm_exe_file - replace a reference to the mm's executable file
1434 * @mm: The mm to change.
1435 * @new_exe_file: The new file to use.
1436 *
1437 * This changes mm's executable file (shown as symlink /proc/[pid]/exe).
1438 *
1439 * Main user is sys_prctl(PR_SET_MM_MAP/EXE_FILE).
1440 */
1441int replace_mm_exe_file(struct mm_struct *mm, struct file *new_exe_file)
1442{
1443	struct vm_area_struct *vma;
1444	struct file *old_exe_file;
1445	int ret = 0;
1446
1447	/* Forbid mm->exe_file change if old file still mapped. */
1448	old_exe_file = get_mm_exe_file(mm);
1449	if (old_exe_file) {
1450		VMA_ITERATOR(vmi, mm, 0);
1451		mmap_read_lock(mm);
1452		for_each_vma(vmi, vma) {
1453			if (!vma->vm_file)
1454				continue;
1455			if (path_equal(&vma->vm_file->f_path,
1456				       &old_exe_file->f_path)) {
1457				ret = -EBUSY;
1458				break;
1459			}
1460		}
1461		mmap_read_unlock(mm);
1462		fput(old_exe_file);
1463		if (ret)
1464			return ret;
1465	}
1466
1467	ret = deny_write_access(new_exe_file);
1468	if (ret)
1469		return -EACCES;
1470	get_file(new_exe_file);
1471
1472	/* set the new file */
1473	mmap_write_lock(mm);
1474	old_exe_file = rcu_dereference_raw(mm->exe_file);
1475	rcu_assign_pointer(mm->exe_file, new_exe_file);
1476	mmap_write_unlock(mm);
1477
1478	if (old_exe_file) {
1479		allow_write_access(old_exe_file);
1480		fput(old_exe_file);
1481	}
1482	return 0;
1483}
1484
1485/**
1486 * get_mm_exe_file - acquire a reference to the mm's executable file
1487 * @mm: The mm of interest.
1488 *
1489 * Returns %NULL if mm has no associated executable file.
1490 * User must release file via fput().
1491 */
1492struct file *get_mm_exe_file(struct mm_struct *mm)
1493{
1494	struct file *exe_file;
1495
1496	rcu_read_lock();
1497	exe_file = get_file_rcu(&mm->exe_file);
1498	rcu_read_unlock();
1499	return exe_file;
1500}
1501
1502/**
1503 * get_task_exe_file - acquire a reference to the task's executable file
1504 * @task: The task.
1505 *
1506 * Returns %NULL if task's mm (if any) has no associated executable file or
1507 * this is a kernel thread with borrowed mm (see the comment above get_task_mm).
1508 * User must release file via fput().
1509 */
1510struct file *get_task_exe_file(struct task_struct *task)
1511{
1512	struct file *exe_file = NULL;
1513	struct mm_struct *mm;
1514
1515	task_lock(task);
1516	mm = task->mm;
1517	if (mm) {
1518		if (!(task->flags & PF_KTHREAD))
1519			exe_file = get_mm_exe_file(mm);
1520	}
1521	task_unlock(task);
1522	return exe_file;
1523}
1524
1525/**
1526 * get_task_mm - acquire a reference to the task's mm
1527 * @task: The task.
1528 *
1529 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
1530 * this kernel workthread has transiently adopted a user mm with use_mm,
1531 * to do its AIO) is not set and if so returns a reference to it, after
1532 * bumping up the use count.  User must release the mm via mmput()
1533 * after use.  Typically used by /proc and ptrace.
1534 */
1535struct mm_struct *get_task_mm(struct task_struct *task)
1536{
1537	struct mm_struct *mm;
1538
1539	task_lock(task);
1540	mm = task->mm;
1541	if (mm) {
1542		if (task->flags & PF_KTHREAD)
1543			mm = NULL;
1544		else
1545			mmget(mm);
1546	}
1547	task_unlock(task);
1548	return mm;
1549}
1550EXPORT_SYMBOL_GPL(get_task_mm);
1551
1552struct mm_struct *mm_access(struct task_struct *task, unsigned int mode)
1553{
1554	struct mm_struct *mm;
1555	int err;
1556
1557	err =  down_read_killable(&task->signal->exec_update_lock);
1558	if (err)
1559		return ERR_PTR(err);
1560
1561	mm = get_task_mm(task);
1562	if (mm && mm != current->mm &&
1563			!ptrace_may_access(task, mode)) {
1564		mmput(mm);
1565		mm = ERR_PTR(-EACCES);
1566	}
1567	up_read(&task->signal->exec_update_lock);
1568
1569	return mm;
1570}
1571
1572static void complete_vfork_done(struct task_struct *tsk)
1573{
1574	struct completion *vfork;
1575
1576	task_lock(tsk);
1577	vfork = tsk->vfork_done;
1578	if (likely(vfork)) {
1579		tsk->vfork_done = NULL;
1580		complete(vfork);
1581	}
1582	task_unlock(tsk);
1583}
1584
1585static int wait_for_vfork_done(struct task_struct *child,
1586				struct completion *vfork)
1587{
1588	unsigned int state = TASK_KILLABLE|TASK_FREEZABLE;
1589	int killed;
1590
1591	cgroup_enter_frozen();
1592	killed = wait_for_completion_state(vfork, state);
1593	cgroup_leave_frozen(false);
1594
1595	if (killed) {
1596		task_lock(child);
1597		child->vfork_done = NULL;
1598		task_unlock(child);
1599	}
1600
1601	put_task_struct(child);
1602	return killed;
1603}
1604
1605/* Please note the differences between mmput and mm_release.
1606 * mmput is called whenever we stop holding onto a mm_struct,
1607 * error success whatever.
1608 *
1609 * mm_release is called after a mm_struct has been removed
1610 * from the current process.
1611 *
1612 * This difference is important for error handling, when we
1613 * only half set up a mm_struct for a new process and need to restore
1614 * the old one.  Because we mmput the new mm_struct before
1615 * restoring the old one. . .
1616 * Eric Biederman 10 January 1998
1617 */
1618static void mm_release(struct task_struct *tsk, struct mm_struct *mm)
1619{
1620	uprobe_free_utask(tsk);
1621
1622	/* Get rid of any cached register state */
1623	deactivate_mm(tsk, mm);
1624
1625	/*
1626	 * Signal userspace if we're not exiting with a core dump
1627	 * because we want to leave the value intact for debugging
1628	 * purposes.
1629	 */
1630	if (tsk->clear_child_tid) {
1631		if (atomic_read(&mm->mm_users) > 1) {
1632			/*
1633			 * We don't check the error code - if userspace has
1634			 * not set up a proper pointer then tough luck.
1635			 */
1636			put_user(0, tsk->clear_child_tid);
1637			do_futex(tsk->clear_child_tid, FUTEX_WAKE,
1638					1, NULL, NULL, 0, 0);
1639		}
1640		tsk->clear_child_tid = NULL;
1641	}
1642
1643	/*
1644	 * All done, finally we can wake up parent and return this mm to him.
1645	 * Also kthread_stop() uses this completion for synchronization.
1646	 */
1647	if (tsk->vfork_done)
1648		complete_vfork_done(tsk);
1649}
1650
1651void exit_mm_release(struct task_struct *tsk, struct mm_struct *mm)
1652{
1653	futex_exit_release(tsk);
1654	mm_release(tsk, mm);
1655}
1656
1657void exec_mm_release(struct task_struct *tsk, struct mm_struct *mm)
1658{
1659	futex_exec_release(tsk);
1660	mm_release(tsk, mm);
1661}
1662
1663/**
1664 * dup_mm() - duplicates an existing mm structure
1665 * @tsk: the task_struct with which the new mm will be associated.
1666 * @oldmm: the mm to duplicate.
1667 *
1668 * Allocates a new mm structure and duplicates the provided @oldmm structure
1669 * content into it.
1670 *
1671 * Return: the duplicated mm or NULL on failure.
1672 */
1673static struct mm_struct *dup_mm(struct task_struct *tsk,
1674				struct mm_struct *oldmm)
1675{
1676	struct mm_struct *mm;
1677	int err;
1678
1679	mm = allocate_mm();
1680	if (!mm)
1681		goto fail_nomem;
1682
1683	memcpy(mm, oldmm, sizeof(*mm));
1684
1685	if (!mm_init(mm, tsk, mm->user_ns))
1686		goto fail_nomem;
1687
1688	err = dup_mmap(mm, oldmm);
1689	if (err)
1690		goto free_pt;
1691
1692	mm->hiwater_rss = get_mm_rss(mm);
1693	mm->hiwater_vm = mm->total_vm;
1694
1695	if (mm->binfmt && !try_module_get(mm->binfmt->module))
1696		goto free_pt;
1697
1698	return mm;
1699
1700free_pt:
1701	/* don't put binfmt in mmput, we haven't got module yet */
1702	mm->binfmt = NULL;
1703	mm_init_owner(mm, NULL);
1704	mmput(mm);
1705
1706fail_nomem:
1707	return NULL;
1708}
1709
1710static int copy_mm(unsigned long clone_flags, struct task_struct *tsk)
1711{
1712	struct mm_struct *mm, *oldmm;
1713
1714	tsk->min_flt = tsk->maj_flt = 0;
1715	tsk->nvcsw = tsk->nivcsw = 0;
1716#ifdef CONFIG_DETECT_HUNG_TASK
1717	tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw;
1718	tsk->last_switch_time = 0;
1719#endif
1720
1721	tsk->mm = NULL;
1722	tsk->active_mm = NULL;
1723
1724	/*
1725	 * Are we cloning a kernel thread?
1726	 *
1727	 * We need to steal a active VM for that..
1728	 */
1729	oldmm = current->mm;
1730	if (!oldmm)
1731		return 0;
1732
1733	if (clone_flags & CLONE_VM) {
1734		mmget(oldmm);
1735		mm = oldmm;
1736	} else {
1737		mm = dup_mm(tsk, current->mm);
1738		if (!mm)
1739			return -ENOMEM;
1740	}
1741
1742	tsk->mm = mm;
1743	tsk->active_mm = mm;
1744	sched_mm_cid_fork(tsk);
1745	return 0;
1746}
1747
1748static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
1749{
1750	struct fs_struct *fs = current->fs;
1751	if (clone_flags & CLONE_FS) {
1752		/* tsk->fs is already what we want */
1753		spin_lock(&fs->lock);
1754		/* "users" and "in_exec" locked for check_unsafe_exec() */
1755		if (fs->in_exec) {
1756			spin_unlock(&fs->lock);
1757			return -EAGAIN;
1758		}
1759		fs->users++;
1760		spin_unlock(&fs->lock);
1761		return 0;
1762	}
1763	tsk->fs = copy_fs_struct(fs);
1764	if (!tsk->fs)
1765		return -ENOMEM;
1766	return 0;
1767}
1768
1769static int copy_files(unsigned long clone_flags, struct task_struct *tsk,
1770		      int no_files)
1771{
1772	struct files_struct *oldf, *newf;
1773	int error = 0;
1774
1775	/*
1776	 * A background process may not have any files ...
1777	 */
1778	oldf = current->files;
1779	if (!oldf)
1780		goto out;
1781
1782	if (no_files) {
1783		tsk->files = NULL;
1784		goto out;
1785	}
1786
1787	if (clone_flags & CLONE_FILES) {
1788		atomic_inc(&oldf->count);
1789		goto out;
1790	}
1791
1792	newf = dup_fd(oldf, NR_OPEN_MAX, &error);
1793	if (!newf)
1794		goto out;
1795
1796	tsk->files = newf;
1797	error = 0;
1798out:
1799	return error;
1800}
1801
1802static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
1803{
1804	struct sighand_struct *sig;
1805
1806	if (clone_flags & CLONE_SIGHAND) {
1807		refcount_inc(&current->sighand->count);
1808		return 0;
1809	}
1810	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
1811	RCU_INIT_POINTER(tsk->sighand, sig);
1812	if (!sig)
1813		return -ENOMEM;
1814
1815	refcount_set(&sig->count, 1);
1816	spin_lock_irq(&current->sighand->siglock);
1817	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
1818	spin_unlock_irq(&current->sighand->siglock);
1819
1820	/* Reset all signal handler not set to SIG_IGN to SIG_DFL. */
1821	if (clone_flags & CLONE_CLEAR_SIGHAND)
1822		flush_signal_handlers(tsk, 0);
1823
1824	return 0;
1825}
1826
1827void __cleanup_sighand(struct sighand_struct *sighand)
1828{
1829	if (refcount_dec_and_test(&sighand->count)) {
1830		signalfd_cleanup(sighand);
1831		/*
1832		 * sighand_cachep is SLAB_TYPESAFE_BY_RCU so we can free it
1833		 * without an RCU grace period, see __lock_task_sighand().
1834		 */
1835		kmem_cache_free(sighand_cachep, sighand);
1836	}
1837}
1838
1839/*
1840 * Initialize POSIX timer handling for a thread group.
1841 */
1842static void posix_cpu_timers_init_group(struct signal_struct *sig)
1843{
1844	struct posix_cputimers *pct = &sig->posix_cputimers;
1845	unsigned long cpu_limit;
1846
1847	cpu_limit = READ_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur);
1848	posix_cputimers_group_init(pct, cpu_limit);
1849}
1850
1851static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
1852{
1853	struct signal_struct *sig;
1854
1855	if (clone_flags & CLONE_THREAD)
1856		return 0;
1857
1858	sig = kmem_cache_zalloc(signal_cachep, GFP_KERNEL);
1859	tsk->signal = sig;
1860	if (!sig)
1861		return -ENOMEM;
1862
1863	sig->nr_threads = 1;
1864	sig->quick_threads = 1;
1865	atomic_set(&sig->live, 1);
1866	refcount_set(&sig->sigcnt, 1);
1867
1868	/* list_add(thread_node, thread_head) without INIT_LIST_HEAD() */
1869	sig->thread_head = (struct list_head)LIST_HEAD_INIT(tsk->thread_node);
1870	tsk->thread_node = (struct list_head)LIST_HEAD_INIT(sig->thread_head);
1871
1872	init_waitqueue_head(&sig->wait_chldexit);
1873	sig->curr_target = tsk;
1874	init_sigpending(&sig->shared_pending);
1875	INIT_HLIST_HEAD(&sig->multiprocess);
1876	seqlock_init(&sig->stats_lock);
1877	prev_cputime_init(&sig->prev_cputime);
1878
1879#ifdef CONFIG_POSIX_TIMERS
1880	INIT_LIST_HEAD(&sig->posix_timers);
1881	hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
1882	sig->real_timer.function = it_real_fn;
1883#endif
1884
1885	task_lock(current->group_leader);
1886	memcpy(sig->rlim, current->signal->rlim, sizeof sig->rlim);
1887	task_unlock(current->group_leader);
1888
1889	posix_cpu_timers_init_group(sig);
1890
1891	tty_audit_fork(sig);
1892	sched_autogroup_fork(sig);
1893
1894	sig->oom_score_adj = current->signal->oom_score_adj;
1895	sig->oom_score_adj_min = current->signal->oom_score_adj_min;
1896
1897	mutex_init(&sig->cred_guard_mutex);
1898	init_rwsem(&sig->exec_update_lock);
1899
1900	return 0;
1901}
1902
1903static void copy_seccomp(struct task_struct *p)
1904{
1905#ifdef CONFIG_SECCOMP
1906	/*
1907	 * Must be called with sighand->lock held, which is common to
1908	 * all threads in the group. Holding cred_guard_mutex is not
1909	 * needed because this new task is not yet running and cannot
1910	 * be racing exec.
1911	 */
1912	assert_spin_locked(&current->sighand->siglock);
1913
1914	/* Ref-count the new filter user, and assign it. */
1915	get_seccomp_filter(current);
1916	p->seccomp = current->seccomp;
1917
1918	/*
1919	 * Explicitly enable no_new_privs here in case it got set
1920	 * between the task_struct being duplicated and holding the
1921	 * sighand lock. The seccomp state and nnp must be in sync.
1922	 */
1923	if (task_no_new_privs(current))
1924		task_set_no_new_privs(p);
1925
1926	/*
1927	 * If the parent gained a seccomp mode after copying thread
1928	 * flags and between before we held the sighand lock, we have
1929	 * to manually enable the seccomp thread flag here.
1930	 */
1931	if (p->seccomp.mode != SECCOMP_MODE_DISABLED)
1932		set_task_syscall_work(p, SECCOMP);
1933#endif
1934}
1935
1936SYSCALL_DEFINE1(set_tid_address, int __user *, tidptr)
1937{
1938	current->clear_child_tid = tidptr;
1939
1940	return task_pid_vnr(current);
1941}
1942
1943static void rt_mutex_init_task(struct task_struct *p)
1944{
1945	raw_spin_lock_init(&p->pi_lock);
1946#ifdef CONFIG_RT_MUTEXES
1947	p->pi_waiters = RB_ROOT_CACHED;
1948	p->pi_top_task = NULL;
1949	p->pi_blocked_on = NULL;
1950#endif
1951}
1952
1953static inline void init_task_pid_links(struct task_struct *task)
1954{
1955	enum pid_type type;
1956
1957	for (type = PIDTYPE_PID; type < PIDTYPE_MAX; ++type)
1958		INIT_HLIST_NODE(&task->pid_links[type]);
1959}
1960
1961static inline void
1962init_task_pid(struct task_struct *task, enum pid_type type, struct pid *pid)
1963{
1964	if (type == PIDTYPE_PID)
1965		task->thread_pid = pid;
1966	else
1967		task->signal->pids[type] = pid;
1968}
1969
1970static inline void rcu_copy_process(struct task_struct *p)
1971{
1972#ifdef CONFIG_PREEMPT_RCU
1973	p->rcu_read_lock_nesting = 0;
1974	p->rcu_read_unlock_special.s = 0;
1975	p->rcu_blocked_node = NULL;
1976	INIT_LIST_HEAD(&p->rcu_node_entry);
1977#endif /* #ifdef CONFIG_PREEMPT_RCU */
1978#ifdef CONFIG_TASKS_RCU
1979	p->rcu_tasks_holdout = false;
1980	INIT_LIST_HEAD(&p->rcu_tasks_holdout_list);
1981	p->rcu_tasks_idle_cpu = -1;
1982	INIT_LIST_HEAD(&p->rcu_tasks_exit_list);
1983#endif /* #ifdef CONFIG_TASKS_RCU */
1984#ifdef CONFIG_TASKS_TRACE_RCU
1985	p->trc_reader_nesting = 0;
1986	p->trc_reader_special.s = 0;
1987	INIT_LIST_HEAD(&p->trc_holdout_list);
1988	INIT_LIST_HEAD(&p->trc_blkd_node);
1989#endif /* #ifdef CONFIG_TASKS_TRACE_RCU */
1990}
1991
1992/**
1993 * __pidfd_prepare - allocate a new pidfd_file and reserve a pidfd
1994 * @pid:   the struct pid for which to create a pidfd
1995 * @flags: flags of the new @pidfd
1996 * @ret: Where to return the file for the pidfd.
1997 *
1998 * Allocate a new file that stashes @pid and reserve a new pidfd number in the
1999 * caller's file descriptor table. The pidfd is reserved but not installed yet.
2000 *
2001 * The helper doesn't perform checks on @pid which makes it useful for pidfds
2002 * created via CLONE_PIDFD where @pid has no task attached when the pidfd and
2003 * pidfd file are prepared.
2004 *
2005 * If this function returns successfully the caller is responsible to either
2006 * call fd_install() passing the returned pidfd and pidfd file as arguments in
2007 * order to install the pidfd into its file descriptor table or they must use
2008 * put_unused_fd() and fput() on the returned pidfd and pidfd file
2009 * respectively.
2010 *
2011 * This function is useful when a pidfd must already be reserved but there
2012 * might still be points of failure afterwards and the caller wants to ensure
2013 * that no pidfd is leaked into its file descriptor table.
2014 *
2015 * Return: On success, a reserved pidfd is returned from the function and a new
2016 *         pidfd file is returned in the last argument to the function. On
2017 *         error, a negative error code is returned from the function and the
2018 *         last argument remains unchanged.
2019 */
2020static int __pidfd_prepare(struct pid *pid, unsigned int flags, struct file **ret)
2021{
2022	int pidfd;
2023	struct file *pidfd_file;
2024
2025	pidfd = get_unused_fd_flags(O_CLOEXEC);
2026	if (pidfd < 0)
2027		return pidfd;
2028
2029	pidfd_file = pidfs_alloc_file(pid, flags | O_RDWR);
2030	if (IS_ERR(pidfd_file)) {
2031		put_unused_fd(pidfd);
2032		return PTR_ERR(pidfd_file);
2033	}
2034	/*
2035	 * anon_inode_getfile() ignores everything outside of the
2036	 * O_ACCMODE | O_NONBLOCK mask, set PIDFD_THREAD manually.
2037	 */
2038	pidfd_file->f_flags |= (flags & PIDFD_THREAD);
2039	*ret = pidfd_file;
2040	return pidfd;
2041}
2042
2043/**
2044 * pidfd_prepare - allocate a new pidfd_file and reserve a pidfd
2045 * @pid:   the struct pid for which to create a pidfd
2046 * @flags: flags of the new @pidfd
2047 * @ret: Where to return the pidfd.
2048 *
2049 * Allocate a new file that stashes @pid and reserve a new pidfd number in the
2050 * caller's file descriptor table. The pidfd is reserved but not installed yet.
2051 *
2052 * The helper verifies that @pid is still in use, without PIDFD_THREAD the
2053 * task identified by @pid must be a thread-group leader.
2054 *
2055 * If this function returns successfully the caller is responsible to either
2056 * call fd_install() passing the returned pidfd and pidfd file as arguments in
2057 * order to install the pidfd into its file descriptor table or they must use
2058 * put_unused_fd() and fput() on the returned pidfd and pidfd file
2059 * respectively.
2060 *
2061 * This function is useful when a pidfd must already be reserved but there
2062 * might still be points of failure afterwards and the caller wants to ensure
2063 * that no pidfd is leaked into its file descriptor table.
2064 *
2065 * Return: On success, a reserved pidfd is returned from the function and a new
2066 *         pidfd file is returned in the last argument to the function. On
2067 *         error, a negative error code is returned from the function and the
2068 *         last argument remains unchanged.
2069 */
2070int pidfd_prepare(struct pid *pid, unsigned int flags, struct file **ret)
2071{
2072	bool thread = flags & PIDFD_THREAD;
2073
2074	if (!pid || !pid_has_task(pid, thread ? PIDTYPE_PID : PIDTYPE_TGID))
2075		return -EINVAL;
2076
2077	return __pidfd_prepare(pid, flags, ret);
2078}
2079
2080static void __delayed_free_task(struct rcu_head *rhp)
2081{
2082	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);
2083
2084	free_task(tsk);
2085}
2086
2087static __always_inline void delayed_free_task(struct task_struct *tsk)
2088{
2089	if (IS_ENABLED(CONFIG_MEMCG))
2090		call_rcu(&tsk->rcu, __delayed_free_task);
2091	else
2092		free_task(tsk);
2093}
2094
2095static void copy_oom_score_adj(u64 clone_flags, struct task_struct *tsk)
2096{
2097	/* Skip if kernel thread */
2098	if (!tsk->mm)
2099		return;
2100
2101	/* Skip if spawning a thread or using vfork */
2102	if ((clone_flags & (CLONE_VM | CLONE_THREAD | CLONE_VFORK)) != CLONE_VM)
2103		return;
2104
2105	/* We need to synchronize with __set_oom_adj */
2106	mutex_lock(&oom_adj_mutex);
2107	set_bit(MMF_MULTIPROCESS, &tsk->mm->flags);
2108	/* Update the values in case they were changed after copy_signal */
2109	tsk->signal->oom_score_adj = current->signal->oom_score_adj;
2110	tsk->signal->oom_score_adj_min = current->signal->oom_score_adj_min;
2111	mutex_unlock(&oom_adj_mutex);
2112}
2113
2114#ifdef CONFIG_RV
2115static void rv_task_fork(struct task_struct *p)
2116{
2117	int i;
2118
2119	for (i = 0; i < RV_PER_TASK_MONITORS; i++)
2120		p->rv[i].da_mon.monitoring = false;
2121}
2122#else
2123#define rv_task_fork(p) do {} while (0)
2124#endif
2125
2126/*
2127 * This creates a new process as a copy of the old one,
2128 * but does not actually start it yet.
2129 *
2130 * It copies the registers, and all the appropriate
2131 * parts of the process environment (as per the clone
2132 * flags). The actual kick-off is left to the caller.
2133 */
2134__latent_entropy struct task_struct *copy_process(
2135					struct pid *pid,
2136					int trace,
2137					int node,
2138					struct kernel_clone_args *args)
2139{
2140	int pidfd = -1, retval;
2141	struct task_struct *p;
2142	struct multiprocess_signals delayed;
2143	struct file *pidfile = NULL;
2144	const u64 clone_flags = args->flags;
2145	struct nsproxy *nsp = current->nsproxy;
2146
2147	/*
2148	 * Don't allow sharing the root directory with processes in a different
2149	 * namespace
2150	 */
2151	if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
2152		return ERR_PTR(-EINVAL);
2153
2154	if ((clone_flags & (CLONE_NEWUSER|CLONE_FS)) == (CLONE_NEWUSER|CLONE_FS))
2155		return ERR_PTR(-EINVAL);
2156
2157	/*
2158	 * Thread groups must share signals as well, and detached threads
2159	 * can only be started up within the thread group.
2160	 */
2161	if ((clone_flags & CLONE_THREAD) && !(clone_flags & CLONE_SIGHAND))
2162		return ERR_PTR(-EINVAL);
2163
2164	/*
2165	 * Shared signal handlers imply shared VM. By way of the above,
2166	 * thread groups also imply shared VM. Blocking this case allows
2167	 * for various simplifications in other code.
2168	 */
2169	if ((clone_flags & CLONE_SIGHAND) && !(clone_flags & CLONE_VM))
2170		return ERR_PTR(-EINVAL);
2171
2172	/*
2173	 * Siblings of global init remain as zombies on exit since they are
2174	 * not reaped by their parent (swapper). To solve this and to avoid
2175	 * multi-rooted process trees, prevent global and container-inits
2176	 * from creating siblings.
2177	 */
2178	if ((clone_flags & CLONE_PARENT) &&
2179				current->signal->flags & SIGNAL_UNKILLABLE)
2180		return ERR_PTR(-EINVAL);
2181
2182	/*
2183	 * If the new process will be in a different pid or user namespace
2184	 * do not allow it to share a thread group with the forking task.
2185	 */
2186	if (clone_flags & CLONE_THREAD) {
2187		if ((clone_flags & (CLONE_NEWUSER | CLONE_NEWPID)) ||
2188		    (task_active_pid_ns(current) != nsp->pid_ns_for_children))
2189			return ERR_PTR(-EINVAL);
2190	}
2191
2192	if (clone_flags & CLONE_PIDFD) {
2193		/*
2194		 * - CLONE_DETACHED is blocked so that we can potentially
2195		 *   reuse it later for CLONE_PIDFD.
2196		 */
2197		if (clone_flags & CLONE_DETACHED)
2198			return ERR_PTR(-EINVAL);
2199	}
2200
2201	/*
2202	 * Force any signals received before this point to be delivered
2203	 * before the fork happens.  Collect up signals sent to multiple
2204	 * processes that happen during the fork and delay them so that
2205	 * they appear to happen after the fork.
2206	 */
2207	sigemptyset(&delayed.signal);
2208	INIT_HLIST_NODE(&delayed.node);
2209
2210	spin_lock_irq(&current->sighand->siglock);
2211	if (!(clone_flags & CLONE_THREAD))
2212		hlist_add_head(&delayed.node, &current->signal->multiprocess);
2213	recalc_sigpending();
2214	spin_unlock_irq(&current->sighand->siglock);
2215	retval = -ERESTARTNOINTR;
2216	if (task_sigpending(current))
2217		goto fork_out;
2218
2219	retval = -ENOMEM;
2220	p = dup_task_struct(current, node);
2221	if (!p)
2222		goto fork_out;
2223	p->flags &= ~PF_KTHREAD;
2224	if (args->kthread)
2225		p->flags |= PF_KTHREAD;
2226	if (args->user_worker) {
2227		/*
2228		 * Mark us a user worker, and block any signal that isn't
2229		 * fatal or STOP
2230		 */
2231		p->flags |= PF_USER_WORKER;
2232		siginitsetinv(&p->blocked, sigmask(SIGKILL)|sigmask(SIGSTOP));
2233	}
2234	if (args->io_thread)
2235		p->flags |= PF_IO_WORKER;
2236
2237	if (args->name)
2238		strscpy_pad(p->comm, args->name, sizeof(p->comm));
2239
2240	p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? args->child_tid : NULL;
2241	/*
2242	 * Clear TID on mm_release()?
2243	 */
2244	p->clear_child_tid = (clone_flags & CLONE_CHILD_CLEARTID) ? args->child_tid : NULL;
2245
2246	ftrace_graph_init_task(p);
2247
2248	rt_mutex_init_task(p);
2249
2250	lockdep_assert_irqs_enabled();
2251#ifdef CONFIG_PROVE_LOCKING
2252	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
2253#endif
2254	retval = copy_creds(p, clone_flags);
2255	if (retval < 0)
2256		goto bad_fork_free;
2257
2258	retval = -EAGAIN;
2259	if (is_rlimit_overlimit(task_ucounts(p), UCOUNT_RLIMIT_NPROC, rlimit(RLIMIT_NPROC))) {
2260		if (p->real_cred->user != INIT_USER &&
2261		    !capable(CAP_SYS_RESOURCE) && !capable(CAP_SYS_ADMIN))
2262			goto bad_fork_cleanup_count;
2263	}
2264	current->flags &= ~PF_NPROC_EXCEEDED;
2265
2266	/*
2267	 * If multiple threads are within copy_process(), then this check
2268	 * triggers too late. This doesn't hurt, the check is only there
2269	 * to stop root fork bombs.
2270	 */
2271	retval = -EAGAIN;
2272	if (data_race(nr_threads >= max_threads))
2273		goto bad_fork_cleanup_count;
2274
2275	delayacct_tsk_init(p);	/* Must remain after dup_task_struct() */
2276	p->flags &= ~(PF_SUPERPRIV | PF_WQ_WORKER | PF_IDLE | PF_NO_SETAFFINITY);
2277	p->flags |= PF_FORKNOEXEC;
2278	INIT_LIST_HEAD(&p->children);
2279	INIT_LIST_HEAD(&p->sibling);
2280	rcu_copy_process(p);
2281	p->vfork_done = NULL;
2282	spin_lock_init(&p->alloc_lock);
2283
2284	init_sigpending(&p->pending);
2285
2286	p->utime = p->stime = p->gtime = 0;
2287#ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME
2288	p->utimescaled = p->stimescaled = 0;
2289#endif
2290	prev_cputime_init(&p->prev_cputime);
2291
2292#ifdef CONFIG_VIRT_CPU_ACCOUNTING_GEN
2293	seqcount_init(&p->vtime.seqcount);
2294	p->vtime.starttime = 0;
2295	p->vtime.state = VTIME_INACTIVE;
2296#endif
2297
2298#ifdef CONFIG_IO_URING
2299	p->io_uring = NULL;
2300#endif
2301
2302	p->default_timer_slack_ns = current->timer_slack_ns;
2303
2304#ifdef CONFIG_PSI
2305	p->psi_flags = 0;
2306#endif
2307
2308	task_io_accounting_init(&p->ioac);
2309	acct_clear_integrals(p);
2310
2311	posix_cputimers_init(&p->posix_cputimers);
2312
2313	p->io_context = NULL;
2314	audit_set_context(p, NULL);
2315	cgroup_fork(p);
2316	if (args->kthread) {
2317		if (!set_kthread_struct(p))
2318			goto bad_fork_cleanup_delayacct;
2319	}
2320#ifdef CONFIG_NUMA
2321	p->mempolicy = mpol_dup(p->mempolicy);
2322	if (IS_ERR(p->mempolicy)) {
2323		retval = PTR_ERR(p->mempolicy);
2324		p->mempolicy = NULL;
2325		goto bad_fork_cleanup_delayacct;
2326	}
2327#endif
2328#ifdef CONFIG_CPUSETS
2329	p->cpuset_mem_spread_rotor = NUMA_NO_NODE;
2330	p->cpuset_slab_spread_rotor = NUMA_NO_NODE;
2331	seqcount_spinlock_init(&p->mems_allowed_seq, &p->alloc_lock);
2332#endif
2333#ifdef CONFIG_TRACE_IRQFLAGS
2334	memset(&p->irqtrace, 0, sizeof(p->irqtrace));
2335	p->irqtrace.hardirq_disable_ip	= _THIS_IP_;
2336	p->irqtrace.softirq_enable_ip	= _THIS_IP_;
2337	p->softirqs_enabled		= 1;
2338	p->softirq_context		= 0;
2339#endif
2340
2341	p->pagefault_disabled = 0;
2342
2343#ifdef CONFIG_LOCKDEP
2344	lockdep_init_task(p);
2345#endif
2346
2347#ifdef CONFIG_DEBUG_MUTEXES
2348	p->blocked_on = NULL; /* not blocked yet */
2349#endif
2350#ifdef CONFIG_BCACHE
2351	p->sequential_io	= 0;
2352	p->sequential_io_avg	= 0;
2353#endif
2354#ifdef CONFIG_BPF_SYSCALL
2355	RCU_INIT_POINTER(p->bpf_storage, NULL);
2356	p->bpf_ctx = NULL;
2357#endif
2358
2359	/* Perform scheduler related setup. Assign this task to a CPU. */
2360	retval = sched_fork(clone_flags, p);
2361	if (retval)
2362		goto bad_fork_cleanup_policy;
2363
2364	retval = perf_event_init_task(p, clone_flags);
2365	if (retval)
2366		goto bad_fork_cleanup_policy;
2367	retval = audit_alloc(p);
2368	if (retval)
2369		goto bad_fork_cleanup_perf;
2370	/* copy all the process information */
2371	shm_init_task(p);
2372	retval = security_task_alloc(p, clone_flags);
2373	if (retval)
2374		goto bad_fork_cleanup_audit;
2375	retval = copy_semundo(clone_flags, p);
2376	if (retval)
2377		goto bad_fork_cleanup_security;
2378	retval = copy_files(clone_flags, p, args->no_files);
2379	if (retval)
2380		goto bad_fork_cleanup_semundo;
2381	retval = copy_fs(clone_flags, p);
2382	if (retval)
2383		goto bad_fork_cleanup_files;
2384	retval = copy_sighand(clone_flags, p);
2385	if (retval)
2386		goto bad_fork_cleanup_fs;
2387	retval = copy_signal(clone_flags, p);
2388	if (retval)
2389		goto bad_fork_cleanup_sighand;
2390	retval = copy_mm(clone_flags, p);
2391	if (retval)
2392		goto bad_fork_cleanup_signal;
2393	retval = copy_namespaces(clone_flags, p);
2394	if (retval)
2395		goto bad_fork_cleanup_mm;
2396	retval = copy_io(clone_flags, p);
2397	if (retval)
2398		goto bad_fork_cleanup_namespaces;
2399	retval = copy_thread(p, args);
2400	if (retval)
2401		goto bad_fork_cleanup_io;
2402
2403	stackleak_task_init(p);
2404
2405	if (pid != &init_struct_pid) {
2406		pid = alloc_pid(p->nsproxy->pid_ns_for_children, args->set_tid,
2407				args->set_tid_size);
2408		if (IS_ERR(pid)) {
2409			retval = PTR_ERR(pid);
2410			goto bad_fork_cleanup_thread;
2411		}
2412	}
2413
2414	/*
2415	 * This has to happen after we've potentially unshared the file
2416	 * descriptor table (so that the pidfd doesn't leak into the child
2417	 * if the fd table isn't shared).
2418	 */
2419	if (clone_flags & CLONE_PIDFD) {
2420		int flags = (clone_flags & CLONE_THREAD) ? PIDFD_THREAD : 0;
2421
2422		/* Note that no task has been attached to @pid yet. */
2423		retval = __pidfd_prepare(pid, flags, &pidfile);
2424		if (retval < 0)
2425			goto bad_fork_free_pid;
2426		pidfd = retval;
2427
2428		retval = put_user(pidfd, args->pidfd);
2429		if (retval)
2430			goto bad_fork_put_pidfd;
2431	}
2432
2433#ifdef CONFIG_BLOCK
2434	p->plug = NULL;
2435#endif
2436	futex_init_task(p);
2437
2438	/*
2439	 * sigaltstack should be cleared when sharing the same VM
2440	 */
2441	if ((clone_flags & (CLONE_VM|CLONE_VFORK)) == CLONE_VM)
2442		sas_ss_reset(p);
2443
2444	/*
2445	 * Syscall tracing and stepping should be turned off in the
2446	 * child regardless of CLONE_PTRACE.
2447	 */
2448	user_disable_single_step(p);
2449	clear_task_syscall_work(p, SYSCALL_TRACE);
2450#if defined(CONFIG_GENERIC_ENTRY) || defined(TIF_SYSCALL_EMU)
2451	clear_task_syscall_work(p, SYSCALL_EMU);
2452#endif
2453	clear_tsk_latency_tracing(p);
2454
2455	/* ok, now we should be set up.. */
2456	p->pid = pid_nr(pid);
2457	if (clone_flags & CLONE_THREAD) {
2458		p->group_leader = current->group_leader;
2459		p->tgid = current->tgid;
2460	} else {
2461		p->group_leader = p;
2462		p->tgid = p->pid;
2463	}
2464
2465	p->nr_dirtied = 0;
2466	p->nr_dirtied_pause = 128 >> (PAGE_SHIFT - 10);
2467	p->dirty_paused_when = 0;
2468
2469	p->pdeath_signal = 0;
2470	p->task_works = NULL;
2471	clear_posix_cputimers_work(p);
2472
2473#ifdef CONFIG_KRETPROBES
2474	p->kretprobe_instances.first = NULL;
2475#endif
2476#ifdef CONFIG_RETHOOK
2477	p->rethooks.first = NULL;
2478#endif
2479
2480	/*
2481	 * Ensure that the cgroup subsystem policies allow the new process to be
2482	 * forked. It should be noted that the new process's css_set can be changed
2483	 * between here and cgroup_post_fork() if an organisation operation is in
2484	 * progress.
2485	 */
2486	retval = cgroup_can_fork(p, args);
2487	if (retval)
2488		goto bad_fork_put_pidfd;
2489
2490	/*
2491	 * Now that the cgroups are pinned, re-clone the parent cgroup and put
2492	 * the new task on the correct runqueue. All this *before* the task
2493	 * becomes visible.
2494	 *
2495	 * This isn't part of ->can_fork() because while the re-cloning is
2496	 * cgroup specific, it unconditionally needs to place the task on a
2497	 * runqueue.
2498	 */
2499	sched_cgroup_fork(p, args);
2500
2501	/*
2502	 * From this point on we must avoid any synchronous user-space
2503	 * communication until we take the tasklist-lock. In particular, we do
2504	 * not want user-space to be able to predict the process start-time by
2505	 * stalling fork(2) after we recorded the start_time but before it is
2506	 * visible to the system.
2507	 */
2508
2509	p->start_time = ktime_get_ns();
2510	p->start_boottime = ktime_get_boottime_ns();
2511
2512	/*
2513	 * Make it visible to the rest of the system, but dont wake it up yet.
2514	 * Need tasklist lock for parent etc handling!
2515	 */
2516	write_lock_irq(&tasklist_lock);
2517
2518	/* CLONE_PARENT re-uses the old parent */
2519	if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) {
2520		p->real_parent = current->real_parent;
2521		p->parent_exec_id = current->parent_exec_id;
2522		if (clone_flags & CLONE_THREAD)
2523			p->exit_signal = -1;
2524		else
2525			p->exit_signal = current->group_leader->exit_signal;
2526	} else {
2527		p->real_parent = current;
2528		p->parent_exec_id = current->self_exec_id;
2529		p->exit_signal = args->exit_signal;
2530	}
2531
2532	klp_copy_process(p);
2533
2534	sched_core_fork(p);
2535
2536	spin_lock(&current->sighand->siglock);
2537
2538	rv_task_fork(p);
2539
2540	rseq_fork(p, clone_flags);
2541
2542	/* Don't start children in a dying pid namespace */
2543	if (unlikely(!(ns_of_pid(pid)->pid_allocated & PIDNS_ADDING))) {
2544		retval = -ENOMEM;
2545		goto bad_fork_cancel_cgroup;
2546	}
2547
2548	/* Let kill terminate clone/fork in the middle */
2549	if (fatal_signal_pending(current)) {
2550		retval = -EINTR;
2551		goto bad_fork_cancel_cgroup;
2552	}
2553
2554	/* No more failure paths after this point. */
2555
2556	/*
2557	 * Copy seccomp details explicitly here, in case they were changed
2558	 * before holding sighand lock.
2559	 */
2560	copy_seccomp(p);
2561
2562	init_task_pid_links(p);
2563	if (likely(p->pid)) {
2564		ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace);
2565
2566		init_task_pid(p, PIDTYPE_PID, pid);
2567		if (thread_group_leader(p)) {
2568			init_task_pid(p, PIDTYPE_TGID, pid);
2569			init_task_pid(p, PIDTYPE_PGID, task_pgrp(current));
2570			init_task_pid(p, PIDTYPE_SID, task_session(current));
2571
2572			if (is_child_reaper(pid)) {
2573				ns_of_pid(pid)->child_reaper = p;
2574				p->signal->flags |= SIGNAL_UNKILLABLE;
2575			}
2576			p->signal->shared_pending.signal = delayed.signal;
2577			p->signal->tty = tty_kref_get(current->signal->tty);
2578			/*
2579			 * Inherit has_child_subreaper flag under the same
2580			 * tasklist_lock with adding child to the process tree
2581			 * for propagate_has_child_subreaper optimization.
2582			 */
2583			p->signal->has_child_subreaper = p->real_parent->signal->has_child_subreaper ||
2584							 p->real_parent->signal->is_child_subreaper;
2585			list_add_tail(&p->sibling, &p->real_parent->children);
2586			list_add_tail_rcu(&p->tasks, &init_task.tasks);
2587			attach_pid(p, PIDTYPE_TGID);
2588			attach_pid(p, PIDTYPE_PGID);
2589			attach_pid(p, PIDTYPE_SID);
2590			__this_cpu_inc(process_counts);
2591		} else {
2592			current->signal->nr_threads++;
2593			current->signal->quick_threads++;
2594			atomic_inc(&current->signal->live);
2595			refcount_inc(&current->signal->sigcnt);
2596			task_join_group_stop(p);
2597			list_add_tail_rcu(&p->thread_node,
2598					  &p->signal->thread_head);
2599		}
2600		attach_pid(p, PIDTYPE_PID);
2601		nr_threads++;
2602	}
2603	total_forks++;
2604	hlist_del_init(&delayed.node);
2605	spin_unlock(&current->sighand->siglock);
2606	syscall_tracepoint_update(p);
2607	write_unlock_irq(&tasklist_lock);
2608
2609	if (pidfile)
2610		fd_install(pidfd, pidfile);
2611
2612	proc_fork_connector(p);
2613	sched_post_fork(p);
2614	cgroup_post_fork(p, args);
2615	perf_event_fork(p);
2616
2617	trace_task_newtask(p, clone_flags);
2618	uprobe_copy_process(p, clone_flags);
2619	user_events_fork(p, clone_flags);
2620
2621	copy_oom_score_adj(clone_flags, p);
2622
2623	return p;
2624
2625bad_fork_cancel_cgroup:
2626	sched_core_free(p);
2627	spin_unlock(&current->sighand->siglock);
2628	write_unlock_irq(&tasklist_lock);
2629	cgroup_cancel_fork(p, args);
2630bad_fork_put_pidfd:
2631	if (clone_flags & CLONE_PIDFD) {
2632		fput(pidfile);
2633		put_unused_fd(pidfd);
2634	}
2635bad_fork_free_pid:
2636	if (pid != &init_struct_pid)
2637		free_pid(pid);
2638bad_fork_cleanup_thread:
2639	exit_thread(p);
2640bad_fork_cleanup_io:
2641	if (p->io_context)
2642		exit_io_context(p);
2643bad_fork_cleanup_namespaces:
2644	exit_task_namespaces(p);
2645bad_fork_cleanup_mm:
2646	if (p->mm) {
2647		mm_clear_owner(p->mm, p);
2648		mmput(p->mm);
2649	}
2650bad_fork_cleanup_signal:
2651	if (!(clone_flags & CLONE_THREAD))
2652		free_signal_struct(p->signal);
2653bad_fork_cleanup_sighand:
2654	__cleanup_sighand(p->sighand);
2655bad_fork_cleanup_fs:
2656	exit_fs(p); /* blocking */
2657bad_fork_cleanup_files:
2658	exit_files(p); /* blocking */
2659bad_fork_cleanup_semundo:
2660	exit_sem(p);
2661bad_fork_cleanup_security:
2662	security_task_free(p);
2663bad_fork_cleanup_audit:
2664	audit_free(p);
2665bad_fork_cleanup_perf:
2666	perf_event_free_task(p);
2667bad_fork_cleanup_policy:
2668	lockdep_free_task(p);
2669#ifdef CONFIG_NUMA
2670	mpol_put(p->mempolicy);
2671#endif
2672bad_fork_cleanup_delayacct:
2673	delayacct_tsk_free(p);
2674bad_fork_cleanup_count:
2675	dec_rlimit_ucounts(task_ucounts(p), UCOUNT_RLIMIT_NPROC, 1);
2676	exit_creds(p);
2677bad_fork_free:
2678	WRITE_ONCE(p->__state, TASK_DEAD);
2679	exit_task_stack_account(p);
2680	put_task_stack(p);
2681	delayed_free_task(p);
2682fork_out:
2683	spin_lock_irq(&current->sighand->siglock);
2684	hlist_del_init(&delayed.node);
2685	spin_unlock_irq(&current->sighand->siglock);
2686	return ERR_PTR(retval);
2687}
2688
2689static inline void init_idle_pids(struct task_struct *idle)
2690{
2691	enum pid_type type;
2692
2693	for (type = PIDTYPE_PID; type < PIDTYPE_MAX; ++type) {
2694		INIT_HLIST_NODE(&idle->pid_links[type]); /* not really needed */
2695		init_task_pid(idle, type, &init_struct_pid);
2696	}
2697}
2698
2699static int idle_dummy(void *dummy)
2700{
2701	/* This function is never called */
2702	return 0;
2703}
2704
2705struct task_struct * __init fork_idle(int cpu)
2706{
2707	struct task_struct *task;
2708	struct kernel_clone_args args = {
2709		.flags		= CLONE_VM,
2710		.fn		= &idle_dummy,
2711		.fn_arg		= NULL,
2712		.kthread	= 1,
2713		.idle		= 1,
2714	};
2715
2716	task = copy_process(&init_struct_pid, 0, cpu_to_node(cpu), &args);
2717	if (!IS_ERR(task)) {
2718		init_idle_pids(task);
2719		init_idle(task, cpu);
2720	}
2721
2722	return task;
2723}
2724
2725/*
2726 * This is like kernel_clone(), but shaved down and tailored to just
2727 * creating io_uring workers. It returns a created task, or an error pointer.
2728 * The returned task is inactive, and the caller must fire it up through
2729 * wake_up_new_task(p). All signals are blocked in the created task.
2730 */
2731struct task_struct *create_io_thread(int (*fn)(void *), void *arg, int node)
2732{
2733	unsigned long flags = CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|
2734				CLONE_IO;
2735	struct kernel_clone_args args = {
2736		.flags		= ((lower_32_bits(flags) | CLONE_VM |
2737				    CLONE_UNTRACED) & ~CSIGNAL),
2738		.exit_signal	= (lower_32_bits(flags) & CSIGNAL),
2739		.fn		= fn,
2740		.fn_arg		= arg,
2741		.io_thread	= 1,
2742		.user_worker	= 1,
2743	};
2744
2745	return copy_process(NULL, 0, node, &args);
2746}
2747
2748/*
2749 *  Ok, this is the main fork-routine.
2750 *
2751 * It copies the process, and if successful kick-starts
2752 * it and waits for it to finish using the VM if required.
2753 *
2754 * args->exit_signal is expected to be checked for sanity by the caller.
2755 */
2756pid_t kernel_clone(struct kernel_clone_args *args)
2757{
2758	u64 clone_flags = args->flags;
2759	struct completion vfork;
2760	struct pid *pid;
2761	struct task_struct *p;
2762	int trace = 0;
2763	pid_t nr;
2764
2765	/*
2766	 * For legacy clone() calls, CLONE_PIDFD uses the parent_tid argument
2767	 * to return the pidfd. Hence, CLONE_PIDFD and CLONE_PARENT_SETTID are
2768	 * mutually exclusive. With clone3() CLONE_PIDFD has grown a separate
2769	 * field in struct clone_args and it still doesn't make sense to have
2770	 * them both point at the same memory location. Performing this check
2771	 * here has the advantage that we don't need to have a separate helper
2772	 * to check for legacy clone().
2773	 */
2774	if ((clone_flags & CLONE_PIDFD) &&
2775	    (clone_flags & CLONE_PARENT_SETTID) &&
2776	    (args->pidfd == args->parent_tid))
2777		return -EINVAL;
2778
2779	/*
2780	 * Determine whether and which event to report to ptracer.  When
2781	 * called from kernel_thread or CLONE_UNTRACED is explicitly
2782	 * requested, no event is reported; otherwise, report if the event
2783	 * for the type of forking is enabled.
2784	 */
2785	if (!(clone_flags & CLONE_UNTRACED)) {
2786		if (clone_flags & CLONE_VFORK)
2787			trace = PTRACE_EVENT_VFORK;
2788		else if (args->exit_signal != SIGCHLD)
2789			trace = PTRACE_EVENT_CLONE;
2790		else
2791			trace = PTRACE_EVENT_FORK;
2792
2793		if (likely(!ptrace_event_enabled(current, trace)))
2794			trace = 0;
2795	}
2796
2797	p = copy_process(NULL, trace, NUMA_NO_NODE, args);
2798	add_latent_entropy();
2799
2800	if (IS_ERR(p))
2801		return PTR_ERR(p);
2802
2803	/*
2804	 * Do this prior waking up the new thread - the thread pointer
2805	 * might get invalid after that point, if the thread exits quickly.
2806	 */
2807	trace_sched_process_fork(current, p);
2808
2809	pid = get_task_pid(p, PIDTYPE_PID);
2810	nr = pid_vnr(pid);
2811
2812	if (clone_flags & CLONE_PARENT_SETTID)
2813		put_user(nr, args->parent_tid);
2814
2815	if (clone_flags & CLONE_VFORK) {
2816		p->vfork_done = &vfork;
2817		init_completion(&vfork);
2818		get_task_struct(p);
2819	}
2820
2821	if (IS_ENABLED(CONFIG_LRU_GEN_WALKS_MMU) && !(clone_flags & CLONE_VM)) {
2822		/* lock the task to synchronize with memcg migration */
2823		task_lock(p);
2824		lru_gen_add_mm(p->mm);
2825		task_unlock(p);
2826	}
2827
2828	wake_up_new_task(p);
2829
2830	/* forking complete and child started to run, tell ptracer */
2831	if (unlikely(trace))
2832		ptrace_event_pid(trace, pid);
2833
2834	if (clone_flags & CLONE_VFORK) {
2835		if (!wait_for_vfork_done(p, &vfork))
2836			ptrace_event_pid(PTRACE_EVENT_VFORK_DONE, pid);
2837	}
2838
2839	put_pid(pid);
2840	return nr;
2841}
2842
2843/*
2844 * Create a kernel thread.
2845 */
2846pid_t kernel_thread(int (*fn)(void *), void *arg, const char *name,
2847		    unsigned long flags)
2848{
2849	struct kernel_clone_args args = {
2850		.flags		= ((lower_32_bits(flags) | CLONE_VM |
2851				    CLONE_UNTRACED) & ~CSIGNAL),
2852		.exit_signal	= (lower_32_bits(flags) & CSIGNAL),
2853		.fn		= fn,
2854		.fn_arg		= arg,
2855		.name		= name,
2856		.kthread	= 1,
2857	};
2858
2859	return kernel_clone(&args);
2860}
2861
2862/*
2863 * Create a user mode thread.
2864 */
2865pid_t user_mode_thread(int (*fn)(void *), void *arg, unsigned long flags)
2866{
2867	struct kernel_clone_args args = {
2868		.flags		= ((lower_32_bits(flags) | CLONE_VM |
2869				    CLONE_UNTRACED) & ~CSIGNAL),
2870		.exit_signal	= (lower_32_bits(flags) & CSIGNAL),
2871		.fn		= fn,
2872		.fn_arg		= arg,
2873	};
2874
2875	return kernel_clone(&args);
2876}
2877
2878#ifdef __ARCH_WANT_SYS_FORK
2879SYSCALL_DEFINE0(fork)
2880{
2881#ifdef CONFIG_MMU
2882	struct kernel_clone_args args = {
2883		.exit_signal = SIGCHLD,
2884	};
2885
2886	return kernel_clone(&args);
2887#else
2888	/* can not support in nommu mode */
2889	return -EINVAL;
2890#endif
2891}
2892#endif
2893
2894#ifdef __ARCH_WANT_SYS_VFORK
2895SYSCALL_DEFINE0(vfork)
2896{
2897	struct kernel_clone_args args = {
2898		.flags		= CLONE_VFORK | CLONE_VM,
2899		.exit_signal	= SIGCHLD,
2900	};
2901
2902	return kernel_clone(&args);
2903}
2904#endif
2905
2906#ifdef __ARCH_WANT_SYS_CLONE
2907#ifdef CONFIG_CLONE_BACKWARDS
2908SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp,
2909		 int __user *, parent_tidptr,
2910		 unsigned long, tls,
2911		 int __user *, child_tidptr)
2912#elif defined(CONFIG_CLONE_BACKWARDS2)
2913SYSCALL_DEFINE5(clone, unsigned long, newsp, unsigned long, clone_flags,
2914		 int __user *, parent_tidptr,
2915		 int __user *, child_tidptr,
2916		 unsigned long, tls)
2917#elif defined(CONFIG_CLONE_BACKWARDS3)
2918SYSCALL_DEFINE6(clone, unsigned long, clone_flags, unsigned long, newsp,
2919		int, stack_size,
2920		int __user *, parent_tidptr,
2921		int __user *, child_tidptr,
2922		unsigned long, tls)
2923#else
2924SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp,
2925		 int __user *, parent_tidptr,
2926		 int __user *, child_tidptr,
2927		 unsigned long, tls)
2928#endif
2929{
2930	struct kernel_clone_args args = {
2931		.flags		= (lower_32_bits(clone_flags) & ~CSIGNAL),
2932		.pidfd		= parent_tidptr,
2933		.child_tid	= child_tidptr,
2934		.parent_tid	= parent_tidptr,
2935		.exit_signal	= (lower_32_bits(clone_flags) & CSIGNAL),
2936		.stack		= newsp,
2937		.tls		= tls,
2938	};
2939
2940	return kernel_clone(&args);
2941}
2942#endif
2943
2944#ifdef __ARCH_WANT_SYS_CLONE3
2945
2946noinline static int copy_clone_args_from_user(struct kernel_clone_args *kargs,
2947					      struct clone_args __user *uargs,
2948					      size_t usize)
2949{
2950	int err;
2951	struct clone_args args;
2952	pid_t *kset_tid = kargs->set_tid;
2953
2954	BUILD_BUG_ON(offsetofend(struct clone_args, tls) !=
2955		     CLONE_ARGS_SIZE_VER0);
2956	BUILD_BUG_ON(offsetofend(struct clone_args, set_tid_size) !=
2957		     CLONE_ARGS_SIZE_VER1);
2958	BUILD_BUG_ON(offsetofend(struct clone_args, cgroup) !=
2959		     CLONE_ARGS_SIZE_VER2);
2960	BUILD_BUG_ON(sizeof(struct clone_args) != CLONE_ARGS_SIZE_VER2);
2961
2962	if (unlikely(usize > PAGE_SIZE))
2963		return -E2BIG;
2964	if (unlikely(usize < CLONE_ARGS_SIZE_VER0))
2965		return -EINVAL;
2966
2967	err = copy_struct_from_user(&args, sizeof(args), uargs, usize);
2968	if (err)
2969		return err;
2970
2971	if (unlikely(args.set_tid_size > MAX_PID_NS_LEVEL))
2972		return -EINVAL;
2973
2974	if (unlikely(!args.set_tid && args.set_tid_size > 0))
2975		return -EINVAL;
2976
2977	if (unlikely(args.set_tid && args.set_tid_size == 0))
2978		return -EINVAL;
2979
2980	/*
2981	 * Verify that higher 32bits of exit_signal are unset and that
2982	 * it is a valid signal
2983	 */
2984	if (unlikely((args.exit_signal & ~((u64)CSIGNAL)) ||
2985		     !valid_signal(args.exit_signal)))
2986		return -EINVAL;
2987
2988	if ((args.flags & CLONE_INTO_CGROUP) &&
2989	    (args.cgroup > INT_MAX || usize < CLONE_ARGS_SIZE_VER2))
2990		return -EINVAL;
2991
2992	*kargs = (struct kernel_clone_args){
2993		.flags		= args.flags,
2994		.pidfd		= u64_to_user_ptr(args.pidfd),
2995		.child_tid	= u64_to_user_ptr(args.child_tid),
2996		.parent_tid	= u64_to_user_ptr(args.parent_tid),
2997		.exit_signal	= args.exit_signal,
2998		.stack		= args.stack,
2999		.stack_size	= args.stack_size,
3000		.tls		= args.tls,
3001		.set_tid_size	= args.set_tid_size,
3002		.cgroup		= args.cgroup,
3003	};
3004
3005	if (args.set_tid &&
3006		copy_from_user(kset_tid, u64_to_user_ptr(args.set_tid),
3007			(kargs->set_tid_size * sizeof(pid_t))))
3008		return -EFAULT;
3009
3010	kargs->set_tid = kset_tid;
3011
3012	return 0;
3013}
3014
3015/**
3016 * clone3_stack_valid - check and prepare stack
3017 * @kargs: kernel clone args
3018 *
3019 * Verify that the stack arguments userspace gave us are sane.
3020 * In addition, set the stack direction for userspace since it's easy for us to
3021 * determine.
3022 */
3023static inline bool clone3_stack_valid(struct kernel_clone_args *kargs)
3024{
3025	if (kargs->stack == 0) {
3026		if (kargs->stack_size > 0)
3027			return false;
3028	} else {
3029		if (kargs->stack_size == 0)
3030			return false;
3031
3032		if (!access_ok((void __user *)kargs->stack, kargs->stack_size))
3033			return false;
3034
3035#if !defined(CONFIG_STACK_GROWSUP)
3036		kargs->stack += kargs->stack_size;
3037#endif
3038	}
3039
3040	return true;
3041}
3042
3043static bool clone3_args_valid(struct kernel_clone_args *kargs)
3044{
3045	/* Verify that no unknown flags are passed along. */
3046	if (kargs->flags &
3047	    ~(CLONE_LEGACY_FLAGS | CLONE_CLEAR_SIGHAND | CLONE_INTO_CGROUP))
3048		return false;
3049
3050	/*
3051	 * - make the CLONE_DETACHED bit reusable for clone3
3052	 * - make the CSIGNAL bits reusable for clone3
3053	 */
3054	if (kargs->flags & (CLONE_DETACHED | (CSIGNAL & (~CLONE_NEWTIME))))
3055		return false;
3056
3057	if ((kargs->flags & (CLONE_SIGHAND | CLONE_CLEAR_SIGHAND)) ==
3058	    (CLONE_SIGHAND | CLONE_CLEAR_SIGHAND))
3059		return false;
3060
3061	if ((kargs->flags & (CLONE_THREAD | CLONE_PARENT)) &&
3062	    kargs->exit_signal)
3063		return false;
3064
3065	if (!clone3_stack_valid(kargs))
3066		return false;
3067
3068	return true;
3069}
3070
3071/**
3072 * sys_clone3 - create a new process with specific properties
3073 * @uargs: argument structure
3074 * @size:  size of @uargs
3075 *
3076 * clone3() is the extensible successor to clone()/clone2().
3077 * It takes a struct as argument that is versioned by its size.
3078 *
3079 * Return: On success, a positive PID for the child process.
3080 *         On error, a negative errno number.
3081 */
3082SYSCALL_DEFINE2(clone3, struct clone_args __user *, uargs, size_t, size)
3083{
3084	int err;
3085
3086	struct kernel_clone_args kargs;
3087	pid_t set_tid[MAX_PID_NS_LEVEL];
3088
3089	kargs.set_tid = set_tid;
3090
3091	err = copy_clone_args_from_user(&kargs, uargs, size);
3092	if (err)
3093		return err;
3094
3095	if (!clone3_args_valid(&kargs))
3096		return -EINVAL;
3097
3098	return kernel_clone(&kargs);
3099}
3100#endif
3101
3102void walk_process_tree(struct task_struct *top, proc_visitor visitor, void *data)
3103{
3104	struct task_struct *leader, *parent, *child;
3105	int res;
3106
3107	read_lock(&tasklist_lock);
3108	leader = top = top->group_leader;
3109down:
3110	for_each_thread(leader, parent) {
3111		list_for_each_entry(child, &parent->children, sibling) {
3112			res = visitor(child, data);
3113			if (res) {
3114				if (res < 0)
3115					goto out;
3116				leader = child;
3117				goto down;
3118			}
3119up:
3120			;
3121		}
3122	}
3123
3124	if (leader != top) {
3125		child = leader;
3126		parent = child->real_parent;
3127		leader = parent->group_leader;
3128		goto up;
3129	}
3130out:
3131	read_unlock(&tasklist_lock);
3132}
3133
3134#ifndef ARCH_MIN_MMSTRUCT_ALIGN
3135#define ARCH_MIN_MMSTRUCT_ALIGN 0
3136#endif
3137
3138static void sighand_ctor(void *data)
3139{
3140	struct sighand_struct *sighand = data;
3141
3142	spin_lock_init(&sighand->siglock);
3143	init_waitqueue_head(&sighand->signalfd_wqh);
3144}
3145
3146void __init mm_cache_init(void)
3147{
3148	unsigned int mm_size;
3149
3150	/*
3151	 * The mm_cpumask is located at the end of mm_struct, and is
3152	 * dynamically sized based on the maximum CPU number this system
3153	 * can have, taking hotplug into account (nr_cpu_ids).
3154	 */
3155	mm_size = sizeof(struct mm_struct) + cpumask_size() + mm_cid_size();
3156
3157	mm_cachep = kmem_cache_create_usercopy("mm_struct",
3158			mm_size, ARCH_MIN_MMSTRUCT_ALIGN,
3159			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
3160			offsetof(struct mm_struct, saved_auxv),
3161			sizeof_field(struct mm_struct, saved_auxv),
3162			NULL);
3163}
3164
3165void __init proc_caches_init(void)
3166{
3167	sighand_cachep = kmem_cache_create("sighand_cache",
3168			sizeof(struct sighand_struct), 0,
3169			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_TYPESAFE_BY_RCU|
3170			SLAB_ACCOUNT, sighand_ctor);
3171	signal_cachep = kmem_cache_create("signal_cache",
3172			sizeof(struct signal_struct), 0,
3173			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
3174			NULL);
3175	files_cachep = kmem_cache_create("files_cache",
3176			sizeof(struct files_struct), 0,
3177			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
3178			NULL);
3179	fs_cachep = kmem_cache_create("fs_cache",
3180			sizeof(struct fs_struct), 0,
3181			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_ACCOUNT,
3182			NULL);
3183
3184	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC|SLAB_ACCOUNT);
3185#ifdef CONFIG_PER_VMA_LOCK
3186	vma_lock_cachep = KMEM_CACHE(vma_lock, SLAB_PANIC|SLAB_ACCOUNT);
3187#endif
3188	mmap_init();
3189	nsproxy_cache_init();
3190}
3191
3192/*
3193 * Check constraints on flags passed to the unshare system call.
3194 */
3195static int check_unshare_flags(unsigned long unshare_flags)
3196{
3197	if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND|
3198				CLONE_VM|CLONE_FILES|CLONE_SYSVSEM|
3199				CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET|
3200				CLONE_NEWUSER|CLONE_NEWPID|CLONE_NEWCGROUP|
3201				CLONE_NEWTIME))
3202		return -EINVAL;
3203	/*
3204	 * Not implemented, but pretend it works if there is nothing
3205	 * to unshare.  Note that unsharing the address space or the
3206	 * signal handlers also need to unshare the signal queues (aka
3207	 * CLONE_THREAD).
3208	 */
3209	if (unshare_flags & (CLONE_THREAD | CLONE_SIGHAND | CLONE_VM)) {
3210		if (!thread_group_empty(current))
3211			return -EINVAL;
3212	}
3213	if (unshare_flags & (CLONE_SIGHAND | CLONE_VM)) {
3214		if (refcount_read(&current->sighand->count) > 1)
3215			return -EINVAL;
3216	}
3217	if (unshare_flags & CLONE_VM) {
3218		if (!current_is_single_threaded())
3219			return -EINVAL;
3220	}
3221
3222	return 0;
3223}
3224
3225/*
3226 * Unshare the filesystem structure if it is being shared
3227 */
3228static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp)
3229{
3230	struct fs_struct *fs = current->fs;
3231
3232	if (!(unshare_flags & CLONE_FS) || !fs)
3233		return 0;
3234
3235	/* don't need lock here; in the worst case we'll do useless copy */
3236	if (fs->users == 1)
3237		return 0;
3238
3239	*new_fsp = copy_fs_struct(fs);
3240	if (!*new_fsp)
3241		return -ENOMEM;
3242
3243	return 0;
3244}
3245
3246/*
3247 * Unshare file descriptor table if it is being shared
3248 */
3249int unshare_fd(unsigned long unshare_flags, unsigned int max_fds,
3250	       struct files_struct **new_fdp)
3251{
3252	struct files_struct *fd = current->files;
3253	int error = 0;
3254
3255	if ((unshare_flags & CLONE_FILES) &&
3256	    (fd && atomic_read(&fd->count) > 1)) {
3257		*new_fdp = dup_fd(fd, max_fds, &error);
3258		if (!*new_fdp)
3259			return error;
3260	}
3261
3262	return 0;
3263}
3264
3265/*
3266 * unshare allows a process to 'unshare' part of the process
3267 * context which was originally shared using clone.  copy_*
3268 * functions used by kernel_clone() cannot be used here directly
3269 * because they modify an inactive task_struct that is being
3270 * constructed. Here we are modifying the current, active,
3271 * task_struct.
3272 */
3273int ksys_unshare(unsigned long unshare_flags)
3274{
3275	struct fs_struct *fs, *new_fs = NULL;
3276	struct files_struct *new_fd = NULL;
3277	struct cred *new_cred = NULL;
3278	struct nsproxy *new_nsproxy = NULL;
3279	int do_sysvsem = 0;
3280	int err;
3281
3282	/*
3283	 * If unsharing a user namespace must also unshare the thread group
3284	 * and unshare the filesystem root and working directories.
3285	 */
3286	if (unshare_flags & CLONE_NEWUSER)
3287		unshare_flags |= CLONE_THREAD | CLONE_FS;
3288	/*
3289	 * If unsharing vm, must also unshare signal handlers.
3290	 */
3291	if (unshare_flags & CLONE_VM)
3292		unshare_flags |= CLONE_SIGHAND;
3293	/*
3294	 * If unsharing a signal handlers, must also unshare the signal queues.
3295	 */
3296	if (unshare_flags & CLONE_SIGHAND)
3297		unshare_flags |= CLONE_THREAD;
3298	/*
3299	 * If unsharing namespace, must also unshare filesystem information.
3300	 */
3301	if (unshare_flags & CLONE_NEWNS)
3302		unshare_flags |= CLONE_FS;
3303
3304	err = check_unshare_flags(unshare_flags);
3305	if (err)
3306		goto bad_unshare_out;
3307	/*
3308	 * CLONE_NEWIPC must also detach from the undolist: after switching
3309	 * to a new ipc namespace, the semaphore arrays from the old
3310	 * namespace are unreachable.
3311	 */
3312	if (unshare_flags & (CLONE_NEWIPC|CLONE_SYSVSEM))
3313		do_sysvsem = 1;
3314	err = unshare_fs(unshare_flags, &new_fs);
3315	if (err)
3316		goto bad_unshare_out;
3317	err = unshare_fd(unshare_flags, NR_OPEN_MAX, &new_fd);
3318	if (err)
3319		goto bad_unshare_cleanup_fs;
3320	err = unshare_userns(unshare_flags, &new_cred);
3321	if (err)
3322		goto bad_unshare_cleanup_fd;
3323	err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy,
3324					 new_cred, new_fs);
3325	if (err)
3326		goto bad_unshare_cleanup_cred;
3327
3328	if (new_cred) {
3329		err = set_cred_ucounts(new_cred);
3330		if (err)
3331			goto bad_unshare_cleanup_cred;
3332	}
3333
3334	if (new_fs || new_fd || do_sysvsem || new_cred || new_nsproxy) {
3335		if (do_sysvsem) {
3336			/*
3337			 * CLONE_SYSVSEM is equivalent to sys_exit().
3338			 */
3339			exit_sem(current);
3340		}
3341		if (unshare_flags & CLONE_NEWIPC) {
3342			/* Orphan segments in old ns (see sem above). */
3343			exit_shm(current);
3344			shm_init_task(current);
3345		}
3346
3347		if (new_nsproxy)
3348			switch_task_namespaces(current, new_nsproxy);
3349
3350		task_lock(current);
3351
3352		if (new_fs) {
3353			fs = current->fs;
3354			spin_lock(&fs->lock);
3355			current->fs = new_fs;
3356			if (--fs->users)
3357				new_fs = NULL;
3358			else
3359				new_fs = fs;
3360			spin_unlock(&fs->lock);
3361		}
3362
3363		if (new_fd)
3364			swap(current->files, new_fd);
3365
3366		task_unlock(current);
3367
3368		if (new_cred) {
3369			/* Install the new user namespace */
3370			commit_creds(new_cred);
3371			new_cred = NULL;
3372		}
3373	}
3374
3375	perf_event_namespaces(current);
3376
3377bad_unshare_cleanup_cred:
3378	if (new_cred)
3379		put_cred(new_cred);
3380bad_unshare_cleanup_fd:
3381	if (new_fd)
3382		put_files_struct(new_fd);
3383
3384bad_unshare_cleanup_fs:
3385	if (new_fs)
3386		free_fs_struct(new_fs);
3387
3388bad_unshare_out:
3389	return err;
3390}
3391
3392SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags)
3393{
3394	return ksys_unshare(unshare_flags);
3395}
3396
3397/*
3398 *	Helper to unshare the files of the current task.
3399 *	We don't want to expose copy_files internals to
3400 *	the exec layer of the kernel.
3401 */
3402
3403int unshare_files(void)
3404{
3405	struct task_struct *task = current;
3406	struct files_struct *old, *copy = NULL;
3407	int error;
3408
3409	error = unshare_fd(CLONE_FILES, NR_OPEN_MAX, &copy);
3410	if (error || !copy)
3411		return error;
3412
3413	old = task->files;
3414	task_lock(task);
3415	task->files = copy;
3416	task_unlock(task);
3417	put_files_struct(old);
3418	return 0;
3419}
3420
3421int sysctl_max_threads(struct ctl_table *table, int write,
3422		       void *buffer, size_t *lenp, loff_t *ppos)
3423{
3424	struct ctl_table t;
3425	int ret;
3426	int threads = max_threads;
3427	int min = 1;
3428	int max = MAX_THREADS;
3429
3430	t = *table;
3431	t.data = &threads;
3432	t.extra1 = &min;
3433	t.extra2 = &max;
3434
3435	ret = proc_dointvec_minmax(&t, write, buffer, lenp, ppos);
3436	if (ret || !write)
3437		return ret;
3438
3439	max_threads = threads;
3440
3441	return 0;
3442}
   1/*
   2 *  linux/kernel/fork.c
   3 *
   4 *  Copyright (C) 1991, 1992  Linus Torvalds
   5 */
   6
   7/*
   8 *  'fork.c' contains the help-routines for the 'fork' system call
   9 * (see also entry.S and others).
  10 * Fork is rather simple, once you get the hang of it, but the memory
  11 * management can be a bitch. See 'mm/memory.c': 'copy_page_range()'
  12 */
  13
  14#include <linux/slab.h>
  15#include <linux/init.h>
  16#include <linux/unistd.h>
  17#include <linux/module.h>
  18#include <linux/vmalloc.h>
  19#include <linux/completion.h>
  20#include <linux/personality.h>
  21#include <linux/mempolicy.h>
  22#include <linux/sem.h>
  23#include <linux/file.h>
  24#include <linux/fdtable.h>
  25#include <linux/iocontext.h>
  26#include <linux/key.h>
  27#include <linux/binfmts.h>
  28#include <linux/mman.h>
  29#include <linux/mmu_notifier.h>
  30#include <linux/fs.h>
  31#include <linux/nsproxy.h>
  32#include <linux/capability.h>
  33#include <linux/cpu.h>
  34#include <linux/cgroup.h>
  35#include <linux/security.h>
  36#include <linux/hugetlb.h>
  37#include <linux/seccomp.h>
  38#include <linux/swap.h>
  39#include <linux/syscalls.h>
  40#include <linux/jiffies.h>
  41#include <linux/futex.h>
  42#include <linux/compat.h>
  43#include <linux/kthread.h>
  44#include <linux/task_io_accounting_ops.h>
  45#include <linux/rcupdate.h>
  46#include <linux/ptrace.h>
  47#include <linux/mount.h>
  48#include <linux/audit.h>
  49#include <linux/memcontrol.h>
  50#include <linux/ftrace.h>
  51#include <linux/proc_fs.h>
  52#include <linux/profile.h>
  53#include <linux/rmap.h>
  54#include <linux/ksm.h>
  55#include <linux/acct.h>
  56#include <linux/tsacct_kern.h>
  57#include <linux/cn_proc.h>
  58#include <linux/freezer.h>
  59#include <linux/delayacct.h>
  60#include <linux/taskstats_kern.h>
  61#include <linux/random.h>
  62#include <linux/tty.h>
  63#include <linux/blkdev.h>
  64#include <linux/fs_struct.h>
  65#include <linux/magic.h>
  66#include <linux/perf_event.h>
  67#include <linux/posix-timers.h>
  68#include <linux/user-return-notifier.h>
  69#include <linux/oom.h>
  70#include <linux/khugepaged.h>
  71#include <linux/signalfd.h>
  72#include <linux/uprobes.h>
  73
  74#include <asm/pgtable.h>
  75#include <asm/pgalloc.h>
  76#include <asm/uaccess.h>
  77#include <asm/mmu_context.h>
  78#include <asm/cacheflush.h>
  79#include <asm/tlbflush.h>
  80
  81#include <trace/events/sched.h>
  82
  83#define CREATE_TRACE_POINTS
  84#include <trace/events/task.h>
  85
  86/*
  87 * Protected counters by write_lock_irq(&tasklist_lock)
  88 */
  89unsigned long total_forks;	/* Handle normal Linux uptimes. */
  90int nr_threads;			/* The idle threads do not count.. */
  91
  92int max_threads;		/* tunable limit on nr_threads */
  93
  94DEFINE_PER_CPU(unsigned long, process_counts) = 0;
  95
  96__cacheline_aligned DEFINE_RWLOCK(tasklist_lock);  /* outer */
  97
  98#ifdef CONFIG_PROVE_RCU
  99int lockdep_tasklist_lock_is_held(void)
 100{
 101	return lockdep_is_held(&tasklist_lock);
 102}
 103EXPORT_SYMBOL_GPL(lockdep_tasklist_lock_is_held);
 104#endif /* #ifdef CONFIG_PROVE_RCU */
 105
 106int nr_processes(void)
 107{
 108	int cpu;
 109	int total = 0;
 110
 111	for_each_possible_cpu(cpu)
 112		total += per_cpu(process_counts, cpu);
 113
 114	return total;
 115}
 116
 117#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
 118static struct kmem_cache *task_struct_cachep;
 119
 120static inline struct task_struct *alloc_task_struct_node(int node)
 121{
 122	return kmem_cache_alloc_node(task_struct_cachep, GFP_KERNEL, node);
 123}
 124
 125void __weak arch_release_task_struct(struct task_struct *tsk) { }
 126
 127static inline void free_task_struct(struct task_struct *tsk)
 128{
 129	arch_release_task_struct(tsk);
 130	kmem_cache_free(task_struct_cachep, tsk);
 131}
 132#endif
 133
 134#ifndef CONFIG_ARCH_THREAD_INFO_ALLOCATOR
 135void __weak arch_release_thread_info(struct thread_info *ti) { }
 136
 137/*
 138 * Allocate pages if THREAD_SIZE is >= PAGE_SIZE, otherwise use a
 139 * kmemcache based allocator.
 140 */
 141# if THREAD_SIZE >= PAGE_SIZE
 142static struct thread_info *alloc_thread_info_node(struct task_struct *tsk,
 143						  int node)
 144{
 145	struct page *page = alloc_pages_node(node, THREADINFO_GFP,
 146					     THREAD_SIZE_ORDER);
 147
 148	return page ? page_address(page) : NULL;
 149}
 150
 151static inline void free_thread_info(struct thread_info *ti)
 152{
 153	arch_release_thread_info(ti);
 154	free_pages((unsigned long)ti, THREAD_SIZE_ORDER);
 155}
 156# else
 157static struct kmem_cache *thread_info_cache;
 158
 159static struct thread_info *alloc_thread_info_node(struct task_struct *tsk,
 160						  int node)
 161{
 162	return kmem_cache_alloc_node(thread_info_cache, THREADINFO_GFP, node);
 163}
 164
 165static void free_thread_info(struct thread_info *ti)
 166{
 167	arch_release_thread_info(ti);
 168	kmem_cache_free(thread_info_cache, ti);
 169}
 170
 171void thread_info_cache_init(void)
 172{
 173	thread_info_cache = kmem_cache_create("thread_info", THREAD_SIZE,
 174					      THREAD_SIZE, 0, NULL);
 175	BUG_ON(thread_info_cache == NULL);
 176}
 177# endif
 178#endif
 179
 180/* SLAB cache for signal_struct structures (tsk->signal) */
 181static struct kmem_cache *signal_cachep;
 182
 183/* SLAB cache for sighand_struct structures (tsk->sighand) */
 184struct kmem_cache *sighand_cachep;
 185
 186/* SLAB cache for files_struct structures (tsk->files) */
 187struct kmem_cache *files_cachep;
 188
 189/* SLAB cache for fs_struct structures (tsk->fs) */
 190struct kmem_cache *fs_cachep;
 191
 192/* SLAB cache for vm_area_struct structures */
 193struct kmem_cache *vm_area_cachep;
 194
 195/* SLAB cache for mm_struct structures (tsk->mm) */
 196static struct kmem_cache *mm_cachep;
 197
 198static void account_kernel_stack(struct thread_info *ti, int account)
 199{
 200	struct zone *zone = page_zone(virt_to_page(ti));
 201
 202	mod_zone_page_state(zone, NR_KERNEL_STACK, account);
 203}
 204
 205void free_task(struct task_struct *tsk)
 206{
 207	account_kernel_stack(tsk->stack, -1);
 208	free_thread_info(tsk->stack);
 209	rt_mutex_debug_task_free(tsk);
 210	ftrace_graph_exit_task(tsk);
 211	put_seccomp_filter(tsk);
 212	free_task_struct(tsk);
 213}
 214EXPORT_SYMBOL(free_task);
 215
 216static inline void free_signal_struct(struct signal_struct *sig)
 217{
 218	taskstats_tgid_free(sig);
 219	sched_autogroup_exit(sig);
 220	kmem_cache_free(signal_cachep, sig);
 221}
 222
 223static inline void put_signal_struct(struct signal_struct *sig)
 224{
 225	if (atomic_dec_and_test(&sig->sigcnt))
 226		free_signal_struct(sig);
 227}
 228
 229void __put_task_struct(struct task_struct *tsk)
 230{
 231	WARN_ON(!tsk->exit_state);
 232	WARN_ON(atomic_read(&tsk->usage));
 233	WARN_ON(tsk == current);
 234
 235	security_task_free(tsk);
 236	exit_creds(tsk);
 237	delayacct_tsk_free(tsk);
 238	put_signal_struct(tsk->signal);
 239
 240	if (!profile_handoff_task(tsk))
 241		free_task(tsk);
 242}
 243EXPORT_SYMBOL_GPL(__put_task_struct);
 244
 245void __init __weak arch_task_cache_init(void) { }
 246
 247void __init fork_init(unsigned long mempages)
 248{
 249#ifndef CONFIG_ARCH_TASK_STRUCT_ALLOCATOR
 250#ifndef ARCH_MIN_TASKALIGN
 251#define ARCH_MIN_TASKALIGN	L1_CACHE_BYTES
 252#endif
 253	/* create a slab on which task_structs can be allocated */
 254	task_struct_cachep =
 255		kmem_cache_create("task_struct", sizeof(struct task_struct),
 256			ARCH_MIN_TASKALIGN, SLAB_PANIC | SLAB_NOTRACK, NULL);
 257#endif
 258
 259	/* do the arch specific task caches init */
 260	arch_task_cache_init();
 261
 262	/*
 263	 * The default maximum number of threads is set to a safe
 264	 * value: the thread structures can take up at most half
 265	 * of memory.
 266	 */
 267	max_threads = mempages / (8 * THREAD_SIZE / PAGE_SIZE);
 268
 269	/*
 270	 * we need to allow at least 20 threads to boot a system
 271	 */
 272	if (max_threads < 20)
 273		max_threads = 20;
 274
 275	init_task.signal->rlim[RLIMIT_NPROC].rlim_cur = max_threads/2;
 276	init_task.signal->rlim[RLIMIT_NPROC].rlim_max = max_threads/2;
 277	init_task.signal->rlim[RLIMIT_SIGPENDING] =
 278		init_task.signal->rlim[RLIMIT_NPROC];
 279}
 280
 281int __attribute__((weak)) arch_dup_task_struct(struct task_struct *dst,
 282					       struct task_struct *src)
 283{
 284	*dst = *src;
 285	return 0;
 286}
 287
 288static struct task_struct *dup_task_struct(struct task_struct *orig)
 289{
 290	struct task_struct *tsk;
 291	struct thread_info *ti;
 292	unsigned long *stackend;
 293	int node = tsk_fork_get_node(orig);
 294	int err;
 295
 296	tsk = alloc_task_struct_node(node);
 297	if (!tsk)
 298		return NULL;
 299
 300	ti = alloc_thread_info_node(tsk, node);
 301	if (!ti) {
 302		free_task_struct(tsk);
 303		return NULL;
 304	}
 305
 306	err = arch_dup_task_struct(tsk, orig);
 307
 308	/*
 309	 * We defer looking at err, because we will need this setup
 310	 * for the clean up path to work correctly.
 311	 */
 312	tsk->stack = ti;
 313	setup_thread_stack(tsk, orig);
 314
 315	if (err)
 316		goto out;
 317
 318	clear_user_return_notifier(tsk);
 319	clear_tsk_need_resched(tsk);
 320	stackend = end_of_stack(tsk);
 321	*stackend = STACK_END_MAGIC;	/* for overflow detection */
 322
 323#ifdef CONFIG_CC_STACKPROTECTOR
 324	tsk->stack_canary = get_random_int();
 325#endif
 326
 327	/*
 328	 * One for us, one for whoever does the "release_task()" (usually
 329	 * parent)
 330	 */
 331	atomic_set(&tsk->usage, 2);
 332#ifdef CONFIG_BLK_DEV_IO_TRACE
 333	tsk->btrace_seq = 0;
 334#endif
 335	tsk->splice_pipe = NULL;
 336
 337	account_kernel_stack(ti, 1);
 338
 339	return tsk;
 340
 341out:
 342	free_thread_info(ti);
 343	free_task_struct(tsk);
 344	return NULL;
 345}
 346
 347#ifdef CONFIG_MMU
 348static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm)
 349{
 350	struct vm_area_struct *mpnt, *tmp, *prev, **pprev;
 351	struct rb_node **rb_link, *rb_parent;
 352	int retval;
 353	unsigned long charge;
 354	struct mempolicy *pol;
 355
 356	down_write(&oldmm->mmap_sem);
 357	flush_cache_dup_mm(oldmm);
 358	/*
 359	 * Not linked in yet - no deadlock potential:
 360	 */
 361	down_write_nested(&mm->mmap_sem, SINGLE_DEPTH_NESTING);
 362
 363	mm->locked_vm = 0;
 364	mm->mmap = NULL;
 365	mm->mmap_cache = NULL;
 366	mm->free_area_cache = oldmm->mmap_base;
 367	mm->cached_hole_size = ~0UL;
 368	mm->map_count = 0;
 369	cpumask_clear(mm_cpumask(mm));
 370	mm->mm_rb = RB_ROOT;
 371	rb_link = &mm->mm_rb.rb_node;
 372	rb_parent = NULL;
 373	pprev = &mm->mmap;
 374	retval = ksm_fork(mm, oldmm);
 375	if (retval)
 376		goto out;
 377	retval = khugepaged_fork(mm, oldmm);
 378	if (retval)
 379		goto out;
 380
 381	prev = NULL;
 382	for (mpnt = oldmm->mmap; mpnt; mpnt = mpnt->vm_next) {
 383		struct file *file;
 384
 385		if (mpnt->vm_flags & VM_DONTCOPY) {
 386			long pages = vma_pages(mpnt);
 387			mm->total_vm -= pages;
 388			vm_stat_account(mm, mpnt->vm_flags, mpnt->vm_file,
 389								-pages);
 390			continue;
 391		}
 392		charge = 0;
 393		if (mpnt->vm_flags & VM_ACCOUNT) {
 394			unsigned long len;
 395			len = (mpnt->vm_end - mpnt->vm_start) >> PAGE_SHIFT;
 396			if (security_vm_enough_memory_mm(oldmm, len)) /* sic */
 397				goto fail_nomem;
 398			charge = len;
 399		}
 400		tmp = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
 401		if (!tmp)
 402			goto fail_nomem;
 403		*tmp = *mpnt;
 404		INIT_LIST_HEAD(&tmp->anon_vma_chain);
 405		pol = mpol_dup(vma_policy(mpnt));
 406		retval = PTR_ERR(pol);
 407		if (IS_ERR(pol))
 408			goto fail_nomem_policy;
 409		vma_set_policy(tmp, pol);
 410		tmp->vm_mm = mm;
 411		if (anon_vma_fork(tmp, mpnt))
 412			goto fail_nomem_anon_vma_fork;
 413		tmp->vm_flags &= ~VM_LOCKED;
 414		tmp->vm_next = tmp->vm_prev = NULL;
 415		file = tmp->vm_file;
 416		if (file) {
 417			struct inode *inode = file->f_path.dentry->d_inode;
 418			struct address_space *mapping = file->f_mapping;
 419
 420			get_file(file);
 421			if (tmp->vm_flags & VM_DENYWRITE)
 422				atomic_dec(&inode->i_writecount);
 423			mutex_lock(&mapping->i_mmap_mutex);
 424			if (tmp->vm_flags & VM_SHARED)
 425				mapping->i_mmap_writable++;
 426			flush_dcache_mmap_lock(mapping);
 427			/* insert tmp into the share list, just after mpnt */
 428			vma_prio_tree_add(tmp, mpnt);
 429			flush_dcache_mmap_unlock(mapping);
 430			mutex_unlock(&mapping->i_mmap_mutex);
 431		}
 432
 433		/*
 434		 * Clear hugetlb-related page reserves for children. This only
 435		 * affects MAP_PRIVATE mappings. Faults generated by the child
 436		 * are not guaranteed to succeed, even if read-only
 437		 */
 438		if (is_vm_hugetlb_page(tmp))
 439			reset_vma_resv_huge_pages(tmp);
 440
 441		/*
 442		 * Link in the new vma and copy the page table entries.
 443		 */
 444		*pprev = tmp;
 445		pprev = &tmp->vm_next;
 446		tmp->vm_prev = prev;
 447		prev = tmp;
 448
 449		__vma_link_rb(mm, tmp, rb_link, rb_parent);
 450		rb_link = &tmp->vm_rb.rb_right;
 451		rb_parent = &tmp->vm_rb;
 452
 453		mm->map_count++;
 454		retval = copy_page_range(mm, oldmm, mpnt);
 455
 456		if (tmp->vm_ops && tmp->vm_ops->open)
 457			tmp->vm_ops->open(tmp);
 458
 459		if (retval)
 460			goto out;
 461
 462		if (file)
 463			uprobe_mmap(tmp);
 464	}
 465	/* a new mm has just been created */
 466	arch_dup_mmap(oldmm, mm);
 467	retval = 0;
 468out:
 469	up_write(&mm->mmap_sem);
 470	flush_tlb_mm(oldmm);
 471	up_write(&oldmm->mmap_sem);
 472	return retval;
 473fail_nomem_anon_vma_fork:
 474	mpol_put(pol);
 475fail_nomem_policy:
 476	kmem_cache_free(vm_area_cachep, tmp);
 477fail_nomem:
 478	retval = -ENOMEM;
 479	vm_unacct_memory(charge);
 480	goto out;
 481}
 482
 483static inline int mm_alloc_pgd(struct mm_struct *mm)
 484{
 485	mm->pgd = pgd_alloc(mm);
 486	if (unlikely(!mm->pgd))
 487		return -ENOMEM;
 488	return 0;
 489}
 490
 491static inline void mm_free_pgd(struct mm_struct *mm)
 492{
 493	pgd_free(mm, mm->pgd);
 494}
 495#else
 496#define dup_mmap(mm, oldmm)	(0)
 497#define mm_alloc_pgd(mm)	(0)
 498#define mm_free_pgd(mm)
 499#endif /* CONFIG_MMU */
 500
 501__cacheline_aligned_in_smp DEFINE_SPINLOCK(mmlist_lock);
 502
 503#define allocate_mm()	(kmem_cache_alloc(mm_cachep, GFP_KERNEL))
 504#define free_mm(mm)	(kmem_cache_free(mm_cachep, (mm)))
 505
 506static unsigned long default_dump_filter = MMF_DUMP_FILTER_DEFAULT;
 507
 508static int __init coredump_filter_setup(char *s)
 509{
 510	default_dump_filter =
 511		(simple_strtoul(s, NULL, 0) << MMF_DUMP_FILTER_SHIFT) &
 512		MMF_DUMP_FILTER_MASK;
 513	return 1;
 514}
 515
 516__setup("coredump_filter=", coredump_filter_setup);
 517
 518#include <linux/init_task.h>
 519
 520static void mm_init_aio(struct mm_struct *mm)
 521{
 522#ifdef CONFIG_AIO
 523	spin_lock_init(&mm->ioctx_lock);
 524	INIT_HLIST_HEAD(&mm->ioctx_list);
 525#endif
 526}
 527
 528static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p)
 529{
 530	atomic_set(&mm->mm_users, 1);
 531	atomic_set(&mm->mm_count, 1);
 532	init_rwsem(&mm->mmap_sem);
 533	INIT_LIST_HEAD(&mm->mmlist);
 534	mm->flags = (current->mm) ?
 535		(current->mm->flags & MMF_INIT_MASK) : default_dump_filter;
 536	mm->core_state = NULL;
 537	mm->nr_ptes = 0;
 538	memset(&mm->rss_stat, 0, sizeof(mm->rss_stat));
 539	spin_lock_init(&mm->page_table_lock);
 540	mm->free_area_cache = TASK_UNMAPPED_BASE;
 541	mm->cached_hole_size = ~0UL;
 542	mm_init_aio(mm);
 543	mm_init_owner(mm, p);
 544
 545	if (likely(!mm_alloc_pgd(mm))) {
 546		mm->def_flags = 0;
 547		mmu_notifier_mm_init(mm);
 548		return mm;
 549	}
 550
 551	free_mm(mm);
 552	return NULL;
 553}
 554
 555static void check_mm(struct mm_struct *mm)
 556{
 557	int i;
 558
 559	for (i = 0; i < NR_MM_COUNTERS; i++) {
 560		long x = atomic_long_read(&mm->rss_stat.count[i]);
 561
 562		if (unlikely(x))
 563			printk(KERN_ALERT "BUG: Bad rss-counter state "
 564					  "mm:%p idx:%d val:%ld\n", mm, i, x);
 565	}
 566
 567#ifdef CONFIG_TRANSPARENT_HUGEPAGE
 568	VM_BUG_ON(mm->pmd_huge_pte);
 569#endif
 570}
 571
 572/*
 573 * Allocate and initialize an mm_struct.
 574 */
 575struct mm_struct *mm_alloc(void)
 576{
 577	struct mm_struct *mm;
 578
 579	mm = allocate_mm();
 580	if (!mm)
 581		return NULL;
 582
 583	memset(mm, 0, sizeof(*mm));
 584	mm_init_cpumask(mm);
 585	return mm_init(mm, current);
 586}
 587
 588/*
 589 * Called when the last reference to the mm
 590 * is dropped: either by a lazy thread or by
 591 * mmput. Free the page directory and the mm.
 592 */
 593void __mmdrop(struct mm_struct *mm)
 594{
 595	BUG_ON(mm == &init_mm);
 596	mm_free_pgd(mm);
 597	destroy_context(mm);
 598	mmu_notifier_mm_destroy(mm);
 599	check_mm(mm);
 600	free_mm(mm);
 601}
 602EXPORT_SYMBOL_GPL(__mmdrop);
 603
 604/*
 605 * Decrement the use count and release all resources for an mm.
 606 */
 607void mmput(struct mm_struct *mm)
 608{
 609	might_sleep();
 610
 611	if (atomic_dec_and_test(&mm->mm_users)) {
 612		uprobe_clear_state(mm);
 613		exit_aio(mm);
 614		ksm_exit(mm);
 615		khugepaged_exit(mm); /* must run before exit_mmap */
 616		exit_mmap(mm);
 617		set_mm_exe_file(mm, NULL);
 618		if (!list_empty(&mm->mmlist)) {
 619			spin_lock(&mmlist_lock);
 620			list_del(&mm->mmlist);
 621			spin_unlock(&mmlist_lock);
 622		}
 623		if (mm->binfmt)
 624			module_put(mm->binfmt->module);
 625		mmdrop(mm);
 626	}
 627}
 628EXPORT_SYMBOL_GPL(mmput);
 629
 630/*
 631 * We added or removed a vma mapping the executable. The vmas are only mapped
 632 * during exec and are not mapped with the mmap system call.
 633 * Callers must hold down_write() on the mm's mmap_sem for these
 634 */
 635void added_exe_file_vma(struct mm_struct *mm)
 636{
 637	mm->num_exe_file_vmas++;
 638}
 639
 640void removed_exe_file_vma(struct mm_struct *mm)
 641{
 642	mm->num_exe_file_vmas--;
 643	if ((mm->num_exe_file_vmas == 0) && mm->exe_file) {
 644		fput(mm->exe_file);
 645		mm->exe_file = NULL;
 646	}
 647
 648}
 649
 650void set_mm_exe_file(struct mm_struct *mm, struct file *new_exe_file)
 651{
 652	if (new_exe_file)
 653		get_file(new_exe_file);
 654	if (mm->exe_file)
 655		fput(mm->exe_file);
 656	mm->exe_file = new_exe_file;
 657	mm->num_exe_file_vmas = 0;
 658}
 659
 660struct file *get_mm_exe_file(struct mm_struct *mm)
 661{
 662	struct file *exe_file;
 663
 664	/* We need mmap_sem to protect against races with removal of
 665	 * VM_EXECUTABLE vmas */
 666	down_read(&mm->mmap_sem);
 667	exe_file = mm->exe_file;
 668	if (exe_file)
 669		get_file(exe_file);
 670	up_read(&mm->mmap_sem);
 671	return exe_file;
 672}
 673
 674static void dup_mm_exe_file(struct mm_struct *oldmm, struct mm_struct *newmm)
 675{
 676	/* It's safe to write the exe_file pointer without exe_file_lock because
 677	 * this is called during fork when the task is not yet in /proc */
 678	newmm->exe_file = get_mm_exe_file(oldmm);
 679}
 680
 681/**
 682 * get_task_mm - acquire a reference to the task's mm
 683 *
 684 * Returns %NULL if the task has no mm.  Checks PF_KTHREAD (meaning
 685 * this kernel workthread has transiently adopted a user mm with use_mm,
 686 * to do its AIO) is not set and if so returns a reference to it, after
 687 * bumping up the use count.  User must release the mm via mmput()
 688 * after use.  Typically used by /proc and ptrace.
 689 */
 690struct mm_struct *get_task_mm(struct task_struct *task)
 691{
 692	struct mm_struct *mm;
 693
 694	task_lock(task);
 695	mm = task->mm;
 696	if (mm) {
 697		if (task->flags & PF_KTHREAD)
 698			mm = NULL;
 699		else
 700			atomic_inc(&mm->mm_users);
 701	}
 702	task_unlock(task);
 703	return mm;
 704}
 705EXPORT_SYMBOL_GPL(get_task_mm);
 706
 707struct mm_struct *mm_access(struct task_struct *task, unsigned int mode)
 708{
 709	struct mm_struct *mm;
 710	int err;
 711
 712	err =  mutex_lock_killable(&task->signal->cred_guard_mutex);
 713	if (err)
 714		return ERR_PTR(err);
 715
 716	mm = get_task_mm(task);
 717	if (mm && mm != current->mm &&
 718			!ptrace_may_access(task, mode)) {
 719		mmput(mm);
 720		mm = ERR_PTR(-EACCES);
 721	}
 722	mutex_unlock(&task->signal->cred_guard_mutex);
 723
 724	return mm;
 725}
 726
 727static void complete_vfork_done(struct task_struct *tsk)
 728{
 729	struct completion *vfork;
 730
 731	task_lock(tsk);
 732	vfork = tsk->vfork_done;
 733	if (likely(vfork)) {
 734		tsk->vfork_done = NULL;
 735		complete(vfork);
 736	}
 737	task_unlock(tsk);
 738}
 739
 740static int wait_for_vfork_done(struct task_struct *child,
 741				struct completion *vfork)
 742{
 743	int killed;
 744
 745	freezer_do_not_count();
 746	killed = wait_for_completion_killable(vfork);
 747	freezer_count();
 748
 749	if (killed) {
 750		task_lock(child);
 751		child->vfork_done = NULL;
 752		task_unlock(child);
 753	}
 754
 755	put_task_struct(child);
 756	return killed;
 757}
 758
 759/* Please note the differences between mmput and mm_release.
 760 * mmput is called whenever we stop holding onto a mm_struct,
 761 * error success whatever.
 762 *
 763 * mm_release is called after a mm_struct has been removed
 764 * from the current process.
 765 *
 766 * This difference is important for error handling, when we
 767 * only half set up a mm_struct for a new process and need to restore
 768 * the old one.  Because we mmput the new mm_struct before
 769 * restoring the old one. . .
 770 * Eric Biederman 10 January 1998
 771 */
 772void mm_release(struct task_struct *tsk, struct mm_struct *mm)
 773{
 774	/* Get rid of any futexes when releasing the mm */
 775#ifdef CONFIG_FUTEX
 776	if (unlikely(tsk->robust_list)) {
 777		exit_robust_list(tsk);
 778		tsk->robust_list = NULL;
 779	}
 780#ifdef CONFIG_COMPAT
 781	if (unlikely(tsk->compat_robust_list)) {
 782		compat_exit_robust_list(tsk);
 783		tsk->compat_robust_list = NULL;
 784	}
 785#endif
 786	if (unlikely(!list_empty(&tsk->pi_state_list)))
 787		exit_pi_state_list(tsk);
 788#endif
 789
 790	uprobe_free_utask(tsk);
 791
 792	/* Get rid of any cached register state */
 793	deactivate_mm(tsk, mm);
 794
 795	/*
 796	 * If we're exiting normally, clear a user-space tid field if
 797	 * requested.  We leave this alone when dying by signal, to leave
 798	 * the value intact in a core dump, and to save the unnecessary
 799	 * trouble, say, a killed vfork parent shouldn't touch this mm.
 800	 * Userland only wants this done for a sys_exit.
 801	 */
 802	if (tsk->clear_child_tid) {
 803		if (!(tsk->flags & PF_SIGNALED) &&
 804		    atomic_read(&mm->mm_users) > 1) {
 805			/*
 806			 * We don't check the error code - if userspace has
 807			 * not set up a proper pointer then tough luck.
 808			 */
 809			put_user(0, tsk->clear_child_tid);
 810			sys_futex(tsk->clear_child_tid, FUTEX_WAKE,
 811					1, NULL, NULL, 0);
 812		}
 813		tsk->clear_child_tid = NULL;
 814	}
 815
 816	/*
 817	 * All done, finally we can wake up parent and return this mm to him.
 818	 * Also kthread_stop() uses this completion for synchronization.
 819	 */
 820	if (tsk->vfork_done)
 821		complete_vfork_done(tsk);
 822}
 823
 824/*
 825 * Allocate a new mm structure and copy contents from the
 826 * mm structure of the passed in task structure.
 827 */
 828struct mm_struct *dup_mm(struct task_struct *tsk)
 829{
 830	struct mm_struct *mm, *oldmm = current->mm;
 831	int err;
 832
 833	if (!oldmm)
 834		return NULL;
 835
 836	mm = allocate_mm();
 837	if (!mm)
 838		goto fail_nomem;
 839
 840	memcpy(mm, oldmm, sizeof(*mm));
 841	mm_init_cpumask(mm);
 842
 843#ifdef CONFIG_TRANSPARENT_HUGEPAGE
 844	mm->pmd_huge_pte = NULL;
 845#endif
 846	uprobe_reset_state(mm);
 847
 848	if (!mm_init(mm, tsk))
 849		goto fail_nomem;
 850
 851	if (init_new_context(tsk, mm))
 852		goto fail_nocontext;
 853
 854	dup_mm_exe_file(oldmm, mm);
 855
 856	err = dup_mmap(mm, oldmm);
 857	if (err)
 858		goto free_pt;
 859
 860	mm->hiwater_rss = get_mm_rss(mm);
 861	mm->hiwater_vm = mm->total_vm;
 862
 863	if (mm->binfmt && !try_module_get(mm->binfmt->module))
 864		goto free_pt;
 865
 866	return mm;
 867
 868free_pt:
 869	/* don't put binfmt in mmput, we haven't got module yet */
 870	mm->binfmt = NULL;
 871	mmput(mm);
 872
 873fail_nomem:
 874	return NULL;
 875
 876fail_nocontext:
 877	/*
 878	 * If init_new_context() failed, we cannot use mmput() to free the mm
 879	 * because it calls destroy_context()
 880	 */
 881	mm_free_pgd(mm);
 882	free_mm(mm);
 883	return NULL;
 884}
 885
 886static int copy_mm(unsigned long clone_flags, struct task_struct *tsk)
 887{
 888	struct mm_struct *mm, *oldmm;
 889	int retval;
 890
 891	tsk->min_flt = tsk->maj_flt = 0;
 892	tsk->nvcsw = tsk->nivcsw = 0;
 893#ifdef CONFIG_DETECT_HUNG_TASK
 894	tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw;
 895#endif
 896
 897	tsk->mm = NULL;
 898	tsk->active_mm = NULL;
 899
 900	/*
 901	 * Are we cloning a kernel thread?
 902	 *
 903	 * We need to steal a active VM for that..
 904	 */
 905	oldmm = current->mm;
 906	if (!oldmm)
 907		return 0;
 908
 909	if (clone_flags & CLONE_VM) {
 910		atomic_inc(&oldmm->mm_users);
 911		mm = oldmm;
 912		goto good_mm;
 913	}
 914
 915	retval = -ENOMEM;
 916	mm = dup_mm(tsk);
 917	if (!mm)
 918		goto fail_nomem;
 919
 920good_mm:
 921	tsk->mm = mm;
 922	tsk->active_mm = mm;
 923	return 0;
 924
 925fail_nomem:
 926	return retval;
 927}
 928
 929static int copy_fs(unsigned long clone_flags, struct task_struct *tsk)
 930{
 931	struct fs_struct *fs = current->fs;
 932	if (clone_flags & CLONE_FS) {
 933		/* tsk->fs is already what we want */
 934		spin_lock(&fs->lock);
 935		if (fs->in_exec) {
 936			spin_unlock(&fs->lock);
 937			return -EAGAIN;
 938		}
 939		fs->users++;
 940		spin_unlock(&fs->lock);
 941		return 0;
 942	}
 943	tsk->fs = copy_fs_struct(fs);
 944	if (!tsk->fs)
 945		return -ENOMEM;
 946	return 0;
 947}
 948
 949static int copy_files(unsigned long clone_flags, struct task_struct *tsk)
 950{
 951	struct files_struct *oldf, *newf;
 952	int error = 0;
 953
 954	/*
 955	 * A background process may not have any files ...
 956	 */
 957	oldf = current->files;
 958	if (!oldf)
 959		goto out;
 960
 961	if (clone_flags & CLONE_FILES) {
 962		atomic_inc(&oldf->count);
 963		goto out;
 964	}
 965
 966	newf = dup_fd(oldf, &error);
 967	if (!newf)
 968		goto out;
 969
 970	tsk->files = newf;
 971	error = 0;
 972out:
 973	return error;
 974}
 975
 976static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
 977{
 978#ifdef CONFIG_BLOCK
 979	struct io_context *ioc = current->io_context;
 980	struct io_context *new_ioc;
 981
 982	if (!ioc)
 983		return 0;
 984	/*
 985	 * Share io context with parent, if CLONE_IO is set
 986	 */
 987	if (clone_flags & CLONE_IO) {
 988		ioc_task_link(ioc);
 989		tsk->io_context = ioc;
 990	} else if (ioprio_valid(ioc->ioprio)) {
 991		new_ioc = get_task_io_context(tsk, GFP_KERNEL, NUMA_NO_NODE);
 992		if (unlikely(!new_ioc))
 993			return -ENOMEM;
 994
 995		new_ioc->ioprio = ioc->ioprio;
 996		put_io_context(new_ioc);
 997	}
 998#endif
 999	return 0;
1000}
1001
1002static int copy_sighand(unsigned long clone_flags, struct task_struct *tsk)
1003{
1004	struct sighand_struct *sig;
1005
1006	if (clone_flags & CLONE_SIGHAND) {
1007		atomic_inc(&current->sighand->count);
1008		return 0;
1009	}
1010	sig = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
1011	rcu_assign_pointer(tsk->sighand, sig);
1012	if (!sig)
1013		return -ENOMEM;
1014	atomic_set(&sig->count, 1);
1015	memcpy(sig->action, current->sighand->action, sizeof(sig->action));
1016	return 0;
1017}
1018
1019void __cleanup_sighand(struct sighand_struct *sighand)
1020{
1021	if (atomic_dec_and_test(&sighand->count)) {
1022		signalfd_cleanup(sighand);
1023		kmem_cache_free(sighand_cachep, sighand);
1024	}
1025}
1026
1027
1028/*
1029 * Initialize POSIX timer handling for a thread group.
1030 */
1031static void posix_cpu_timers_init_group(struct signal_struct *sig)
1032{
1033	unsigned long cpu_limit;
1034
1035	/* Thread group counters. */
1036	thread_group_cputime_init(sig);
1037
1038	cpu_limit = ACCESS_ONCE(sig->rlim[RLIMIT_CPU].rlim_cur);
1039	if (cpu_limit != RLIM_INFINITY) {
1040		sig->cputime_expires.prof_exp = secs_to_cputime(cpu_limit);
1041		sig->cputimer.running = 1;
1042	}
1043
1044	/* The timer lists. */
1045	INIT_LIST_HEAD(&sig->cpu_timers[0]);
1046	INIT_LIST_HEAD(&sig->cpu_timers[1]);
1047	INIT_LIST_HEAD(&sig->cpu_timers[2]);
1048}
1049
1050static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
1051{
1052	struct signal_struct *sig;
1053
1054	if (clone_flags & CLONE_THREAD)
1055		return 0;
1056
1057	sig = kmem_cache_zalloc(signal_cachep, GFP_KERNEL);
1058	tsk->signal = sig;
1059	if (!sig)
1060		return -ENOMEM;
1061
1062	sig->nr_threads = 1;
1063	atomic_set(&sig->live, 1);
1064	atomic_set(&sig->sigcnt, 1);
1065	init_waitqueue_head(&sig->wait_chldexit);
1066	if (clone_flags & CLONE_NEWPID)
1067		sig->flags |= SIGNAL_UNKILLABLE;
1068	sig->curr_target = tsk;
1069	init_sigpending(&sig->shared_pending);
1070	INIT_LIST_HEAD(&sig->posix_timers);
1071
1072	hrtimer_init(&sig->real_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
1073	sig->real_timer.function = it_real_fn;
1074
1075	task_lock(current->group_leader);
1076	memcpy(sig->rlim, current->signal->rlim, sizeof sig->rlim);
1077	task_unlock(current->group_leader);
1078
1079	posix_cpu_timers_init_group(sig);
1080
1081	tty_audit_fork(sig);
1082	sched_autogroup_fork(sig);
1083
1084#ifdef CONFIG_CGROUPS
1085	init_rwsem(&sig->group_rwsem);
1086#endif
1087
1088	sig->oom_adj = current->signal->oom_adj;
1089	sig->oom_score_adj = current->signal->oom_score_adj;
1090	sig->oom_score_adj_min = current->signal->oom_score_adj_min;
1091
1092	sig->has_child_subreaper = current->signal->has_child_subreaper ||
1093				   current->signal->is_child_subreaper;
1094
1095	mutex_init(&sig->cred_guard_mutex);
1096
1097	return 0;
1098}
1099
1100static void copy_flags(unsigned long clone_flags, struct task_struct *p)
1101{
1102	unsigned long new_flags = p->flags;
1103
1104	new_flags &= ~(PF_SUPERPRIV | PF_WQ_WORKER);
1105	new_flags |= PF_FORKNOEXEC;
1106	p->flags = new_flags;
1107}
1108
1109SYSCALL_DEFINE1(set_tid_address, int __user *, tidptr)
1110{
1111	current->clear_child_tid = tidptr;
1112
1113	return task_pid_vnr(current);
1114}
1115
1116static void rt_mutex_init_task(struct task_struct *p)
1117{
1118	raw_spin_lock_init(&p->pi_lock);
1119#ifdef CONFIG_RT_MUTEXES
1120	plist_head_init(&p->pi_waiters);
1121	p->pi_blocked_on = NULL;
1122#endif
1123}
1124
1125#ifdef CONFIG_MM_OWNER
1126void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
1127{
1128	mm->owner = p;
1129}
1130#endif /* CONFIG_MM_OWNER */
1131
1132/*
1133 * Initialize POSIX timer handling for a single task.
1134 */
1135static void posix_cpu_timers_init(struct task_struct *tsk)
1136{
1137	tsk->cputime_expires.prof_exp = 0;
1138	tsk->cputime_expires.virt_exp = 0;
1139	tsk->cputime_expires.sched_exp = 0;
1140	INIT_LIST_HEAD(&tsk->cpu_timers[0]);
1141	INIT_LIST_HEAD(&tsk->cpu_timers[1]);
1142	INIT_LIST_HEAD(&tsk->cpu_timers[2]);
1143}
1144
1145/*
1146 * This creates a new process as a copy of the old one,
1147 * but does not actually start it yet.
1148 *
1149 * It copies the registers, and all the appropriate
1150 * parts of the process environment (as per the clone
1151 * flags). The actual kick-off is left to the caller.
1152 */
1153static struct task_struct *copy_process(unsigned long clone_flags,
1154					unsigned long stack_start,
1155					struct pt_regs *regs,
1156					unsigned long stack_size,
1157					int __user *child_tidptr,
1158					struct pid *pid,
1159					int trace)
1160{
1161	int retval;
1162	struct task_struct *p;
1163	int cgroup_callbacks_done = 0;
1164
1165	if ((clone_flags & (CLONE_NEWNS|CLONE_FS)) == (CLONE_NEWNS|CLONE_FS))
1166		return ERR_PTR(-EINVAL);
1167
1168	/*
1169	 * Thread groups must share signals as well, and detached threads
1170	 * can only be started up within the thread group.
1171	 */
1172	if ((clone_flags & CLONE_THREAD) && !(clone_flags & CLONE_SIGHAND))
1173		return ERR_PTR(-EINVAL);
1174
1175	/*
1176	 * Shared signal handlers imply shared VM. By way of the above,
1177	 * thread groups also imply shared VM. Blocking this case allows
1178	 * for various simplifications in other code.
1179	 */
1180	if ((clone_flags & CLONE_SIGHAND) && !(clone_flags & CLONE_VM))
1181		return ERR_PTR(-EINVAL);
1182
1183	/*
1184	 * Siblings of global init remain as zombies on exit since they are
1185	 * not reaped by their parent (swapper). To solve this and to avoid
1186	 * multi-rooted process trees, prevent global and container-inits
1187	 * from creating siblings.
1188	 */
1189	if ((clone_flags & CLONE_PARENT) &&
1190				current->signal->flags & SIGNAL_UNKILLABLE)
1191		return ERR_PTR(-EINVAL);
1192
1193	retval = security_task_create(clone_flags);
1194	if (retval)
1195		goto fork_out;
1196
1197	retval = -ENOMEM;
1198	p = dup_task_struct(current);
1199	if (!p)
1200		goto fork_out;
1201
1202	ftrace_graph_init_task(p);
1203	get_seccomp_filter(p);
1204
1205	rt_mutex_init_task(p);
1206
1207#ifdef CONFIG_PROVE_LOCKING
1208	DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
1209	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
1210#endif
1211	retval = -EAGAIN;
1212	if (atomic_read(&p->real_cred->user->processes) >=
1213			task_rlimit(p, RLIMIT_NPROC)) {
1214		if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_RESOURCE) &&
1215		    p->real_cred->user != INIT_USER)
1216			goto bad_fork_free;
1217	}
1218	current->flags &= ~PF_NPROC_EXCEEDED;
1219
1220	retval = copy_creds(p, clone_flags);
1221	if (retval < 0)
1222		goto bad_fork_free;
1223
1224	/*
1225	 * If multiple threads are within copy_process(), then this check
1226	 * triggers too late. This doesn't hurt, the check is only there
1227	 * to stop root fork bombs.
1228	 */
1229	retval = -EAGAIN;
1230	if (nr_threads >= max_threads)
1231		goto bad_fork_cleanup_count;
1232
1233	if (!try_module_get(task_thread_info(p)->exec_domain->module))
1234		goto bad_fork_cleanup_count;
1235
1236	p->did_exec = 0;
1237	delayacct_tsk_init(p);	/* Must remain after dup_task_struct() */
1238	copy_flags(clone_flags, p);
1239	INIT_LIST_HEAD(&p->children);
1240	INIT_LIST_HEAD(&p->sibling);
1241	rcu_copy_process(p);
1242	p->vfork_done = NULL;
1243	spin_lock_init(&p->alloc_lock);
1244
1245	init_sigpending(&p->pending);
1246
1247	p->utime = p->stime = p->gtime = 0;
1248	p->utimescaled = p->stimescaled = 0;
1249#ifndef CONFIG_VIRT_CPU_ACCOUNTING
1250	p->prev_utime = p->prev_stime = 0;
1251#endif
1252#if defined(SPLIT_RSS_COUNTING)
1253	memset(&p->rss_stat, 0, sizeof(p->rss_stat));
1254#endif
1255
1256	p->default_timer_slack_ns = current->timer_slack_ns;
1257
1258	task_io_accounting_init(&p->ioac);
1259	acct_clear_integrals(p);
1260
1261	posix_cpu_timers_init(p);
1262
1263	do_posix_clock_monotonic_gettime(&p->start_time);
1264	p->real_start_time = p->start_time;
1265	monotonic_to_bootbased(&p->real_start_time);
1266	p->io_context = NULL;
1267	p->audit_context = NULL;
1268	if (clone_flags & CLONE_THREAD)
1269		threadgroup_change_begin(current);
1270	cgroup_fork(p);
1271#ifdef CONFIG_NUMA
1272	p->mempolicy = mpol_dup(p->mempolicy);
1273	if (IS_ERR(p->mempolicy)) {
1274		retval = PTR_ERR(p->mempolicy);
1275		p->mempolicy = NULL;
1276		goto bad_fork_cleanup_cgroup;
1277	}
1278	mpol_fix_fork_child_flag(p);
1279#endif
1280#ifdef CONFIG_CPUSETS
1281	p->cpuset_mem_spread_rotor = NUMA_NO_NODE;
1282	p->cpuset_slab_spread_rotor = NUMA_NO_NODE;
1283	seqcount_init(&p->mems_allowed_seq);
1284#endif
1285#ifdef CONFIG_TRACE_IRQFLAGS
1286	p->irq_events = 0;
1287#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
1288	p->hardirqs_enabled = 1;
1289#else
1290	p->hardirqs_enabled = 0;
1291#endif
1292	p->hardirq_enable_ip = 0;
1293	p->hardirq_enable_event = 0;
1294	p->hardirq_disable_ip = _THIS_IP_;
1295	p->hardirq_disable_event = 0;
1296	p->softirqs_enabled = 1;
1297	p->softirq_enable_ip = _THIS_IP_;
1298	p->softirq_enable_event = 0;
1299	p->softirq_disable_ip = 0;
1300	p->softirq_disable_event = 0;
1301	p->hardirq_context = 0;
1302	p->softirq_context = 0;
1303#endif
1304#ifdef CONFIG_LOCKDEP
1305	p->lockdep_depth = 0; /* no locks held yet */
1306	p->curr_chain_key = 0;
1307	p->lockdep_recursion = 0;
1308#endif
1309
1310#ifdef CONFIG_DEBUG_MUTEXES
1311	p->blocked_on = NULL; /* not blocked yet */
1312#endif
1313#ifdef CONFIG_CGROUP_MEM_RES_CTLR
1314	p->memcg_batch.do_batch = 0;
1315	p->memcg_batch.memcg = NULL;
1316#endif
1317
1318	/* Perform scheduler related setup. Assign this task to a CPU. */
1319	sched_fork(p);
1320
1321	retval = perf_event_init_task(p);
1322	if (retval)
1323		goto bad_fork_cleanup_policy;
1324	retval = audit_alloc(p);
1325	if (retval)
1326		goto bad_fork_cleanup_policy;
1327	/* copy all the process information */
1328	retval = copy_semundo(clone_flags, p);
1329	if (retval)
1330		goto bad_fork_cleanup_audit;
1331	retval = copy_files(clone_flags, p);
1332	if (retval)
1333		goto bad_fork_cleanup_semundo;
1334	retval = copy_fs(clone_flags, p);
1335	if (retval)
1336		goto bad_fork_cleanup_files;
1337	retval = copy_sighand(clone_flags, p);
1338	if (retval)
1339		goto bad_fork_cleanup_fs;
1340	retval = copy_signal(clone_flags, p);
1341	if (retval)
1342		goto bad_fork_cleanup_sighand;
1343	retval = copy_mm(clone_flags, p);
1344	if (retval)
1345		goto bad_fork_cleanup_signal;
1346	retval = copy_namespaces(clone_flags, p);
1347	if (retval)
1348		goto bad_fork_cleanup_mm;
1349	retval = copy_io(clone_flags, p);
1350	if (retval)
1351		goto bad_fork_cleanup_namespaces;
1352	retval = copy_thread(clone_flags, stack_start, stack_size, p, regs);
1353	if (retval)
1354		goto bad_fork_cleanup_io;
1355
1356	if (pid != &init_struct_pid) {
1357		retval = -ENOMEM;
1358		pid = alloc_pid(p->nsproxy->pid_ns);
1359		if (!pid)
1360			goto bad_fork_cleanup_io;
1361	}
1362
1363	p->pid = pid_nr(pid);
1364	p->tgid = p->pid;
1365	if (clone_flags & CLONE_THREAD)
1366		p->tgid = current->tgid;
1367
1368	p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL;
1369	/*
1370	 * Clear TID on mm_release()?
1371	 */
1372	p->clear_child_tid = (clone_flags & CLONE_CHILD_CLEARTID) ? child_tidptr : NULL;
1373#ifdef CONFIG_BLOCK
1374	p->plug = NULL;
1375#endif
1376#ifdef CONFIG_FUTEX
1377	p->robust_list = NULL;
1378#ifdef CONFIG_COMPAT
1379	p->compat_robust_list = NULL;
1380#endif
1381	INIT_LIST_HEAD(&p->pi_state_list);
1382	p->pi_state_cache = NULL;
1383#endif
1384	uprobe_copy_process(p);
1385	/*
1386	 * sigaltstack should be cleared when sharing the same VM
1387	 */
1388	if ((clone_flags & (CLONE_VM|CLONE_VFORK)) == CLONE_VM)
1389		p->sas_ss_sp = p->sas_ss_size = 0;
1390
1391	/*
1392	 * Syscall tracing and stepping should be turned off in the
1393	 * child regardless of CLONE_PTRACE.
1394	 */
1395	user_disable_single_step(p);
1396	clear_tsk_thread_flag(p, TIF_SYSCALL_TRACE);
1397#ifdef TIF_SYSCALL_EMU
1398	clear_tsk_thread_flag(p, TIF_SYSCALL_EMU);
1399#endif
1400	clear_all_latency_tracing(p);
1401
1402	/* ok, now we should be set up.. */
1403	if (clone_flags & CLONE_THREAD)
1404		p->exit_signal = -1;
1405	else if (clone_flags & CLONE_PARENT)
1406		p->exit_signal = current->group_leader->exit_signal;
1407	else
1408		p->exit_signal = (clone_flags & CSIGNAL);
1409
1410	p->pdeath_signal = 0;
1411	p->exit_state = 0;
1412
1413	p->nr_dirtied = 0;
1414	p->nr_dirtied_pause = 128 >> (PAGE_SHIFT - 10);
1415	p->dirty_paused_when = 0;
1416
1417	/*
1418	 * Ok, make it visible to the rest of the system.
1419	 * We dont wake it up yet.
1420	 */
1421	p->group_leader = p;
1422	INIT_LIST_HEAD(&p->thread_group);
1423	INIT_HLIST_HEAD(&p->task_works);
1424
1425	/* Now that the task is set up, run cgroup callbacks if
1426	 * necessary. We need to run them before the task is visible
1427	 * on the tasklist. */
1428	cgroup_fork_callbacks(p);
1429	cgroup_callbacks_done = 1;
1430
1431	/* Need tasklist lock for parent etc handling! */
1432	write_lock_irq(&tasklist_lock);
1433
1434	/* CLONE_PARENT re-uses the old parent */
1435	if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) {
1436		p->real_parent = current->real_parent;
1437		p->parent_exec_id = current->parent_exec_id;
1438	} else {
1439		p->real_parent = current;
1440		p->parent_exec_id = current->self_exec_id;
1441	}
1442
1443	spin_lock(&current->sighand->siglock);
1444
1445	/*
1446	 * Process group and session signals need to be delivered to just the
1447	 * parent before the fork or both the parent and the child after the
1448	 * fork. Restart if a signal comes in before we add the new process to
1449	 * it's process group.
1450	 * A fatal signal pending means that current will exit, so the new
1451	 * thread can't slip out of an OOM kill (or normal SIGKILL).
1452	*/
1453	recalc_sigpending();
1454	if (signal_pending(current)) {
1455		spin_unlock(&current->sighand->siglock);
1456		write_unlock_irq(&tasklist_lock);
1457		retval = -ERESTARTNOINTR;
1458		goto bad_fork_free_pid;
1459	}
1460
1461	if (clone_flags & CLONE_THREAD) {
1462		current->signal->nr_threads++;
1463		atomic_inc(&current->signal->live);
1464		atomic_inc(&current->signal->sigcnt);
1465		p->group_leader = current->group_leader;
1466		list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group);
1467	}
1468
1469	if (likely(p->pid)) {
1470		ptrace_init_task(p, (clone_flags & CLONE_PTRACE) || trace);
1471
1472		if (thread_group_leader(p)) {
1473			if (is_child_reaper(pid))
1474				p->nsproxy->pid_ns->child_reaper = p;
1475
1476			p->signal->leader_pid = pid;
1477			p->signal->tty = tty_kref_get(current->signal->tty);
1478			attach_pid(p, PIDTYPE_PGID, task_pgrp(current));
1479			attach_pid(p, PIDTYPE_SID, task_session(current));
1480			list_add_tail(&p->sibling, &p->real_parent->children);
1481			list_add_tail_rcu(&p->tasks, &init_task.tasks);
1482			__this_cpu_inc(process_counts);
1483		}
1484		attach_pid(p, PIDTYPE_PID, pid);
1485		nr_threads++;
1486	}
1487
1488	total_forks++;
1489	spin_unlock(&current->sighand->siglock);
1490	write_unlock_irq(&tasklist_lock);
1491	proc_fork_connector(p);
1492	cgroup_post_fork(p);
1493	if (clone_flags & CLONE_THREAD)
1494		threadgroup_change_end(current);
1495	perf_event_fork(p);
1496
1497	trace_task_newtask(p, clone_flags);
1498
1499	return p;
1500
1501bad_fork_free_pid:
1502	if (pid != &init_struct_pid)
1503		free_pid(pid);
1504bad_fork_cleanup_io:
1505	if (p->io_context)
1506		exit_io_context(p);
1507bad_fork_cleanup_namespaces:
1508	if (unlikely(clone_flags & CLONE_NEWPID))
1509		pid_ns_release_proc(p->nsproxy->pid_ns);
1510	exit_task_namespaces(p);
1511bad_fork_cleanup_mm:
1512	if (p->mm)
1513		mmput(p->mm);
1514bad_fork_cleanup_signal:
1515	if (!(clone_flags & CLONE_THREAD))
1516		free_signal_struct(p->signal);
1517bad_fork_cleanup_sighand:
1518	__cleanup_sighand(p->sighand);
1519bad_fork_cleanup_fs:
1520	exit_fs(p); /* blocking */
1521bad_fork_cleanup_files:
1522	exit_files(p); /* blocking */
1523bad_fork_cleanup_semundo:
1524	exit_sem(p);
1525bad_fork_cleanup_audit:
1526	audit_free(p);
1527bad_fork_cleanup_policy:
1528	perf_event_free_task(p);
1529#ifdef CONFIG_NUMA
1530	mpol_put(p->mempolicy);
1531bad_fork_cleanup_cgroup:
1532#endif
1533	if (clone_flags & CLONE_THREAD)
1534		threadgroup_change_end(current);
1535	cgroup_exit(p, cgroup_callbacks_done);
1536	delayacct_tsk_free(p);
1537	module_put(task_thread_info(p)->exec_domain->module);
1538bad_fork_cleanup_count:
1539	atomic_dec(&p->cred->user->processes);
1540	exit_creds(p);
1541bad_fork_free:
1542	free_task(p);
1543fork_out:
1544	return ERR_PTR(retval);
1545}
1546
1547noinline struct pt_regs * __cpuinit __attribute__((weak)) idle_regs(struct pt_regs *regs)
1548{
1549	memset(regs, 0, sizeof(struct pt_regs));
1550	return regs;
1551}
1552
1553static inline void init_idle_pids(struct pid_link *links)
1554{
1555	enum pid_type type;
1556
1557	for (type = PIDTYPE_PID; type < PIDTYPE_MAX; ++type) {
1558		INIT_HLIST_NODE(&links[type].node); /* not really needed */
1559		links[type].pid = &init_struct_pid;
1560	}
1561}
1562
1563struct task_struct * __cpuinit fork_idle(int cpu)
1564{
1565	struct task_struct *task;
1566	struct pt_regs regs;
1567
1568	task = copy_process(CLONE_VM, 0, idle_regs(&regs), 0, NULL,
1569			    &init_struct_pid, 0);
1570	if (!IS_ERR(task)) {
1571		init_idle_pids(task->pids);
1572		init_idle(task, cpu);
1573	}
1574
1575	return task;
1576}
1577
1578/*
1579 *  Ok, this is the main fork-routine.
1580 *
1581 * It copies the process, and if successful kick-starts
1582 * it and waits for it to finish using the VM if required.
1583 */
1584long do_fork(unsigned long clone_flags,
1585	      unsigned long stack_start,
1586	      struct pt_regs *regs,
1587	      unsigned long stack_size,
1588	      int __user *parent_tidptr,
1589	      int __user *child_tidptr)
1590{
1591	struct task_struct *p;
1592	int trace = 0;
1593	long nr;
1594
1595	/*
1596	 * Do some preliminary argument and permissions checking before we
1597	 * actually start allocating stuff
1598	 */
1599	if (clone_flags & CLONE_NEWUSER) {
1600		if (clone_flags & CLONE_THREAD)
1601			return -EINVAL;
1602		/* hopefully this check will go away when userns support is
1603		 * complete
1604		 */
1605		if (!capable(CAP_SYS_ADMIN) || !capable(CAP_SETUID) ||
1606				!capable(CAP_SETGID))
1607			return -EPERM;
1608	}
1609
1610	/*
1611	 * Determine whether and which event to report to ptracer.  When
1612	 * called from kernel_thread or CLONE_UNTRACED is explicitly
1613	 * requested, no event is reported; otherwise, report if the event
1614	 * for the type of forking is enabled.
1615	 */
1616	if (likely(user_mode(regs)) && !(clone_flags & CLONE_UNTRACED)) {
1617		if (clone_flags & CLONE_VFORK)
1618			trace = PTRACE_EVENT_VFORK;
1619		else if ((clone_flags & CSIGNAL) != SIGCHLD)
1620			trace = PTRACE_EVENT_CLONE;
1621		else
1622			trace = PTRACE_EVENT_FORK;
1623
1624		if (likely(!ptrace_event_enabled(current, trace)))
1625			trace = 0;
1626	}
1627
1628	p = copy_process(clone_flags, stack_start, regs, stack_size,
1629			 child_tidptr, NULL, trace);
1630	/*
1631	 * Do this prior waking up the new thread - the thread pointer
1632	 * might get invalid after that point, if the thread exits quickly.
1633	 */
1634	if (!IS_ERR(p)) {
1635		struct completion vfork;
1636
1637		trace_sched_process_fork(current, p);
1638
1639		nr = task_pid_vnr(p);
1640
1641		if (clone_flags & CLONE_PARENT_SETTID)
1642			put_user(nr, parent_tidptr);
1643
1644		if (clone_flags & CLONE_VFORK) {
1645			p->vfork_done = &vfork;
1646			init_completion(&vfork);
1647			get_task_struct(p);
1648		}
1649
1650		wake_up_new_task(p);
1651
1652		/* forking complete and child started to run, tell ptracer */
1653		if (unlikely(trace))
1654			ptrace_event(trace, nr);
1655
1656		if (clone_flags & CLONE_VFORK) {
1657			if (!wait_for_vfork_done(p, &vfork))
1658				ptrace_event(PTRACE_EVENT_VFORK_DONE, nr);
1659		}
1660	} else {
1661		nr = PTR_ERR(p);
1662	}
1663	return nr;
1664}
1665
1666#ifndef ARCH_MIN_MMSTRUCT_ALIGN
1667#define ARCH_MIN_MMSTRUCT_ALIGN 0
1668#endif
1669
1670static void sighand_ctor(void *data)
1671{
1672	struct sighand_struct *sighand = data;
1673
1674	spin_lock_init(&sighand->siglock);
1675	init_waitqueue_head(&sighand->signalfd_wqh);
1676}
1677
1678void __init proc_caches_init(void)
1679{
1680	sighand_cachep = kmem_cache_create("sighand_cache",
1681			sizeof(struct sighand_struct), 0,
1682			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_DESTROY_BY_RCU|
1683			SLAB_NOTRACK, sighand_ctor);
1684	signal_cachep = kmem_cache_create("signal_cache",
1685			sizeof(struct signal_struct), 0,
1686			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1687	files_cachep = kmem_cache_create("files_cache",
1688			sizeof(struct files_struct), 0,
1689			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1690	fs_cachep = kmem_cache_create("fs_cache",
1691			sizeof(struct fs_struct), 0,
1692			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1693	/*
1694	 * FIXME! The "sizeof(struct mm_struct)" currently includes the
1695	 * whole struct cpumask for the OFFSTACK case. We could change
1696	 * this to *only* allocate as much of it as required by the
1697	 * maximum number of CPU's we can ever have.  The cpumask_allocation
1698	 * is at the end of the structure, exactly for that reason.
1699	 */
1700	mm_cachep = kmem_cache_create("mm_struct",
1701			sizeof(struct mm_struct), ARCH_MIN_MMSTRUCT_ALIGN,
1702			SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL);
1703	vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC);
1704	mmap_init();
1705	nsproxy_cache_init();
1706}
1707
1708/*
1709 * Check constraints on flags passed to the unshare system call.
1710 */
1711static int check_unshare_flags(unsigned long unshare_flags)
1712{
1713	if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND|
1714				CLONE_VM|CLONE_FILES|CLONE_SYSVSEM|
1715				CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET))
1716		return -EINVAL;
1717	/*
1718	 * Not implemented, but pretend it works if there is nothing to
1719	 * unshare. Note that unsharing CLONE_THREAD or CLONE_SIGHAND
1720	 * needs to unshare vm.
1721	 */
1722	if (unshare_flags & (CLONE_THREAD | CLONE_SIGHAND | CLONE_VM)) {
1723		/* FIXME: get_task_mm() increments ->mm_users */
1724		if (atomic_read(&current->mm->mm_users) > 1)
1725			return -EINVAL;
1726	}
1727
1728	return 0;
1729}
1730
1731/*
1732 * Unshare the filesystem structure if it is being shared
1733 */
1734static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp)
1735{
1736	struct fs_struct *fs = current->fs;
1737
1738	if (!(unshare_flags & CLONE_FS) || !fs)
1739		return 0;
1740
1741	/* don't need lock here; in the worst case we'll do useless copy */
1742	if (fs->users == 1)
1743		return 0;
1744
1745	*new_fsp = copy_fs_struct(fs);
1746	if (!*new_fsp)
1747		return -ENOMEM;
1748
1749	return 0;
1750}
1751
1752/*
1753 * Unshare file descriptor table if it is being shared
1754 */
1755static int unshare_fd(unsigned long unshare_flags, struct files_struct **new_fdp)
1756{
1757	struct files_struct *fd = current->files;
1758	int error = 0;
1759
1760	if ((unshare_flags & CLONE_FILES) &&
1761	    (fd && atomic_read(&fd->count) > 1)) {
1762		*new_fdp = dup_fd(fd, &error);
1763		if (!*new_fdp)
1764			return error;
1765	}
1766
1767	return 0;
1768}
1769
1770/*
1771 * unshare allows a process to 'unshare' part of the process
1772 * context which was originally shared using clone.  copy_*
1773 * functions used by do_fork() cannot be used here directly
1774 * because they modify an inactive task_struct that is being
1775 * constructed. Here we are modifying the current, active,
1776 * task_struct.
1777 */
1778SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags)
1779{
1780	struct fs_struct *fs, *new_fs = NULL;
1781	struct files_struct *fd, *new_fd = NULL;
1782	struct nsproxy *new_nsproxy = NULL;
1783	int do_sysvsem = 0;
1784	int err;
1785
1786	err = check_unshare_flags(unshare_flags);
1787	if (err)
1788		goto bad_unshare_out;
1789
1790	/*
1791	 * If unsharing namespace, must also unshare filesystem information.
1792	 */
1793	if (unshare_flags & CLONE_NEWNS)
1794		unshare_flags |= CLONE_FS;
1795	/*
1796	 * CLONE_NEWIPC must also detach from the undolist: after switching
1797	 * to a new ipc namespace, the semaphore arrays from the old
1798	 * namespace are unreachable.
1799	 */
1800	if (unshare_flags & (CLONE_NEWIPC|CLONE_SYSVSEM))
1801		do_sysvsem = 1;
1802	err = unshare_fs(unshare_flags, &new_fs);
1803	if (err)
1804		goto bad_unshare_out;
1805	err = unshare_fd(unshare_flags, &new_fd);
1806	if (err)
1807		goto bad_unshare_cleanup_fs;
1808	err = unshare_nsproxy_namespaces(unshare_flags, &new_nsproxy, new_fs);
1809	if (err)
1810		goto bad_unshare_cleanup_fd;
1811
1812	if (new_fs || new_fd || do_sysvsem || new_nsproxy) {
1813		if (do_sysvsem) {
1814			/*
1815			 * CLONE_SYSVSEM is equivalent to sys_exit().
1816			 */
1817			exit_sem(current);
1818		}
1819
1820		if (new_nsproxy) {
1821			switch_task_namespaces(current, new_nsproxy);
1822			new_nsproxy = NULL;
1823		}
1824
1825		task_lock(current);
1826
1827		if (new_fs) {
1828			fs = current->fs;
1829			spin_lock(&fs->lock);
1830			current->fs = new_fs;
1831			if (--fs->users)
1832				new_fs = NULL;
1833			else
1834				new_fs = fs;
1835			spin_unlock(&fs->lock);
1836		}
1837
1838		if (new_fd) {
1839			fd = current->files;
1840			current->files = new_fd;
1841			new_fd = fd;
1842		}
1843
1844		task_unlock(current);
1845	}
1846
1847	if (new_nsproxy)
1848		put_nsproxy(new_nsproxy);
1849
1850bad_unshare_cleanup_fd:
1851	if (new_fd)
1852		put_files_struct(new_fd);
1853
1854bad_unshare_cleanup_fs:
1855	if (new_fs)
1856		free_fs_struct(new_fs);
1857
1858bad_unshare_out:
1859	return err;
1860}
1861
1862/*
1863 *	Helper to unshare the files of the current task.
1864 *	We don't want to expose copy_files internals to
1865 *	the exec layer of the kernel.
1866 */
1867
1868int unshare_files(struct files_struct **displaced)
1869{
1870	struct task_struct *task = current;
1871	struct files_struct *copy = NULL;
1872	int error;
1873
1874	error = unshare_fd(CLONE_FILES, &copy);
1875	if (error || !copy) {
1876		*displaced = NULL;
1877		return error;
1878	}
1879	*displaced = task->files;
1880	task_lock(task);
1881	task->files = copy;
1882	task_unlock(task);
1883	return 0;
1884}