blob: 1f2ddaaf3c03bd02ad43a4d20dfac1e0fe48ed04 [file] [log] [blame]
Andrea Arcangeli86039bd2015-09-04 15:46:31 -07001/*
2 * fs/userfaultfd.c
3 *
4 * Copyright (C) 2007 Davide Libenzi <davidel@xmailserver.org>
5 * Copyright (C) 2008-2009 Red Hat, Inc.
6 * Copyright (C) 2015 Red Hat, Inc.
7 *
8 * This work is licensed under the terms of the GNU GPL, version 2. See
9 * the COPYING file in the top-level directory.
10 *
11 * Some part derived from fs/eventfd.c (anon inode setup) and
12 * mm/ksm.c (mm hashing).
13 */
14
15#include <linux/hashtable.h>
16#include <linux/sched.h>
17#include <linux/mm.h>
18#include <linux/poll.h>
19#include <linux/slab.h>
20#include <linux/seq_file.h>
21#include <linux/file.h>
22#include <linux/bug.h>
23#include <linux/anon_inodes.h>
24#include <linux/syscalls.h>
25#include <linux/userfaultfd_k.h>
26#include <linux/mempolicy.h>
27#include <linux/ioctl.h>
28#include <linux/security.h>
29
30enum userfaultfd_state {
31 UFFD_STATE_WAIT_API,
32 UFFD_STATE_RUNNING,
33};
34
35struct userfaultfd_ctx {
36 /* pseudo fd refcounting */
37 atomic_t refcount;
38 /* waitqueue head for the userfaultfd page faults */
39 wait_queue_head_t fault_wqh;
40 /* waitqueue head for the pseudo fd to wakeup poll/read */
41 wait_queue_head_t fd_wqh;
42 /* userfaultfd syscall flags */
43 unsigned int flags;
44 /* state machine */
45 enum userfaultfd_state state;
46 /* released */
47 bool released;
48 /* mm with one ore more vmas attached to this userfaultfd_ctx */
49 struct mm_struct *mm;
50};
51
52struct userfaultfd_wait_queue {
Andrea Arcangelia9b85f92015-09-04 15:46:37 -070053 struct uffd_msg msg;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -070054 wait_queue_t wq;
55 bool pending;
56 struct userfaultfd_ctx *ctx;
57};
58
59struct userfaultfd_wake_range {
60 unsigned long start;
61 unsigned long len;
62};
63
64static int userfaultfd_wake_function(wait_queue_t *wq, unsigned mode,
65 int wake_flags, void *key)
66{
67 struct userfaultfd_wake_range *range = key;
68 int ret;
69 struct userfaultfd_wait_queue *uwq;
70 unsigned long start, len;
71
72 uwq = container_of(wq, struct userfaultfd_wait_queue, wq);
73 ret = 0;
74 /* don't wake the pending ones to avoid reads to block */
75 if (uwq->pending && !ACCESS_ONCE(uwq->ctx->released))
76 goto out;
77 /* len == 0 means wake all */
78 start = range->start;
79 len = range->len;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -070080 if (len && (start > uwq->msg.arg.pagefault.address ||
81 start + len <= uwq->msg.arg.pagefault.address))
Andrea Arcangeli86039bd2015-09-04 15:46:31 -070082 goto out;
83 ret = wake_up_state(wq->private, mode);
84 if (ret)
85 /*
86 * Wake only once, autoremove behavior.
87 *
88 * After the effect of list_del_init is visible to the
89 * other CPUs, the waitqueue may disappear from under
90 * us, see the !list_empty_careful() in
91 * handle_userfault(). try_to_wake_up() has an
92 * implicit smp_mb__before_spinlock, and the
93 * wq->private is read before calling the extern
94 * function "wake_up_state" (which in turns calls
95 * try_to_wake_up). While the spin_lock;spin_unlock;
96 * wouldn't be enough, the smp_mb__before_spinlock is
97 * enough to avoid an explicit smp_mb() here.
98 */
99 list_del_init(&wq->task_list);
100out:
101 return ret;
102}
103
104/**
105 * userfaultfd_ctx_get - Acquires a reference to the internal userfaultfd
106 * context.
107 * @ctx: [in] Pointer to the userfaultfd context.
108 *
109 * Returns: In case of success, returns not zero.
110 */
111static void userfaultfd_ctx_get(struct userfaultfd_ctx *ctx)
112{
113 if (!atomic_inc_not_zero(&ctx->refcount))
114 BUG();
115}
116
117/**
118 * userfaultfd_ctx_put - Releases a reference to the internal userfaultfd
119 * context.
120 * @ctx: [in] Pointer to userfaultfd context.
121 *
122 * The userfaultfd context reference must have been previously acquired either
123 * with userfaultfd_ctx_get() or userfaultfd_ctx_fdget().
124 */
125static void userfaultfd_ctx_put(struct userfaultfd_ctx *ctx)
126{
127 if (atomic_dec_and_test(&ctx->refcount)) {
128 VM_BUG_ON(spin_is_locked(&ctx->fault_pending_wqh.lock));
129 VM_BUG_ON(waitqueue_active(&ctx->fault_pending_wqh));
130 VM_BUG_ON(spin_is_locked(&ctx->fault_wqh.lock));
131 VM_BUG_ON(waitqueue_active(&ctx->fault_wqh));
132 VM_BUG_ON(spin_is_locked(&ctx->fd_wqh.lock));
133 VM_BUG_ON(waitqueue_active(&ctx->fd_wqh));
134 mmput(ctx->mm);
135 kfree(ctx);
136 }
137}
138
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700139static inline void msg_init(struct uffd_msg *msg)
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700140{
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700141 BUILD_BUG_ON(sizeof(struct uffd_msg) != 32);
142 /*
143 * Must use memset to zero out the paddings or kernel data is
144 * leaked to userland.
145 */
146 memset(msg, 0, sizeof(struct uffd_msg));
147}
148
149static inline struct uffd_msg userfault_msg(unsigned long address,
150 unsigned int flags,
151 unsigned long reason)
152{
153 struct uffd_msg msg;
154 msg_init(&msg);
155 msg.event = UFFD_EVENT_PAGEFAULT;
156 msg.arg.pagefault.address = address;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700157 if (flags & FAULT_FLAG_WRITE)
158 /*
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700159 * If UFFD_FEATURE_PAGEFAULT_FLAG_WRITE was set in the
160 * uffdio_api.features and UFFD_PAGEFAULT_FLAG_WRITE
161 * was not set in a UFFD_EVENT_PAGEFAULT, it means it
162 * was a read fault, otherwise if set it means it's
163 * a write fault.
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700164 */
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700165 msg.arg.pagefault.flags |= UFFD_PAGEFAULT_FLAG_WRITE;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700166 if (reason & VM_UFFD_WP)
167 /*
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700168 * If UFFD_FEATURE_PAGEFAULT_FLAG_WP was set in the
169 * uffdio_api.features and UFFD_PAGEFAULT_FLAG_WP was
170 * not set in a UFFD_EVENT_PAGEFAULT, it means it was
171 * a missing fault, otherwise if set it means it's a
172 * write protect fault.
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700173 */
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700174 msg.arg.pagefault.flags |= UFFD_PAGEFAULT_FLAG_WP;
175 return msg;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700176}
177
178/*
179 * The locking rules involved in returning VM_FAULT_RETRY depending on
180 * FAULT_FLAG_ALLOW_RETRY, FAULT_FLAG_RETRY_NOWAIT and
181 * FAULT_FLAG_KILLABLE are not straightforward. The "Caution"
182 * recommendation in __lock_page_or_retry is not an understatement.
183 *
184 * If FAULT_FLAG_ALLOW_RETRY is set, the mmap_sem must be released
185 * before returning VM_FAULT_RETRY only if FAULT_FLAG_RETRY_NOWAIT is
186 * not set.
187 *
188 * If FAULT_FLAG_ALLOW_RETRY is set but FAULT_FLAG_KILLABLE is not
189 * set, VM_FAULT_RETRY can still be returned if and only if there are
190 * fatal_signal_pending()s, and the mmap_sem must be released before
191 * returning it.
192 */
193int handle_userfault(struct vm_area_struct *vma, unsigned long address,
194 unsigned int flags, unsigned long reason)
195{
196 struct mm_struct *mm = vma->vm_mm;
197 struct userfaultfd_ctx *ctx;
198 struct userfaultfd_wait_queue uwq;
199
200 BUG_ON(!rwsem_is_locked(&mm->mmap_sem));
201
202 ctx = vma->vm_userfaultfd_ctx.ctx;
203 if (!ctx)
204 return VM_FAULT_SIGBUS;
205
206 BUG_ON(ctx->mm != mm);
207
208 VM_BUG_ON(reason & ~(VM_UFFD_MISSING|VM_UFFD_WP));
209 VM_BUG_ON(!(reason & VM_UFFD_MISSING) ^ !!(reason & VM_UFFD_WP));
210
211 /*
212 * If it's already released don't get it. This avoids to loop
213 * in __get_user_pages if userfaultfd_release waits on the
214 * caller of handle_userfault to release the mmap_sem.
215 */
216 if (unlikely(ACCESS_ONCE(ctx->released)))
217 return VM_FAULT_SIGBUS;
218
219 /*
220 * Check that we can return VM_FAULT_RETRY.
221 *
222 * NOTE: it should become possible to return VM_FAULT_RETRY
223 * even if FAULT_FLAG_TRIED is set without leading to gup()
224 * -EBUSY failures, if the userfaultfd is to be extended for
225 * VM_UFFD_WP tracking and we intend to arm the userfault
226 * without first stopping userland access to the memory. For
227 * VM_UFFD_MISSING userfaults this is enough for now.
228 */
229 if (unlikely(!(flags & FAULT_FLAG_ALLOW_RETRY))) {
230 /*
231 * Validate the invariant that nowait must allow retry
232 * to be sure not to return SIGBUS erroneously on
233 * nowait invocations.
234 */
235 BUG_ON(flags & FAULT_FLAG_RETRY_NOWAIT);
236#ifdef CONFIG_DEBUG_VM
237 if (printk_ratelimit()) {
238 printk(KERN_WARNING
239 "FAULT_FLAG_ALLOW_RETRY missing %x\n", flags);
240 dump_stack();
241 }
242#endif
243 return VM_FAULT_SIGBUS;
244 }
245
246 /*
247 * Handle nowait, not much to do other than tell it to retry
248 * and wait.
249 */
250 if (flags & FAULT_FLAG_RETRY_NOWAIT)
251 return VM_FAULT_RETRY;
252
253 /* take the reference before dropping the mmap_sem */
254 userfaultfd_ctx_get(ctx);
255
256 /* be gentle and immediately relinquish the mmap_sem */
257 up_read(&mm->mmap_sem);
258
259 init_waitqueue_func_entry(&uwq.wq, userfaultfd_wake_function);
260 uwq.wq.private = current;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700261 uwq.msg = userfault_msg(address, flags, reason);
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700262 uwq.pending = true;
263 uwq.ctx = ctx;
264
265 spin_lock(&ctx->fault_wqh.lock);
266 /*
267 * After the __add_wait_queue the uwq is visible to userland
268 * through poll/read().
269 */
270 __add_wait_queue(&ctx->fault_wqh, &uwq.wq);
271 for (;;) {
272 set_current_state(TASK_KILLABLE);
273 if (!uwq.pending || ACCESS_ONCE(ctx->released) ||
274 fatal_signal_pending(current))
275 break;
276 spin_unlock(&ctx->fault_wqh.lock);
277
278 wake_up_poll(&ctx->fd_wqh, POLLIN);
279 schedule();
280
281 spin_lock(&ctx->fault_wqh.lock);
282 }
283 __remove_wait_queue(&ctx->fault_wqh, &uwq.wq);
284 __set_current_state(TASK_RUNNING);
285 spin_unlock(&ctx->fault_wqh.lock);
286
287 /*
288 * ctx may go away after this if the userfault pseudo fd is
289 * already released.
290 */
291 userfaultfd_ctx_put(ctx);
292
293 return VM_FAULT_RETRY;
294}
295
296static int userfaultfd_release(struct inode *inode, struct file *file)
297{
298 struct userfaultfd_ctx *ctx = file->private_data;
299 struct mm_struct *mm = ctx->mm;
300 struct vm_area_struct *vma, *prev;
301 /* len == 0 means wake all */
302 struct userfaultfd_wake_range range = { .len = 0, };
303 unsigned long new_flags;
304
305 ACCESS_ONCE(ctx->released) = true;
306
307 /*
308 * Flush page faults out of all CPUs. NOTE: all page faults
309 * must be retried without returning VM_FAULT_SIGBUS if
310 * userfaultfd_ctx_get() succeeds but vma->vma_userfault_ctx
311 * changes while handle_userfault released the mmap_sem. So
312 * it's critical that released is set to true (above), before
313 * taking the mmap_sem for writing.
314 */
315 down_write(&mm->mmap_sem);
316 prev = NULL;
317 for (vma = mm->mmap; vma; vma = vma->vm_next) {
318 cond_resched();
319 BUG_ON(!!vma->vm_userfaultfd_ctx.ctx ^
320 !!(vma->vm_flags & (VM_UFFD_MISSING | VM_UFFD_WP)));
321 if (vma->vm_userfaultfd_ctx.ctx != ctx) {
322 prev = vma;
323 continue;
324 }
325 new_flags = vma->vm_flags & ~(VM_UFFD_MISSING | VM_UFFD_WP);
326 prev = vma_merge(mm, prev, vma->vm_start, vma->vm_end,
327 new_flags, vma->anon_vma,
328 vma->vm_file, vma->vm_pgoff,
329 vma_policy(vma),
330 NULL_VM_UFFD_CTX);
331 if (prev)
332 vma = prev;
333 else
334 prev = vma;
335 vma->vm_flags = new_flags;
336 vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
337 }
338 up_write(&mm->mmap_sem);
339
340 /*
341 * After no new page faults can wait on this fault_wqh, flush
342 * the last page faults that may have been already waiting on
343 * the fault_wqh.
344 */
345 spin_lock(&ctx->fault_wqh.lock);
346 __wake_up_locked_key(&ctx->fault_wqh, TASK_NORMAL, 0, &range);
347 spin_unlock(&ctx->fault_wqh.lock);
348
349 wake_up_poll(&ctx->fd_wqh, POLLHUP);
350 userfaultfd_ctx_put(ctx);
351 return 0;
352}
353
354/* fault_wqh.lock must be hold by the caller */
355static inline unsigned int find_userfault(struct userfaultfd_ctx *ctx,
356 struct userfaultfd_wait_queue **uwq)
357{
358 wait_queue_t *wq;
359 struct userfaultfd_wait_queue *_uwq;
360 unsigned int ret = 0;
361
362 VM_BUG_ON(!spin_is_locked(&ctx->fault_wqh.lock));
363
364 list_for_each_entry(wq, &ctx->fault_wqh.task_list, task_list) {
365 _uwq = container_of(wq, struct userfaultfd_wait_queue, wq);
366 if (_uwq->pending) {
367 ret = POLLIN;
368 if (!uwq)
369 /*
370 * If there's at least a pending and
371 * we don't care which one it is,
372 * break immediately and leverage the
373 * efficiency of the LIFO walk.
374 */
375 break;
376 /*
377 * If we need to find which one was pending we
378 * keep walking until we find the first not
379 * pending one, so we read() them in FIFO order.
380 */
381 *uwq = _uwq;
382 } else
383 /*
384 * break the loop at the first not pending
385 * one, there cannot be pending userfaults
386 * after the first not pending one, because
387 * all new pending ones are inserted at the
388 * head and we walk it in LIFO.
389 */
390 break;
391 }
392
393 return ret;
394}
395
396static unsigned int userfaultfd_poll(struct file *file, poll_table *wait)
397{
398 struct userfaultfd_ctx *ctx = file->private_data;
399 unsigned int ret;
400
401 poll_wait(file, &ctx->fd_wqh, wait);
402
403 switch (ctx->state) {
404 case UFFD_STATE_WAIT_API:
405 return POLLERR;
406 case UFFD_STATE_RUNNING:
407 spin_lock(&ctx->fault_wqh.lock);
408 ret = find_userfault(ctx, NULL);
409 spin_unlock(&ctx->fault_wqh.lock);
410 return ret;
411 default:
412 BUG();
413 }
414}
415
416static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait,
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700417 struct uffd_msg *msg)
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700418{
419 ssize_t ret;
420 DECLARE_WAITQUEUE(wait, current);
421 struct userfaultfd_wait_queue *uwq = NULL;
422
423 /* always take the fd_wqh lock before the fault_wqh lock */
424 spin_lock(&ctx->fd_wqh.lock);
425 __add_wait_queue(&ctx->fd_wqh, &wait);
426 for (;;) {
427 set_current_state(TASK_INTERRUPTIBLE);
428 spin_lock(&ctx->fault_wqh.lock);
429 if (find_userfault(ctx, &uwq)) {
430 /*
431 * The fault_wqh.lock prevents the uwq to
432 * disappear from under us.
433 */
434 uwq->pending = false;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700435 /* careful to always initialize msg if ret == 0 */
436 *msg = uwq->msg;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700437 spin_unlock(&ctx->fault_wqh.lock);
438 ret = 0;
439 break;
440 }
441 spin_unlock(&ctx->fault_wqh.lock);
442 if (signal_pending(current)) {
443 ret = -ERESTARTSYS;
444 break;
445 }
446 if (no_wait) {
447 ret = -EAGAIN;
448 break;
449 }
450 spin_unlock(&ctx->fd_wqh.lock);
451 schedule();
452 spin_lock(&ctx->fd_wqh.lock);
453 }
454 __remove_wait_queue(&ctx->fd_wqh, &wait);
455 __set_current_state(TASK_RUNNING);
456 spin_unlock(&ctx->fd_wqh.lock);
457
458 return ret;
459}
460
461static ssize_t userfaultfd_read(struct file *file, char __user *buf,
462 size_t count, loff_t *ppos)
463{
464 struct userfaultfd_ctx *ctx = file->private_data;
465 ssize_t _ret, ret = 0;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700466 struct uffd_msg msg;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700467 int no_wait = file->f_flags & O_NONBLOCK;
468
469 if (ctx->state == UFFD_STATE_WAIT_API)
470 return -EINVAL;
471 BUG_ON(ctx->state != UFFD_STATE_RUNNING);
472
473 for (;;) {
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700474 if (count < sizeof(msg))
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700475 return ret ? ret : -EINVAL;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700476 _ret = userfaultfd_ctx_read(ctx, no_wait, &msg);
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700477 if (_ret < 0)
478 return ret ? ret : _ret;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700479 if (copy_to_user((__u64 __user *) buf, &msg, sizeof(msg)))
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700480 return ret ? ret : -EFAULT;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700481 ret += sizeof(msg);
482 buf += sizeof(msg);
483 count -= sizeof(msg);
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700484 /*
485 * Allow to read more than one fault at time but only
486 * block if waiting for the very first one.
487 */
488 no_wait = O_NONBLOCK;
489 }
490}
491
492static void __wake_userfault(struct userfaultfd_ctx *ctx,
493 struct userfaultfd_wake_range *range)
494{
495 unsigned long start, end;
496
497 start = range->start;
498 end = range->start + range->len;
499
500 spin_lock(&ctx->fault_wqh.lock);
501 /* wake all in the range and autoremove */
502 __wake_up_locked_key(&ctx->fault_wqh, TASK_NORMAL, 0, range);
503 spin_unlock(&ctx->fault_wqh.lock);
504}
505
506static __always_inline void wake_userfault(struct userfaultfd_ctx *ctx,
507 struct userfaultfd_wake_range *range)
508{
509 /*
510 * To be sure waitqueue_active() is not reordered by the CPU
511 * before the pagetable update, use an explicit SMP memory
512 * barrier here. PT lock release or up_read(mmap_sem) still
513 * have release semantics that can allow the
514 * waitqueue_active() to be reordered before the pte update.
515 */
516 smp_mb();
517
518 /*
519 * Use waitqueue_active because it's very frequent to
520 * change the address space atomically even if there are no
521 * userfaults yet. So we take the spinlock only when we're
522 * sure we've userfaults to wake.
523 */
524 if (waitqueue_active(&ctx->fault_wqh))
525 __wake_userfault(ctx, range);
526}
527
528static __always_inline int validate_range(struct mm_struct *mm,
529 __u64 start, __u64 len)
530{
531 __u64 task_size = mm->task_size;
532
533 if (start & ~PAGE_MASK)
534 return -EINVAL;
535 if (len & ~PAGE_MASK)
536 return -EINVAL;
537 if (!len)
538 return -EINVAL;
539 if (start < mmap_min_addr)
540 return -EINVAL;
541 if (start >= task_size)
542 return -EINVAL;
543 if (len > task_size - start)
544 return -EINVAL;
545 return 0;
546}
547
548static int userfaultfd_register(struct userfaultfd_ctx *ctx,
549 unsigned long arg)
550{
551 struct mm_struct *mm = ctx->mm;
552 struct vm_area_struct *vma, *prev, *cur;
553 int ret;
554 struct uffdio_register uffdio_register;
555 struct uffdio_register __user *user_uffdio_register;
556 unsigned long vm_flags, new_flags;
557 bool found;
558 unsigned long start, end, vma_end;
559
560 user_uffdio_register = (struct uffdio_register __user *) arg;
561
562 ret = -EFAULT;
563 if (copy_from_user(&uffdio_register, user_uffdio_register,
564 sizeof(uffdio_register)-sizeof(__u64)))
565 goto out;
566
567 ret = -EINVAL;
568 if (!uffdio_register.mode)
569 goto out;
570 if (uffdio_register.mode & ~(UFFDIO_REGISTER_MODE_MISSING|
571 UFFDIO_REGISTER_MODE_WP))
572 goto out;
573 vm_flags = 0;
574 if (uffdio_register.mode & UFFDIO_REGISTER_MODE_MISSING)
575 vm_flags |= VM_UFFD_MISSING;
576 if (uffdio_register.mode & UFFDIO_REGISTER_MODE_WP) {
577 vm_flags |= VM_UFFD_WP;
578 /*
579 * FIXME: remove the below error constraint by
580 * implementing the wprotect tracking mode.
581 */
582 ret = -EINVAL;
583 goto out;
584 }
585
586 ret = validate_range(mm, uffdio_register.range.start,
587 uffdio_register.range.len);
588 if (ret)
589 goto out;
590
591 start = uffdio_register.range.start;
592 end = start + uffdio_register.range.len;
593
594 down_write(&mm->mmap_sem);
595 vma = find_vma_prev(mm, start, &prev);
596
597 ret = -ENOMEM;
598 if (!vma)
599 goto out_unlock;
600
601 /* check that there's at least one vma in the range */
602 ret = -EINVAL;
603 if (vma->vm_start >= end)
604 goto out_unlock;
605
606 /*
607 * Search for not compatible vmas.
608 *
609 * FIXME: this shall be relaxed later so that it doesn't fail
610 * on tmpfs backed vmas (in addition to the current allowance
611 * on anonymous vmas).
612 */
613 found = false;
614 for (cur = vma; cur && cur->vm_start < end; cur = cur->vm_next) {
615 cond_resched();
616
617 BUG_ON(!!cur->vm_userfaultfd_ctx.ctx ^
618 !!(cur->vm_flags & (VM_UFFD_MISSING | VM_UFFD_WP)));
619
620 /* check not compatible vmas */
621 ret = -EINVAL;
622 if (cur->vm_ops)
623 goto out_unlock;
624
625 /*
626 * Check that this vma isn't already owned by a
627 * different userfaultfd. We can't allow more than one
628 * userfaultfd to own a single vma simultaneously or we
629 * wouldn't know which one to deliver the userfaults to.
630 */
631 ret = -EBUSY;
632 if (cur->vm_userfaultfd_ctx.ctx &&
633 cur->vm_userfaultfd_ctx.ctx != ctx)
634 goto out_unlock;
635
636 found = true;
637 }
638 BUG_ON(!found);
639
640 if (vma->vm_start < start)
641 prev = vma;
642
643 ret = 0;
644 do {
645 cond_resched();
646
647 BUG_ON(vma->vm_ops);
648 BUG_ON(vma->vm_userfaultfd_ctx.ctx &&
649 vma->vm_userfaultfd_ctx.ctx != ctx);
650
651 /*
652 * Nothing to do: this vma is already registered into this
653 * userfaultfd and with the right tracking mode too.
654 */
655 if (vma->vm_userfaultfd_ctx.ctx == ctx &&
656 (vma->vm_flags & vm_flags) == vm_flags)
657 goto skip;
658
659 if (vma->vm_start > start)
660 start = vma->vm_start;
661 vma_end = min(end, vma->vm_end);
662
663 new_flags = (vma->vm_flags & ~vm_flags) | vm_flags;
664 prev = vma_merge(mm, prev, start, vma_end, new_flags,
665 vma->anon_vma, vma->vm_file, vma->vm_pgoff,
666 vma_policy(vma),
667 ((struct vm_userfaultfd_ctx){ ctx }));
668 if (prev) {
669 vma = prev;
670 goto next;
671 }
672 if (vma->vm_start < start) {
673 ret = split_vma(mm, vma, start, 1);
674 if (ret)
675 break;
676 }
677 if (vma->vm_end > end) {
678 ret = split_vma(mm, vma, end, 0);
679 if (ret)
680 break;
681 }
682 next:
683 /*
684 * In the vma_merge() successful mprotect-like case 8:
685 * the next vma was merged into the current one and
686 * the current one has not been updated yet.
687 */
688 vma->vm_flags = new_flags;
689 vma->vm_userfaultfd_ctx.ctx = ctx;
690
691 skip:
692 prev = vma;
693 start = vma->vm_end;
694 vma = vma->vm_next;
695 } while (vma && vma->vm_start < end);
696out_unlock:
697 up_write(&mm->mmap_sem);
698 if (!ret) {
699 /*
700 * Now that we scanned all vmas we can already tell
701 * userland which ioctls methods are guaranteed to
702 * succeed on this range.
703 */
704 if (put_user(UFFD_API_RANGE_IOCTLS,
705 &user_uffdio_register->ioctls))
706 ret = -EFAULT;
707 }
708out:
709 return ret;
710}
711
712static int userfaultfd_unregister(struct userfaultfd_ctx *ctx,
713 unsigned long arg)
714{
715 struct mm_struct *mm = ctx->mm;
716 struct vm_area_struct *vma, *prev, *cur;
717 int ret;
718 struct uffdio_range uffdio_unregister;
719 unsigned long new_flags;
720 bool found;
721 unsigned long start, end, vma_end;
722 const void __user *buf = (void __user *)arg;
723
724 ret = -EFAULT;
725 if (copy_from_user(&uffdio_unregister, buf, sizeof(uffdio_unregister)))
726 goto out;
727
728 ret = validate_range(mm, uffdio_unregister.start,
729 uffdio_unregister.len);
730 if (ret)
731 goto out;
732
733 start = uffdio_unregister.start;
734 end = start + uffdio_unregister.len;
735
736 down_write(&mm->mmap_sem);
737 vma = find_vma_prev(mm, start, &prev);
738
739 ret = -ENOMEM;
740 if (!vma)
741 goto out_unlock;
742
743 /* check that there's at least one vma in the range */
744 ret = -EINVAL;
745 if (vma->vm_start >= end)
746 goto out_unlock;
747
748 /*
749 * Search for not compatible vmas.
750 *
751 * FIXME: this shall be relaxed later so that it doesn't fail
752 * on tmpfs backed vmas (in addition to the current allowance
753 * on anonymous vmas).
754 */
755 found = false;
756 ret = -EINVAL;
757 for (cur = vma; cur && cur->vm_start < end; cur = cur->vm_next) {
758 cond_resched();
759
760 BUG_ON(!!cur->vm_userfaultfd_ctx.ctx ^
761 !!(cur->vm_flags & (VM_UFFD_MISSING | VM_UFFD_WP)));
762
763 /*
764 * Check not compatible vmas, not strictly required
765 * here as not compatible vmas cannot have an
766 * userfaultfd_ctx registered on them, but this
767 * provides for more strict behavior to notice
768 * unregistration errors.
769 */
770 if (cur->vm_ops)
771 goto out_unlock;
772
773 found = true;
774 }
775 BUG_ON(!found);
776
777 if (vma->vm_start < start)
778 prev = vma;
779
780 ret = 0;
781 do {
782 cond_resched();
783
784 BUG_ON(vma->vm_ops);
785
786 /*
787 * Nothing to do: this vma is already registered into this
788 * userfaultfd and with the right tracking mode too.
789 */
790 if (!vma->vm_userfaultfd_ctx.ctx)
791 goto skip;
792
793 if (vma->vm_start > start)
794 start = vma->vm_start;
795 vma_end = min(end, vma->vm_end);
796
797 new_flags = vma->vm_flags & ~(VM_UFFD_MISSING | VM_UFFD_WP);
798 prev = vma_merge(mm, prev, start, vma_end, new_flags,
799 vma->anon_vma, vma->vm_file, vma->vm_pgoff,
800 vma_policy(vma),
801 NULL_VM_UFFD_CTX);
802 if (prev) {
803 vma = prev;
804 goto next;
805 }
806 if (vma->vm_start < start) {
807 ret = split_vma(mm, vma, start, 1);
808 if (ret)
809 break;
810 }
811 if (vma->vm_end > end) {
812 ret = split_vma(mm, vma, end, 0);
813 if (ret)
814 break;
815 }
816 next:
817 /*
818 * In the vma_merge() successful mprotect-like case 8:
819 * the next vma was merged into the current one and
820 * the current one has not been updated yet.
821 */
822 vma->vm_flags = new_flags;
823 vma->vm_userfaultfd_ctx = NULL_VM_UFFD_CTX;
824
825 skip:
826 prev = vma;
827 start = vma->vm_end;
828 vma = vma->vm_next;
829 } while (vma && vma->vm_start < end);
830out_unlock:
831 up_write(&mm->mmap_sem);
832out:
833 return ret;
834}
835
836/*
837 * This is mostly needed to re-wakeup those userfaults that were still
838 * pending when userland wake them up the first time. We don't wake
839 * the pending one to avoid blocking reads to block, or non blocking
840 * read to return -EAGAIN, if used with POLLIN, to avoid userland
841 * doubts on why POLLIN wasn't reliable.
842 */
843static int userfaultfd_wake(struct userfaultfd_ctx *ctx,
844 unsigned long arg)
845{
846 int ret;
847 struct uffdio_range uffdio_wake;
848 struct userfaultfd_wake_range range;
849 const void __user *buf = (void __user *)arg;
850
851 ret = -EFAULT;
852 if (copy_from_user(&uffdio_wake, buf, sizeof(uffdio_wake)))
853 goto out;
854
855 ret = validate_range(ctx->mm, uffdio_wake.start, uffdio_wake.len);
856 if (ret)
857 goto out;
858
859 range.start = uffdio_wake.start;
860 range.len = uffdio_wake.len;
861
862 /*
863 * len == 0 means wake all and we don't want to wake all here,
864 * so check it again to be sure.
865 */
866 VM_BUG_ON(!range.len);
867
868 wake_userfault(ctx, &range);
869 ret = 0;
870
871out:
872 return ret;
873}
874
875/*
876 * userland asks for a certain API version and we return which bits
877 * and ioctl commands are implemented in this kernel for such API
878 * version or -EINVAL if unknown.
879 */
880static int userfaultfd_api(struct userfaultfd_ctx *ctx,
881 unsigned long arg)
882{
883 struct uffdio_api uffdio_api;
884 void __user *buf = (void __user *)arg;
885 int ret;
886
887 ret = -EINVAL;
888 if (ctx->state != UFFD_STATE_WAIT_API)
889 goto out;
890 ret = -EFAULT;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700891 if (copy_from_user(&uffdio_api, buf, sizeof(uffdio_api)))
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700892 goto out;
Andrea Arcangelia9b85f92015-09-04 15:46:37 -0700893 if (uffdio_api.api != UFFD_API || uffdio_api.features) {
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700894 memset(&uffdio_api, 0, sizeof(uffdio_api));
895 if (copy_to_user(buf, &uffdio_api, sizeof(uffdio_api)))
896 goto out;
897 ret = -EINVAL;
898 goto out;
899 }
Pavel Emelyanov3f602d22015-09-04 15:46:34 -0700900 uffdio_api.features = UFFD_API_FEATURES;
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700901 uffdio_api.ioctls = UFFD_API_IOCTLS;
902 ret = -EFAULT;
903 if (copy_to_user(buf, &uffdio_api, sizeof(uffdio_api)))
904 goto out;
905 ctx->state = UFFD_STATE_RUNNING;
906 ret = 0;
907out:
908 return ret;
909}
910
911static long userfaultfd_ioctl(struct file *file, unsigned cmd,
912 unsigned long arg)
913{
914 int ret = -EINVAL;
915 struct userfaultfd_ctx *ctx = file->private_data;
916
917 switch(cmd) {
918 case UFFDIO_API:
919 ret = userfaultfd_api(ctx, arg);
920 break;
921 case UFFDIO_REGISTER:
922 ret = userfaultfd_register(ctx, arg);
923 break;
924 case UFFDIO_UNREGISTER:
925 ret = userfaultfd_unregister(ctx, arg);
926 break;
927 case UFFDIO_WAKE:
928 ret = userfaultfd_wake(ctx, arg);
929 break;
930 }
931 return ret;
932}
933
934#ifdef CONFIG_PROC_FS
935static void userfaultfd_show_fdinfo(struct seq_file *m, struct file *f)
936{
937 struct userfaultfd_ctx *ctx = f->private_data;
938 wait_queue_t *wq;
939 struct userfaultfd_wait_queue *uwq;
940 unsigned long pending = 0, total = 0;
941
942 spin_lock(&ctx->fault_wqh.lock);
943 list_for_each_entry(wq, &ctx->fault_wqh.task_list, task_list) {
944 uwq = container_of(wq, struct userfaultfd_wait_queue, wq);
945 if (uwq->pending)
946 pending++;
947 total++;
948 }
949 spin_unlock(&ctx->fault_wqh.lock);
950
951 /*
952 * If more protocols will be added, there will be all shown
953 * separated by a space. Like this:
954 * protocols: aa:... bb:...
955 */
956 seq_printf(m, "pending:\t%lu\ntotal:\t%lu\nAPI:\t%Lx:%x:%Lx\n",
Pavel Emelyanov3f602d22015-09-04 15:46:34 -0700957 pending, total, UFFD_API, UFFD_API_FEATURES,
Andrea Arcangeli86039bd2015-09-04 15:46:31 -0700958 UFFD_API_IOCTLS|UFFD_API_RANGE_IOCTLS);
959}
960#endif
961
962static const struct file_operations userfaultfd_fops = {
963#ifdef CONFIG_PROC_FS
964 .show_fdinfo = userfaultfd_show_fdinfo,
965#endif
966 .release = userfaultfd_release,
967 .poll = userfaultfd_poll,
968 .read = userfaultfd_read,
969 .unlocked_ioctl = userfaultfd_ioctl,
970 .compat_ioctl = userfaultfd_ioctl,
971 .llseek = noop_llseek,
972};
973
974/**
975 * userfaultfd_file_create - Creates an userfaultfd file pointer.
976 * @flags: Flags for the userfaultfd file.
977 *
978 * This function creates an userfaultfd file pointer, w/out installing
979 * it into the fd table. This is useful when the userfaultfd file is
980 * used during the initialization of data structures that require
981 * extra setup after the userfaultfd creation. So the userfaultfd
982 * creation is split into the file pointer creation phase, and the
983 * file descriptor installation phase. In this way races with
984 * userspace closing the newly installed file descriptor can be
985 * avoided. Returns an userfaultfd file pointer, or a proper error
986 * pointer.
987 */
988static struct file *userfaultfd_file_create(int flags)
989{
990 struct file *file;
991 struct userfaultfd_ctx *ctx;
992
993 BUG_ON(!current->mm);
994
995 /* Check the UFFD_* constants for consistency. */
996 BUILD_BUG_ON(UFFD_CLOEXEC != O_CLOEXEC);
997 BUILD_BUG_ON(UFFD_NONBLOCK != O_NONBLOCK);
998
999 file = ERR_PTR(-EINVAL);
1000 if (flags & ~UFFD_SHARED_FCNTL_FLAGS)
1001 goto out;
1002
1003 file = ERR_PTR(-ENOMEM);
1004 ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
1005 if (!ctx)
1006 goto out;
1007
1008 atomic_set(&ctx->refcount, 1);
1009 init_waitqueue_head(&ctx->fault_wqh);
1010 init_waitqueue_head(&ctx->fd_wqh);
1011 ctx->flags = flags;
1012 ctx->state = UFFD_STATE_WAIT_API;
1013 ctx->released = false;
1014 ctx->mm = current->mm;
1015 /* prevent the mm struct to be freed */
1016 atomic_inc(&ctx->mm->mm_users);
1017
1018 file = anon_inode_getfile("[userfaultfd]", &userfaultfd_fops, ctx,
1019 O_RDWR | (flags & UFFD_SHARED_FCNTL_FLAGS));
1020 if (IS_ERR(file))
1021 kfree(ctx);
1022out:
1023 return file;
1024}
1025
1026SYSCALL_DEFINE1(userfaultfd, int, flags)
1027{
1028 int fd, error;
1029 struct file *file;
1030
1031 error = get_unused_fd_flags(flags & UFFD_SHARED_FCNTL_FLAGS);
1032 if (error < 0)
1033 return error;
1034 fd = error;
1035
1036 file = userfaultfd_file_create(flags);
1037 if (IS_ERR(file)) {
1038 error = PTR_ERR(file);
1039 goto err_put_unused_fd;
1040 }
1041 fd_install(fd, file);
1042
1043 return fd;
1044
1045err_put_unused_fd:
1046 put_unused_fd(fd);
1047
1048 return error;
1049}