summaryrefslogtreecommitdiff
path: root/io_uring/waitid.c
blob: 6f851978606d988cfe2486d725d905b558b4a066 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
// SPDX-License-Identifier: GPL-2.0
/*
 * Support for async notification of waitid
 */
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/compat.h>
#include <linux/io_uring.h>

#include <uapi/linux/io_uring.h>

#include "io_uring.h"
#include "cancel.h"
#include "waitid.h"
#include "../kernel/exit.h"

static void io_waitid_cb(struct io_kiocb *req, struct io_tw_state *ts);

#define IO_WAITID_CANCEL_FLAG	BIT(31)
#define IO_WAITID_REF_MASK	GENMASK(30, 0)

struct io_waitid {
	struct file *file;
	int which;
	pid_t upid;
	int options;
	atomic_t refs;
	struct wait_queue_head *head;
	struct siginfo __user *infop;
	struct waitid_info info;
};

static void io_waitid_free(struct io_kiocb *req)
{
	struct io_waitid_async *iwa = req->async_data;

	put_pid(iwa->wo.wo_pid);
	kfree(req->async_data);
	req->async_data = NULL;
	req->flags &= ~REQ_F_ASYNC_DATA;
}

#ifdef CONFIG_COMPAT
static bool io_waitid_compat_copy_si(struct io_waitid *iw, int signo)
{
	struct compat_siginfo __user *infop;
	bool ret;

	infop = (struct compat_siginfo __user *) iw->infop;

	if (!user_write_access_begin(infop, sizeof(*infop)))
		return false;

	unsafe_put_user(signo, &infop->si_signo, Efault);
	unsafe_put_user(0, &infop->si_errno, Efault);
	unsafe_put_user(iw->info.cause, &infop->si_code, Efault);
	unsafe_put_user(iw->info.pid, &infop->si_pid, Efault);
	unsafe_put_user(iw->info.uid, &infop->si_uid, Efault);
	unsafe_put_user(iw->info.status, &infop->si_status, Efault);
	ret = true;
done:
	user_write_access_end();
	return ret;
Efault:
	ret = false;
	goto done;
}
#endif

static bool io_waitid_copy_si(struct io_kiocb *req, int signo)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	bool ret;

	if (!iw->infop)
		return true;

#ifdef CONFIG_COMPAT
	if (req->ctx->compat)
		return io_waitid_compat_copy_si(iw, signo);
#endif

	if (!user_write_access_begin(iw->infop, sizeof(*iw->infop)))
		return false;

	unsafe_put_user(signo, &iw->infop->si_signo, Efault);
	unsafe_put_user(0, &iw->infop->si_errno, Efault);
	unsafe_put_user(iw->info.cause, &iw->infop->si_code, Efault);
	unsafe_put_user(iw->info.pid, &iw->infop->si_pid, Efault);
	unsafe_put_user(iw->info.uid, &iw->infop->si_uid, Efault);
	unsafe_put_user(iw->info.status, &iw->infop->si_status, Efault);
	ret = true;
done:
	user_write_access_end();
	return ret;
Efault:
	ret = false;
	goto done;
}

static int io_waitid_finish(struct io_kiocb *req, int ret)
{
	int signo = 0;

	if (ret > 0) {
		signo = SIGCHLD;
		ret = 0;
	}

	if (!io_waitid_copy_si(req, signo))
		ret = -EFAULT;
	io_waitid_free(req);
	return ret;
}

static void io_waitid_complete(struct io_kiocb *req, int ret)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	struct io_tw_state ts = { .locked = true };

	/* anyone completing better be holding a reference */
	WARN_ON_ONCE(!(atomic_read(&iw->refs) & IO_WAITID_REF_MASK));

	lockdep_assert_held(&req->ctx->uring_lock);

	/*
	 * Did cancel find it meanwhile?
	 */
	if (hlist_unhashed(&req->hash_node))
		return;

	hlist_del_init(&req->hash_node);

	ret = io_waitid_finish(req, ret);
	if (ret < 0)
		req_set_fail(req);
	io_req_set_res(req, ret, 0);
	io_req_task_complete(req, &ts);
}

static bool __io_waitid_cancel(struct io_ring_ctx *ctx, struct io_kiocb *req)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	struct io_waitid_async *iwa = req->async_data;

	/*
	 * Mark us canceled regardless of ownership. This will prevent a
	 * potential retry from a spurious wakeup.
	 */
	atomic_or(IO_WAITID_CANCEL_FLAG, &iw->refs);

	/* claim ownership */
	if (atomic_fetch_inc(&iw->refs) & IO_WAITID_REF_MASK)
		return false;

	spin_lock_irq(&iw->head->lock);
	list_del_init(&iwa->wo.child_wait.entry);
	spin_unlock_irq(&iw->head->lock);
	io_waitid_complete(req, -ECANCELED);
	return true;
}

int io_waitid_cancel(struct io_ring_ctx *ctx, struct io_cancel_data *cd,
		     unsigned int issue_flags)
{
	struct hlist_node *tmp;
	struct io_kiocb *req;
	int nr = 0;

	if (cd->flags & (IORING_ASYNC_CANCEL_FD|IORING_ASYNC_CANCEL_FD_FIXED))
		return -ENOENT;

	io_ring_submit_lock(ctx, issue_flags);
	hlist_for_each_entry_safe(req, tmp, &ctx->waitid_list, hash_node) {
		if (req->cqe.user_data != cd->data &&
		    !(cd->flags & IORING_ASYNC_CANCEL_ANY))
			continue;
		if (__io_waitid_cancel(ctx, req))
			nr++;
		if (!(cd->flags & IORING_ASYNC_CANCEL_ALL))
			break;
	}
	io_ring_submit_unlock(ctx, issue_flags);

	if (nr)
		return nr;

	return -ENOENT;
}

bool io_waitid_remove_all(struct io_ring_ctx *ctx, struct task_struct *task,
			  bool cancel_all)
{
	struct hlist_node *tmp;
	struct io_kiocb *req;
	bool found = false;

	lockdep_assert_held(&ctx->uring_lock);

	hlist_for_each_entry_safe(req, tmp, &ctx->waitid_list, hash_node) {
		if (!io_match_task_safe(req, task, cancel_all))
			continue;
		__io_waitid_cancel(ctx, req);
		found = true;
	}

	return found;
}

static inline bool io_waitid_drop_issue_ref(struct io_kiocb *req)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	struct io_waitid_async *iwa = req->async_data;

	if (!atomic_sub_return(1, &iw->refs))
		return false;

	/*
	 * Wakeup triggered, racing with us. It was prevented from
	 * completing because of that, queue up the tw to do that.
	 */
	req->io_task_work.func = io_waitid_cb;
	io_req_task_work_add(req);
	remove_wait_queue(iw->head, &iwa->wo.child_wait);
	return true;
}

static void io_waitid_cb(struct io_kiocb *req, struct io_tw_state *ts)
{
	struct io_waitid_async *iwa = req->async_data;
	struct io_ring_ctx *ctx = req->ctx;
	int ret;

	io_tw_lock(ctx, ts);

	ret = __do_wait(&iwa->wo);

	/*
	 * If we get -ERESTARTSYS here, we need to re-arm and check again
	 * to ensure we get another callback. If the retry works, then we can
	 * just remove ourselves from the waitqueue again and finish the
	 * request.
	 */
	if (unlikely(ret == -ERESTARTSYS)) {
		struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);

		/* Don't retry if cancel found it meanwhile */
		ret = -ECANCELED;
		if (!(atomic_read(&iw->refs) & IO_WAITID_CANCEL_FLAG)) {
			iw->head = &current->signal->wait_chldexit;
			add_wait_queue(iw->head, &iwa->wo.child_wait);
			ret = __do_wait(&iwa->wo);
			if (ret == -ERESTARTSYS) {
				/* retry armed, drop our ref */
				io_waitid_drop_issue_ref(req);
				return;
			}

			remove_wait_queue(iw->head, &iwa->wo.child_wait);
		}
	}

	io_waitid_complete(req, ret);
}

static int io_waitid_wait(struct wait_queue_entry *wait, unsigned mode,
			  int sync, void *key)
{
	struct wait_opts *wo = container_of(wait, struct wait_opts, child_wait);
	struct io_waitid_async *iwa = container_of(wo, struct io_waitid_async, wo);
	struct io_kiocb *req = iwa->req;
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	struct task_struct *p = key;

	if (!pid_child_should_wake(wo, p))
		return 0;

	/* cancel is in progress */
	if (atomic_fetch_inc(&iw->refs) & IO_WAITID_REF_MASK)
		return 1;

	req->io_task_work.func = io_waitid_cb;
	io_req_task_work_add(req);
	list_del_init(&wait->entry);
	return 1;
}

int io_waitid_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);

	if (sqe->addr || sqe->buf_index || sqe->addr3 || sqe->waitid_flags)
		return -EINVAL;

	iw->which = READ_ONCE(sqe->len);
	iw->upid = READ_ONCE(sqe->fd);
	iw->options = READ_ONCE(sqe->file_index);
	iw->infop = u64_to_user_ptr(READ_ONCE(sqe->addr2));
	return 0;
}

int io_waitid(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_waitid *iw = io_kiocb_to_cmd(req, struct io_waitid);
	struct io_ring_ctx *ctx = req->ctx;
	struct io_waitid_async *iwa;
	int ret;

	if (io_alloc_async_data(req))
		return -ENOMEM;

	iwa = req->async_data;
	iwa->req = req;

	ret = kernel_waitid_prepare(&iwa->wo, iw->which, iw->upid, &iw->info,
					iw->options, NULL);
	if (ret)
		goto done;

	/*
	 * Mark the request as busy upfront, in case we're racing with the
	 * wakeup. If we are, then we'll notice when we drop this initial
	 * reference again after arming.
	 */
	atomic_set(&iw->refs, 1);

	/*
	 * Cancel must hold the ctx lock, so there's no risk of cancelation
	 * finding us until a) we remain on the list, and b) the lock is
	 * dropped. We only need to worry about racing with the wakeup
	 * callback.
	 */
	io_ring_submit_lock(ctx, issue_flags);
	hlist_add_head(&req->hash_node, &ctx->waitid_list);

	init_waitqueue_func_entry(&iwa->wo.child_wait, io_waitid_wait);
	iwa->wo.child_wait.private = req->task;
	iw->head = &current->signal->wait_chldexit;
	add_wait_queue(iw->head, &iwa->wo.child_wait);

	ret = __do_wait(&iwa->wo);
	if (ret == -ERESTARTSYS) {
		/*
		 * Nobody else grabbed a reference, it'll complete when we get
		 * a waitqueue callback, or if someone cancels it.
		 */
		if (!io_waitid_drop_issue_ref(req)) {
			io_ring_submit_unlock(ctx, issue_flags);
			return IOU_ISSUE_SKIP_COMPLETE;
		}

		/*
		 * Wakeup triggered, racing with us. It was prevented from
		 * completing because of that, queue up the tw to do that.
		 */
		io_ring_submit_unlock(ctx, issue_flags);
		return IOU_ISSUE_SKIP_COMPLETE;
	}

	hlist_del_init(&req->hash_node);
	remove_wait_queue(iw->head, &iwa->wo.child_wait);
	ret = io_waitid_finish(req, ret);

	io_ring_submit_unlock(ctx, issue_flags);
done:
	if (ret < 0)
		req_set_fail(req);
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}