1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /* In-kernel rxperf server for testing purposes.
3 *
4 * Copyright (C) 2022 Red Hat, Inc. All Rights Reserved.
5 * Written by David Howells (dhowells@redhat.com)
6 */
7
8 #define pr_fmt(fmt) "rxperf: " fmt
9 #include <linux/module.h>
10 #include <linux/slab.h>
11 #include <net/sock.h>
12 #include <net/af_rxrpc.h>
13 #define RXRPC_TRACE_ONLY_DEFINE_ENUMS
14 #include <trace/events/rxrpc.h>
15
16 MODULE_DESCRIPTION("rxperf test server (afs)");
17 MODULE_AUTHOR("Red Hat, Inc.");
18 MODULE_LICENSE("GPL");
19
20 #define RXPERF_PORT 7009
21 #define RX_PERF_SERVICE 147
22 #define RX_PERF_VERSION 3
23 #define RX_PERF_SEND 0
24 #define RX_PERF_RECV 1
25 #define RX_PERF_RPC 3
26 #define RX_PERF_FILE 4
27 #define RX_PERF_MAGIC_COOKIE 0x4711
28
29 struct rxperf_proto_params {
30 __be32 version;
31 __be32 type;
32 __be32 rsize;
33 __be32 wsize;
34 } __packed;
35
36 static const u8 rxperf_magic_cookie[] = { 0x00, 0x00, 0x47, 0x11 };
37 static const u8 secret[8] = { 0xa7, 0x83, 0x8a, 0xcb, 0xc7, 0x83, 0xec, 0x94 };
38
39 enum rxperf_call_state {
40 RXPERF_CALL_SV_AWAIT_PARAMS, /* Server: Awaiting parameter block */
41 RXPERF_CALL_SV_AWAIT_REQUEST, /* Server: Awaiting request data */
42 RXPERF_CALL_SV_REPLYING, /* Server: Replying */
43 RXPERF_CALL_SV_AWAIT_ACK, /* Server: Awaiting final ACK */
44 RXPERF_CALL_COMPLETE, /* Completed or failed */
45 };
46
47 struct rxperf_call {
48 struct rxrpc_call *rxcall;
49 struct iov_iter iter;
50 struct kvec kvec[1];
51 struct work_struct work;
52 const char *type;
53 size_t iov_len;
54 size_t req_len; /* Size of request blob */
55 size_t reply_len; /* Size of reply blob */
56 unsigned int debug_id;
57 unsigned int operation_id;
58 struct rxperf_proto_params params;
59 __be32 tmp[2];
60 s32 abort_code;
61 enum rxperf_call_state state;
62 short error;
63 unsigned short unmarshal;
64 u16 service_id;
65 int (*deliver)(struct rxperf_call *call);
66 void (*processor)(struct work_struct *work);
67 };
68
69 static struct socket *rxperf_socket;
70 static struct key *rxperf_sec_keyring; /* Ring of security/crypto keys */
71 static struct workqueue_struct *rxperf_workqueue;
72
73 static void rxperf_deliver_to_call(struct work_struct *work);
74 static int rxperf_deliver_param_block(struct rxperf_call *call);
75 static int rxperf_deliver_request(struct rxperf_call *call);
76 static int rxperf_process_call(struct rxperf_call *call);
77 static void rxperf_charge_preallocation(struct work_struct *work);
78
79 static DECLARE_WORK(rxperf_charge_preallocation_work,
80 rxperf_charge_preallocation);
81
rxperf_set_call_state(struct rxperf_call * call,enum rxperf_call_state to)82 static inline void rxperf_set_call_state(struct rxperf_call *call,
83 enum rxperf_call_state to)
84 {
85 call->state = to;
86 }
87
rxperf_set_call_complete(struct rxperf_call * call,int error,s32 remote_abort)88 static inline void rxperf_set_call_complete(struct rxperf_call *call,
89 int error, s32 remote_abort)
90 {
91 if (call->state != RXPERF_CALL_COMPLETE) {
92 call->abort_code = remote_abort;
93 call->error = error;
94 call->state = RXPERF_CALL_COMPLETE;
95 }
96 }
97
rxperf_rx_discard_new_call(struct rxrpc_call * rxcall,unsigned long user_call_ID)98 static void rxperf_rx_discard_new_call(struct rxrpc_call *rxcall,
99 unsigned long user_call_ID)
100 {
101 kfree((struct rxperf_call *)user_call_ID);
102 }
103
rxperf_rx_new_call(struct sock * sk,struct rxrpc_call * rxcall,unsigned long user_call_ID)104 static void rxperf_rx_new_call(struct sock *sk, struct rxrpc_call *rxcall,
105 unsigned long user_call_ID)
106 {
107 queue_work(rxperf_workqueue, &rxperf_charge_preallocation_work);
108 }
109
rxperf_queue_call_work(struct rxperf_call * call)110 static void rxperf_queue_call_work(struct rxperf_call *call)
111 {
112 queue_work(rxperf_workqueue, &call->work);
113 }
114
rxperf_notify_rx(struct sock * sk,struct rxrpc_call * rxcall,unsigned long call_user_ID)115 static void rxperf_notify_rx(struct sock *sk, struct rxrpc_call *rxcall,
116 unsigned long call_user_ID)
117 {
118 struct rxperf_call *call = (struct rxperf_call *)call_user_ID;
119
120 if (call->state != RXPERF_CALL_COMPLETE)
121 rxperf_queue_call_work(call);
122 }
123
rxperf_rx_attach(struct rxrpc_call * rxcall,unsigned long user_call_ID)124 static void rxperf_rx_attach(struct rxrpc_call *rxcall, unsigned long user_call_ID)
125 {
126 struct rxperf_call *call = (struct rxperf_call *)user_call_ID;
127
128 call->rxcall = rxcall;
129 }
130
rxperf_notify_end_reply_tx(struct sock * sock,struct rxrpc_call * rxcall,unsigned long call_user_ID)131 static void rxperf_notify_end_reply_tx(struct sock *sock,
132 struct rxrpc_call *rxcall,
133 unsigned long call_user_ID)
134 {
135 rxperf_set_call_state((struct rxperf_call *)call_user_ID,
136 RXPERF_CALL_SV_AWAIT_ACK);
137 }
138
139 /*
140 * Charge the incoming call preallocation.
141 */
rxperf_charge_preallocation(struct work_struct * work)142 static void rxperf_charge_preallocation(struct work_struct *work)
143 {
144 struct rxperf_call *call;
145
146 for (;;) {
147 call = kzalloc(sizeof(*call), GFP_KERNEL);
148 if (!call)
149 break;
150
151 call->type = "unset";
152 call->debug_id = atomic_inc_return(&rxrpc_debug_id);
153 call->deliver = rxperf_deliver_param_block;
154 call->state = RXPERF_CALL_SV_AWAIT_PARAMS;
155 call->service_id = RX_PERF_SERVICE;
156 call->iov_len = sizeof(call->params);
157 call->kvec[0].iov_len = sizeof(call->params);
158 call->kvec[0].iov_base = &call->params;
159 iov_iter_kvec(&call->iter, READ, call->kvec, 1, call->iov_len);
160 INIT_WORK(&call->work, rxperf_deliver_to_call);
161
162 if (rxrpc_kernel_charge_accept(rxperf_socket,
163 rxperf_notify_rx,
164 rxperf_rx_attach,
165 (unsigned long)call,
166 GFP_KERNEL,
167 call->debug_id) < 0)
168 break;
169 call = NULL;
170 }
171
172 kfree(call);
173 }
174
175 /*
176 * Open an rxrpc socket and bind it to be a server for callback notifications
177 * - the socket is left in blocking mode and non-blocking ops use MSG_DONTWAIT
178 */
rxperf_open_socket(void)179 static int rxperf_open_socket(void)
180 {
181 struct sockaddr_rxrpc srx;
182 struct socket *socket;
183 int ret;
184
185 ret = sock_create_kern(&init_net, AF_RXRPC, SOCK_DGRAM, PF_INET6,
186 &socket);
187 if (ret < 0)
188 goto error_1;
189
190 socket->sk->sk_allocation = GFP_NOFS;
191
192 /* bind the callback manager's address to make this a server socket */
193 memset(&srx, 0, sizeof(srx));
194 srx.srx_family = AF_RXRPC;
195 srx.srx_service = RX_PERF_SERVICE;
196 srx.transport_type = SOCK_DGRAM;
197 srx.transport_len = sizeof(srx.transport.sin6);
198 srx.transport.sin6.sin6_family = AF_INET6;
199 srx.transport.sin6.sin6_port = htons(RXPERF_PORT);
200
201 ret = rxrpc_sock_set_min_security_level(socket->sk,
202 RXRPC_SECURITY_ENCRYPT);
203 if (ret < 0)
204 goto error_2;
205
206 ret = rxrpc_sock_set_security_keyring(socket->sk, rxperf_sec_keyring);
207
208 ret = kernel_bind(socket, (struct sockaddr *)&srx, sizeof(srx));
209 if (ret < 0)
210 goto error_2;
211
212 rxrpc_kernel_new_call_notification(socket, rxperf_rx_new_call,
213 rxperf_rx_discard_new_call);
214
215 ret = kernel_listen(socket, INT_MAX);
216 if (ret < 0)
217 goto error_2;
218
219 rxperf_socket = socket;
220 rxperf_charge_preallocation(&rxperf_charge_preallocation_work);
221 return 0;
222
223 error_2:
224 sock_release(socket);
225 error_1:
226 pr_err("Can't set up rxperf socket: %d\n", ret);
227 return ret;
228 }
229
230 /*
231 * close the rxrpc socket rxperf was using
232 */
rxperf_close_socket(void)233 static void rxperf_close_socket(void)
234 {
235 kernel_listen(rxperf_socket, 0);
236 kernel_sock_shutdown(rxperf_socket, SHUT_RDWR);
237 flush_workqueue(rxperf_workqueue);
238 sock_release(rxperf_socket);
239 }
240
241 /*
242 * Log remote abort codes that indicate that we have a protocol disagreement
243 * with the server.
244 */
rxperf_log_error(struct rxperf_call * call,s32 remote_abort)245 static void rxperf_log_error(struct rxperf_call *call, s32 remote_abort)
246 {
247 static int max = 0;
248 const char *msg;
249 int m;
250
251 switch (remote_abort) {
252 case RX_EOF: msg = "unexpected EOF"; break;
253 case RXGEN_CC_MARSHAL: msg = "client marshalling"; break;
254 case RXGEN_CC_UNMARSHAL: msg = "client unmarshalling"; break;
255 case RXGEN_SS_MARSHAL: msg = "server marshalling"; break;
256 case RXGEN_SS_UNMARSHAL: msg = "server unmarshalling"; break;
257 case RXGEN_DECODE: msg = "opcode decode"; break;
258 case RXGEN_SS_XDRFREE: msg = "server XDR cleanup"; break;
259 case RXGEN_CC_XDRFREE: msg = "client XDR cleanup"; break;
260 case -32: msg = "insufficient data"; break;
261 default:
262 return;
263 }
264
265 m = max;
266 if (m < 3) {
267 max = m + 1;
268 pr_info("Peer reported %s failure on %s\n", msg, call->type);
269 }
270 }
271
272 /*
273 * deliver messages to a call
274 */
rxperf_deliver_to_call(struct work_struct * work)275 static void rxperf_deliver_to_call(struct work_struct *work)
276 {
277 struct rxperf_call *call = container_of(work, struct rxperf_call, work);
278 enum rxperf_call_state state;
279 u32 abort_code, remote_abort = 0;
280 int ret = 0;
281
282 if (call->state == RXPERF_CALL_COMPLETE)
283 return;
284
285 while (state = call->state,
286 state == RXPERF_CALL_SV_AWAIT_PARAMS ||
287 state == RXPERF_CALL_SV_AWAIT_REQUEST ||
288 state == RXPERF_CALL_SV_AWAIT_ACK
289 ) {
290 if (state == RXPERF_CALL_SV_AWAIT_ACK) {
291 if (!rxrpc_kernel_check_life(rxperf_socket, call->rxcall))
292 goto call_complete;
293 return;
294 }
295
296 ret = call->deliver(call);
297 if (ret == 0)
298 ret = rxperf_process_call(call);
299
300 switch (ret) {
301 case 0:
302 continue;
303 case -EINPROGRESS:
304 case -EAGAIN:
305 return;
306 case -ECONNABORTED:
307 rxperf_log_error(call, call->abort_code);
308 goto call_complete;
309 case -EOPNOTSUPP:
310 abort_code = RXGEN_OPCODE;
311 rxrpc_kernel_abort_call(rxperf_socket, call->rxcall,
312 abort_code, ret,
313 rxperf_abort_op_not_supported);
314 goto call_complete;
315 case -ENOTSUPP:
316 abort_code = RX_USER_ABORT;
317 rxrpc_kernel_abort_call(rxperf_socket, call->rxcall,
318 abort_code, ret,
319 rxperf_abort_op_not_supported);
320 goto call_complete;
321 case -EIO:
322 pr_err("Call %u in bad state %u\n",
323 call->debug_id, call->state);
324 fallthrough;
325 case -ENODATA:
326 case -EBADMSG:
327 case -EMSGSIZE:
328 case -ENOMEM:
329 case -EFAULT:
330 rxrpc_kernel_abort_call(rxperf_socket, call->rxcall,
331 RXGEN_SS_UNMARSHAL, ret,
332 rxperf_abort_unmarshal_error);
333 goto call_complete;
334 default:
335 rxrpc_kernel_abort_call(rxperf_socket, call->rxcall,
336 RX_CALL_DEAD, ret,
337 rxperf_abort_general_error);
338 goto call_complete;
339 }
340 }
341
342 call_complete:
343 rxperf_set_call_complete(call, ret, remote_abort);
344 /* The call may have been requeued */
345 rxrpc_kernel_end_call(rxperf_socket, call->rxcall);
346 cancel_work(&call->work);
347 kfree(call);
348 }
349
350 /*
351 * Extract a piece of data from the received data socket buffers.
352 */
rxperf_extract_data(struct rxperf_call * call,bool want_more)353 static int rxperf_extract_data(struct rxperf_call *call, bool want_more)
354 {
355 u32 remote_abort = 0;
356 int ret;
357
358 ret = rxrpc_kernel_recv_data(rxperf_socket, call->rxcall, &call->iter,
359 &call->iov_len, want_more, &remote_abort,
360 &call->service_id);
361 pr_debug("Extract i=%zu l=%zu m=%u ret=%d\n",
362 iov_iter_count(&call->iter), call->iov_len, want_more, ret);
363 if (ret == 0 || ret == -EAGAIN)
364 return ret;
365
366 if (ret == 1) {
367 switch (call->state) {
368 case RXPERF_CALL_SV_AWAIT_REQUEST:
369 rxperf_set_call_state(call, RXPERF_CALL_SV_REPLYING);
370 break;
371 case RXPERF_CALL_COMPLETE:
372 pr_debug("premature completion %d", call->error);
373 return call->error;
374 default:
375 break;
376 }
377 return 0;
378 }
379
380 rxperf_set_call_complete(call, ret, remote_abort);
381 return ret;
382 }
383
384 /*
385 * Grab the operation ID from an incoming manager call.
386 */
rxperf_deliver_param_block(struct rxperf_call * call)387 static int rxperf_deliver_param_block(struct rxperf_call *call)
388 {
389 u32 version;
390 int ret;
391
392 /* Extract the parameter block */
393 ret = rxperf_extract_data(call, true);
394 if (ret < 0)
395 return ret;
396
397 version = ntohl(call->params.version);
398 call->operation_id = ntohl(call->params.type);
399 call->deliver = rxperf_deliver_request;
400
401 if (version != RX_PERF_VERSION) {
402 pr_info("Version mismatch %x\n", version);
403 return -ENOTSUPP;
404 }
405
406 switch (call->operation_id) {
407 case RX_PERF_SEND:
408 call->type = "send";
409 call->reply_len = 0;
410 call->iov_len = 4; /* Expect req size */
411 break;
412 case RX_PERF_RECV:
413 call->type = "recv";
414 call->req_len = 0;
415 call->iov_len = 4; /* Expect reply size */
416 break;
417 case RX_PERF_RPC:
418 call->type = "rpc";
419 call->iov_len = 8; /* Expect req size and reply size */
420 break;
421 case RX_PERF_FILE:
422 call->type = "file";
423 fallthrough;
424 default:
425 return -EOPNOTSUPP;
426 }
427
428 rxperf_set_call_state(call, RXPERF_CALL_SV_AWAIT_REQUEST);
429 return call->deliver(call);
430 }
431
432 /*
433 * Deliver the request data.
434 */
rxperf_deliver_request(struct rxperf_call * call)435 static int rxperf_deliver_request(struct rxperf_call *call)
436 {
437 int ret;
438
439 switch (call->unmarshal) {
440 case 0:
441 call->kvec[0].iov_len = call->iov_len;
442 call->kvec[0].iov_base = call->tmp;
443 iov_iter_kvec(&call->iter, READ, call->kvec, 1, call->iov_len);
444 call->unmarshal++;
445 fallthrough;
446 case 1:
447 ret = rxperf_extract_data(call, true);
448 if (ret < 0)
449 return ret;
450
451 switch (call->operation_id) {
452 case RX_PERF_SEND:
453 call->type = "send";
454 call->req_len = ntohl(call->tmp[0]);
455 call->reply_len = 0;
456 break;
457 case RX_PERF_RECV:
458 call->type = "recv";
459 call->req_len = 0;
460 call->reply_len = ntohl(call->tmp[0]);
461 break;
462 case RX_PERF_RPC:
463 call->type = "rpc";
464 call->req_len = ntohl(call->tmp[0]);
465 call->reply_len = ntohl(call->tmp[1]);
466 break;
467 default:
468 pr_info("Can't parse extra params\n");
469 return -EIO;
470 }
471
472 pr_debug("CALL op=%s rq=%zx rp=%zx\n",
473 call->type, call->req_len, call->reply_len);
474
475 call->iov_len = call->req_len;
476 iov_iter_discard(&call->iter, READ, call->req_len);
477 call->unmarshal++;
478 fallthrough;
479 case 2:
480 ret = rxperf_extract_data(call, false);
481 if (ret < 0)
482 return ret;
483 call->unmarshal++;
484 fallthrough;
485 default:
486 return 0;
487 }
488 }
489
490 /*
491 * Process a call for which we've received the request.
492 */
rxperf_process_call(struct rxperf_call * call)493 static int rxperf_process_call(struct rxperf_call *call)
494 {
495 struct msghdr msg = {};
496 struct bio_vec bv;
497 struct kvec iov[1];
498 ssize_t n;
499 size_t reply_len = call->reply_len, len;
500
501 rxrpc_kernel_set_tx_length(rxperf_socket, call->rxcall,
502 reply_len + sizeof(rxperf_magic_cookie));
503
504 while (reply_len > 0) {
505 len = min_t(size_t, reply_len, PAGE_SIZE);
506 bvec_set_page(&bv, ZERO_PAGE(0), len, 0);
507 iov_iter_bvec(&msg.msg_iter, WRITE, &bv, 1, len);
508 msg.msg_flags = MSG_MORE;
509 n = rxrpc_kernel_send_data(rxperf_socket, call->rxcall, &msg,
510 len, rxperf_notify_end_reply_tx);
511 if (n < 0)
512 return n;
513 if (n == 0)
514 return -EIO;
515 reply_len -= n;
516 }
517
518 len = sizeof(rxperf_magic_cookie);
519 iov[0].iov_base = (void *)rxperf_magic_cookie;
520 iov[0].iov_len = len;
521 iov_iter_kvec(&msg.msg_iter, WRITE, iov, 1, len);
522 msg.msg_flags = 0;
523 n = rxrpc_kernel_send_data(rxperf_socket, call->rxcall, &msg, len,
524 rxperf_notify_end_reply_tx);
525 if (n >= 0)
526 return 0; /* Success */
527
528 if (n == -ENOMEM)
529 rxrpc_kernel_abort_call(rxperf_socket, call->rxcall,
530 RXGEN_SS_MARSHAL, -ENOMEM,
531 rxperf_abort_oom);
532 return n;
533 }
534
535 /*
536 * Add a key to the security keyring.
537 */
rxperf_add_key(struct key * keyring)538 static int rxperf_add_key(struct key *keyring)
539 {
540 key_ref_t kref;
541 int ret;
542
543 kref = key_create_or_update(make_key_ref(keyring, true),
544 "rxrpc_s",
545 __stringify(RX_PERF_SERVICE) ":2",
546 secret,
547 sizeof(secret),
548 KEY_POS_VIEW | KEY_POS_READ | KEY_POS_SEARCH
549 | KEY_USR_VIEW,
550 KEY_ALLOC_NOT_IN_QUOTA);
551
552 if (IS_ERR(kref)) {
553 pr_err("Can't allocate rxperf server key: %ld\n", PTR_ERR(kref));
554 return PTR_ERR(kref);
555 }
556
557 ret = key_link(keyring, key_ref_to_ptr(kref));
558 if (ret < 0)
559 pr_err("Can't link rxperf server key: %d\n", ret);
560 key_ref_put(kref);
561 return ret;
562 }
563
564 /*
565 * Initialise the rxperf server.
566 */
rxperf_init(void)567 static int __init rxperf_init(void)
568 {
569 struct key *keyring;
570 int ret = -ENOMEM;
571
572 pr_info("Server registering\n");
573
574 rxperf_workqueue = alloc_workqueue("rxperf", 0, 0);
575 if (!rxperf_workqueue)
576 goto error_workqueue;
577
578 keyring = keyring_alloc("rxperf_server",
579 GLOBAL_ROOT_UID, GLOBAL_ROOT_GID, current_cred(),
580 KEY_POS_VIEW | KEY_POS_READ | KEY_POS_SEARCH |
581 KEY_POS_WRITE |
582 KEY_USR_VIEW | KEY_USR_READ | KEY_USR_SEARCH |
583 KEY_USR_WRITE |
584 KEY_OTH_VIEW | KEY_OTH_READ | KEY_OTH_SEARCH,
585 KEY_ALLOC_NOT_IN_QUOTA,
586 NULL, NULL);
587 if (IS_ERR(keyring)) {
588 pr_err("Can't allocate rxperf server keyring: %ld\n",
589 PTR_ERR(keyring));
590 goto error_keyring;
591 }
592 rxperf_sec_keyring = keyring;
593 ret = rxperf_add_key(keyring);
594 if (ret < 0)
595 goto error_key;
596
597 ret = rxperf_open_socket();
598 if (ret < 0)
599 goto error_socket;
600 return 0;
601
602 error_socket:
603 error_key:
604 key_put(rxperf_sec_keyring);
605 error_keyring:
606 destroy_workqueue(rxperf_workqueue);
607 rcu_barrier();
608 error_workqueue:
609 pr_err("Failed to register: %d\n", ret);
610 return ret;
611 }
612 late_initcall(rxperf_init); /* Must be called after net/ to create socket */
613
rxperf_exit(void)614 static void __exit rxperf_exit(void)
615 {
616 pr_info("Server unregistering.\n");
617
618 rxperf_close_socket();
619 key_put(rxperf_sec_keyring);
620 destroy_workqueue(rxperf_workqueue);
621 rcu_barrier();
622 }
623 module_exit(rxperf_exit);
624
625