1 // SPDX-License-Identifier: GPL-2.0
2 #include <linux/ceph/ceph_debug.h>
4 #include <linux/bvec.h>
5 #include <linux/crc32c.h>
7 #include <linux/socket.h>
10 #include <linux/ceph/ceph_features.h>
11 #include <linux/ceph/decode.h>
12 #include <linux/ceph/libceph.h>
13 #include <linux/ceph/messenger.h>
15 /* static tag bytes (protocol control messages) */
16 static char tag_msg = CEPH_MSGR_TAG_MSG;
17 static char tag_ack = CEPH_MSGR_TAG_ACK;
18 static char tag_keepalive = CEPH_MSGR_TAG_KEEPALIVE;
19 static char tag_keepalive2 = CEPH_MSGR_TAG_KEEPALIVE2;
22 * If @buf is NULL, discard up to @len bytes.
24 static int ceph_tcp_recvmsg(struct socket *sock, void *buf, size_t len)
26 struct kvec iov = {buf, len};
27 struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
31 msg.msg_flags |= MSG_TRUNC;
33 iov_iter_kvec(&msg.msg_iter, ITER_DEST, &iov, 1, len);
34 r = sock_recvmsg(sock, &msg, msg.msg_flags);
40 static int ceph_tcp_recvpage(struct socket *sock, struct page *page,
41 int page_offset, size_t length)
44 struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
47 BUG_ON(page_offset + length > PAGE_SIZE);
48 bvec_set_page(&bvec, page, length, page_offset);
49 iov_iter_bvec(&msg.msg_iter, ITER_DEST, &bvec, 1, length);
50 r = sock_recvmsg(sock, &msg, msg.msg_flags);
57 * write something. @more is true if caller will be sending more data
60 static int ceph_tcp_sendmsg(struct socket *sock, struct kvec *iov,
61 size_t kvlen, size_t len, bool more)
63 struct msghdr msg = { .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL };
67 msg.msg_flags |= MSG_MORE;
69 msg.msg_flags |= MSG_EOR; /* superfluous, but what the hell */
71 r = kernel_sendmsg(sock, &msg, iov, kvlen, len);
78 * @more: MSG_MORE or 0.
80 static int ceph_tcp_sendpage(struct socket *sock, struct page *page,
81 int offset, size_t size, int more)
84 .msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL | more,
90 * MSG_SPLICE_PAGES cannot properly handle pages with page_count == 0,
91 * we need to fall back to sendmsg if that's the case.
93 * Same goes for slab pages: skb_can_coalesce() allows
94 * coalescing neighboring slab objects into a single frag which
95 * triggers one of hardened usercopy checks.
97 if (sendpage_ok(page))
98 msg.msg_flags |= MSG_SPLICE_PAGES;
100 bvec_set_page(&bvec, page, size, offset);
101 iov_iter_bvec(&msg.msg_iter, ITER_SOURCE, &bvec, 1, size);
103 ret = sock_sendmsg(sock, &msg);
110 static void con_out_kvec_reset(struct ceph_connection *con)
112 BUG_ON(con->v1.out_skip);
114 con->v1.out_kvec_left = 0;
115 con->v1.out_kvec_bytes = 0;
116 con->v1.out_kvec_cur = &con->v1.out_kvec[0];
119 static void con_out_kvec_add(struct ceph_connection *con,
120 size_t size, void *data)
122 int index = con->v1.out_kvec_left;
124 BUG_ON(con->v1.out_skip);
125 BUG_ON(index >= ARRAY_SIZE(con->v1.out_kvec));
127 con->v1.out_kvec[index].iov_len = size;
128 con->v1.out_kvec[index].iov_base = data;
129 con->v1.out_kvec_left++;
130 con->v1.out_kvec_bytes += size;
134 * Chop off a kvec from the end. Return residual number of bytes for
135 * that kvec, i.e. how many bytes would have been written if the kvec
138 static int con_out_kvec_skip(struct ceph_connection *con)
142 if (con->v1.out_kvec_bytes > 0) {
143 skip = con->v1.out_kvec_cur[con->v1.out_kvec_left - 1].iov_len;
144 BUG_ON(con->v1.out_kvec_bytes < skip);
145 BUG_ON(!con->v1.out_kvec_left);
146 con->v1.out_kvec_bytes -= skip;
147 con->v1.out_kvec_left--;
153 static size_t sizeof_footer(struct ceph_connection *con)
155 return (con->peer_features & CEPH_FEATURE_MSG_AUTH) ?
156 sizeof(struct ceph_msg_footer) :
157 sizeof(struct ceph_msg_footer_old);
160 static void prepare_message_data(struct ceph_msg *msg, u32 data_len)
162 /* Initialize data cursor if it's not a sparse read */
163 u64 len = msg->sparse_read_total ? : data_len;
165 ceph_msg_data_cursor_init(&msg->cursor, msg, len);
169 * Prepare footer for currently outgoing message, and finish things
170 * off. Assumes out_kvec* are already valid.. we just add on to the end.
172 static void prepare_write_message_footer(struct ceph_connection *con)
174 struct ceph_msg *m = con->out_msg;
176 m->footer.flags |= CEPH_MSG_FOOTER_COMPLETE;
178 dout("prepare_write_message_footer %p\n", con);
179 con_out_kvec_add(con, sizeof_footer(con), &m->footer);
180 if (con->peer_features & CEPH_FEATURE_MSG_AUTH) {
181 if (con->ops->sign_message)
182 con->ops->sign_message(m);
186 m->old_footer.flags = m->footer.flags;
188 con->v1.out_more = m->more_to_follow;
189 con->v1.out_msg_done = true;
193 * Prepare headers for the next outgoing message.
195 static void prepare_write_message(struct ceph_connection *con)
200 con_out_kvec_reset(con);
201 con->v1.out_msg_done = false;
203 /* Sneak an ack in there first? If we can get it into the same
204 * TCP packet that's a good thing. */
205 if (con->in_seq > con->in_seq_acked) {
206 con->in_seq_acked = con->in_seq;
207 con_out_kvec_add(con, sizeof (tag_ack), &tag_ack);
208 con->v1.out_temp_ack = cpu_to_le64(con->in_seq_acked);
209 con_out_kvec_add(con, sizeof(con->v1.out_temp_ack),
210 &con->v1.out_temp_ack);
213 ceph_con_get_out_msg(con);
216 dout("prepare_write_message %p seq %lld type %d len %d+%d+%zd\n",
217 m, con->out_seq, le16_to_cpu(m->hdr.type),
218 le32_to_cpu(m->hdr.front_len), le32_to_cpu(m->hdr.middle_len),
220 WARN_ON(m->front.iov_len != le32_to_cpu(m->hdr.front_len));
221 WARN_ON(m->data_length != le32_to_cpu(m->hdr.data_len));
223 /* tag + hdr + front + middle */
224 con_out_kvec_add(con, sizeof (tag_msg), &tag_msg);
225 con_out_kvec_add(con, sizeof(con->v1.out_hdr), &con->v1.out_hdr);
226 con_out_kvec_add(con, m->front.iov_len, m->front.iov_base);
229 con_out_kvec_add(con, m->middle->vec.iov_len,
230 m->middle->vec.iov_base);
232 /* fill in hdr crc and finalize hdr */
233 crc = crc32c(0, &m->hdr, offsetof(struct ceph_msg_header, crc));
234 con->out_msg->hdr.crc = cpu_to_le32(crc);
235 memcpy(&con->v1.out_hdr, &con->out_msg->hdr, sizeof(con->v1.out_hdr));
237 /* fill in front and middle crc, footer */
238 crc = crc32c(0, m->front.iov_base, m->front.iov_len);
239 con->out_msg->footer.front_crc = cpu_to_le32(crc);
241 crc = crc32c(0, m->middle->vec.iov_base,
242 m->middle->vec.iov_len);
243 con->out_msg->footer.middle_crc = cpu_to_le32(crc);
245 con->out_msg->footer.middle_crc = 0;
246 dout("%s front_crc %u middle_crc %u\n", __func__,
247 le32_to_cpu(con->out_msg->footer.front_crc),
248 le32_to_cpu(con->out_msg->footer.middle_crc));
249 con->out_msg->footer.flags = 0;
251 /* is there a data payload? */
252 con->out_msg->footer.data_crc = 0;
253 if (m->data_length) {
254 prepare_message_data(con->out_msg, m->data_length);
255 con->v1.out_more = 1; /* data + footer will follow */
257 /* no, queue up footer too and be done */
258 prepare_write_message_footer(con);
261 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
267 static void prepare_write_ack(struct ceph_connection *con)
269 dout("prepare_write_ack %p %llu -> %llu\n", con,
270 con->in_seq_acked, con->in_seq);
271 con->in_seq_acked = con->in_seq;
273 con_out_kvec_reset(con);
275 con_out_kvec_add(con, sizeof (tag_ack), &tag_ack);
277 con->v1.out_temp_ack = cpu_to_le64(con->in_seq_acked);
278 con_out_kvec_add(con, sizeof(con->v1.out_temp_ack),
279 &con->v1.out_temp_ack);
281 con->v1.out_more = 1; /* more will follow.. eventually.. */
282 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
286 * Prepare to share the seq during handshake
288 static void prepare_write_seq(struct ceph_connection *con)
290 dout("prepare_write_seq %p %llu -> %llu\n", con,
291 con->in_seq_acked, con->in_seq);
292 con->in_seq_acked = con->in_seq;
294 con_out_kvec_reset(con);
296 con->v1.out_temp_ack = cpu_to_le64(con->in_seq_acked);
297 con_out_kvec_add(con, sizeof(con->v1.out_temp_ack),
298 &con->v1.out_temp_ack);
300 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
304 * Prepare to write keepalive byte.
306 static void prepare_write_keepalive(struct ceph_connection *con)
308 dout("prepare_write_keepalive %p\n", con);
309 con_out_kvec_reset(con);
310 if (con->peer_features & CEPH_FEATURE_MSGR_KEEPALIVE2) {
311 struct timespec64 now;
313 ktime_get_real_ts64(&now);
314 con_out_kvec_add(con, sizeof(tag_keepalive2), &tag_keepalive2);
315 ceph_encode_timespec64(&con->v1.out_temp_keepalive2, &now);
316 con_out_kvec_add(con, sizeof(con->v1.out_temp_keepalive2),
317 &con->v1.out_temp_keepalive2);
319 con_out_kvec_add(con, sizeof(tag_keepalive), &tag_keepalive);
321 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
325 * Connection negotiation.
328 static int get_connect_authorizer(struct ceph_connection *con)
330 struct ceph_auth_handshake *auth;
333 if (!con->ops->get_authorizer) {
335 con->v1.out_connect.authorizer_protocol = CEPH_AUTH_UNKNOWN;
336 con->v1.out_connect.authorizer_len = 0;
340 auth = con->ops->get_authorizer(con, &auth_proto, con->v1.auth_retry);
342 return PTR_ERR(auth);
345 con->v1.out_connect.authorizer_protocol = cpu_to_le32(auth_proto);
346 con->v1.out_connect.authorizer_len =
347 cpu_to_le32(auth->authorizer_buf_len);
352 * We connected to a peer and are saying hello.
354 static void prepare_write_banner(struct ceph_connection *con)
356 con_out_kvec_add(con, strlen(CEPH_BANNER), CEPH_BANNER);
357 con_out_kvec_add(con, sizeof (con->msgr->my_enc_addr),
358 &con->msgr->my_enc_addr);
360 con->v1.out_more = 0;
361 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
364 static void __prepare_write_connect(struct ceph_connection *con)
366 con_out_kvec_add(con, sizeof(con->v1.out_connect),
367 &con->v1.out_connect);
369 con_out_kvec_add(con, con->v1.auth->authorizer_buf_len,
370 con->v1.auth->authorizer_buf);
372 con->v1.out_more = 0;
373 ceph_con_flag_set(con, CEPH_CON_F_WRITE_PENDING);
376 static int prepare_write_connect(struct ceph_connection *con)
378 unsigned int global_seq = ceph_get_global_seq(con->msgr, 0);
382 switch (con->peer_name.type) {
383 case CEPH_ENTITY_TYPE_MON:
384 proto = CEPH_MONC_PROTOCOL;
386 case CEPH_ENTITY_TYPE_OSD:
387 proto = CEPH_OSDC_PROTOCOL;
389 case CEPH_ENTITY_TYPE_MDS:
390 proto = CEPH_MDSC_PROTOCOL;
396 dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con,
397 con->v1.connect_seq, global_seq, proto);
399 con->v1.out_connect.features =
400 cpu_to_le64(from_msgr(con->msgr)->supported_features);
401 con->v1.out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT);
402 con->v1.out_connect.connect_seq = cpu_to_le32(con->v1.connect_seq);
403 con->v1.out_connect.global_seq = cpu_to_le32(global_seq);
404 con->v1.out_connect.protocol_version = cpu_to_le32(proto);
405 con->v1.out_connect.flags = 0;
407 ret = get_connect_authorizer(con);
411 __prepare_write_connect(con);
416 * write as much of pending kvecs to the socket as we can.
418 * 0 -> socket full, but more to do
421 static int write_partial_kvec(struct ceph_connection *con)
425 dout("write_partial_kvec %p %d left\n", con, con->v1.out_kvec_bytes);
426 while (con->v1.out_kvec_bytes > 0) {
427 ret = ceph_tcp_sendmsg(con->sock, con->v1.out_kvec_cur,
428 con->v1.out_kvec_left,
429 con->v1.out_kvec_bytes,
433 con->v1.out_kvec_bytes -= ret;
434 if (!con->v1.out_kvec_bytes)
437 /* account for full iov entries consumed */
438 while (ret >= con->v1.out_kvec_cur->iov_len) {
439 BUG_ON(!con->v1.out_kvec_left);
440 ret -= con->v1.out_kvec_cur->iov_len;
441 con->v1.out_kvec_cur++;
442 con->v1.out_kvec_left--;
444 /* and for a partially-consumed entry */
446 con->v1.out_kvec_cur->iov_len -= ret;
447 con->v1.out_kvec_cur->iov_base += ret;
450 con->v1.out_kvec_left = 0;
453 dout("write_partial_kvec %p %d left in %d kvecs ret = %d\n", con,
454 con->v1.out_kvec_bytes, con->v1.out_kvec_left, ret);
455 return ret; /* done! */
459 * Write as much message data payload as we can. If we finish, queue
461 * 1 -> done, footer is now queued in out_kvec[].
462 * 0 -> socket full, but more to do
465 static int write_partial_message_data(struct ceph_connection *con)
467 struct ceph_msg *msg = con->out_msg;
468 struct ceph_msg_data_cursor *cursor = &msg->cursor;
469 bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC);
472 dout("%s %p msg %p\n", __func__, con, msg);
474 if (!msg->num_data_items)
478 * Iterate through each page that contains data to be
479 * written, and send as much as possible for each.
481 * If we are calculating the data crc (the default), we will
482 * need to map the page. If we have no pages, they have
483 * been revoked, so use the zero page.
485 crc = do_datacrc ? le32_to_cpu(msg->footer.data_crc) : 0;
486 while (cursor->total_resid) {
492 if (!cursor->resid) {
493 ceph_msg_data_advance(cursor, 0);
497 page = ceph_msg_data_next(cursor, &page_offset, &length);
498 ret = ceph_tcp_sendpage(con->sock, page, page_offset, length,
502 msg->footer.data_crc = cpu_to_le32(crc);
506 if (do_datacrc && cursor->need_crc)
507 crc = ceph_crc32c_page(crc, page, page_offset, length);
508 ceph_msg_data_advance(cursor, (size_t)ret);
511 dout("%s %p msg %p done\n", __func__, con, msg);
513 /* prepare and queue up footer, too */
515 msg->footer.data_crc = cpu_to_le32(crc);
517 msg->footer.flags |= CEPH_MSG_FOOTER_NOCRC;
518 con_out_kvec_reset(con);
519 prepare_write_message_footer(con);
521 return 1; /* must return > 0 to indicate success */
527 static int write_partial_skip(struct ceph_connection *con)
531 dout("%s %p %d left\n", __func__, con, con->v1.out_skip);
532 while (con->v1.out_skip > 0) {
533 size_t size = min(con->v1.out_skip, (int)PAGE_SIZE);
535 ret = ceph_tcp_sendpage(con->sock, ceph_zero_page, 0, size,
539 con->v1.out_skip -= ret;
547 * Prepare to read connection handshake, or an ack.
549 static void prepare_read_banner(struct ceph_connection *con)
551 dout("prepare_read_banner %p\n", con);
552 con->v1.in_base_pos = 0;
555 static void prepare_read_connect(struct ceph_connection *con)
557 dout("prepare_read_connect %p\n", con);
558 con->v1.in_base_pos = 0;
561 static void prepare_read_ack(struct ceph_connection *con)
563 dout("prepare_read_ack %p\n", con);
564 con->v1.in_base_pos = 0;
567 static void prepare_read_seq(struct ceph_connection *con)
569 dout("prepare_read_seq %p\n", con);
570 con->v1.in_base_pos = 0;
571 con->v1.in_tag = CEPH_MSGR_TAG_SEQ;
574 static void prepare_read_tag(struct ceph_connection *con)
576 dout("prepare_read_tag %p\n", con);
577 con->v1.in_base_pos = 0;
578 con->v1.in_tag = CEPH_MSGR_TAG_READY;
581 static void prepare_read_keepalive_ack(struct ceph_connection *con)
583 dout("prepare_read_keepalive_ack %p\n", con);
584 con->v1.in_base_pos = 0;
588 * Prepare to read a message.
590 static int prepare_read_message(struct ceph_connection *con)
592 dout("prepare_read_message %p\n", con);
593 BUG_ON(con->in_msg != NULL);
594 con->v1.in_base_pos = 0;
595 con->in_front_crc = con->in_middle_crc = con->in_data_crc = 0;
599 static int read_partial(struct ceph_connection *con,
600 int end, int size, void *object)
602 while (con->v1.in_base_pos < end) {
603 int left = end - con->v1.in_base_pos;
604 int have = size - left;
605 int ret = ceph_tcp_recvmsg(con->sock, object + have, left);
608 con->v1.in_base_pos += ret;
614 * Read all or part of the connect-side handshake on a new connection
616 static int read_partial_banner(struct ceph_connection *con)
622 dout("read_partial_banner %p at %d\n", con, con->v1.in_base_pos);
625 size = strlen(CEPH_BANNER);
627 ret = read_partial(con, end, size, con->v1.in_banner);
631 size = sizeof(con->v1.actual_peer_addr);
633 ret = read_partial(con, end, size, &con->v1.actual_peer_addr);
636 ceph_decode_banner_addr(&con->v1.actual_peer_addr);
638 size = sizeof(con->v1.peer_addr_for_me);
640 ret = read_partial(con, end, size, &con->v1.peer_addr_for_me);
643 ceph_decode_banner_addr(&con->v1.peer_addr_for_me);
649 static int read_partial_connect(struct ceph_connection *con)
655 dout("read_partial_connect %p at %d\n", con, con->v1.in_base_pos);
657 size = sizeof(con->v1.in_reply);
659 ret = read_partial(con, end, size, &con->v1.in_reply);
664 size = le32_to_cpu(con->v1.in_reply.authorizer_len);
665 if (size > con->v1.auth->authorizer_reply_buf_len) {
666 pr_err("authorizer reply too big: %d > %zu\n", size,
667 con->v1.auth->authorizer_reply_buf_len);
673 ret = read_partial(con, end, size,
674 con->v1.auth->authorizer_reply_buf);
679 dout("read_partial_connect %p tag %d, con_seq = %u, g_seq = %u\n",
680 con, con->v1.in_reply.tag,
681 le32_to_cpu(con->v1.in_reply.connect_seq),
682 le32_to_cpu(con->v1.in_reply.global_seq));
688 * Verify the hello banner looks okay.
690 static int verify_hello(struct ceph_connection *con)
692 if (memcmp(con->v1.in_banner, CEPH_BANNER, strlen(CEPH_BANNER))) {
693 pr_err("connect to %s got bad banner\n",
694 ceph_pr_addr(&con->peer_addr));
695 con->error_msg = "protocol error, bad banner";
701 static int process_banner(struct ceph_connection *con)
703 struct ceph_entity_addr *my_addr = &con->msgr->inst.addr;
705 dout("process_banner on %p\n", con);
707 if (verify_hello(con) < 0)
711 * Make sure the other end is who we wanted. note that the other
712 * end may not yet know their ip address, so if it's 0.0.0.0, give
713 * them the benefit of the doubt.
715 if (memcmp(&con->peer_addr, &con->v1.actual_peer_addr,
716 sizeof(con->peer_addr)) != 0 &&
717 !(ceph_addr_is_blank(&con->v1.actual_peer_addr) &&
718 con->v1.actual_peer_addr.nonce == con->peer_addr.nonce)) {
719 pr_warn("wrong peer, want %s/%u, got %s/%u\n",
720 ceph_pr_addr(&con->peer_addr),
721 le32_to_cpu(con->peer_addr.nonce),
722 ceph_pr_addr(&con->v1.actual_peer_addr),
723 le32_to_cpu(con->v1.actual_peer_addr.nonce));
724 con->error_msg = "wrong peer at address";
729 * did we learn our address?
731 if (ceph_addr_is_blank(my_addr)) {
732 memcpy(&my_addr->in_addr,
733 &con->v1.peer_addr_for_me.in_addr,
734 sizeof(con->v1.peer_addr_for_me.in_addr));
735 ceph_addr_set_port(my_addr, 0);
736 ceph_encode_my_addr(con->msgr);
737 dout("process_banner learned my addr is %s\n",
738 ceph_pr_addr(my_addr));
744 static int process_connect(struct ceph_connection *con)
746 u64 sup_feat = from_msgr(con->msgr)->supported_features;
747 u64 req_feat = from_msgr(con->msgr)->required_features;
748 u64 server_feat = le64_to_cpu(con->v1.in_reply.features);
751 dout("process_connect on %p tag %d\n", con, con->v1.in_tag);
754 int len = le32_to_cpu(con->v1.in_reply.authorizer_len);
757 * Any connection that defines ->get_authorizer()
758 * should also define ->add_authorizer_challenge() and
759 * ->verify_authorizer_reply().
761 * See get_connect_authorizer().
763 if (con->v1.in_reply.tag ==
764 CEPH_MSGR_TAG_CHALLENGE_AUTHORIZER) {
765 ret = con->ops->add_authorizer_challenge(
766 con, con->v1.auth->authorizer_reply_buf, len);
770 con_out_kvec_reset(con);
771 __prepare_write_connect(con);
772 prepare_read_connect(con);
777 ret = con->ops->verify_authorizer_reply(con);
779 con->error_msg = "bad authorize reply";
785 switch (con->v1.in_reply.tag) {
786 case CEPH_MSGR_TAG_FEATURES:
787 pr_err("%s%lld %s feature set mismatch,"
788 " my %llx < server's %llx, missing %llx\n",
789 ENTITY_NAME(con->peer_name),
790 ceph_pr_addr(&con->peer_addr),
791 sup_feat, server_feat, server_feat & ~sup_feat);
792 con->error_msg = "missing required protocol features";
795 case CEPH_MSGR_TAG_BADPROTOVER:
796 pr_err("%s%lld %s protocol version mismatch,"
797 " my %d != server's %d\n",
798 ENTITY_NAME(con->peer_name),
799 ceph_pr_addr(&con->peer_addr),
800 le32_to_cpu(con->v1.out_connect.protocol_version),
801 le32_to_cpu(con->v1.in_reply.protocol_version));
802 con->error_msg = "protocol version mismatch";
805 case CEPH_MSGR_TAG_BADAUTHORIZER:
806 con->v1.auth_retry++;
807 dout("process_connect %p got BADAUTHORIZER attempt %d\n", con,
809 if (con->v1.auth_retry == 2) {
810 con->error_msg = "connect authorization failure";
813 con_out_kvec_reset(con);
814 ret = prepare_write_connect(con);
817 prepare_read_connect(con);
820 case CEPH_MSGR_TAG_RESETSESSION:
822 * If we connected with a large connect_seq but the peer
823 * has no record of a session with us (no connection, or
824 * connect_seq == 0), they will send RESETSESION to indicate
825 * that they must have reset their session, and may have
828 dout("process_connect got RESET peer seq %u\n",
829 le32_to_cpu(con->v1.in_reply.connect_seq));
830 pr_info("%s%lld %s session reset\n",
831 ENTITY_NAME(con->peer_name),
832 ceph_pr_addr(&con->peer_addr));
833 ceph_con_reset_session(con);
834 con_out_kvec_reset(con);
835 ret = prepare_write_connect(con);
838 prepare_read_connect(con);
840 /* Tell ceph about it. */
841 mutex_unlock(&con->mutex);
842 if (con->ops->peer_reset)
843 con->ops->peer_reset(con);
844 mutex_lock(&con->mutex);
845 if (con->state != CEPH_CON_S_V1_CONNECT_MSG)
849 case CEPH_MSGR_TAG_RETRY_SESSION:
851 * If we sent a smaller connect_seq than the peer has, try
852 * again with a larger value.
854 dout("process_connect got RETRY_SESSION my seq %u, peer %u\n",
855 le32_to_cpu(con->v1.out_connect.connect_seq),
856 le32_to_cpu(con->v1.in_reply.connect_seq));
857 con->v1.connect_seq = le32_to_cpu(con->v1.in_reply.connect_seq);
858 con_out_kvec_reset(con);
859 ret = prepare_write_connect(con);
862 prepare_read_connect(con);
865 case CEPH_MSGR_TAG_RETRY_GLOBAL:
867 * If we sent a smaller global_seq than the peer has, try
868 * again with a larger value.
870 dout("process_connect got RETRY_GLOBAL my %u peer_gseq %u\n",
871 con->v1.peer_global_seq,
872 le32_to_cpu(con->v1.in_reply.global_seq));
873 ceph_get_global_seq(con->msgr,
874 le32_to_cpu(con->v1.in_reply.global_seq));
875 con_out_kvec_reset(con);
876 ret = prepare_write_connect(con);
879 prepare_read_connect(con);
882 case CEPH_MSGR_TAG_SEQ:
883 case CEPH_MSGR_TAG_READY:
884 if (req_feat & ~server_feat) {
885 pr_err("%s%lld %s protocol feature mismatch,"
886 " my required %llx > server's %llx, need %llx\n",
887 ENTITY_NAME(con->peer_name),
888 ceph_pr_addr(&con->peer_addr),
889 req_feat, server_feat, req_feat & ~server_feat);
890 con->error_msg = "missing required protocol features";
894 WARN_ON(con->state != CEPH_CON_S_V1_CONNECT_MSG);
895 con->state = CEPH_CON_S_OPEN;
896 con->v1.auth_retry = 0; /* we authenticated; clear flag */
897 con->v1.peer_global_seq =
898 le32_to_cpu(con->v1.in_reply.global_seq);
899 con->v1.connect_seq++;
900 con->peer_features = server_feat;
901 dout("process_connect got READY gseq %d cseq %d (%d)\n",
902 con->v1.peer_global_seq,
903 le32_to_cpu(con->v1.in_reply.connect_seq),
904 con->v1.connect_seq);
905 WARN_ON(con->v1.connect_seq !=
906 le32_to_cpu(con->v1.in_reply.connect_seq));
908 if (con->v1.in_reply.flags & CEPH_MSG_CONNECT_LOSSY)
909 ceph_con_flag_set(con, CEPH_CON_F_LOSSYTX);
911 con->delay = 0; /* reset backoff memory */
913 if (con->v1.in_reply.tag == CEPH_MSGR_TAG_SEQ) {
914 prepare_write_seq(con);
915 prepare_read_seq(con);
917 prepare_read_tag(con);
921 case CEPH_MSGR_TAG_WAIT:
923 * If there is a connection race (we are opening
924 * connections to each other), one of us may just have
925 * to WAIT. This shouldn't happen if we are the
928 con->error_msg = "protocol error, got WAIT as client";
932 con->error_msg = "protocol error, garbage tag during connect";
939 * read (part of) an ack
941 static int read_partial_ack(struct ceph_connection *con)
943 int size = sizeof(con->v1.in_temp_ack);
946 return read_partial(con, end, size, &con->v1.in_temp_ack);
950 * We can finally discard anything that's been acked.
952 static void process_ack(struct ceph_connection *con)
954 u64 ack = le64_to_cpu(con->v1.in_temp_ack);
956 if (con->v1.in_tag == CEPH_MSGR_TAG_ACK)
957 ceph_con_discard_sent(con, ack);
959 ceph_con_discard_requeued(con, ack);
961 prepare_read_tag(con);
964 static int read_partial_message_chunk(struct ceph_connection *con,
965 struct kvec *section,
966 unsigned int sec_len, u32 *crc)
972 while (section->iov_len < sec_len) {
973 BUG_ON(section->iov_base == NULL);
974 left = sec_len - section->iov_len;
975 ret = ceph_tcp_recvmsg(con->sock, (char *)section->iov_base +
976 section->iov_len, left);
979 section->iov_len += ret;
981 if (section->iov_len == sec_len)
982 *crc = crc32c(*crc, section->iov_base, section->iov_len);
987 static inline int read_partial_message_section(struct ceph_connection *con,
988 struct kvec *section,
989 unsigned int sec_len, u32 *crc)
992 return read_partial_message_chunk(con, section, sec_len, crc);
995 static int read_partial_sparse_msg_extent(struct ceph_connection *con, u32 *crc)
997 struct ceph_msg_data_cursor *cursor = &con->in_msg->cursor;
998 bool do_bounce = ceph_test_opt(from_msgr(con->msgr), RXBOUNCE);
1000 if (do_bounce && unlikely(!con->bounce_page)) {
1001 con->bounce_page = alloc_page(GFP_NOIO);
1002 if (!con->bounce_page) {
1003 pr_err("failed to allocate bounce page\n");
1008 while (cursor->sr_resid > 0) {
1009 struct page *page, *rpage;
1013 page = ceph_msg_data_next(cursor, &off, &len);
1014 rpage = do_bounce ? con->bounce_page : page;
1016 /* clamp to what remains in extent */
1017 len = min_t(int, len, cursor->sr_resid);
1018 ret = ceph_tcp_recvpage(con->sock, rpage, (int)off, len);
1021 *crc = ceph_crc32c_page(*crc, rpage, off, ret);
1022 ceph_msg_data_advance(cursor, (size_t)ret);
1023 cursor->sr_resid -= ret;
1025 memcpy_page(page, off, rpage, off, ret);
1030 static int read_partial_sparse_msg_data(struct ceph_connection *con)
1032 struct ceph_msg_data_cursor *cursor = &con->in_msg->cursor;
1033 bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC);
1038 crc = con->in_data_crc;
1040 while (cursor->total_resid) {
1041 if (con->v1.in_sr_kvec.iov_base)
1042 ret = read_partial_message_chunk(con,
1043 &con->v1.in_sr_kvec,
1046 else if (cursor->sr_resid > 0)
1047 ret = read_partial_sparse_msg_extent(con, &crc);
1051 memset(&con->v1.in_sr_kvec, 0, sizeof(con->v1.in_sr_kvec));
1052 ret = con->ops->sparse_read(con, cursor,
1053 (char **)&con->v1.in_sr_kvec.iov_base);
1055 ret = ret ? ret : 1; /* must return > 0 to indicate success */
1058 con->v1.in_sr_len = ret;
1062 con->in_data_crc = crc;
1067 static int read_partial_msg_data(struct ceph_connection *con)
1069 struct ceph_msg_data_cursor *cursor = &con->in_msg->cursor;
1070 bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC);
1078 crc = con->in_data_crc;
1079 while (cursor->total_resid) {
1080 if (!cursor->resid) {
1081 ceph_msg_data_advance(cursor, 0);
1085 page = ceph_msg_data_next(cursor, &page_offset, &length);
1086 ret = ceph_tcp_recvpage(con->sock, page, page_offset, length);
1089 con->in_data_crc = crc;
1095 crc = ceph_crc32c_page(crc, page, page_offset, ret);
1096 ceph_msg_data_advance(cursor, (size_t)ret);
1099 con->in_data_crc = crc;
1101 return 1; /* must return > 0 to indicate success */
1104 static int read_partial_msg_data_bounce(struct ceph_connection *con)
1106 struct ceph_msg_data_cursor *cursor = &con->in_msg->cursor;
1112 if (unlikely(!con->bounce_page)) {
1113 con->bounce_page = alloc_page(GFP_NOIO);
1114 if (!con->bounce_page) {
1115 pr_err("failed to allocate bounce page\n");
1120 crc = con->in_data_crc;
1121 while (cursor->total_resid) {
1122 if (!cursor->resid) {
1123 ceph_msg_data_advance(cursor, 0);
1127 page = ceph_msg_data_next(cursor, &off, &len);
1128 ret = ceph_tcp_recvpage(con->sock, con->bounce_page, 0, len);
1130 con->in_data_crc = crc;
1134 crc = crc32c(crc, page_address(con->bounce_page), ret);
1135 memcpy_to_page(page, off, page_address(con->bounce_page), ret);
1137 ceph_msg_data_advance(cursor, ret);
1139 con->in_data_crc = crc;
1141 return 1; /* must return > 0 to indicate success */
1145 * read (part of) a message.
1147 static int read_partial_message(struct ceph_connection *con)
1149 struct ceph_msg *m = con->in_msg;
1153 unsigned int front_len, middle_len, data_len;
1154 bool do_datacrc = !ceph_test_opt(from_msgr(con->msgr), NOCRC);
1155 bool need_sign = (con->peer_features & CEPH_FEATURE_MSG_AUTH);
1159 dout("read_partial_message con %p msg %p\n", con, m);
1162 size = sizeof(con->v1.in_hdr);
1164 ret = read_partial(con, end, size, &con->v1.in_hdr);
1168 crc = crc32c(0, &con->v1.in_hdr, offsetof(struct ceph_msg_header, crc));
1169 if (cpu_to_le32(crc) != con->v1.in_hdr.crc) {
1170 pr_err("read_partial_message bad hdr crc %u != expected %u\n",
1171 crc, con->v1.in_hdr.crc);
1175 front_len = le32_to_cpu(con->v1.in_hdr.front_len);
1176 if (front_len > CEPH_MSG_MAX_FRONT_LEN)
1178 middle_len = le32_to_cpu(con->v1.in_hdr.middle_len);
1179 if (middle_len > CEPH_MSG_MAX_MIDDLE_LEN)
1181 data_len = le32_to_cpu(con->v1.in_hdr.data_len);
1182 if (data_len > CEPH_MSG_MAX_DATA_LEN)
1186 seq = le64_to_cpu(con->v1.in_hdr.seq);
1187 if ((s64)seq - (s64)con->in_seq < 1) {
1188 pr_info("skipping %s%lld %s seq %lld expected %lld\n",
1189 ENTITY_NAME(con->peer_name),
1190 ceph_pr_addr(&con->peer_addr),
1191 seq, con->in_seq + 1);
1192 con->v1.in_base_pos = -front_len - middle_len - data_len -
1194 con->v1.in_tag = CEPH_MSGR_TAG_READY;
1196 } else if ((s64)seq - (s64)con->in_seq > 1) {
1197 pr_err("read_partial_message bad seq %lld expected %lld\n",
1198 seq, con->in_seq + 1);
1199 con->error_msg = "bad message sequence # for incoming message";
1203 /* allocate message? */
1207 dout("got hdr type %d front %d data %d\n", con->v1.in_hdr.type,
1208 front_len, data_len);
1209 ret = ceph_con_in_msg_alloc(con, &con->v1.in_hdr, &skip);
1213 BUG_ON((!con->in_msg) ^ skip);
1215 /* skip this message */
1216 dout("alloc_msg said skip message\n");
1217 con->v1.in_base_pos = -front_len - middle_len -
1218 data_len - sizeof_footer(con);
1219 con->v1.in_tag = CEPH_MSGR_TAG_READY;
1224 BUG_ON(!con->in_msg);
1225 BUG_ON(con->in_msg->con != con);
1227 m->front.iov_len = 0; /* haven't read it yet */
1229 m->middle->vec.iov_len = 0;
1231 /* prepare for data payload, if any */
1234 prepare_message_data(con->in_msg, data_len);
1238 ret = read_partial_message_section(con, &m->front, front_len,
1239 &con->in_front_crc);
1245 ret = read_partial_message_section(con, &m->middle->vec,
1247 &con->in_middle_crc);
1254 if (!m->num_data_items)
1257 if (m->sparse_read_total)
1258 ret = read_partial_sparse_msg_data(con);
1259 else if (ceph_test_opt(from_msgr(con->msgr), RXBOUNCE))
1260 ret = read_partial_msg_data_bounce(con);
1262 ret = read_partial_msg_data(con);
1268 size = sizeof_footer(con);
1270 ret = read_partial(con, end, size, &m->footer);
1275 m->footer.flags = m->old_footer.flags;
1279 dout("read_partial_message got msg %p %d (%u) + %d (%u) + %d (%u)\n",
1280 m, front_len, m->footer.front_crc, middle_len,
1281 m->footer.middle_crc, data_len, m->footer.data_crc);
1284 if (con->in_front_crc != le32_to_cpu(m->footer.front_crc)) {
1285 pr_err("read_partial_message %p front crc %u != exp. %u\n",
1286 m, con->in_front_crc, m->footer.front_crc);
1289 if (con->in_middle_crc != le32_to_cpu(m->footer.middle_crc)) {
1290 pr_err("read_partial_message %p middle crc %u != exp %u\n",
1291 m, con->in_middle_crc, m->footer.middle_crc);
1295 (m->footer.flags & CEPH_MSG_FOOTER_NOCRC) == 0 &&
1296 con->in_data_crc != le32_to_cpu(m->footer.data_crc)) {
1297 pr_err("read_partial_message %p data crc %u != exp. %u\n", m,
1298 con->in_data_crc, le32_to_cpu(m->footer.data_crc));
1302 if (need_sign && con->ops->check_message_signature &&
1303 con->ops->check_message_signature(m)) {
1304 pr_err("read_partial_message %p signature check failed\n", m);
1308 return 1; /* done! */
1311 static int read_keepalive_ack(struct ceph_connection *con)
1313 struct ceph_timespec ceph_ts;
1314 size_t size = sizeof(ceph_ts);
1315 int ret = read_partial(con, size, size, &ceph_ts);
1318 ceph_decode_timespec64(&con->last_keepalive_ack, &ceph_ts);
1319 prepare_read_tag(con);
1324 * Read what we can from the socket.
1326 int ceph_con_v1_try_read(struct ceph_connection *con)
1331 dout("try_read start %p state %d\n", con, con->state);
1332 if (con->state != CEPH_CON_S_V1_BANNER &&
1333 con->state != CEPH_CON_S_V1_CONNECT_MSG &&
1334 con->state != CEPH_CON_S_OPEN)
1339 dout("try_read tag %d in_base_pos %d\n", con->v1.in_tag,
1340 con->v1.in_base_pos);
1342 if (con->state == CEPH_CON_S_V1_BANNER) {
1343 ret = read_partial_banner(con);
1346 ret = process_banner(con);
1350 con->state = CEPH_CON_S_V1_CONNECT_MSG;
1353 * Received banner is good, exchange connection info.
1354 * Do not reset out_kvec, as sending our banner raced
1355 * with receiving peer banner after connect completed.
1357 ret = prepare_write_connect(con);
1360 prepare_read_connect(con);
1362 /* Send connection info before awaiting response */
1366 if (con->state == CEPH_CON_S_V1_CONNECT_MSG) {
1367 ret = read_partial_connect(con);
1370 ret = process_connect(con);
1376 WARN_ON(con->state != CEPH_CON_S_OPEN);
1378 if (con->v1.in_base_pos < 0) {
1380 * skipping + discarding content.
1382 ret = ceph_tcp_recvmsg(con->sock, NULL, -con->v1.in_base_pos);
1385 dout("skipped %d / %d bytes\n", ret, -con->v1.in_base_pos);
1386 con->v1.in_base_pos += ret;
1387 if (con->v1.in_base_pos)
1390 if (con->v1.in_tag == CEPH_MSGR_TAG_READY) {
1394 ret = ceph_tcp_recvmsg(con->sock, &con->v1.in_tag, 1);
1397 dout("try_read got tag %d\n", con->v1.in_tag);
1398 switch (con->v1.in_tag) {
1399 case CEPH_MSGR_TAG_MSG:
1400 prepare_read_message(con);
1402 case CEPH_MSGR_TAG_ACK:
1403 prepare_read_ack(con);
1405 case CEPH_MSGR_TAG_KEEPALIVE2_ACK:
1406 prepare_read_keepalive_ack(con);
1408 case CEPH_MSGR_TAG_CLOSE:
1409 ceph_con_close_socket(con);
1410 con->state = CEPH_CON_S_CLOSED;
1416 if (con->v1.in_tag == CEPH_MSGR_TAG_MSG) {
1417 ret = read_partial_message(con);
1421 con->error_msg = "bad crc/signature";
1427 con->error_msg = "io error";
1432 if (con->v1.in_tag == CEPH_MSGR_TAG_READY)
1434 ceph_con_process_message(con);
1435 if (con->state == CEPH_CON_S_OPEN)
1436 prepare_read_tag(con);
1439 if (con->v1.in_tag == CEPH_MSGR_TAG_ACK ||
1440 con->v1.in_tag == CEPH_MSGR_TAG_SEQ) {
1442 * the final handshake seq exchange is semantically
1443 * equivalent to an ACK
1445 ret = read_partial_ack(con);
1451 if (con->v1.in_tag == CEPH_MSGR_TAG_KEEPALIVE2_ACK) {
1452 ret = read_keepalive_ack(con);
1459 dout("try_read done on %p ret %d\n", con, ret);
1463 pr_err("try_read bad tag %d\n", con->v1.in_tag);
1464 con->error_msg = "protocol error, garbage tag";
1470 * Write something to the socket. Called in a worker thread when the
1471 * socket appears to be writeable and we have something ready to send.
1473 int ceph_con_v1_try_write(struct ceph_connection *con)
1477 dout("try_write start %p state %d\n", con, con->state);
1478 if (con->state != CEPH_CON_S_PREOPEN &&
1479 con->state != CEPH_CON_S_V1_BANNER &&
1480 con->state != CEPH_CON_S_V1_CONNECT_MSG &&
1481 con->state != CEPH_CON_S_OPEN)
1484 /* open the socket first? */
1485 if (con->state == CEPH_CON_S_PREOPEN) {
1487 con->state = CEPH_CON_S_V1_BANNER;
1489 con_out_kvec_reset(con);
1490 prepare_write_banner(con);
1491 prepare_read_banner(con);
1493 BUG_ON(con->in_msg);
1494 con->v1.in_tag = CEPH_MSGR_TAG_READY;
1495 dout("try_write initiating connect on %p new state %d\n",
1497 ret = ceph_tcp_connect(con);
1499 con->error_msg = "connect error";
1505 dout("try_write out_kvec_bytes %d\n", con->v1.out_kvec_bytes);
1508 /* kvec data queued? */
1509 if (con->v1.out_kvec_left) {
1510 ret = write_partial_kvec(con);
1514 if (con->v1.out_skip) {
1515 ret = write_partial_skip(con);
1522 if (con->v1.out_msg_done) {
1523 ceph_msg_put(con->out_msg);
1524 con->out_msg = NULL; /* we're done with this one */
1528 ret = write_partial_message_data(con);
1530 goto more; /* we need to send the footer, too! */
1534 dout("try_write write_partial_message_data err %d\n",
1541 if (con->state == CEPH_CON_S_OPEN) {
1542 if (ceph_con_flag_test_and_clear(con,
1543 CEPH_CON_F_KEEPALIVE_PENDING)) {
1544 prepare_write_keepalive(con);
1547 /* is anything else pending? */
1548 if (!list_empty(&con->out_queue)) {
1549 prepare_write_message(con);
1552 if (con->in_seq > con->in_seq_acked) {
1553 prepare_write_ack(con);
1558 /* Nothing to do! */
1559 ceph_con_flag_clear(con, CEPH_CON_F_WRITE_PENDING);
1560 dout("try_write nothing else to write.\n");
1563 dout("try_write done on %p ret %d\n", con, ret);
1567 void ceph_con_v1_revoke(struct ceph_connection *con)
1569 struct ceph_msg *msg = con->out_msg;
1571 WARN_ON(con->v1.out_skip);
1573 if (con->v1.out_msg_done) {
1574 con->v1.out_skip += con_out_kvec_skip(con);
1576 WARN_ON(!msg->data_length);
1577 con->v1.out_skip += sizeof_footer(con);
1579 /* data, middle, front */
1580 if (msg->data_length)
1581 con->v1.out_skip += msg->cursor.total_resid;
1583 con->v1.out_skip += con_out_kvec_skip(con);
1584 con->v1.out_skip += con_out_kvec_skip(con);
1586 dout("%s con %p out_kvec_bytes %d out_skip %d\n", __func__, con,
1587 con->v1.out_kvec_bytes, con->v1.out_skip);
1590 void ceph_con_v1_revoke_incoming(struct ceph_connection *con)
1592 unsigned int front_len = le32_to_cpu(con->v1.in_hdr.front_len);
1593 unsigned int middle_len = le32_to_cpu(con->v1.in_hdr.middle_len);
1594 unsigned int data_len = le32_to_cpu(con->v1.in_hdr.data_len);
1596 /* skip rest of message */
1597 con->v1.in_base_pos = con->v1.in_base_pos -
1598 sizeof(struct ceph_msg_header) -
1602 sizeof(struct ceph_msg_footer);
1604 con->v1.in_tag = CEPH_MSGR_TAG_READY;
1607 dout("%s con %p in_base_pos %d\n", __func__, con, con->v1.in_base_pos);
1610 bool ceph_con_v1_opened(struct ceph_connection *con)
1612 return con->v1.connect_seq;
1615 void ceph_con_v1_reset_session(struct ceph_connection *con)
1617 con->v1.connect_seq = 0;
1618 con->v1.peer_global_seq = 0;
1621 void ceph_con_v1_reset_protocol(struct ceph_connection *con)
1623 con->v1.out_skip = 0;