summaryrefslogtreecommitdiffstats
path: root/src/plugins/quic
diff options
context:
space:
mode:
authorNathan Skrzypczak <nathan.skrzypczak@gmail.com>2019-03-19 13:57:31 +0100
committerFlorin Coras <florin.coras@gmail.com>2019-03-29 15:57:30 +0000
commit60f3e6547aa75a31fad1f78bea11c950cc30d63c (patch)
tree4df1f649f10272c5f276754d6b0dbecaaa08784a /src/plugins/quic
parentdb4cae5d438e5114a19df38acbb9e1dd6c99fe3f (diff)
Integrate first QUIC protocol implementation
Currently supports on single stream exposed through standard internal APIs Based on libquicly & picotls by h2o Change-Id: I7bc1ec0e399d1fb02bfd1da91aa7410076d08d14 Signed-off-by: Nathan Skrzypczak <nathan.skrzypczak@gmail.com>
Diffstat (limited to 'src/plugins/quic')
-rw-r--r--src/plugins/quic/CMakeLists.txt43
-rw-r--r--src/plugins/quic/quic.c1592
-rw-r--r--src/plugins/quic/quic.h118
3 files changed, 1753 insertions, 0 deletions
diff --git a/src/plugins/quic/CMakeLists.txt b/src/plugins/quic/CMakeLists.txt
new file mode 100644
index 00000000000..ccd49682e8c
--- /dev/null
+++ b/src/plugins/quic/CMakeLists.txt
@@ -0,0 +1,43 @@
+
+# Copyright (c) 2019 Cisco
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at:
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+message(STATUS "Looking for quicly")
+unset(QUIC_LINK_LIBRARIES)
+
+find_path(QUICLY_INCLUDE_DIR NAMES quicly.h)
+find_library (QUICLY_LIBRARY NAMES "libquicly.a")
+find_library (PICOTLS_CORE_LIBRARY NAMES "libpicotls-core.a")
+find_library (PICOTLS_OPENSSL_LIBRARY NAMES "libpicotls-openssl.a")
+
+list(APPEND QUIC_LINK_LIBRARIES
+ ${QUICLY_LIBRARY}
+ ${PICOTLS_CORE_LIBRARY}
+ ${PICOTLS_OPENSSL_LIBRARY}
+)
+
+if(QUICLY_INCLUDE_DIR AND QUIC_LINK_LIBRARIES)
+ include_directories (${QUICLY_INCLUDE_DIR})
+ add_vpp_plugin(quic
+ SOURCES
+ quic.c
+
+ INSTALL_HEADERS
+ quic.h
+
+ LINK_LIBRARIES ${QUIC_LINK_LIBRARIES}
+ )
+ message(STATUS "Found quicly in ${QUICLY_INCLUDE_DIR}")
+else()
+ message(WARNING "-- quicly not found - quic_plugin disabled")
+endif()
diff --git a/src/plugins/quic/quic.c b/src/plugins/quic/quic.c
new file mode 100644
index 00000000000..82de6de130b
--- /dev/null
+++ b/src/plugins/quic/quic.c
@@ -0,0 +1,1592 @@
+/*
+ * Copyright (c) 2019 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <sys/socket.h>
+
+#include <vnet/session/application.h>
+#include <vnet/session/transport.h>
+#include <vnet/session/session.h>
+#include <vlib/unix/plugin.h>
+#include <vpp/app/version.h>
+#include <openssl/pem.h>
+
+#include <vppinfra/lock.h>
+
+#include <quic/quic.h>
+
+#include <quicly/streambuf.h>
+#include <picotls/openssl.h>
+#include <picotls/pembase64.h>
+
+static quic_main_t quic_main;
+
+static void quic_update_timer (quic_ctx_t * ctx);
+static int64_t quic_get_time (quicly_now_cb * self);
+static void quic_connection_closed (u32 conn_index);
+static void quic_disconnect (u32 ctx_index, u32 thread_index);
+
+#define QUIC_INT_MAX 0x3FFFFFFFFFFFFFFF
+
+u32
+quic_ctx_half_open_alloc (void)
+{
+ quic_main_t *qm = &quic_main;
+ u8 will_expand = 0;
+ quic_ctx_t *ctx;
+ u32 ctx_index;
+
+ pool_get_aligned_will_expand (qm->half_open_ctx_pool, will_expand, 0);
+ if (PREDICT_FALSE (will_expand && vlib_num_workers ()))
+ {
+ clib_rwlock_writer_lock (&qm->half_open_rwlock);
+ pool_get (qm->half_open_ctx_pool, ctx);
+ ctx_index = ctx - qm->half_open_ctx_pool;
+ clib_rwlock_writer_unlock (&qm->half_open_rwlock);
+ }
+ else
+ {
+ /* reader lock assumption: only main thread will call pool_get */
+ clib_rwlock_reader_lock (&qm->half_open_rwlock);
+ pool_get (qm->half_open_ctx_pool, ctx);
+ ctx_index = ctx - qm->half_open_ctx_pool;
+ clib_rwlock_reader_unlock (&qm->half_open_rwlock);
+ }
+ memset (ctx, 0, sizeof (*ctx));
+ return ctx_index;
+}
+
+void
+quic_ctx_half_open_free (u32 ho_index)
+{
+ quic_main_t *qm = &quic_main;
+ clib_rwlock_writer_lock (&qm->half_open_rwlock);
+ pool_put_index (qm->half_open_ctx_pool, ho_index);
+ clib_rwlock_writer_unlock (&qm->half_open_rwlock);
+}
+
+quic_ctx_t *
+quic_ctx_half_open_get (u32 ctx_index)
+{
+ quic_main_t *qm = &quic_main;
+ clib_rwlock_reader_lock (&qm->half_open_rwlock);
+ return pool_elt_at_index (qm->half_open_ctx_pool, ctx_index);
+}
+
+void
+quic_ctx_half_open_reader_unlock ()
+{
+ clib_rwlock_reader_unlock (&quic_main.half_open_rwlock);
+}
+
+u32
+quic_ctx_half_open_index (quic_ctx_t * ctx)
+{
+ return (ctx - quic_main.half_open_ctx_pool);
+}
+
+u32
+quic_ctx_alloc ()
+{
+ u8 thread_index = vlib_get_thread_index ();
+ quic_main_t *qm = &quic_main;
+ quic_ctx_t *ctx;
+
+ pool_get (qm->ctx_pool[thread_index], ctx);
+
+ memset (ctx, 0, sizeof (quic_ctx_t));
+ ctx->c_thread_index = thread_index;
+ return ctx - qm->ctx_pool[thread_index];
+}
+
+static void
+quic_ctx_free (quic_ctx_t * ctx)
+{
+ QUIC_DBG (2, "Free ctx %u", ctx->c_c_index);
+ u32 thread_index = ctx->c_thread_index;
+ if (CLIB_DEBUG)
+ memset (ctx, 0xfb, sizeof (*ctx));
+ pool_put (quic_main.ctx_pool[thread_index], ctx);
+}
+
+static quic_ctx_t *
+quic_ctx_get (u32 ctx_index)
+{
+ return pool_elt_at_index (quic_main.ctx_pool[vlib_get_thread_index ()],
+ ctx_index);
+}
+
+static quic_ctx_t *
+quic_ctx_get_w_thread (u32 ctx_index, u8 thread_index)
+{
+ return pool_elt_at_index (quic_main.ctx_pool[thread_index], ctx_index);
+}
+
+static void
+quic_disconnect_transport (quic_ctx_t * ctx)
+{
+ QUIC_DBG (2, "Called quic_disconnect_transport");
+ vnet_disconnect_args_t a = {
+ .handle = ctx->c_quic_ctx_id.quic_session,
+ .app_index = quic_main.app_index,
+ };
+
+ if (vnet_disconnect_session (&a))
+ clib_warning ("UDP session disconnect errored");
+}
+
+static int
+quic_send_datagram (session_t * session, quicly_datagram_t * packet)
+{
+ QUIC_DBG (2, "Called quic_send_datagram at %ld", quic_get_time (NULL));
+ u32 max_enqueue;
+ session_dgram_hdr_t hdr;
+ int rv;
+ u32 len;
+ svm_fifo_t *f;
+ transport_connection_t *tc;
+
+ len = packet->data.len;
+ f = session->tx_fifo;
+ tc = session_get_transport (session);
+
+ max_enqueue = svm_fifo_max_enqueue (f);
+ if (max_enqueue <= sizeof (session_dgram_hdr_t))
+ return 1;
+
+ max_enqueue -= sizeof (session_dgram_hdr_t);
+
+ if (max_enqueue < len)
+ return 1;
+
+ // Build packet header for fifo
+ hdr.data_length = len;
+ hdr.data_offset = 0;
+ hdr.is_ip4 = tc->is_ip4;
+ clib_memcpy (&hdr.lcl_ip, &tc->lcl_ip, sizeof (ip46_address_t));
+ hdr.lcl_port = tc->lcl_port;
+
+ // Read dest address from quicly-provided sockaddr
+ if (hdr.is_ip4)
+ {
+ ASSERT (packet->sa.sa_family == AF_INET);
+ struct sockaddr_in *sa4 = (struct sockaddr_in *) &packet->sa;
+ hdr.rmt_port = sa4->sin_port;
+ hdr.rmt_ip.ip4.as_u32 = sa4->sin_addr.s_addr;
+ }
+ else
+ {
+ ASSERT (packet->sa.sa_family == AF_INET6);
+ struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *) &packet->sa;
+ hdr.rmt_port = sa6->sin6_port;
+ clib_memcpy (&hdr.rmt_ip.ip6, &sa6->sin6_addr, 16);
+ }
+
+ rv = svm_fifo_enqueue_nowait (f, sizeof (hdr), (u8 *) & hdr);
+ ASSERT (rv == sizeof (hdr));
+ if (svm_fifo_enqueue_nowait (f, len, packet->data.base) != len)
+ return 1;
+ return 0;
+}
+
+static int
+quic_send_packets (quic_ctx_t * ctx)
+{
+ //QUIC_DBG (2, "Called quic_send_packets");
+ quicly_datagram_t *packets[16];
+ session_t *quic_session;
+ quicly_conn_t *conn;
+ size_t num_packets, i;
+ int ret;
+
+ quic_session = session_get_from_handle (ctx->c_quic_ctx_id.quic_session);
+ conn = ctx->c_quic_ctx_id.conn;
+
+ if (!conn)
+ return 0;
+
+ do
+ {
+ num_packets = sizeof (packets) / sizeof (packets[0]);
+ if ((ret = quicly_send (conn, packets, &num_packets)) == 0)
+ {
+ for (i = 0; i != num_packets; ++i)
+ {
+ if (quic_send_datagram (quic_session, packets[i]))
+ {
+ QUIC_DBG (2, "quic_send_datagram failed");
+ goto stop_sending;
+ }
+ ret = 0;
+ quicly_default_free_packet_cb.cb
+ (&quicly_default_free_packet_cb, packets[i]);
+ }
+ }
+ else
+ {
+ QUIC_DBG (2, "quicly_send returned %d, closing connection\n", ret);
+ return ret;
+ }
+ }
+ while (ret == 0 && num_packets == sizeof (packets) / sizeof (packets[0]));
+
+stop_sending:
+ if (svm_fifo_set_event (quic_session->tx_fifo))
+ session_send_io_evt_to_thread (quic_session->tx_fifo, FIFO_EVENT_APP_TX);
+
+ quic_update_timer (ctx);
+ return 0;
+}
+
+/*****************************************************************************
+ * START QUICLY CALLBACKS
+ * Called from QUIC lib
+ *****************************************************************************/
+
+static int
+quic_on_stop_sending (quicly_stream_t * stream, int error_code)
+{
+ QUIC_DBG (2, "received STOP_SENDING: %d", error_code);
+ return 0;
+}
+
+static int
+quic_on_receive_reset (quicly_stream_t * stream, int error_code)
+{
+ QUIC_DBG (2, "received RESET_STREAM: %d", error_code);
+ return 0;
+}
+
+static int
+quic_on_receive (quicly_stream_t * stream, size_t off, const void *src,
+ size_t len)
+{
+ QUIC_DBG (2, "received data: %lu bytes", len);
+ u32 to_enqueue, ctx_id;
+ quic_ctx_t *ctx;
+ session_t *app_session;
+ svm_fifo_t *rx_fifo;
+ app_worker_t *app_wrk;
+
+ ctx_id = (u64) * quicly_get_data (stream->conn);
+ ctx = quic_ctx_get (ctx_id);
+ app_session = session_get_from_handle (ctx->c_quic_ctx_id.app_session);
+ rx_fifo = app_session->rx_fifo;
+ to_enqueue = svm_fifo_max_enqueue (rx_fifo);
+ if (to_enqueue > len)
+ to_enqueue = len;
+ // TODO what happens to the excess bytes?
+
+ svm_fifo_enqueue_nowait (rx_fifo, to_enqueue, src);
+
+ // Notify app
+ app_wrk = app_worker_get_if_valid (app_session->app_wrk_index);
+ if (PREDICT_TRUE (app_wrk != 0))
+ app_worker_lock_and_send_event (app_wrk, app_session, SESSION_IO_EVT_RX);
+ return 0;
+}
+
+static const quicly_stream_callbacks_t quic_stream_callbacks = {
+ .on_destroy = quicly_streambuf_destroy,
+ .on_send_shift = quicly_streambuf_egress_shift,
+ .on_send_emit = quicly_streambuf_egress_emit,
+ .on_send_stop = quic_on_stop_sending,
+ .on_receive = quic_on_receive,
+ .on_receive_reset = quic_on_receive_reset
+};
+
+static int
+quic_on_stream_open (quicly_stream_open_cb * self, quicly_stream_t * stream)
+{
+ QUIC_DBG (2, "on_stream_open called");
+ int ret;
+ if ((ret =
+ quicly_streambuf_create (stream, sizeof (quicly_streambuf_t))) != 0)
+ {
+ return ret;
+ }
+ stream->callbacks = &quic_stream_callbacks;
+ return 0;
+}
+
+static quicly_stream_open_cb on_stream_open = { &quic_on_stream_open };
+
+static void
+quic_on_conn_close (quicly_closed_by_peer_cb * self, quicly_conn_t * conn,
+ int code, uint64_t frame_type,
+ const char *reason, size_t reason_len)
+{
+ QUIC_DBG (2, "connection closed, reason: %s", reason);
+ u32 ctx_index = (u64) * quicly_get_data (conn);
+ quic_connection_closed (ctx_index);
+}
+
+static quicly_closed_by_peer_cb on_closed_by_peer = { &quic_on_conn_close };
+
+
+/*****************************************************************************
+ * END QUICLY CALLBACKS
+ *****************************************************************************/
+
+/* single-entry session cache */
+struct st_util_session_cache_t
+{
+ ptls_encrypt_ticket_t super;
+ uint8_t id[32];
+ ptls_iovec_t data;
+};
+
+static int
+encrypt_ticket_cb (ptls_encrypt_ticket_t * _self, ptls_t * tls,
+ int is_encrypt, ptls_buffer_t * dst, ptls_iovec_t src)
+{
+ struct st_util_session_cache_t *self = (void *) _self;
+ int ret;
+
+ if (is_encrypt)
+ {
+
+ /* replace the cached entry along with a newly generated session id */
+ free (self->data.base);
+ if ((self->data.base = malloc (src.len)) == NULL)
+ return PTLS_ERROR_NO_MEMORY;
+
+ ptls_get_context (tls)->random_bytes (self->id, sizeof (self->id));
+ memcpy (self->data.base, src.base, src.len);
+ self->data.len = src.len;
+
+ /* store the session id in buffer */
+ if ((ret = ptls_buffer_reserve (dst, sizeof (self->id))) != 0)
+ return ret;
+ memcpy (dst->base + dst->off, self->id, sizeof (self->id));
+ dst->off += sizeof (self->id);
+
+ }
+ else
+ {
+
+ /* check if session id is the one stored in cache */
+ if (src.len != sizeof (self->id))
+ return PTLS_ERROR_SESSION_NOT_FOUND;
+ if (memcmp (self->id, src.base, sizeof (self->id)) != 0)
+ return PTLS_ERROR_SESSION_NOT_FOUND;
+
+ /* return the cached value */
+ if ((ret = ptls_buffer_reserve (dst, self->data.len)) != 0)
+ return ret;
+ memcpy (dst->base + dst->off, self->data.base, self->data.len);
+ dst->off += self->data.len;
+ }
+
+ return 0;
+}
+
+static struct st_util_session_cache_t sc = {
+ .super = {
+ .cb = encrypt_ticket_cb,
+ },
+};
+
+/* *INDENT-OFF* */
+static ptls_context_t quic_tlsctx = {
+ .random_bytes = ptls_openssl_random_bytes,
+ .get_time = &ptls_get_time,
+ .key_exchanges = ptls_openssl_key_exchanges,
+ .cipher_suites = ptls_openssl_cipher_suites,
+ .certificates = {
+ .list = NULL,
+ .count = 0
+ },
+ .esni = NULL,
+ .on_client_hello = NULL,
+ .emit_certificate = NULL,
+ .sign_certificate = NULL,
+ .verify_certificate = NULL,
+ .ticket_lifetime = 86400,
+ .max_early_data_size = 8192,
+ .hkdf_label_prefix__obsolete = NULL,
+ .require_dhe_on_psk = 1,
+ .encrypt_ticket = &sc.super,
+};
+/* *INDENT-ON* */
+
+static int
+ptls_compare_separator_line (const char *line, const char *begin_or_end,
+ const char *label)
+{
+ int ret = strncmp (line, "-----", 5);
+ size_t text_index = 5;
+
+ if (ret == 0)
+ {
+ size_t begin_or_end_length = strlen (begin_or_end);
+ ret = strncmp (line + text_index, begin_or_end, begin_or_end_length);
+ text_index += begin_or_end_length;
+ }
+
+ if (ret == 0)
+ {
+ ret = line[text_index] - ' ';
+ text_index++;
+ }
+
+ if (ret == 0)
+ {
+ size_t label_length = strlen (label);
+ ret = strncmp (line + text_index, label, label_length);
+ text_index += label_length;
+ }
+
+ if (ret == 0)
+ {
+ ret = strncmp (line + text_index, "-----", 5);
+ }
+
+ return ret;
+}
+
+static int
+ptls_get_bio_pem_object (BIO * bio, const char *label, ptls_buffer_t * buf)
+{
+ int ret = PTLS_ERROR_PEM_LABEL_NOT_FOUND;
+ char line[256];
+ ptls_base64_decode_state_t state;
+
+ /* Get the label on a line by itself */
+ while (BIO_gets (bio, line, 256))
+ {
+ if (ptls_compare_separator_line (line, "BEGIN", label) == 0)
+ {
+ ret = 0;
+ ptls_base64_decode_init (&state);
+ break;
+ }
+ }
+ /* Get the data in the buffer */
+ while (ret == 0 && BIO_gets (bio, line, 256))
+ {
+ if (ptls_compare_separator_line (line, "END", label) == 0)
+ {
+ if (state.status == PTLS_BASE64_DECODE_DONE
+ || (state.status == PTLS_BASE64_DECODE_IN_PROGRESS
+ && state.nbc == 0))
+ {
+ ret = 0;
+ }
+ else
+ {
+ ret = PTLS_ERROR_INCORRECT_BASE64;
+ }
+ break;
+ }
+ else
+ {
+ ret = ptls_base64_decode (line, &state, buf);
+ }
+ }
+
+ return ret;
+}
+
+int
+ptls_load_bio_pem_objects (BIO * bio, const char *label, ptls_iovec_t * list,
+ size_t list_max, size_t * nb_objects)
+{
+ int ret = 0;
+ size_t count = 0;
+
+ *nb_objects = 0;
+
+ if (ret == 0)
+ {
+ while (count < list_max)
+ {
+ ptls_buffer_t buf;
+
+ ptls_buffer_init (&buf, "", 0);
+
+ ret = ptls_get_bio_pem_object (bio, label, &buf);
+
+ if (ret == 0)
+ {
+ if (buf.off > 0 && buf.is_allocated)
+ {
+ list[count].base = buf.base;
+ list[count].len = buf.off;
+ count++;
+ }
+ else
+ {
+ ptls_buffer_dispose (&buf);
+ }
+ }
+ else
+ {
+ ptls_buffer_dispose (&buf);
+ break;
+ }
+ }
+ }
+
+ if (ret == PTLS_ERROR_PEM_LABEL_NOT_FOUND && count > 0)
+ {
+ ret = 0;
+ }
+
+ *nb_objects = count;
+
+ return ret;
+}
+
+#define PTLS_MAX_CERTS_IN_CONTEXT 16
+
+int
+ptls_load_bio_certificates (ptls_context_t * ctx, BIO * bio)
+{
+ int ret = 0;
+
+ ctx->certificates.list =
+ (ptls_iovec_t *) malloc (PTLS_MAX_CERTS_IN_CONTEXT *
+ sizeof (ptls_iovec_t));
+
+ if (ctx->certificates.list == NULL)
+ {
+ ret = PTLS_ERROR_NO_MEMORY;
+ }
+ else
+ {
+ ret =
+ ptls_load_bio_pem_objects (bio, "CERTIFICATE", ctx->certificates.list,
+ PTLS_MAX_CERTS_IN_CONTEXT,
+ &ctx->certificates.count);
+ }
+
+ return ret;
+}
+
+static inline void
+load_bio_certificate_chain (ptls_context_t * ctx, const char *cert_data)
+{
+ BIO *cert_bio;
+ cert_bio = BIO_new_mem_buf (cert_data, -1);
+ if (ptls_load_bio_certificates (ctx, cert_bio) != 0)
+ {
+ BIO_free (cert_bio);
+ fprintf (stderr, "failed to load certificate:%s\n", strerror (errno));
+ exit (1);
+ }
+ BIO_free (cert_bio);
+}
+
+static inline void
+load_bio_private_key (ptls_context_t * ctx, const char *pk_data)
+{
+ static ptls_openssl_sign_certificate_t sc;
+ EVP_PKEY *pkey;
+ BIO *key_bio;
+
+ key_bio = BIO_new_mem_buf (pk_data, -1);
+ pkey = PEM_read_bio_PrivateKey (key_bio, NULL, NULL, NULL);
+ BIO_free (key_bio);
+
+ if (pkey == NULL)
+ {
+ fprintf (stderr, "failed to read private key from app configuration\n");
+ exit (1);
+ }
+
+ ptls_openssl_init_sign_certificate (&sc, pkey);
+ EVP_PKEY_free (pkey);
+
+ ctx->sign_certificate = &sc.super;
+}
+
+static void
+quic_connection_closed (u32 ctx_index)
+{
+ QUIC_DBG (2, "QUIC connection closed");
+ quic_ctx_t *ctx;
+
+ ctx = quic_ctx_get (ctx_index);
+ // TODO if connection is not established, just delete the session
+
+ // Do not try to send anything anymore
+ ctx->stream = NULL;
+ quicly_free (ctx->c_quic_ctx_id.conn);
+ ctx->c_quic_ctx_id.conn = NULL;
+ session_transport_closing_notify (&ctx->connection);
+}
+
+static int64_t
+quic_get_time (quicly_now_cb * self)
+{
+ // TODO read value set by set_time_now?
+ // (needs to change it not to call this function)
+ vlib_main_t *vlib_main = vlib_get_main ();
+ f64 time = vlib_time_now (vlib_main);
+ return (int64_t) (time * 1000.f);
+}
+quicly_now_cb quicly_vpp_now_cb = { quic_get_time };
+
+static void
+allocate_quicly_ctx (application_t * app, u8 is_client)
+{
+ QUIC_DBG (2, "Called allocate_quicly_ctx");
+ struct
+ {
+ quicly_context_t _;
+ char cid_key[17];
+ } *ctx_data;
+ quicly_context_t *quicly_ctx;
+ char *cid_key;
+
+ ctx_data = malloc (sizeof (*ctx_data));
+ quicly_ctx = &ctx_data->_;
+ app->quicly_ctx = (u64 *) quicly_ctx;
+ memcpy (quicly_ctx, &quicly_default_context, sizeof (quicly_context_t));
+
+ quicly_ctx->tls = &quic_tlsctx;
+ quicly_ctx->stream_open = &on_stream_open;
+ quicly_ctx->closed_by_peer = &on_closed_by_peer;
+ quicly_ctx->now = &quicly_vpp_now_cb;
+
+ quicly_amend_ptls_context (quicly_ctx->tls);
+
+ quicly_ctx->event_log.mask = INT64_MAX;
+ quicly_ctx->event_log.cb = quicly_new_default_event_log_cb (stderr);
+
+ quicly_ctx->transport_params.max_data = QUIC_INT_MAX;
+ quicly_ctx->transport_params.max_streams_uni = QUIC_INT_MAX;
+ quicly_ctx->transport_params.max_streams_bidi = QUIC_INT_MAX;
+ quicly_ctx->transport_params.max_stream_data.bidi_local = QUIC_INT_MAX;
+ quicly_ctx->transport_params.max_stream_data.bidi_remote = QUIC_INT_MAX;
+ quicly_ctx->transport_params.max_stream_data.uni = QUIC_INT_MAX;
+
+ if (!is_client)
+ {
+ load_bio_private_key (quicly_ctx->tls, (char *) app->tls_key);
+ load_bio_certificate_chain (quicly_ctx->tls, (char *) app->tls_cert);
+ cid_key = ctx_data->cid_key;
+ quicly_ctx->tls->random_bytes (cid_key, 16);
+ cid_key[16] = 0;
+ quicly_ctx->encrypt_cid =
+ quicly_new_default_encrypt_cid_cb (&ptls_openssl_bfecb,
+ &ptls_openssl_sha256,
+ ptls_iovec_init (cid_key,
+ strlen
+ (cid_key)));
+ quicly_ctx->decrypt_cid =
+ quicly_new_default_decrypt_cid_cb (&ptls_openssl_bfecb,
+ &ptls_openssl_sha256,
+ ptls_iovec_init (cid_key,
+ strlen
+ (cid_key)));
+ }
+}
+
+
+/*****************************************************************************
+ * BEGIN TIMERS HANDLING
+ *****************************************************************************/
+
+static u32
+quic_set_time_now (u32 thread_index)
+{
+ quic_main.wrk_ctx[thread_index].time_now = quic_get_time (NULL);
+ return quic_main.wrk_ctx[thread_index].time_now;
+}
+
+static void
+quic_timer_expired (u32 conn_index)
+{
+ quic_ctx_t *ctx;
+ QUIC_DBG (2, "Timer expired for conn %u at %ld", conn_index,
+ quic_get_time (NULL));
+ ctx = quic_ctx_get (conn_index);
+ ctx->c_quic_ctx_id.timer_handle = QUIC_TIMER_HANDLE_INVALID;
+ if (quic_send_packets (ctx))
+ {
+ quic_connection_closed (conn_index);
+ }
+}
+
+static void
+quic_update_timer (quic_ctx_t * ctx)
+{
+ tw_timer_wheel_1t_3w_1024sl_ov_t *tw;
+ int64_t next_timeout;
+
+ // This timeout is in ms which is the unit of our timer
+ next_timeout = quicly_get_first_timeout (ctx->c_quic_ctx_id.conn);
+ tw = &quic_main.wrk_ctx[vlib_get_thread_index ()].timer_wheel;
+ f64 next_timeout_f = ((f64) next_timeout) / 1000.f;
+
+ clib_warning ("Timer set to %ld (%lf)", next_timeout, next_timeout_f);
+
+ if (ctx->c_quic_ctx_id.timer_handle == QUIC_TIMER_HANDLE_INVALID)
+ {
+ if (next_timeout == INT64_MAX)
+ return;
+ ctx->c_quic_ctx_id.timer_handle =
+ tw_timer_start_1t_3w_1024sl_ov (tw, ctx->c_c_index, 0,
+ next_timeout_f);
+ }
+ else
+ {
+ if (next_timeout == INT64_MAX)
+ {
+ tw_timer_stop_1t_3w_1024sl_ov (tw, ctx->c_quic_ctx_id.timer_handle);
+ ctx->c_quic_ctx_id.timer_handle = QUIC_TIMER_HANDLE_INVALID;
+ }
+ else
+ tw_timer_update_1t_3w_1024sl_ov (tw, ctx->c_quic_ctx_id.timer_handle,
+ next_timeout_f);
+ }
+}
+
+static void
+quic_expired_timers_dispatch (u32 * expired_timers)
+{
+ int i;
+
+ for (i = 0; i < vec_len (expired_timers); i++)
+ {
+ quic_timer_expired (expired_timers[i]);
+ }
+}
+
+
+/*****************************************************************************
+ * END TIMERS HANDLING
+ *
+ * BEGIN TRANSPORT PROTO FUNCTIONS
+ *****************************************************************************/
+
+int
+quic_connect (transport_endpoint_cfg_t * tep)
+{
+ QUIC_DBG (2, "Called quic_connect");
+ vnet_connect_args_t _cargs = { {}, }, *cargs = &_cargs;
+ session_endpoint_cfg_t *sep;
+ quic_main_t *qm = &quic_main;
+ quic_ctx_t *ctx;
+ app_worker_t *app_wrk;
+ application_t *app;
+ u32 ctx_index;
+ int error;
+
+ sep = (session_endpoint_cfg_t *) tep;
+ ctx_index = quic_ctx_half_open_alloc ();
+ ctx = quic_ctx_half_open_get (ctx_index);
+ ctx->c_quic_ctx_id.parent_app_wrk_idx = sep->app_wrk_index;
+ ctx->c_s_index = 0xFAFAFAFA;
+ ctx->c_quic_ctx_id.udp_is_ip4 = sep->is_ip4;
+ ctx->c_quic_ctx_id.timer_handle = QUIC_TIMER_HANDLE_INVALID;
+ ctx->c_quic_ctx_id.conn_state = QUIC_CONN_STATE_HANDSHAKE;
+ ctx->client_opaque = sep->opaque;
+ if (sep->hostname)
+ {
+ ctx->srv_hostname = format (0, "%v", sep->hostname);
+ vec_terminate_c_string (ctx->srv_hostname);
+ }
+ else
+ {
+ // needed by quic for crypto + determining client / server
+ ctx->srv_hostname =
+ format (0, "%U", format_ip46_address, &sep->ip, sep->is_ip4);
+ }
+
+ quic_ctx_half_open_reader_unlock ();
+
+ clib_memcpy (&cargs->sep, sep, sizeof (session_endpoint_cfg_t));
+ cargs->sep.transport_proto = TRANSPORT_PROTO_UDP;
+ cargs->app_index = qm->app_index;
+ cargs->api_context = ctx_index;
+
+ app_wrk = app_worker_get (sep->app_wrk_index);
+ app = application_get (app_wrk->app_index);
+ ctx->c_quic_ctx_id.parent_app_id = app_wrk->app_index;
+
+ allocate_quicly_ctx (app, 1 /* is client */ );
+
+ if ((error = vnet_connect (cargs)))
+ return error;
+
+ QUIC_DBG (1, "New connect request %u", ctx_index);
+ return 0;
+}
+
+static void
+quic_disconnect (u32 ctx_index, u32 thread_index)
+{
+ QUIC_DBG (2, "Called quic_disconnect");
+ tw_timer_wheel_1t_3w_1024sl_ov_t *tw;
+ quic_ctx_t *ctx;
+
+ QUIC_DBG (1, "Disconnecting %x", ctx_index);
+
+ ctx = quic_ctx_get (ctx_index);
+ if (ctx->c_quic_ctx_id.timer_handle != QUIC_TIMER_HANDLE_INVALID)
+ {
+ tw = &quic_main.wrk_ctx[vlib_get_thread_index ()].timer_wheel;
+ tw_timer_stop_1t_3w_1024sl_ov (tw, ctx->c_quic_ctx_id.timer_handle);
+ }
+ quic_disconnect_transport (ctx);
+ // This removes the session from the lookup table and frees it.
+ session_transport_delete_notify (&ctx->connection);
+ quic_ctx_free (ctx);
+}
+
+u32
+quic_start_listen (u32 app_listen_session_index, transport_endpoint_t * tep)
+{
+ QUIC_DBG (2, "Called quic_start_listen");
+ vnet_listen_args_t _bargs, *args = &_bargs;
+ quic_main_t *qm = &quic_main;
+ session_handle_t udp_handle;
+ session_endpoint_cfg_t *sep;
+ session_t *quic_listen_session, *app_listen_session;
+ app_worker_t *app_wrk;
+ application_t *app;
+ quic_ctx_t *lctx;
+ u32 lctx_index;
+ app_listener_t *app_listener;
+
+ sep = (session_endpoint_cfg_t *) tep;
+ app_wrk = app_worker_get (sep->app_wrk_index);
+ app = application_get (app_wrk->app_index);
+
+ allocate_quicly_ctx (app, 0 /* is_client */ );
+
+ sep->transport_proto = TRANSPORT_PROTO_UDP;
+ memset (args, 0, sizeof (*args));
+ args->app_index = qm->app_index;
+ args->sep_ext = *sep;
+ args->sep_ext.ns_index = app->ns_index;
+ if (vnet_listen (args))
+ return -1;
+
+ lctx_index = quic_ctx_alloc (); // listener
+ udp_handle = args->handle;
+ app_listener = app_listener_get_w_handle (udp_handle);
+ quic_listen_session = app_listener_get_session (app_listener);
+ quic_listen_session->opaque = lctx_index;
+
+ app_listen_session = listen_session_get (app_listen_session_index);
+
+ lctx = quic_ctx_get (lctx_index); // listener
+ lctx->is_listener = 1;
+ lctx->c_quic_ctx_id.parent_app_wrk_idx = sep->app_wrk_index;
+ lctx->c_quic_ctx_id.parent_app_id = app_wrk->app_index;
+ lctx->c_quic_ctx_id.quic_session = udp_handle;
+ lctx->c_quic_ctx_id.app_session =
+ listen_session_get_handle (app_listen_session);
+ lctx->c_quic_ctx_id.udp_is_ip4 = sep->is_ip4;
+
+ QUIC_DBG (1, "Started listening %d", lctx_index);
+ return lctx_index;
+}
+
+u32
+quic_stop_listen (u32 lctx_index)
+{
+ QUIC_DBG (2, "Called quic_stop_listen");
+ quic_ctx_t *lctx;
+
+ lctx = quic_ctx_get (lctx_index); // listener
+ vnet_unlisten_args_t a = {
+ .handle = lctx->c_quic_ctx_id.quic_session,
+ .app_index = quic_main.app_index,
+ .wrk_map_index = 0 /* default wrk */
+ };
+ if (vnet_unlisten (&a))
+ clib_warning ("unlisten errored");
+
+ // TODO: crypto state cleanup
+
+ quic_ctx_free (lctx); // listener
+ return 0;
+}
+
+transport_connection_t *
+quic_connection_get (u32 ctx_index, u32 thread_index)
+{
+ QUIC_DBG (2, "Called quic_connection_get");
+ quic_ctx_t *ctx;
+ ctx = quic_ctx_get_w_thread (ctx_index, thread_index);
+ return &ctx->connection;
+}
+
+transport_connection_t *
+quic_listener_get (u32 listener_index)
+{
+ QUIC_DBG (2, "Called quic_listener_get");
+ quic_ctx_t *ctx;
+ ctx = quic_ctx_get (listener_index);
+ return &ctx->connection;
+}
+
+static void
+quic_update_time (f64 now, u8 thread_index)
+{
+ tw_timer_wheel_1t_3w_1024sl_ov_t *tw;
+
+ tw = &quic_main.wrk_ctx[thread_index].timer_wheel;
+ quic_set_time_now (thread_index);
+ tw_timer_expire_timers_1t_3w_1024sl_ov (tw, now);
+}
+
+static u8 *
+format_quic_connection (u8 * s, va_list * args)
+{
+ s = format (s, "[QUIC] connection");
+ return s;
+}
+
+static u8 *
+format_quic_half_open (u8 * s, va_list * args)
+{
+ u32 qc_index = va_arg (*args, u32);
+ quic_ctx_t *ctx = quic_ctx_half_open_get (qc_index);
+ s = format (s, "[QUIC] half-open app %u", ctx->c_quic_ctx_id.parent_app_id);
+ quic_ctx_half_open_reader_unlock ();
+ return s;
+}
+
+// TODO improve
+static u8 *
+format_quic_listener (u8 * s, va_list * args)
+{
+ s = format (s, "[QUIC] listener");
+ return s;
+}
+
+/*****************************************************************************
+ * END TRANSPORT PROTO FUNCTIONS
+ *
+ * START SESSION CALLBACKS
+ * Called from UDP layer
+ *****************************************************************************/
+
+static inline void
+quic_build_sockaddr (struct sockaddr *sa, socklen_t * salen,
+ ip46_address_t * addr, u16 port, u8 is_ip4)
+{
+ if (is_ip4)
+ {
+ struct sockaddr_in *sa4 = (struct sockaddr_in *) sa;
+ sa4->sin_family = AF_INET;
+ sa4->sin_port = port;
+ sa4->sin_addr.s_addr = addr->ip4.as_u32;
+ *salen = sizeof (struct sockaddr_in);
+ }
+ else
+ {
+ struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *) sa;
+ sa6->sin6_family = AF_INET6;
+ sa6->sin6_port = port;
+ clib_memcpy (&sa6->sin6_addr, &addr->ip6, 16);
+ *salen = sizeof (struct sockaddr_in6);
+ }
+}
+
+static int
+quic_delayed_notify_app_connected (void *ctx_index)
+{
+ QUIC_DBG (1, "quic_notify_app_connected");
+ session_t *app_session;
+ app_worker_t *app_wrk;
+ quic_ctx_t *ctx;
+ ctx = quic_ctx_get ((u32) (u64) ctx_index);
+
+ app_wrk = app_worker_get_if_valid (ctx->c_quic_ctx_id.parent_app_wrk_idx);
+ if (!app_wrk)
+ {
+ quic_disconnect_transport (ctx);
+ return -1;
+ }
+
+ app_session = session_alloc (ctx->c_thread_index);
+ QUIC_DBG (1, "Created app_session, id %u", app_session->session_index);
+ ctx->c_s_index = app_session->session_index;
+ app_session->app_wrk_index = ctx->c_quic_ctx_id.parent_app_wrk_idx;
+ app_session->connection_index = ctx->c_c_index;
+ app_session->session_type =
+ session_type_from_proto_and_ip (TRANSPORT_PROTO_QUIC,
+ ctx->c_quic_ctx_id.udp_is_ip4);
+
+ if (app_worker_init_connected (app_wrk, app_session)) // TODO dont allocate fifos
+ {
+ quic_disconnect (ctx->c_c_index, vlib_get_thread_index ());
+ return app_worker_connect_notify (app_wrk, NULL, ctx->client_opaque);
+ }
+
+ app_session->session_state = SESSION_STATE_CONNECTING;
+ if (app_worker_connect_notify (app_wrk, app_session, ctx->client_opaque))
+ {
+ QUIC_DBG (1, "failed to notify app");
+ quic_disconnect (ctx->c_c_index, vlib_get_thread_index ());
+ return -1;
+ }
+
+ ctx->c_quic_ctx_id.app_session = session_handle (app_session);
+ app_session->session_state = SESSION_STATE_LISTENING;
+ session_lookup_add_connection (&ctx->connection,
+ session_handle (app_session));
+
+ return 0;
+}
+
+int
+quic_session_connected_callback (u32 quic_app_index, u32 ho_ctx_idx,
+ session_t * s, u8 is_fail)
+{
+ QUIC_DBG (2, "Called quic_session_connected_callback");
+ // This should always be called before quic_connect returns since UDP always
+ // connects instantly.
+ struct sockaddr_in6 sa6;
+ struct sockaddr *sa = (struct sockaddr *) &sa6;
+ socklen_t salen;
+ transport_connection_t *tc;
+ quic_ctx_t *ho_ctx, *ctx;
+ u32 ctx_index;
+ int ret;
+ application_t *app;
+ app_worker_t *app_wrk;
+
+ ho_ctx = quic_ctx_half_open_get (ho_ctx_idx);
+ if (is_fail)
+ {
+ int (*cb_fn) (u32, u32, session_t *, u8), rv = 0;
+ u32 wrk_index, api_context;
+
+ wrk_index = ho_ctx->c_quic_ctx_id.parent_app_wrk_idx;
+ app_wrk =
+ app_worker_get_if_valid (ho_ctx->c_quic_ctx_id.parent_app_wrk_idx);
+ if (app_wrk)
+ {
+ api_context = ho_ctx->c_s_index;
+ app = application_get (app_wrk->app_index);
+ cb_fn = app->cb_fns.session_connected_callback;
+ rv = cb_fn (wrk_index, api_context, 0, 1 /* failed */ );
+ }
+ quic_ctx_half_open_reader_unlock ();
+ quic_ctx_half_open_free (ho_ctx_idx);
+ return rv;
+ }
+
+ app_wrk =
+ app_worker_get_if_valid (ho_ctx->c_quic_ctx_id.parent_app_wrk_idx);
+ if (!app_wrk)
+ {
+ QUIC_DBG (1, "Appwrk not found");
+ return -1;
+ }
+ app = application_get (app_wrk->app_index);
+
+ ctx_index = quic_ctx_alloc ();
+ ctx = quic_ctx_get (ctx_index);
+ clib_memcpy (ctx, ho_ctx, sizeof (*ctx));
+ quic_ctx_half_open_reader_unlock (); // TODO: this is a race
+ quic_ctx_half_open_free (ho_ctx_idx);
+
+ ctx->c_thread_index = vlib_get_thread_index ();
+ ctx->c_c_index = ctx_index;
+
+ QUIC_DBG (1, "Quic connect for returned %u. New connection [%u]%x",
+ is_fail, vlib_get_thread_index (), (ctx) ? ctx_index : ~0);
+
+ ctx->c_quic_ctx_id.quic_session = session_handle (s);
+ s->opaque = ctx_index;
+ s->session_state = SESSION_STATE_READY;
+
+ // Init QUIC lib connection
+ // Generate required sockaddr & salen
+ tc = session_get_transport (s);
+ quic_build_sockaddr (sa, &salen, &tc->rmt_ip, tc->rmt_port, tc->is_ip4);
+
+ ret =
+ quicly_connect (&ctx->c_quic_ctx_id.conn,
+ (quicly_context_t *) app->quicly_ctx,
+ (char *) ctx->srv_hostname, sa, salen,
+ &quic_main.next_cid, &quic_main.hs_properties, NULL);
+ ++quic_main.next_cid.master_id;
+ // Save context handle in quicly connection
+ *quicly_get_data (ctx->c_quic_ctx_id.conn) = (void *) (u64) ctx_index;
+ assert (ret == 0);
+
+ if (quic_send_packets (ctx))
+ {
+ quic_connection_closed (ctx_index);
+ }
+ return ret;
+}
+
+void
+quic_session_disconnect_callback (session_t * s)
+{
+ clib_warning ("UDP session disconnected???");
+}
+
+void
+quic_session_reset_callback (session_t * s)
+{
+ clib_warning ("UDP session reset???");
+}
+
+static int
+quic_add_segment_callback (u32 client_index, u64 seg_handle)
+{
+ QUIC_DBG (2, "Called quic_add_segment_callback");
+ QUIC_DBG (2, "NOT IMPLEMENTED");
+ /* No-op for builtin */
+ return 0;
+}
+
+static int
+quic_del_segment_callback (u32 client_index, u64 seg_handle)
+{
+ QUIC_DBG (2, "Called quic_del_segment_callback");
+ QUIC_DBG (2, "NOT IMPLEMENTED");
+ /* No-op for builtin */
+ return 0;
+}
+
+int
+quic_add_vpp_q_builtin_tx_evt (session_t * s)
+{
+ if (svm_fifo_set_event (s->tx_fifo))
+ session_send_io_evt_to_thread_custom (s, s->thread_index,
+ FIFO_EVENT_BUILTIN_TX);
+ return 0;
+}
+
+void
+quic_open_stream_if_ready (quic_ctx_t * ctx)
+{
+ quicly_conn_t *conn = ctx->c_quic_ctx_id.conn;
+ if (ctx->stream)
+ {
+ QUIC_DBG (2, "----------- > FOUND Stream id %d",
+ ctx->stream->stream_id);
+ QUIC_DBG (2, "----------- > FOUND Stream is_open %d",
+ ctx->stream->sendstate.is_open);
+ return;
+ }
+ if (quicly_connection_is_ready (conn))
+ assert (!quicly_open_stream (conn, &ctx->stream, 0));
+ QUIC_DBG (2, "Stream id %d", ctx->stream->stream_id);
+ QUIC_DBG (2, "Stream is_open %d", ctx->stream->sendstate.is_open);
+}
+
+int
+quic_custom_tx_callback (void *session)
+{
+ QUIC_DBG (2, "Called quic_custom_tx_callback");
+ session_t *app_session = (session_t *) session;
+ quic_ctx_t *ctx;
+ svm_fifo_t *f;
+ u32 deq_max;
+ u8 *data;
+
+ if (PREDICT_FALSE
+ (app_session->session_state >= SESSION_STATE_TRANSPORT_CLOSING))
+ return 0;
+ ctx = quic_ctx_get (app_session->connection_index);
+ quic_open_stream_if_ready (ctx);
+ if (!ctx->stream)
+ {
+ quic_add_vpp_q_builtin_tx_evt (app_session);
+ return 0;
+ }
+
+ f = app_session->tx_fifo;
+ deq_max = svm_fifo_max_dequeue (f);
+ if (!deq_max)
+ return 0;
+
+ data = svm_fifo_head (f);
+ if (quicly_streambuf_egress_write (ctx->stream, data, deq_max))
+ {
+ assert (0);
+ return 0;
+ }
+ QUIC_DBG (2, "Sent %u bytes", deq_max);
+ svm_fifo_dequeue_drop (f, deq_max);
+ if (quic_send_packets (ctx))
+ {
+ quic_connection_closed (ctx->c_c_index);
+ }
+ return 0;
+}
+
+int
+quic_find_packet_ctx (quic_ctx_t ** ctx, quicly_conn_t ** conn,
+ struct sockaddr *sa, socklen_t salen,
+ quicly_decoded_packet_t packet)
+{
+ quic_ctx_t *ctx_;
+ quicly_conn_t *conn_;
+ /* *INDENT-OFF* */
+ pool_foreach (ctx_, quic_main.ctx_pool[vlib_get_thread_index()],
+ ({
+ conn_ = ctx_->c_quic_ctx_id.conn;
+ if (conn_ && !ctx_->is_listener)
+ {
+ if (quicly_is_destination(conn_, sa, salen, &packet))
+ {
+ *conn = conn_;
+ *ctx = ctx_;
+ QUIC_DBG (2, "connection_found");
+ return 0;
+ }
+ }
+ }));
+ /* *INDENT-ON* */
+ return 0;
+}
+
+static int
+quic_receive (quic_ctx_t * ctx, quicly_conn_t * conn,
+ quicly_decoded_packet_t packet)
+{
+ quicly_receive (conn, &packet);
+ // Conn may be set to null if the connection is terminated
+ if (ctx->c_quic_ctx_id.conn
+ && ctx->c_quic_ctx_id.conn_state == QUIC_CONN_STATE_HANDSHAKE)
+ {
+ if (quicly_connection_is_ready (conn))
+ {
+ ctx->c_quic_ctx_id.conn_state = QUIC_CONN_STATE_READY;
+ if (quicly_is_client (conn))
+ session_send_rpc_evt_to_thread_force (vlib_get_thread_index (),
+ &quic_delayed_notify_app_connected,
+ (void *) (u64)
+ ctx->c_c_index);
+ }
+ }
+ if (quic_send_packets (ctx))
+ {
+ quic_connection_closed (ctx->c_c_index);
+ }
+ return 0;
+}
+
+static int
+quic_delayed_create_app_session (void *ctx_index)
+{
+ quic_ctx_t *lctx, *ctx;
+ session_t *app_session, *app_listen_session;
+ app_worker_t *app_wrk;
+ int rv;
+
+ ctx = quic_ctx_get ((u32) (u64) ctx_index);
+ app_session = session_alloc (ctx->c_thread_index);
+ app_session->session_state = SESSION_STATE_LISTENING;
+ ctx->c_s_index = app_session->session_index;
+
+ lctx = quic_ctx_get (ctx->c_quic_ctx_id.listener_ctx_id);
+
+ app_listen_session =
+ listen_session_get_from_handle (lctx->c_quic_ctx_id.app_session);
+ app_session->app_wrk_index = lctx->c_quic_ctx_id.parent_app_wrk_idx;
+ app_session->connection_index = ctx->c_c_index;
+ app_session->session_type = app_listen_session->session_type;
+ app_session->listener_index = app_listen_session->session_index;
+ app_session->app_index = quic_main.app_index;
+
+ // TODO: don't alloc fifos when we don't transfer data on this session
+ if ((rv = app_worker_init_accepted (app_session)))
+ {
+ QUIC_DBG (1, "failed to allocate fifos");
+ session_free (app_session);
+ return rv;
+ }
+ ctx->c_quic_ctx_id.app_session = session_handle (app_session);
+ ctx->c_quic_ctx_id.parent_app_id = lctx->c_quic_ctx_id.parent_app_id;
+ ctx->c_quic_ctx_id.udp_is_ip4 = lctx->c_quic_ctx_id.udp_is_ip4;
+ ctx->c_quic_ctx_id.parent_app_wrk_idx = app_session->app_wrk_index;
+ session_lookup_add_connection (&ctx->connection,
+ session_handle (app_session));
+ app_wrk = app_worker_get (app_session->app_wrk_index);
+ rv = app_worker_accept_notify (app_wrk, app_session);
+ if (rv)
+ {
+ QUIC_DBG (1, "failed to notify accept worker app");
+ return rv;
+ }
+ return 0;
+}
+
+static int
+quic_create_connection (quicly_context_t * quicly_ctx,
+ u64 quic_session_handle, u32 lctx_index,
+ quicly_conn_t * conn, struct sockaddr *sa,
+ socklen_t salen, quicly_decoded_packet_t packet)
+{
+ quic_ctx_t *ctx;
+ u32 ctx_index;
+
+ /* new connection, accept and create context if packet is valid */
+ // TODO: check if socket is actually listening?
+ QUIC_DBG (2, "New connection created");
+ if (quicly_accept (&conn, quicly_ctx, sa, salen,
+ &packet, ptls_iovec_init (NULL, 0),
+ &quic_main.next_cid, NULL) != 0)
+ {
+ // Invalid packet, pass
+ assert (conn == NULL);
+ QUIC_DBG (2, "Accept failed");
+ return 0;
+ }
+ assert (conn != NULL);
+
+ ++quic_main.next_cid.master_id;
+ // Create context
+ ctx_index = quic_ctx_alloc ();
+ ctx = quic_ctx_get (ctx_index);
+ // Save ctx handle in quicly connection
+ *quicly_get_data (conn) = (void *) (u64) ctx_index;
+
+ ctx->c_thread_index = vlib_get_thread_index ();
+ ctx->c_c_index = ctx_index;
+ ctx->c_quic_ctx_id.quic_session = quic_session_handle;
+ ctx->c_quic_ctx_id.listener_ctx_id = lctx_index;
+ ctx->c_quic_ctx_id.timer_handle = QUIC_TIMER_HANDLE_INVALID;
+ ctx->c_quic_ctx_id.conn = conn;
+
+ session_send_rpc_evt_to_thread_force (vlib_get_thread_index (),
+ &quic_delayed_create_app_session,
+ (void *) (u64) ctx_index);
+ if (quic_send_packets (ctx))
+ {
+ quic_connection_closed (ctx_index);
+ }
+ return 0;
+}
+
+static int
+quic_reset_connection (quicly_context_t * quicly_ctx, u64 quic_session_handle,
+ struct sockaddr *sa, socklen_t salen,
+ quicly_decoded_packet_t packet)
+{
+ /* short header packet; potentially a dead connection. No need to check the length of the incoming packet,
+ * because loop is prevented by authenticating the CID (by checking node_id and thread_id). If the peer is also
+ * sending a reset, then the next CID is highly likely to contain a non-authenticating CID, ... */
+ QUIC_DBG (2, "Sending stateless reset");
+ quicly_datagram_t *dgram;
+ session_t *quic_session;
+ if (packet.cid.dest.plaintext.node_id == 0
+ && packet.cid.dest.plaintext.thread_id == 0)
+ {
+ dgram = quicly_send_stateless_reset (quicly_ctx, sa, salen,
+ &packet.cid.dest.plaintext);
+ quic_session = session_get_from_handle (quic_session_handle);
+ if (quic_send_datagram (quic_session, dgram)) // TODO : missing event on fifo
+ QUIC_DBG (2, "Send reset failed");
+ }
+ return 0;
+}
+
+int
+quic_app_rx_callback (session_t * quic_session)
+{
+ // Read data from UDP rx_fifo and pass it to the quicly conn.
+ QUIC_DBG (2, "Called quic_app_rx_callback");
+
+ quicly_decoded_packet_t packet;
+ session_dgram_hdr_t ph;
+ application_t *app;
+ quicly_conn_t *conn = NULL;
+ quic_ctx_t *lctx, *ctx = NULL;
+ svm_fifo_t *f;
+ size_t plen;
+ struct sockaddr_in6 sa6;
+ struct sockaddr *sa = (struct sockaddr *) &sa6;
+ socklen_t salen;
+ u32 max_deq, len;
+ u8 *data;
+ u32 lctx_index = quic_session->opaque;
+ u64 quic_session_handle = session_handle (quic_session);
+
+ f = quic_session->rx_fifo;
+
+ do
+ {
+ conn = NULL;
+ max_deq = svm_fifo_max_dequeue (f);
+ if (max_deq < sizeof (session_dgram_hdr_t))
+ {
+ svm_fifo_unset_event (f);
+ return 0;
+ }
+ QUIC_DBG (2, "Processing one packet at %ld", quic_get_time (NULL));
+
+ svm_fifo_unset_event (f);
+ svm_fifo_peek (f, 0, sizeof (ph), (u8 *) & ph);
+ ASSERT (ph.data_length >= ph.data_offset);
+ len = ph.data_length - ph.data_offset;
+
+ quic_build_sockaddr (sa, &salen, &ph.rmt_ip, ph.rmt_port, ph.is_ip4);
+
+ // Quicly can read len bytes from the fifo at offset:
+ // ph.data_offset + SESSION_CONN_HDR_LEN
+ data = svm_fifo_head (f) + ph.data_offset + SESSION_CONN_HDR_LEN;
+
+ lctx = quic_ctx_get (lctx_index);
+ app = application_get (lctx->c_quic_ctx_id.parent_app_id);
+
+ plen =
+ quicly_decode_packet ((quicly_context_t *) app->quicly_ctx, &packet,
+ data, len);
+ if (plen != SIZE_MAX)
+ {
+ quic_find_packet_ctx (&ctx, &conn, sa, salen, packet);
+ if (conn != NULL)
+ quic_receive (ctx, conn, packet);
+ else if (QUICLY_PACKET_IS_LONG_HEADER (packet.octets.base[0]))
+ quic_create_connection ((quicly_context_t *) app->quicly_ctx,
+ quic_session_handle, lctx_index, conn,
+ sa, salen, packet);
+ else if (((quicly_context_t *) app->quicly_ctx)->encrypt_cid)
+ quic_reset_connection ((quicly_context_t *) app->quicly_ctx,
+ quic_session_handle, sa, salen, packet);
+ }
+ svm_fifo_dequeue_drop (f,
+ ph.data_length + ph.data_offset +
+ SESSION_CONN_HDR_LEN);
+ }
+ while (1);
+ return 0;
+}
+
+/*****************************************************************************
+ * END TRANSPORT PROTO FUNCTIONS
+*****************************************************************************/
+
+/* *INDENT-OFF* */
+static session_cb_vft_t quic_app_cb_vft = {
+ .session_accept_callback = NULL,
+ .session_disconnect_callback = quic_session_disconnect_callback,
+ .session_connected_callback = quic_session_connected_callback,
+ .session_reset_callback = quic_session_reset_callback,
+ .add_segment_callback = quic_add_segment_callback,
+ .del_segment_callback = quic_del_segment_callback,
+ .builtin_app_rx_callback = quic_app_rx_callback,
+};
+
+const static transport_proto_vft_t quic_proto = {
+ .connect = quic_connect,
+ .close = quic_disconnect,
+ .start_listen = quic_start_listen,
+ .stop_listen = quic_stop_listen,
+ .get_connection = quic_connection_get,
+ .get_listener = quic_listener_get,
+ .update_time = quic_update_time,
+ .custom_tx = quic_custom_tx_callback,
+ .tx_type = TRANSPORT_TX_INTERNAL,
+ .service_type = TRANSPORT_SERVICE_APP,
+ .format_connection = format_quic_connection,
+ .format_half_open = format_quic_half_open,
+ .format_listener = format_quic_listener,
+};
+/* *INDENT-ON* */
+
+static clib_error_t *
+quic_init (vlib_main_t * vm)
+{
+ QUIC_DBG (2, "Called quic_init");
+ vlib_thread_main_t *vtm = vlib_get_thread_main ();
+ vnet_app_attach_args_t _a, *a = &_a;
+ u64 options[APP_OPTIONS_N_OPTIONS];
+ u32 segment_size = 512 << 20;
+ quic_main_t *qm = &quic_main;
+ u32 fifo_size = 64 << 10;
+ u32 num_threads;
+
+ num_threads = 1 /* main thread */ + vtm->n_threads;
+
+ memset (a, 0, sizeof (*a));
+ memset (options, 0, sizeof (options));
+
+ a->session_cb_vft = &quic_app_cb_vft;
+ a->api_client_index = APP_INVALID_INDEX;
+ a->options = options;
+ a->name = format (0, "quic");
+ a->options[APP_OPTIONS_SEGMENT_SIZE] = segment_size;
+ a->options[APP_OPTIONS_RX_FIFO_SIZE] = fifo_size;
+ a->options[APP_OPTIONS_TX_FIFO_SIZE] = fifo_size;
+ a->options[APP_OPTIONS_FLAGS] = APP_OPTIONS_FLAGS_IS_BUILTIN;
+ a->options[APP_OPTIONS_FLAGS] |= APP_OPTIONS_FLAGS_USE_GLOBAL_SCOPE;
+ a->options[APP_OPTIONS_FLAGS] |= APP_OPTIONS_FLAGS_IS_TRANSPORT_APP;
+
+ if (vnet_application_attach (a))
+ {
+ clib_warning ("failed to attach quic app");
+ return clib_error_return (0, "failed to attach quic app");
+ }
+
+ vec_validate (qm->ctx_pool, num_threads - 1);
+ vec_validate (qm->wrk_ctx, num_threads - 1);
+ // Timers, one per thread.
+ tw_timer_wheel_1t_3w_1024sl_ov_t *tw;
+ /* *INDENT-OFF* */
+ foreach_vlib_main (({
+ tw = &qm->wrk_ctx[ii].timer_wheel;
+ tw_timer_wheel_init_1t_3w_1024sl_ov (tw, quic_expired_timers_dispatch,
+ 10e-3 /* timer period 1ms */ , ~0);
+ tw->last_run_time = vlib_time_now (this_vlib_main);
+ }));
+ /* *INDENT-ON* */
+
+ if (!qm->ca_cert_path)
+ qm->ca_cert_path = QUIC_DEFAULT_CA_CERT_PATH;
+
+ qm->app_index = a->app_index;
+ clib_rwlock_init (&qm->half_open_rwlock);
+ qm->tstamp_ticks_per_clock = vm->clib_time.seconds_per_clock
+ / QUIC_TSTAMP_RESOLUTION;
+
+ transport_register_protocol (TRANSPORT_PROTO_QUIC, &quic_proto,
+ FIB_PROTOCOL_IP4, ~0);
+ transport_register_protocol (TRANSPORT_PROTO_QUIC, &quic_proto,
+ FIB_PROTOCOL_IP6, ~0);
+
+ vec_free (a->name);
+ return 0;
+}
+
+quic_main_t *
+vnet_quic_get_main (void)
+{
+ return &quic_main;
+}
+
+VLIB_INIT_FUNCTION (quic_init);
+
+/* *INDENT-OFF* */
+VLIB_PLUGIN_REGISTER () =
+{
+ .version = VPP_BUILD_VER,
+ .description = "Quic transport protocol",
+};
+/* *INDENT-ON* */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */
diff --git a/src/plugins/quic/quic.h b/src/plugins/quic/quic.h
new file mode 100644
index 00000000000..f574b29a583
--- /dev/null
+++ b/src/plugins/quic/quic.h
@@ -0,0 +1,118 @@
+/*
+ * Copyright (c) 2019 Cisco and/or its affiliates.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at:
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef __included_quic_h__
+#define __included_quic_h__
+
+#include <vnet/session/application_interface.h>
+
+#include <vppinfra/lock.h>
+#include <vppinfra/tw_timer_1t_3w_1024sl_ov.h>
+
+#include <quicly.h>
+
+#define QUIC_DEBUG 4
+#define QUIC_DEBUG_LEVEL_CLIENT 0
+#define QUIC_DEBUG_LEVEL_SERVER 0
+
+#define QUIC_DEFAULT_CA_CERT_PATH "/etc/ssl/certs/ca-certificates.crt"
+
+#define QUIC_TIMER_HANDLE_INVALID ((u32) ~0)
+
+#define QUIC_TSTAMP_RESOLUTION 0.001 /* QUIC tick resolution (1ms) */
+
+
+#if QUIC_DEBUG
+#define QUIC_DBG(_lvl, _fmt, _args...) \
+ if (_lvl <= QUIC_DEBUG) \
+ clib_warning (_fmt, ##_args)
+#else
+#define QUIC_DBG(_lvl, _fmt, _args...)
+#endif
+
+#define QUIC_CONN_STATE_HANDSHAKE 0
+#define QUIC_CONN_STATE_READY 1
+
+
+/* *INDENT-OFF* */
+typedef CLIB_PACKED (struct quic_ctx_id_
+{
+ session_handle_t app_session;
+ session_handle_t quic_session;
+ u32 parent_app_wrk_idx;
+ u32 parent_app_id;
+ u32 listener_ctx_id;
+ u32 timer_handle;
+ quicly_conn_t *conn;
+ u8 udp_is_ip4;
+ u8 conn_state;
+}) quic_ctx_id_t;
+/* *INDENT-ON* */
+
+STATIC_ASSERT (sizeof (quic_ctx_id_t) <= 42, "ctx id must be less than 42");
+
+typedef struct quic_ctx_
+{
+ union
+ {
+ transport_connection_t connection;
+ quic_ctx_id_t c_quic_ctx_id;
+ };
+
+ quicly_stream_t *stream;
+ u8 *srv_hostname;
+ u32 client_opaque;
+ u8 is_listener;
+} quic_ctx_t;
+
+typedef struct quic_worker_ctx_
+{
+ CLIB_CACHE_LINE_ALIGN_MARK (cacheline0);
+ u32 time_now; /**< worker time */
+ tw_timer_wheel_1t_3w_1024sl_ov_t timer_wheel; /**< worker timer wheel */
+ u32 *tx_buffers; /**< tx buffer free list */
+} quic_worker_ctx_t;
+
+typedef struct quic_main_
+{
+ u32 app_index;
+ quic_ctx_t *half_open_ctx_pool;
+ quic_ctx_t **ctx_pool;
+ clib_rwlock_t half_open_rwlock;
+ quic_worker_ctx_t *wrk_ctx;
+ f64 tstamp_ticks_per_clock;
+
+
+ /*
+ * Config
+ */
+ quicly_context_t quicly_ctx;
+ ptls_handshake_properties_t hs_properties;
+ quicly_cid_plaintext_t next_cid;
+ u8 use_test_cert_in_ca;
+ char *ca_cert_path;
+} quic_main_t;
+
+quic_main_t *vnet_quic_get_main (void);
+
+#endif /* __included_quic_h__ */
+
+/*
+ * fd.io coding-style-patch-verification: ON
+ *
+ * Local Variables:
+ * eval: (c-set-style "gnu")
+ * End:
+ */