zephyr/modusocket: Implement recv() for UDP sockets.

The foundation of recv() support is per-socket queue of incoming packets,
implemented using Zephyr FIFO object. This patch implements just recv()
for UDP, because TCP recv() requires much more fine-grained control of
network fragments and handling other issues, like EOF condition, etc.

zephyr/modusocket.c

@@ -29,13 +29,21 @@
 
 #include "py/runtime.h"
 
+#include <stdio.h>
 #include <zephyr.h>
 #include <net/net_context.h>
 #include <net/nbuf.h>
 
+#if 0 // print debugging info
+#define DEBUG_printf printf
+#else // don't print debugging info
+#define DEBUG_printf(...) (void)0
+#endif
+
 typedef struct _socket_obj_t {
     mp_obj_base_t base;
     struct net_context *ctx;
+    struct k_fifo recv_q;
 } socket_obj_t;
 
 STATIC const mp_obj_type_t socket_type;
@@ -62,6 +70,44 @@ STATIC void parse_inet_addr(socket_obj_t *socket, mp_obj_t addr_in, struct socka
     sockaddr_in->sin_port = htons(mp_obj_get_int(addr_items[1]));
 }
 
+// Copy data from Zephyr net_buf chain into linear buffer.
+// We don't use net_nbuf_read(), because it's weird (e.g., we'd like to
+// free processed data fragment ASAP, while net_nbuf_read() holds onto
+// the whole fragment chain to do its deeds, and that's minor comparing
+// to the fact that it copies data byte by byte).
+static char *net_buf_gather(struct net_buf *buf, char *to, unsigned max_len) {
+    struct net_buf *tmp = buf->frags;
+    unsigned header_len = net_nbuf_appdata(buf) - tmp->data;
+    net_buf_pull(tmp, header_len);
+
+    while (tmp && max_len) {
+        unsigned len = tmp->len;
+        if (len > max_len) {
+            len = max_len;
+        }
+        memcpy(to, tmp->data, len);
+        to += len;
+        max_len -= len;
+        tmp = net_buf_frag_del(buf, tmp);
+    }
+
+    return to;
+}
+
+// Callback for incoming packets.
+static void sock_received_cb(struct net_context *context, struct net_buf *net_buf, int status, void *user_data) {
+    socket_obj_t *socket = (socket_obj_t*)user_data;
+    DEBUG_printf("recv cb: context: %p, status: %d, buf: %p", context, status, net_buf);
+    if (net_buf) {
+        DEBUG_printf(" (sz=%d, l=%d), token: %p", net_buf->size, net_buf->len, net_nbuf_token(net_buf));
+    }
+    DEBUG_printf("\n");
+
+    // net_buf->frags will be overwritten by fifo, so save it
+    net_nbuf_set_token(net_buf, net_buf->frags);
+    k_fifo_put(&socket->recv_q, net_buf);
+}
+
 // Methods
 
 STATIC void socket_print(const mp_print_t *print, mp_obj_t self_in, mp_print_kind_t kind) {
@@ -79,6 +125,7 @@ STATIC mp_obj_t socket_make_new(const mp_obj_type_t *type, size_t n_args, size_t
 
     socket_obj_t *socket = m_new_obj_with_finaliser(socket_obj_t);
     socket->base.type = type;
+    k_fifo_init(&socket->recv_q);
 
     int family = AF_INET;
     int socktype = SOCK_STREAM;
@@ -114,6 +161,8 @@ STATIC mp_obj_t socket_bind(mp_obj_t self_in, mp_obj_t addr_in) {
     parse_inet_addr(socket, addr_in, &sockaddr);
 
     RAISE_ERRNO(net_context_bind(socket->ctx, &sockaddr, sizeof(sockaddr)));
+    DEBUG_printf("Setting recv cb after bind\n");
+    RAISE_ERRNO(net_context_recv(socket->ctx, sock_received_cb, K_NO_WAIT, socket));
     return mp_const_none;
 }
 STATIC MP_DEFINE_CONST_FUN_OBJ_2(socket_bind_obj, socket_bind);
@@ -126,6 +175,8 @@ STATIC mp_obj_t socket_connect(mp_obj_t self_in, mp_obj_t addr_in) {
     parse_inet_addr(socket, addr_in, &sockaddr);
 
     RAISE_ERRNO(net_context_connect(socket->ctx, &sockaddr, sizeof(sockaddr), NULL, K_FOREVER, NULL));
+    DEBUG_printf("Setting recv cb after connect()\n");
+    RAISE_ERRNO(net_context_recv(socket->ctx, sock_received_cb, K_NO_WAIT, socket));
     return mp_const_none;
 }
 STATIC MP_DEFINE_CONST_FUN_OBJ_2(socket_connect_obj, socket_connect);
@@ -149,6 +200,41 @@ STATIC mp_obj_t socket_send(mp_obj_t self_in, mp_obj_t buf_in) {
 }
 STATIC MP_DEFINE_CONST_FUN_OBJ_2(socket_send_obj, socket_send);
 
+STATIC mp_obj_t socket_recv(mp_obj_t self_in, mp_obj_t len_in) {
+    socket_obj_t *socket = self_in;
+    socket_check_closed(socket);
+
+    enum net_sock_type sock_type = net_context_get_type(socket->ctx);
+    mp_int_t max_len = mp_obj_get_int(len_in);
+    unsigned recv_len;
+    vstr_t vstr;
+
+    if (sock_type == SOCK_DGRAM) {
+
+        struct net_buf *net_buf = k_fifo_get(&socket->recv_q, K_FOREVER);
+        // Restore ->frags overwritten by fifo
+        net_buf->frags = net_nbuf_token(net_buf);
+
+        recv_len = net_nbuf_appdatalen(net_buf);
+        DEBUG_printf("recv: net_buf=%p, appdatalen: %d\n", net_buf, recv_len);
+
+        if (recv_len > max_len) {
+            recv_len = max_len;
+        }
+
+        vstr_init_len(&vstr, recv_len);
+        net_buf_gather(net_buf, vstr.buf, recv_len);
+        net_nbuf_unref(net_buf);
+
+    } else {
+        mp_not_implemented("");
+    }
+
+    mp_obj_t ret = mp_obj_new_str_from_vstr(&mp_type_bytes, &vstr);
+    return ret;
+}
+STATIC MP_DEFINE_CONST_FUN_OBJ_2(socket_recv_obj, socket_recv);
+
 STATIC mp_obj_t socket_close(mp_obj_t self_in) {
     socket_obj_t *socket = self_in;
     if (socket->ctx != NULL) {
@@ -165,6 +251,7 @@ STATIC const mp_map_elem_t socket_locals_dict_table[] = {
     { MP_OBJ_NEW_QSTR(MP_QSTR_bind), (mp_obj_t)&socket_bind_obj },
     { MP_OBJ_NEW_QSTR(MP_QSTR_connect), (mp_obj_t)&socket_connect_obj },
     { MP_OBJ_NEW_QSTR(MP_QSTR_send), (mp_obj_t)&socket_send_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_recv), (mp_obj_t)&socket_recv_obj },
 };
 STATIC MP_DEFINE_CONST_DICT(socket_locals_dict, socket_locals_dict_table);