You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@nuttx.apache.org by xi...@apache.org on 2023/01/11 07:14:10 UTC
[nuttx] branch master updated: move usrsock to kernel space
This is an automated email from the ASF dual-hosted git repository.
xiaoxiang pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/nuttx.git
The following commit(s) were added to refs/heads/master by this push:
new aca1a065a8 move usrsock to kernel space
aca1a065a8 is described below
commit aca1a065a8ebef28600794fc6315c7a73e90cb9d
Author: 梁超众 <li...@xiaomi.com>
AuthorDate: Tue Jan 3 10:54:07 2023 +0800
move usrsock to kernel space
Signed-off-by: 梁超众 <li...@xiaomi.com>
---
boards/sim/sim/sim/configs/rpserver/defconfig | 2 +-
drivers/drivers_initialize.c | 7 +
drivers/usrsock/Kconfig | 24 +
drivers/usrsock/Make.defs | 10 +
drivers/usrsock/usrsock_rpmsg.c | 19 +-
drivers/usrsock/usrsock_rpmsg_server.c | 1135 ++++++++++++++++++++
{drivers => include/nuttx}/usrsock/usrsock_rpmsg.h | 55 +-
net/usrsock/usrsock_devif.c | 4 +-
8 files changed, 1242 insertions(+), 14 deletions(-)
diff --git a/boards/sim/sim/sim/configs/rpserver/defconfig b/boards/sim/sim/sim/configs/rpserver/defconfig
index b94cba812b..e3020a2a5a 100644
--- a/boards/sim/sim/sim/configs/rpserver/defconfig
+++ b/boards/sim/sim/sim/configs/rpserver/defconfig
@@ -45,7 +45,6 @@ CONFIG_NETDEV_STATISTICS=y
CONFIG_NETINIT_DHCPC=y
CONFIG_NETINIT_THREAD=y
CONFIG_NETUTILS_DHCPC=y
-CONFIG_NETUTILS_USRSOCK_RPMSG=y
CONFIG_NET_ARP_SEND=y
CONFIG_NET_BROADCAST=y
CONFIG_NET_ICMP=y
@@ -59,6 +58,7 @@ CONFIG_NET_TCPBACKLOG=y
CONFIG_NET_TCP_WRITE_BUFFERS=y
CONFIG_NET_UDP=y
CONFIG_NET_UDP_WRITE_BUFFERS=y
+CONFIG_NET_USRSOCK_RPMSG_SERVER=y
CONFIG_NSH_ARCHINIT=y
CONFIG_NSH_BUILTIN_APPS=y
CONFIG_NSH_PROMPT_STRING="server> "
diff --git a/drivers/drivers_initialize.c b/drivers/drivers_initialize.c
index f0cf0495bc..9a1377edc2 100644
--- a/drivers/drivers_initialize.c
+++ b/drivers/drivers_initialize.c
@@ -40,6 +40,7 @@
#include <nuttx/serial/pty.h>
#include <nuttx/syslog/syslog.h>
#include <nuttx/syslog/syslog_console.h>
+#include <nuttx/usrsock/usrsock_rpmsg.h>
/****************************************************************************
* Public Functions
@@ -177,4 +178,10 @@ void drivers_initialize(void)
#ifdef CONFIG_RPMSGMTD_SERVER
rpmsgmtd_server_init();
#endif
+
+#ifdef CONFIG_NET_USRSOCK_RPMSG_SERVER
+ /* Initialize the user socket rpmsg server */
+
+ usrsock_rpmsg_server_initialize();
+#endif
}
diff --git a/drivers/usrsock/Kconfig b/drivers/usrsock/Kconfig
index dd005afd42..0a43f1304b 100644
--- a/drivers/usrsock/Kconfig
+++ b/drivers/usrsock/Kconfig
@@ -35,3 +35,27 @@ config NET_USRSOCK_RPMSG_CPUNAME
endmenu
endif # NET_USRSOCK
+
+config NET_USRSOCK_RPMSG_SERVER
+ bool "RPMSG usrsock"
+ default n
+ depends on NET && RPTUN
+ ---help---
+ Enable usrsock through rpmsg channel.
+
+if NET_USRSOCK_RPMSG_SERVER
+
+config NET_USRSOCK_RPMSG_SERVER_NSOCKS
+ int "The maximum number of socket descriptors for usrsock monitoring"
+ default 64
+ ---help---
+ The maximum number of socket description for usrsosck monitoring.
+
+config NET_USRSOCK_RPMSG_SERVER_NIOVEC
+ int "The maximum number of I/O vector for reassemble buffer"
+ default 8
+ ---help---
+ The maximum number of I/O vector for reassemble buffer.
+
+endif # NET_USRSOCK_RPMSG_SERVER
+
diff --git a/drivers/usrsock/Make.defs b/drivers/usrsock/Make.defs
index e7231a5356..e08d90d368 100644
--- a/drivers/usrsock/Make.defs
+++ b/drivers/usrsock/Make.defs
@@ -36,3 +36,13 @@ DEPPATH += --dep-path usrsock
VPATH += :usrsock
endif
+
+ifeq ($(CONFIG_NET_USRSOCK_RPMSG_SERVER),y)
+
+CSRCS += usrsock_rpmsg_server.c
+
+# Include User Socket Driver build support
+
+DEPPATH += --dep-path usrsock
+VPATH += :usrsock
+endif
diff --git a/drivers/usrsock/usrsock_rpmsg.c b/drivers/usrsock/usrsock_rpmsg.c
index 8b2a584dc1..dc5f3da53e 100644
--- a/drivers/usrsock/usrsock_rpmsg.c
+++ b/drivers/usrsock/usrsock_rpmsg.c
@@ -25,8 +25,7 @@
#include <nuttx/net/dns.h>
#include <nuttx/rptun/openamp.h>
#include <nuttx/semaphore.h>
-
-#include "usrsock_rpmsg.h"
+#include <nuttx/usrsock/usrsock_rpmsg.h>
/****************************************************************************
* Private Types
@@ -85,6 +84,7 @@ static int usrsock_rpmsg_send_dns_request(FAR void *arg,
FAR struct rpmsg_endpoint *ept = &priv->ept;
FAR struct usrsock_rpmsg_dns_request_s *dns;
uint32_t len;
+ int ret;
dns = rpmsg_get_tx_payload_buffer(ept, &len, true);
if (dns == NULL)
@@ -98,7 +98,11 @@ static int usrsock_rpmsg_send_dns_request(FAR void *arg,
dns->addrlen = addrlen;
memcpy(dns + 1, addr, addrlen);
- return rpmsg_send_nocopy(ept, dns, sizeof(*dns) + addrlen);
+ net_lock();
+ ret = rpmsg_send_nocopy(ept, dns, sizeof(*dns) + addrlen);
+ net_unlock();
+
+ return ret;
}
#endif
@@ -141,10 +145,14 @@ static int usrsock_rpmsg_ept_cb(FAR struct rpmsg_endpoint *ept,
{
return usrsock_rpmsg_dns_handler(ept, data);
}
- else
+
+ if (common->msgid == USRSOCK_RPMSG_FRAG_RESPONSE)
{
- return usrsock_response(data, len, NULL);
+ data = (FAR char *)data + sizeof(struct usrsock_message_frag_ack_s);
+ len -= sizeof(struct usrsock_message_frag_ack_s);
}
+
+ return usrsock_response(data, len, NULL);
}
static void usrsock_rpmsg_device_created(FAR struct rpmsg_device *rdev,
@@ -223,6 +231,7 @@ int usrsock_request(FAR struct iovec *iov, unsigned int iovcnt)
if (done)
{
+ ret = OK;
break;
}
}
diff --git a/drivers/usrsock/usrsock_rpmsg_server.c b/drivers/usrsock/usrsock_rpmsg_server.c
new file mode 100644
index 0000000000..2795a55e3c
--- /dev/null
+++ b/drivers/usrsock/usrsock_rpmsg_server.c
@@ -0,0 +1,1135 @@
+/****************************************************************************
+ * drivers/usrsock/usrsock_rpmsg_server.c
+ *
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership. The
+ * ASF licenses this file to you 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.
+ *
+ ****************************************************************************/
+
+/****************************************************************************
+ * Included Files
+ ****************************************************************************/
+
+#include <nuttx/config.h>
+
+#include <errno.h>
+#include <poll.h>
+#include <string.h>
+
+#include <sys/ioctl.h>
+
+#include <nuttx/mutex.h>
+#include <nuttx/net/dns.h>
+#include <nuttx/net/net.h>
+#include <nuttx/rptun/openamp.h>
+#include <nuttx/usrsock/usrsock_rpmsg.h>
+#ifdef CONFIG_NETDEV_WIRELESS_IOCTL
+#include <nuttx/wireless/wireless.h>
+#endif
+
+/****************************************************************************
+ * Private Types
+ ****************************************************************************/
+
+struct usrsock_rpmsg_s
+{
+ rmutex_t mutex;
+ struct iovec iov[CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC];
+ struct socket socks[CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS];
+ FAR struct rpmsg_endpoint *epts[CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS];
+ struct pollfd pfds[CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS];
+};
+
+/****************************************************************************
+ * Private Function Prototypes
+ ****************************************************************************/
+
+static bool usrsock_rpmsg_available(FAR struct socket *psock, int cmd);
+
+static int usrsock_rpmsg_send_ack(FAR struct rpmsg_endpoint *ept,
+ uint16_t events,
+ uint32_t xid, int32_t result);
+static int usrsock_rpmsg_send_data_ack(FAR struct rpmsg_endpoint *ept,
+ FAR struct usrsock_message_datareq_ack_s *ack,
+ uint16_t events,
+ uint32_t xid, int32_t result,
+ uint16_t valuelen,
+ uint16_t valuelen_nontrunc,
+ int32_t datalen);
+static int usrsock_rpmsg_send_event(FAR struct rpmsg_endpoint *ept,
+ int16_t usockid, uint16_t events);
+
+static int usrsock_rpmsg_socket_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_close_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_connect_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_sendto_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_recvfrom_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_setsockopt_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_getsockopt_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_getsockname_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_getpeername_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_bind_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_listen_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_accept_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_ioctl_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+static int usrsock_rpmsg_dns_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_);
+
+static bool usrsock_rpmsg_ns_match(FAR struct rpmsg_device *rdev,
+ FAR void *priv_, FAR const char *name,
+ uint32_t dest);
+static void usrsock_rpmsg_ns_bind(FAR struct rpmsg_device *rdev,
+ FAR void *priv_, FAR const char *name,
+ uint32_t dest);
+static void usrsock_rpmsg_ns_unbind(FAR struct rpmsg_endpoint *ept);
+static int usrsock_rpmsg_ept_cb(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len, uint32_t src,
+ FAR void *priv);
+
+static void usrsock_rpmsg_poll_cb(FAR struct pollfd *pfds);
+static int usrsock_rpmsg_poll_setup(FAR struct pollfd *pfds,
+ pollevent_t events);
+
+/****************************************************************************
+ * Private Data
+ ****************************************************************************/
+
+static const rpmsg_ept_cb g_usrsock_rpmsg_handler[] =
+{
+ usrsock_rpmsg_socket_handler,
+ usrsock_rpmsg_close_handler,
+ usrsock_rpmsg_connect_handler,
+ usrsock_rpmsg_sendto_handler,
+ usrsock_rpmsg_recvfrom_handler,
+ usrsock_rpmsg_setsockopt_handler,
+ usrsock_rpmsg_getsockopt_handler,
+ usrsock_rpmsg_getsockname_handler,
+ usrsock_rpmsg_getpeername_handler,
+ usrsock_rpmsg_bind_handler,
+ usrsock_rpmsg_listen_handler,
+ usrsock_rpmsg_accept_handler,
+ usrsock_rpmsg_ioctl_handler,
+ usrsock_rpmsg_dns_handler,
+};
+
+/****************************************************************************
+ * Private Functions
+ ****************************************************************************/
+
+static bool usrsock_rpmsg_available(FAR struct socket *psock, int cmd)
+{
+ int len;
+
+ if (psock_ioctl(psock, cmd, &len, sizeof(len)) == 0)
+ {
+ if (len > 0)
+ {
+ return true;
+ }
+ }
+
+ return false;
+}
+
+static int usrsock_rpmsg_send_ack(FAR struct rpmsg_endpoint *ept,
+ uint16_t events,
+ uint32_t xid, int32_t result)
+{
+ struct usrsock_message_req_ack_s ack;
+
+ ack.head.msgid = USRSOCK_MESSAGE_RESPONSE_ACK;
+ ack.head.flags = (result == -EINPROGRESS);
+ ack.head.events = events;
+
+ ack.xid = xid;
+ ack.result = result == -EINPROGRESS ? 0 : result;
+
+ return rpmsg_send(ept, &ack, sizeof(ack));
+}
+
+static int usrsock_rpmsg_send_data_ack(FAR struct rpmsg_endpoint *ept,
+ FAR struct usrsock_message_datareq_ack_s *ack,
+ uint16_t events,
+ uint32_t xid, int32_t result,
+ uint16_t valuelen,
+ uint16_t valuelen_nontrunc,
+ int32_t datalen)
+{
+ ack->reqack.head.msgid = USRSOCK_MESSAGE_RESPONSE_DATA_ACK;
+ ack->reqack.head.flags = 0;
+ ack->reqack.head.events = events;
+
+ ack->reqack.xid = xid;
+ ack->reqack.result = result;
+
+ if (result < 0)
+ {
+ valuelen = 0;
+ valuelen_nontrunc = 0;
+ datalen = 0;
+ }
+ else if (valuelen > valuelen_nontrunc)
+ {
+ valuelen = valuelen_nontrunc;
+ }
+
+ ack->valuelen = valuelen;
+ ack->valuelen_nontrunc = valuelen_nontrunc;
+
+ return rpmsg_send_nocopy(ept, ack, sizeof(*ack) + valuelen + datalen);
+}
+
+static int usrsock_rpmsg_send_frag_ack(FAR struct rpmsg_endpoint *ept,
+ FAR struct usrsock_message_frag_ack_s *ack,
+ uint16_t events,
+ uint32_t xid, int32_t result,
+ uint32_t datalen)
+{
+ ack->reqack.head.msgid = USRSOCK_RPMSG_FRAG_RESPONSE;
+ ack->reqack.head.flags = 0;
+ ack->reqack.head.events = events;
+ ack->reqack.xid = xid;
+ ack->reqack.result = result;
+ ack->datalen = datalen;
+
+ return rpmsg_send_nocopy(ept, ack, sizeof(*ack) + datalen);
+}
+
+static int usrsock_rpmsg_send_event(FAR struct rpmsg_endpoint *ept,
+ int16_t usockid, uint16_t events)
+{
+ struct usrsock_message_socket_event_s event;
+
+ event.head.msgid = USRSOCK_MESSAGE_SOCKET_EVENT;
+ event.head.flags = USRSOCK_MESSAGE_FLAG_EVENT;
+ event.head.events = events;
+
+ event.usockid = usockid;
+
+ return rpmsg_send(ept, &event, sizeof(event));
+}
+
+static int usrsock_rpmsg_socket_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_socket_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ uint16_t events = 0;
+ int ret = -ENFILE;
+ int retr;
+ int i;
+
+ nxrmutex_lock(&priv->mutex);
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS; i++)
+ {
+ if (priv->epts[i] == NULL)
+ {
+ priv->epts[i] = ept;
+ nxrmutex_unlock(&priv->mutex);
+ ret = psock_socket(req->domain, req->type | SOCK_NONBLOCK,
+ req->protocol, &priv->socks[i]);
+ if (ret >= 0)
+ {
+ ret = i; /* Return index as the usockid */
+ if (req->type != SOCK_STREAM && req->type != SOCK_SEQPACKET)
+ {
+ events = USRSOCK_EVENT_SENDTO_READY;
+ }
+ }
+ else
+ {
+ priv->epts[i] = NULL;
+ }
+
+ break;
+ }
+ }
+
+ if (i == CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ nxrmutex_unlock(&priv->mutex);
+ }
+
+ retr = usrsock_rpmsg_send_ack(ept, events, req->head.xid, ret);
+ if (retr >= 0 && ret >= 0 &&
+ req->type != SOCK_STREAM && req->type != SOCK_SEQPACKET)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[ret], POLLIN);
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_close_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_close_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ int ret = -EBADF;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], 0);
+
+ /* It's safe to close sock here */
+
+ ret = psock_close(&priv->socks[req->usockid]);
+ nxrmutex_lock(&priv->mutex);
+ priv->epts[req->usockid] = NULL;
+ nxrmutex_unlock(&priv->mutex);
+ }
+
+ return usrsock_rpmsg_send_ack(ept, 0, req->head.xid, ret);
+}
+
+static int usrsock_rpmsg_connect_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_connect_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ bool inprogress = false;
+ int ret = -EBADF;
+ int retr;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_connect(&priv->socks[req->usockid],
+ (FAR const struct sockaddr *)(req + 1), req->addrlen);
+ }
+
+ retr = usrsock_rpmsg_send_ack(ept, 0, req->head.xid, ret);
+ if (ret == -EINPROGRESS)
+ {
+ inprogress = true;
+ ret = 0;
+ }
+
+ if (retr >= 0 && ret >= 0)
+ {
+ pollevent_t events = POLLIN;
+
+ if (inprogress)
+ {
+ events |= POLLOUT;
+ }
+
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], events);
+ if (!inprogress)
+ {
+ retr = usrsock_rpmsg_send_event(ept, req->usockid,
+ USRSOCK_EVENT_SENDTO_READY);
+ }
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_sendto_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_sendto_s *req;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ uint16_t events = 0;
+ ssize_t ret = -EBADF;
+ size_t total;
+ int retr;
+ int i;
+
+ if (priv->iov[0].iov_base)
+ {
+ size_t hlen;
+ struct msghdr msg =
+ {
+ };
+
+ req = priv->iov[0].iov_base;
+ hlen = sizeof(*req) + req->addrlen;
+
+ total = len;
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC; i++)
+ {
+ if (!priv->iov[i].iov_base)
+ {
+ priv->iov[i].iov_base = data;
+ priv->iov[i].iov_len = len;
+ rpmsg_hold_rx_buffer(ept, data);
+ break;
+ }
+
+ total += priv->iov[i].iov_len;
+ }
+
+ if (i == CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC)
+ {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ /* Partial packet ? continue to fetch */
+
+ if (req->buflen > total - hlen)
+ {
+ return 0;
+ }
+ else if (req->buflen < total - hlen)
+ {
+ ret = -EINVAL;
+ goto out;
+ }
+
+ /* Skip the sendto header from I/O vector */
+
+ priv->iov[0].iov_base = (FAR char *)priv->iov[0].iov_base + hlen;
+ priv->iov[0].iov_len -= hlen;
+
+ msg.msg_name = req->addrlen ? (FAR void *)(req + 1) : NULL;
+ msg.msg_namelen = req->addrlen;
+ msg.msg_iov = priv->iov;
+ msg.msg_iovlen = i + 1;
+
+ ret = psock_sendmsg(&priv->socks[req->usockid], &msg, req->flags);
+
+ /* Recover the I/O vector */
+
+ priv->iov[0].iov_base = (FAR char *)priv->iov[0].iov_base - hlen;
+ priv->iov[0].iov_len += hlen;
+ }
+ else
+ {
+ req = data;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ total = sizeof(*req) + req->addrlen + req->buflen;
+ if (total > len)
+ {
+ priv->iov[0].iov_base = data;
+ priv->iov[0].iov_len = len;
+
+ rpmsg_hold_rx_buffer(ept, data);
+ return 0;
+ }
+ else
+ {
+ ret = psock_sendto(&priv->socks[req->usockid],
+ (FAR const void *)(req + 1) + req->addrlen, req->buflen,
+ req->flags,
+ req->addrlen ?
+ (FAR const struct sockaddr *)(req + 1) : NULL,
+ req->addrlen);
+ }
+ }
+ }
+
+out:
+ if (ret > 0 &&
+ usrsock_rpmsg_available(&priv->socks[req->usockid], FIONSPACE))
+ {
+ events |= USRSOCK_EVENT_SENDTO_READY;
+ }
+
+ retr = usrsock_rpmsg_send_ack(ept, events, req->head.xid, ret);
+ if (retr >= 0 && events == 0)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], POLLOUT);
+ }
+
+ if (priv->iov[0].iov_base)
+ {
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC; i++)
+ {
+ if (priv->iov[i].iov_base == NULL)
+ {
+ break;
+ }
+
+ rpmsg_release_rx_buffer(ept, priv->iov[i].iov_base);
+ priv->iov[i].iov_base = NULL;
+ priv->iov[i].iov_len = 0;
+ }
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_recvfrom_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ struct iovec iov[CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC];
+ FAR struct usrsock_request_recvfrom_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ socklen_t outaddrlen = req->max_addrlen;
+ socklen_t inaddrlen = req->max_addrlen;
+ size_t buflen = req->max_buflen;
+ size_t totlen = 0;
+ ssize_t ret = -EBADF;
+ uint32_t len = buflen;
+ uint16_t events = 0;
+ uint8_t i = 0;
+ int retr;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (sizeof(*ack) + inaddrlen + buflen < len)
+ {
+ len = sizeof(*ack) + inaddrlen + buflen;
+ }
+
+ memset(iov, 0, sizeof(iov));
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_recvfrom(&priv->socks[req->usockid],
+ (FAR void *)(ack + 1) + inaddrlen,
+ len - sizeof(*ack) - inaddrlen,
+ req->flags,
+ outaddrlen ? (FAR struct sockaddr *)(ack + 1) : NULL,
+ outaddrlen ? &outaddrlen : NULL);
+ totlen = ret;
+ if (ret > 0)
+ {
+ if (outaddrlen < inaddrlen)
+ {
+ memcpy((FAR void *)(ack + 1) + outaddrlen,
+ (FAR void *)(ack + 1) + inaddrlen, ret);
+ }
+
+ while (totlen < buflen &&
+ i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC)
+ {
+ uint32_t offset = sizeof(struct usrsock_message_frag_ack_s);
+
+ if (!usrsock_rpmsg_available(&priv->socks[req->usockid],
+ FIONREAD))
+ {
+ break;
+ }
+
+ iov[i].iov_base = rpmsg_get_tx_payload_buffer(ept,
+ &len,
+ false);
+ if (!iov[i].iov_base)
+ {
+ events |= USRSOCK_EVENT_RECVFROM_AVAIL;
+ break;
+ }
+
+ DEBUGASSERT(len > offset);
+ if (buflen - totlen < len - offset)
+ {
+ len = buflen - totlen + offset;
+ }
+
+ /* Should never wait */
+
+ iov[i].iov_len =
+ psock_recvfrom(&priv->socks[req->usockid],
+ (FAR char *)iov[i].iov_base + offset,
+ len - offset,
+ req->flags | MSG_DONTWAIT,
+ NULL, NULL);
+ if ((ssize_t)iov[i].iov_len > 0)
+ {
+ totlen += iov[i].iov_len;
+ if (iov[i].iov_len < len - offset)
+ {
+ break;
+ }
+ }
+ else
+ {
+ iov[i].iov_len = 0;
+ events |= USRSOCK_EVENT_RECVFROM_AVAIL;
+ break;
+ }
+
+ i++;
+ }
+ }
+ }
+
+ retr = usrsock_rpmsg_send_data_ack(ept,
+ ack, events, req->head.xid,
+ totlen, inaddrlen, outaddrlen,
+ ret);
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NIOVEC; i++)
+ {
+ if (!iov[i].iov_base)
+ {
+ break;
+ }
+
+ if (!iov[i].iov_len || retr <= 0)
+ {
+ rpmsg_release_tx_buffer(ept, iov[i].iov_base);
+ iov[i].iov_base = NULL;
+ continue;
+ }
+
+ usrsock_rpmsg_send_frag_ack(ept,
+ (FAR struct usrsock_message_frag_ack_s *)iov[i].iov_base,
+ events, req->head.xid, totlen, iov[i].iov_len);
+ }
+
+ if (retr >= 0 && events == 0)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], POLLIN);
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_setsockopt_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_setsockopt_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ int ret = -EBADF;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_setsockopt(&priv->socks[req->usockid],
+ req->level, req->option, req + 1, req->valuelen);
+ }
+
+ return usrsock_rpmsg_send_ack(ept, 0, req->head.xid, ret);
+}
+
+static int usrsock_rpmsg_getsockopt_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_getsockopt_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ socklen_t optlen = req->max_valuelen;
+ int ret = -EBADF;
+ uint32_t len;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_getsockopt(&priv->socks[req->usockid],
+ req->level, req->option, ack + 1, &optlen);
+ }
+
+ return usrsock_rpmsg_send_data_ack(ept,
+ ack, 0, req->head.xid, ret, optlen, optlen, ret);
+}
+
+static int usrsock_rpmsg_getsockname_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_getsockname_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ socklen_t outaddrlen = req->max_addrlen;
+ socklen_t inaddrlen = req->max_addrlen;
+ int ret = -EBADF;
+ uint32_t len;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_getsockname(&priv->socks[req->usockid],
+ (FAR struct sockaddr *)(ack + 1), &outaddrlen);
+ }
+
+ return usrsock_rpmsg_send_data_ack(ept,
+ ack, 0, req->head.xid, ret, inaddrlen, outaddrlen, ret);
+}
+
+static int usrsock_rpmsg_getpeername_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_getpeername_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ socklen_t outaddrlen = req->max_addrlen;
+ socklen_t inaddrlen = req->max_addrlen;
+ int ret = -EBADF;
+ uint32_t len;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_getpeername(&priv->socks[req->usockid],
+ (FAR struct sockaddr *)(ack + 1), &outaddrlen);
+ }
+
+ return usrsock_rpmsg_send_data_ack(ept,
+ ack, 0, req->head.xid, ret, inaddrlen, outaddrlen, ret);
+}
+
+static int usrsock_rpmsg_bind_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_bind_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ int ret = -EBADF;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_bind(&priv->socks[req->usockid],
+ (FAR const struct sockaddr *)(req + 1), req->addrlen);
+ }
+
+ return usrsock_rpmsg_send_ack(ept, 0, req->head.xid, ret);
+}
+
+static int usrsock_rpmsg_listen_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_listen_s *req = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ int ret = -EBADF;
+ int retr;
+
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = psock_listen(&priv->socks[req->usockid], req->backlog);
+ }
+
+ retr = usrsock_rpmsg_send_ack(ept, 0, req->head.xid, ret);
+ if (retr >= 0 && ret >= 0)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], POLLIN);
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_accept_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_accept_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ socklen_t outaddrlen = req->max_addrlen;
+ socklen_t inaddrlen = req->max_addrlen;
+ uint32_t len;
+ int ret = -EBADF;
+ int i = 0;
+ int retr;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ ret = -ENFILE; /* Assume no free socket handler */
+ nxrmutex_lock(&priv->mutex);
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS; i++)
+ {
+ if (priv->epts[i] == NULL)
+ {
+ priv->epts[i] = ept;
+ nxrmutex_unlock(&priv->mutex);
+ ret = psock_accept(&priv->socks[req->usockid],
+ outaddrlen ? (FAR struct sockaddr *)(ack + 1) : NULL,
+ outaddrlen ? &outaddrlen : NULL, &priv->socks[i]);
+ if (ret >= 0)
+ {
+ int nonblock = 1;
+
+ psock_ioctl(&priv->socks[i], FIONBIO, &nonblock);
+
+ /* Append index as usockid to the payload */
+
+ if (outaddrlen <= inaddrlen)
+ {
+ *(FAR int16_t *)
+ ((FAR void *)(ack + 1) + outaddrlen) = i;
+ }
+ else
+ {
+ *(FAR int16_t *)
+ ((FAR void *)(ack + 1) + inaddrlen) = i;
+ }
+
+ ret = sizeof(int16_t); /* Return usockid size */
+ }
+ else
+ {
+ priv->epts[i] = NULL;
+ }
+
+ break;
+ }
+ }
+
+ if (i == CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ nxrmutex_unlock(&priv->mutex);
+ }
+ }
+
+ retr = usrsock_rpmsg_send_data_ack(ept,
+ ack, 0, req->head.xid, ret, inaddrlen, outaddrlen, ret);
+ if (retr >= 0 && ret >= 0)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[req->usockid], POLLIN);
+ usrsock_rpmsg_poll_setup(&priv->pfds[i], POLLIN);
+ usrsock_rpmsg_send_event(ept, i, USRSOCK_EVENT_SENDTO_READY);
+ }
+
+ return retr;
+}
+
+static int usrsock_rpmsg_ioctl_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len_,
+ uint32_t src, FAR void *priv_)
+{
+ FAR struct usrsock_request_ioctl_s *req = data;
+ FAR struct usrsock_message_datareq_ack_s *ack;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+#ifdef CONFIG_NETDEV_WIRELESS_IOCTL
+ FAR struct iwreq *wlreq;
+ FAR struct iwreq *wlack;
+#endif
+ int ret = -EBADF;
+ uint32_t len;
+
+ ack = rpmsg_get_tx_payload_buffer(ept, &len, true);
+ if (req->usockid >= 0 &&
+ req->usockid < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS)
+ {
+ memcpy(ack + 1, req + 1, len_ - sizeof(*req));
+#ifdef CONFIG_NETDEV_WIRELESS_IOCTL
+ wlreq = (FAR struct iwreq *)(req + 1);
+ wlack = (FAR struct iwreq *)(ack + 1);
+ if (WL_IS80211POINTERCMD(req->cmd) && wlreq->u.data.pointer)
+ {
+ wlack->u.data.pointer = wlack + 1;
+ }
+#endif
+
+ ret = psock_ioctl(&priv->socks[req->usockid],
+ req->cmd, (unsigned long)(ack + 1));
+
+#ifdef CONFIG_NETDEV_WIRELESS_IOCTL
+ if (WL_IS80211POINTERCMD(req->cmd) && wlreq->u.data.pointer)
+ {
+ if (ret >= 0)
+ {
+ ret = wlreq->u.data.length;
+ }
+
+ wlack->u.data.pointer = wlreq->u.data.pointer;
+ }
+#endif
+ }
+
+ return usrsock_rpmsg_send_data_ack(ept,
+ ack, 0, req->head.xid, ret, req->arglen, req->arglen, ret);
+}
+
+static int usrsock_rpmsg_dns_handler(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len,
+ uint32_t src, FAR void *priv_)
+{
+#ifdef CONFIG_NETDB_DNSCLIENT
+ FAR struct usrsock_rpmsg_dns_request_s *dns = data;
+
+ dns_add_nameserver((FAR struct sockaddr *)(dns + 1), dns->addrlen);
+#endif
+
+ return 0;
+}
+
+#ifdef CONFIG_NETDB_DNSCLIENT
+static int usrsock_rpmsg_send_dns_event(FAR void *arg,
+ FAR struct sockaddr *addr,
+ socklen_t addrlen)
+{
+ FAR struct rpmsg_endpoint *ept = arg;
+ FAR struct usrsock_rpmsg_dns_event_s *dns;
+ uint32_t len;
+
+ dns = rpmsg_get_tx_payload_buffer(ept, &len, true);
+
+ dns->head.msgid = USRSOCK_RPMSG_DNS_EVENT;
+ dns->head.flags = USRSOCK_MESSAGE_FLAG_EVENT;
+
+ dns->addrlen = addrlen;
+ memcpy(dns + 1, addr, addrlen);
+
+ return rpmsg_send_nocopy(ept, dns, sizeof(*dns) + addrlen);
+}
+#endif
+
+static bool usrsock_rpmsg_ns_match(FAR struct rpmsg_device *rdev,
+ FAR void *priv_, FAR const char *name,
+ uint32_t dest)
+{
+ return !strcmp(name, USRSOCK_RPMSG_EPT_NAME);
+}
+
+static void usrsock_rpmsg_ns_bind(FAR struct rpmsg_device *rdev,
+ FAR void *priv_, FAR const char *name,
+ uint32_t dest)
+{
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+ FAR struct rpmsg_endpoint *ept;
+ int ret;
+
+ ept = kmm_zalloc(sizeof(struct rpmsg_endpoint));
+ if (!ept)
+ {
+ return;
+ }
+
+ ept->priv = priv;
+
+ ret = rpmsg_create_ept(ept, rdev, USRSOCK_RPMSG_EPT_NAME,
+ RPMSG_ADDR_ANY, dest,
+ usrsock_rpmsg_ept_cb, usrsock_rpmsg_ns_unbind);
+ if (ret < 0)
+ {
+ kmm_free(ept);
+ return;
+ }
+
+#ifdef CONFIG_NETDB_DNSCLIENT
+ dns_register_notify(usrsock_rpmsg_send_dns_event, ept);
+#endif
+}
+
+static void usrsock_rpmsg_ns_unbind(FAR struct rpmsg_endpoint *ept)
+{
+ FAR struct usrsock_rpmsg_s *priv = ept->priv;
+ int i;
+
+#ifdef CONFIG_NETDB_DNSCLIENT
+ dns_unregister_notify(usrsock_rpmsg_send_dns_event, ept);
+#endif
+
+ /* Collect all socks belong to the dead client */
+
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS; i++)
+ {
+ if (priv->epts[i] == ept)
+ {
+ usrsock_rpmsg_poll_setup(&priv->pfds[i], 0);
+
+ /* It's safe to close socks here */
+
+ psock_close(&priv->socks[i]);
+ nxrmutex_lock(&priv->mutex);
+ priv->epts[i] = NULL;
+ nxrmutex_unlock(&priv->mutex);
+ }
+ }
+
+ rpmsg_destroy_ept(ept);
+}
+
+static int usrsock_rpmsg_ept_cb(FAR struct rpmsg_endpoint *ept,
+ FAR void *data, size_t len, uint32_t src,
+ FAR void *priv_)
+{
+ FAR struct usrsock_request_common_s *common = data;
+ FAR struct usrsock_rpmsg_s *priv = priv_;
+
+ if (priv->iov[0].iov_base)
+ {
+ return usrsock_rpmsg_sendto_handler(ept, data, len, src, priv);
+ }
+ else if (common->reqid >= 0 && common->reqid <= USRSOCK_REQUEST__MAX)
+ {
+ return g_usrsock_rpmsg_handler[common->reqid](ept, data, len,
+ src, priv);
+ }
+
+ return -EINVAL;
+}
+
+static int usrsock_rpmsg_poll_setup(FAR struct pollfd *pfds,
+ pollevent_t events)
+{
+ FAR struct usrsock_rpmsg_s *priv = (FAR struct usrsock_rpmsg_s *)pfds->arg;
+ int ret = 0;
+
+ net_lock();
+
+ if (events)
+ {
+ if (!pfds->events)
+ {
+ pfds->revents = 0;
+ pfds->events = events;
+ ret = psock_poll(&priv->socks[pfds->fd], pfds, true);
+ }
+ else
+ {
+ pfds->events = events;
+ }
+ }
+ else
+ {
+ pfds->events = 0;
+ ret = psock_poll(&priv->socks[pfds->fd], pfds, false);
+ }
+
+ net_unlock();
+
+ return ret;
+}
+
+static void usrsock_rpmsg_poll_cb(FAR struct pollfd *pfds)
+{
+ FAR struct usrsock_rpmsg_s *priv = (FAR struct usrsock_rpmsg_s *)pfds->arg;
+ int events = 0;
+
+ nxrmutex_lock(&priv->mutex);
+
+ if (!priv->epts[pfds->fd])
+ {
+ nxrmutex_unlock(&priv->mutex);
+ return;
+ }
+
+ if (pfds->revents & POLLIN)
+ {
+ events |= USRSOCK_EVENT_RECVFROM_AVAIL;
+
+ /* Stop poll in until recv get called */
+
+ pfds->events &= ~POLLIN;
+ pfds->revents &= ~POLLIN;
+ }
+
+ if (pfds->revents & POLLOUT)
+ {
+ events |= USRSOCK_EVENT_SENDTO_READY;
+
+ /* Stop poll out until send get called */
+
+ pfds->events &= ~POLLOUT;
+ pfds->revents &= ~POLLOUT;
+ }
+
+ if (pfds->revents & (POLLHUP | POLLERR))
+ {
+ events |= USRSOCK_EVENT_REMOTE_CLOSED;
+
+ /* Check data that has not been recv */
+
+ if (usrsock_rpmsg_available(&priv->socks[pfds->fd], FIONREAD))
+ {
+ events |= USRSOCK_EVENT_RECVFROM_AVAIL;
+ }
+ }
+
+ if (!(pfds->events & (POLLIN | POLLOUT)))
+ {
+ usrsock_rpmsg_poll_setup(pfds, 0);
+ }
+
+ if (events != 0)
+ {
+ usrsock_rpmsg_send_event(priv->epts[pfds->fd], pfds->fd, events);
+ }
+
+ nxrmutex_unlock(&priv->mutex);
+}
+
+/****************************************************************************
+ * Public Functions
+ ****************************************************************************/
+
+int usrsock_rpmsg_server_initialize(void)
+{
+ FAR struct usrsock_rpmsg_s *priv;
+ int ret;
+ int i;
+
+ priv = kmm_calloc(1, sizeof(*priv));
+ if (priv == NULL)
+ {
+ return -ENOMEM;
+ }
+
+ nxrmutex_init(&priv->mutex);
+
+ /* Setup poll callback settings */
+
+ for (i = 0; i < CONFIG_NET_USRSOCK_RPMSG_SERVER_NSOCKS; i++)
+ {
+ priv->pfds[i].fd = i;
+ priv->pfds[i].arg = priv;
+ priv->pfds[i].cb = usrsock_rpmsg_poll_cb;
+ }
+
+ ret = rpmsg_register_callback(priv,
+ NULL,
+ NULL,
+ usrsock_rpmsg_ns_match,
+ usrsock_rpmsg_ns_bind);
+ if (ret >= 0)
+ {
+ return ret;
+ }
+
+ nxrmutex_destroy(&priv->mutex);
+ kmm_free(priv);
+ return ret;
+}
diff --git a/drivers/usrsock/usrsock_rpmsg.h b/include/nuttx/usrsock/usrsock_rpmsg.h
similarity index 58%
rename from drivers/usrsock/usrsock_rpmsg.h
rename to include/nuttx/usrsock/usrsock_rpmsg.h
index 541bbe2ae7..e950906aa4 100644
--- a/drivers/usrsock/usrsock_rpmsg.h
+++ b/include/nuttx/usrsock/usrsock_rpmsg.h
@@ -1,5 +1,5 @@
/****************************************************************************
- * drivers/usrsock/usrsock_rpmsg.h
+ * include/nuttx/usrsock/usrsock_rpmsg.h
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
@@ -18,8 +18,8 @@
*
****************************************************************************/
-#ifndef __DRIVERS_USRSOCK_USRSOCK_RPMSG_H
-#define __DRIVERS_USRSOCK_USRSOCK_RPMSG_H
+#ifndef __INCLUDE_NUTTX_USRSOCK_USRSOCK_RPMSG_H
+#define __INCLUDE_NUTTX_USRSOCK_USRSOCK_RPMSG_H
/****************************************************************************
* Included Files
@@ -33,8 +33,10 @@
#define USRSOCK_RPMSG_EPT_NAME "rpmsg-usrsock"
-#define USRSOCK_RPMSG_DNS_REQUEST USRSOCK_REQUEST__MAX
-#define USRSOCK_RPMSG_DNS_EVENT 127
+#define USRSOCK_RPMSG_DNS_REQUEST USRSOCK_REQUEST__MAX
+
+#define USRSOCK_RPMSG_DNS_EVENT 126
+#define USRSOCK_RPMSG_FRAG_RESPONSE 127
/****************************************************************************
* Public Types
@@ -58,4 +60,45 @@ begin_packed_struct struct usrsock_rpmsg_dns_event_s
uint16_t addrlen;
} end_packed_struct;
-#endif /* __DRIVERS_USRSOCK_USRSOCK_RPMSG_H */
+/* fragemented ack message */
+
+begin_packed_struct struct usrsock_message_frag_ack_s
+{
+ struct usrsock_message_req_ack_s reqack;
+
+ /* head.result => positive buflen, negative error-code. */
+
+ uint32_t datalen; /* fragment's payload length */
+} end_packed_struct;
+
+/****************************************************************************
+ * Public Function Prototypes
+ ****************************************************************************/
+
+#ifdef __cplusplus
+#define EXTERN extern "C"
+extern "C"
+{
+#else
+#define EXTERN extern
+#endif
+
+/****************************************************************************
+ * Name: usrsock_rpmsg_server_initialize()
+ *
+ * Description:
+ * Initialize the User Socket rpmsg server. Called once and only
+ * from the driver layer.
+ *
+ ****************************************************************************/
+
+#ifdef CONFIG_NET_USRSOCK_RPMSG_SERVER
+int usrsock_rpmsg_server_initialize(void);
+#endif
+
+#undef EXTERN
+#if defined(__cplusplus)
+}
+#endif
+
+#endif /* __INCLUDE_NUTTX_USRSOCK_USRSOCK_RPMSG_H */
diff --git a/net/usrsock/usrsock_devif.c b/net/usrsock/usrsock_devif.c
index 7f40a29f83..354fe40b3e 100644
--- a/net/usrsock/usrsock_devif.c
+++ b/net/usrsock/usrsock_devif.c
@@ -625,7 +625,7 @@ ssize_t usrsock_iovec_put(FAR struct iovec *iov, int iovcnt, size_t pos,
}
/****************************************************************************
- * Name: usrsock_request() - finish usrsock's request
+ * Name: usrsock_do_request() - finish usrsock's request
****************************************************************************/
int usrsock_do_request(FAR struct usrsock_conn_s *conn,
@@ -657,7 +657,7 @@ int usrsock_do_request(FAR struct usrsock_conn_s *conn,
req->ackxid = req_head->xid;
ret = usrsock_request(iov, iovcnt);
- if (ret == OK)
+ if (ret >= 0)
{
/* Wait ack for request. */