nuttx/net/netdev/netdev_iob.c

233 lines
6.2 KiB
C
Raw Normal View History

net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
/****************************************************************************
* net/netdev/netdev_iob.c
*
* SPDX-License-Identifier: Apache-2.0
*
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
* 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 <debug.h>
#include <errno.h>
#include <nuttx/net/netdev.h>
/****************************************************************************
* Public Functions
****************************************************************************/
/****************************************************************************
* Name: netdev_iob_prepare
*
* Description:
* Prepare data buffer for a given NIC
* The iob offset will be updated to l2 gruard size by default:
* ----------------------------------------------------------------
* | iob entry |
* ---------------------------------------------------------------|
* |<--- CONFIG_NET_LL_GUARDSIZE -->|<--- io_len/io_pktlen(0) --->|
* ---------------------------------------------------------------|
*
* Assumptions:
* The caller has locked the network.
*
* Returned Value:
* A non-zero copy is returned on success.
*
****************************************************************************/
int netdev_iob_prepare(FAR struct net_driver_s *dev, bool throttled,
unsigned int timeout)
{
/* Prepare iob buffer */
if (dev->d_iob == NULL)
{
dev->d_iob = net_iobtimedalloc(false, timeout);
if (dev->d_iob == NULL && throttled)
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
{
dev->d_iob = net_iobtimedalloc(true, timeout);
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
}
}
if (dev->d_iob == NULL)
{
nwarn("WARNING: IOB Prepare failed for dev %s!\n", dev->d_ifname);
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
return -ENOMEM;
}
/* Update l2 gruard size */
iob_reserve(dev->d_iob, CONFIG_NET_LL_GUARDSIZE);
/* Set the device buffer to l2 */
dev->d_buf = NETLLBUF;
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
return OK;
}
/****************************************************************************
* Name: netdev_iob_prepare_dynamic
*
* Description:
* Pre-alloc the iob for the data to be sent.
*
* Assumptions:
* The caller has locked the network.
*
****************************************************************************/
#ifdef CONFIG_IOB_ALLOC
void netdev_iob_prepare_dynamic(FAR struct net_driver_s *dev, uint16_t size)
{
FAR struct iob_s *iob;
size += CONFIG_NET_LL_GUARDSIZE;
if (dev->d_iob && size <= IOB_BUFSIZE(dev->d_iob))
{
return;
}
/* alloc new iob for jumbo frame */
iob = iob_alloc_dynamic(size);
if (iob == NULL)
{
nerr("ERROR: Failed to allocate an I/O buffer.");
return;
}
iob_reserve(iob, CONFIG_NET_LL_GUARDSIZE);
netdev_iob_replace(dev, iob);
}
#endif
/****************************************************************************
* Name: netdev_iob_replace
*
* Description:
* Replace buffer resources for a given NIC
*
* Assumptions:
* The caller has locked the network and new iob is prepared with
* l2 gruard size as offset.
*
****************************************************************************/
void netdev_iob_replace(FAR struct net_driver_s *dev, FAR struct iob_s *iob)
{
/* Release previous buffer */
netdev_iob_release(dev);
/* Set new buffer */
dev->d_iob = iob;
dev->d_buf = NETLLBUF;
dev->d_len = iob->io_pktlen;
}
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
/****************************************************************************
* Name: netdev_iob_clear
*
* Description:
* Clean up buffer resources for a given NIC
*
* Assumptions:
* The caller has locked the network and dev->d_iob has been
* released or taken away.
*
****************************************************************************/
void netdev_iob_clear(FAR struct net_driver_s *dev)
{
/* Clear the device buffer */
dev->d_iob = NULL;
dev->d_buf = NULL;
dev->d_len = 0;
}
/****************************************************************************
* Name: netdev_iob_release
*
* Description:
* Release buffer resources for a given NIC
*
* Assumptions:
* The caller has locked the network.
*
****************************************************************************/
void netdev_iob_release(FAR struct net_driver_s *dev)
{
/* Release device buffer */
if (dev->d_iob != NULL)
{
iob_free_chain(dev->d_iob);
dev->d_iob = NULL;
}
dev->d_buf = NULL;
net/l2/l3/l4: add support of iob offload 1. Add new config CONFIG_NET_LL_GUARDSIZE to isolation of l2 stack, which will benefit l3(IP) layer for multi-MAC(l2) implementation, especially in some NICs such as celluler net driver. new configuration options: CONFIG_NET_LL_GUARDSIZE CONFIG_NET_LL_GUARDSIZE will reserved l2 buffer header size of network buffer to isolate the L2/L3 (MAC/IP) data on network layer, which will be beneficial to L3 network layer protocol transparent transmission and forwarding ------------------------------------------------------------ Layout of frist iob entry: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- iob | Reserved | io_len | ------------------------------------------------- ------------------------------------------------------------- Layout of different NICs implementation: iob_data (aligned by CONFIG_IOB_ALIGNMENT) | | io_offset(CONFIG_NET_LL_GUARDSIZE) | | ------------------------------------------------- Ethernet | Reserved | ETH_HDRLEN | io_len | ---------------------------------|--------------- 8021Q | Reserved | ETH_8021Q_HDRLEN | io_len | ---------------------------------|--------------- ipforward | Reserved | io_len | ------------------------------------------------- -------------------------------------------------------------------- 2. Support iob offload to l2 driver to avoid unnecessary memory copy Support send/receive iob vectors directly between the NICs and l3/l4 stack to avoid unnecessary memory copies, especially on hardware that supports Scatter/gather, which can greatly improve performance. new interface to support iob offload: ------------------------------------------ | IOB version | original | |----------------------------------------| | devif_iob_poll() | devif_poll() | | ... | ... | ------------------------------------------ -------------------------------------------------------------------- 1> NIC hardware support Scatter/gather transfer TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | dev->d_iob: | --------------- --------------- io_data iob1 | | | iob3 | | | \ --------------- --------------- --------------- | --------------- | iob0 | | | | iob2 | | | | --------------- | --------------- | \ | / / \ | / / ---------------------------------------------- NICs io vector | | | | | | | | | | ---------------------------------------------- RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to each iobs) -------------------------------------------------------------------- 2> CONFIG_IOB_BUFSIZE is greater than MTU: TX: "(CONFIG_IOB_BUFSIZE) > (MAX_NETDEV_PKTSIZE + CONFIG_NET_GUARDSIZE + CONFIG_NET_LL_GUARDSIZE)" tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll("NIC"_txpoll) callback() // "NIC"_txpoll | "NIC"_send() (dev->d_iob->io_data[CONFIG_NET_LL_GUARDSIZE - NET_LL_HDRLEN(dev)]) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | pkt/ipv[4/6]_input()/... | | NICs io vector receive(iov_base to io_data) -------------------------------------------------------------------- 3> Compatible with all old flat buffer NICs TX: tcp_poll()/udp_poll()/pkt_poll()/...(l3|l4) / \ / \ devif_poll_[l3|l4]_connections() devif_iob_send() (nocopy:udp/icmp/...) / \ (copy:tcp) / \ devif_iob_poll(devif_poll_callback()) devif_poll_callback() /* new interface, gather iobs to flat buffer */ / \ / \ devif_poll("NIC"_txpoll) "NIC"_send()(dev->d_buf) RX: [tcp|udp|icmp|...]ipv[4|6]_data_handler()(iob_concat/append to readahead) | | [tcp|udp|icmp|...]_ipv[4|6]_in()/... | | netdev_input() /* new interface, Scatter/gather flat/iob buffer */ | | pkt/ipv[4|6]_input()/... | | NICs io vector receive(Orignal flat buffer) 3. Iperf passthrough on NuttX simulator: ------------------------------------------------- | Protocol | Server | Client | | |-----------------------------------------------| | TCP | 813 | 834 | Mbits/sec | | TCP(Offload) | 1720 | 1100 | Mbits/sec | | UDP | 22 | 757 | Mbits/sec | | UDP(Offload) | 25 | 1250 | Mbits/sec | ------------------------------------------------- Signed-off-by: chao an <anchao@xiaomi.com>
2022-11-23 12:39:40 +01:00
}
/****************************************************************************
* Name: netdev_iob_clone
*
* Description:
* Backup the current iob buffer for a given NIC by cloning it.
*
* Assumptions:
* The caller has locked the network.
*
****************************************************************************/
FAR struct iob_s *netdev_iob_clone(FAR struct net_driver_s *dev,
bool throttled)
{
FAR struct iob_s *iob;
int ret;
iob = iob_tryalloc(throttled);
if (iob == NULL)
{
nwarn("WARNING: IOB alloc failed for dev %s!\n", dev->d_ifname);
return NULL;
}
iob_reserve(iob, CONFIG_NET_LL_GUARDSIZE);
ret = iob_clone_partial(dev->d_iob, dev->d_iob->io_pktlen, 0,
iob, 0, throttled, false);
if (ret < 0)
{
iob_free_chain(iob);
nwarn("WARNING: IOB clone failed for dev %s, ret=%d!\n",
dev->d_ifname, ret);
return NULL;
}
return iob;
}