Commit Graph

24 Commits

Author SHA1 Message Date
chao an
34d2cde8a8 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>
2022-12-03 11:47:04 +08:00
liyi
391b501639 net: extract l3 header build code into new functions
Signed-off-by: liyi <liyi25@xiaomi.com>
2022-11-29 18:36:15 +08:00
chao an
a8d3286258 net: move device buffer define to common header
Signed-off-by: chao an <anchao@xiaomi.com>
2022-10-28 00:32:16 -04:00
Petro Karashchenko
08043fb5bc net: unify FAR keyword usage for all net buffer memory mapped buffers
Signed-off-by: Petro Karashchenko <petro.karashchenko@gmail.com>
2022-01-20 01:42:56 +08:00
Gregory Nutt
04c9079954 Back in 2007, an early network implementation was developed for NuttX. This early development was inspired largely by uIP 1.0 and recognition of that was noted in the then BSD license headers. Over the next 14 years, a new, much more advanced, original network was developed for NuttX. However, some references to Adam Dunkels were still present in the file headers.
Because of this, it will take some time to detangle the licensing under net/.  Many new features, original features were added to the NuttX network.  Clearly, any references to Adam Dunkels in the files that implement these new features that have no counterpart in uIP 1.0 are errors.

This PR removes the references and converts the license headers to Apache 2.0 where possible.  The affected files include only (1) the implementation of IPv6 (including neighbor support under ICMPv6) and (2) Raw sockets.  Neither of these features are present in uIP 1.0 and the licenses can be freely updated.
2021-09-17 21:49:44 -05:00
patacongo
51cbc85535 net/: Run current nxstyle against all .c files under net/. Correct new problems identified by nxstyle. This is primarily a verification of the nxstyle changes (which all look good). (#89)
Co-authored-by: Gregory Nutt <gnutt@nuttx.org>
2020-01-12 21:17:29 +01:00
Gregory Nutt
98f431d8d8 net/: Run all .c files under net/ through tools/nxstyle and fix all resulting complaints. 2020-01-11 13:41:33 -03:00
Xiang Xiao
71e1056e6d net/icmpv6, neighbor, and netdev: Rename netdev_dev_lladdrsize() to netdev_lladdrsize() and move the prototype to include/nuttx/nex/netdev.h, giving is global scope within the OS. Reviewer: This degrades the integrity of the modular architecture by introducing a new coupling via C function call. Not a good thing at all. 2018-11-09 16:05:11 -06:00
Xiang Xiao
4f25b83f8f et/icmpv6: Remove the coupling between ICMPv6 and Ethernet by moving all Ethernet stuff to neighbor_ethernet_out.c and make the mac address as first field. 2018-11-09 15:41:23 -06:00
Xiang Xiao
a3cda3867b net/icmpv6/icmpv6_linkipaddr.c: Encapsulate the link scope address generation into icmpv6_linkipaddr() 2018-11-09 15:28:58 -06:00
Gregory Nutt
33ed27ae65 Squashed commit of the following:
net/mld: Checksum calculation needs to know the full size of the IPv6 header, including the size of the extension headers.  The payload size in the IPv64 header must include the extension headers (not considered part of the header in this case).  Fixes a few additional errors in size, endian-ness and checksum calculations.  Wireshark now reports the the outgoing Report has a good checksum.

    net/mld/mld_query.c:  Add a cast to assure that the left shift does not overflow.
2018-11-05 08:56:07 -06:00
Gregory Nutt
22cd0d47fa This commit attempts remove some long standard confusion in naming and some actual problems that result from the naming confusion. The basic problem is the standard MTU does not include the size of the Ethernet header. For clarity, I changed the naming of most things called MTU to PKTSIZE. For example, CONFIG_NET_ETH_MTU is now CONFIG_NET_ETH_PKTSIZE.
This makes the user interface a little hostile.  People thing of an MTU of 1500 bytes, but the corresponding packet is really 1514 bytes (including the 14 byte Ethernet header).  A more friendly solution would configure the MTU (as before), but then derive the packet buffer size by adding the MAC header length.  Instead, we define the packet buffer size then derive the MTU.

The MTU is not common currency in networking.  On the wire, the only real issue is the MSS which is derived from MTU by subtracting the IP header and TCP header sizes (for the case of TCP).  Now it is derived for the PKTSIZE by subtracting the IP header, the TCP header, and the MAC header sizes.  So we should be all good and without the recurring 14 byte error in MTU's and MSS's.

Squashed commit of the following:

    Trivial update to fix some spacing issues.
    net/: Rename several macros containing _MTU to _PKTSIZE.
    net/: Rename CONFIG_NET_SLIP_MTU to CONFIG_NET_SLIP_PKTSIZE and similarly for CONFIG_NET_TUN_MTU.  These are not the MTU which does not include the size of the link layer header.  These are the full size of the packet buffer memory (minus any GUARD bytes).
    net/: Rename CONFIG_NET_6LOWPAN_MTU to CONFIG_NET_6LOWPAN_PKTSIZE and similarly for CONFIG_NET_TUN_MTU.  These are not the MTU which does not include the size of the link layer header.  These are the full size of the packet buffer memory (minus any GUARD bytes).
    net/: Rename CONFIG_NET_ETH_MTU to CONFIG_NET_ETH_PKTSIZE.  This is not the MTU which does not include the size of the link layer header.  This is the full size of the packet buffer memory (minus any GUARD bytes).
    net/: Rename the file d_mtu in the network driver structure to d_pktsize.  That value saved there is not the MTU.  The packetsize is the memory large enough to hold the maximum packet PLUS the size of the link layer header.  The MTU does not include the link layer header.
2018-07-04 14:10:40 -06:00
Gregory Nutt
d741185c91 net/icmpv6/icmpv6_radvertise.c: set the prefix length to preconfigured value 2017-11-08 07:50:36 -06:00
Sakari Kapanen
795e884aea net/icmpv6/icmpv6_radvertise.c: Add option to manually specify router prefix 2017-11-08 07:47:41 -06:00
Sakari Kapanen
d0f72617ef From b7b1826930c7da10416bceebce5d63dc8e4086ca Mon Sep 17 00:00:00 2001
From: Sakari Kapanen <sakari.kapanen@optofidelity.com>
Date: Sun, 24 Sep 2017 13:19:25 +0300
Subject: [PATCH] ICMPv6: fix router advertisement

Fixes several errors preventing icmpv6_radvertise.c from being compiled. Fixes conversions to network byte order (namely vlifetime, plifetime, mtu).

IPv6 source address is set to link-local IP address instead of the address in the netdev structure. This is in compliance to RFC 4861. RA didn't work on Linux before this change.

Finally, router prefix and prefix length are derived from the IPv6 address and netmask in the netdev structure. This seems to make more sense than using a predefined, separate prefix from the config.
2017-09-24 06:33:05 -06:00
Gregory Nutt
aeb59383cf Remove CONFIG_NET_MULTILINK. This increases code size by a little, but greatly reduces the complexity of the network code. 2017-08-08 15:26:09 -06:00
Gregory Nutt
2ada7d5892 IPv6: Remove comparisons to the address with all ones set. IPv6 does not support broadcast addresses and certainly not in that form. Replace with multicast addresses beginning with 0xff02. 2017-08-07 11:50:50 -06:00
Gregory Nutt
b4a0ac53a8 Networking: Improve naming and simplify some logic of previous commit. 2017-07-07 06:51:52 -06:00
Gregory Nutt
70c6b52132 ICMPv6: Fix an address size calculation that was bungled in a recent commit. 2017-07-06 19:37:01 -06:00
Gregory Nutt
a5f361e984 ICMPv6: Fix so that ICMPv6 can be used with 6LoWPAN. 2017-07-06 12:00:59 -06:00
Gregory Nutt
79256573e1 net: network drver now retains Ethernet MAC address in a union so that other link layer addresses may be used in a MULTILINK environment. 2017-04-22 11:10:30 -06:00
Gregory Nutt
43eb04bb8f Without lowsyslog() *llinfo() is not useful. Eliminate and replace with *info(). 2016-06-20 11:59:15 -06:00
Gregory Nutt
fc3540cffe Replace all occurrences of vdbg with vinfo 2016-06-11 11:59:51 -06:00
Gregory Nutt
111f4a45a3 Forgot to add this file before a previous commit 2015-02-05 09:43:07 -06:00