/linuxptp

My Reading code of linuxptp

Primary LanguageCGNU General Public License v2.0GPL-2.0

Introduction

This software is an implementation of the Precision Time Protocol (PTP) according to IEEE standard 1588 for Linux. The dual design goals are to provide a robust implementation of the standard and to use the most relevant and modern Application Programming Interfaces (API) offered by the Linux kernel. Supporting legacy APIs and other platforms is not a goal.

License

The software is copyrighted by the authors and is licensed under the GNU General Public License. See the file, COPYING, for details of the license terms.

Features

  • Supports hardware and software time stamping via the Linux SO_TIMESTAMPING socket option.
  • Supports the Linux PTP Hardware Clock (PHC) subsystem by using the clock_gettime family of calls, including the clock_adjtimex system call.
  • Implements Boundary Clock (BC), Ordinary Clock (OC) and Transparent Clock (TC).
  • Transport over UDP/IPv4, UDP/IPv6, and raw Ethernet (Layer 2).
  • Supports IEEE 802.1AS-2011 in the role of end station.
  • Modular design allowing painless addition of new transports and clock servos.
  • Implements unicast operation.
  • Supports a number of profiles, including:
    • The automotive profile
    • The default 1588 profile.
    • The enterprise profile.
    • The telecom profiles G.8265.1, G.8275.1, and G.8275.2.
  • Supports the NetSync Monitor protocol.
  • Implements Peer to peer one-step.
  • Supports bonded, IPoIB, and vlan interfaces.

Getting the Code

You can download the latest released version at Source Forge.

http://sourceforge.net/projects/linuxptp/files/latest/download

The source code is managed using the git version control system. To get your own copy of the project sources, use the following command.

git clone git://git.code.sf.net/p/linuxptp/code linuxptp

If the git protocol is blocked by your local area network, then you can use the alternative HTTP protocol instead.

git clone http://git.code.sf.net/p/linuxptp/code linuxptp

System Requirements

In order to run this software, you need Linux kernel version 3.0 or newer. Check whether your network interface supports PTP with the following command.

ethtool -T eth0

This command shows whether a MAC supports hardware or software time stamping. The following example output indicates support for hardware time stamping.

Time stamping parameters for eth6:
Capabilities:
        hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
        software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
        hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
        software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
        software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
        hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
PTP Hardware Clock: 1
Hardware Transmit Timestamp Modes:
        off                   (HWTSTAMP_TX_OFF)
        on                    (HWTSTAMP_TX_ON)
Hardware Receive Filter Modes:
        none                  (HWTSTAMP_FILTER_NONE)
        all                   (HWTSTAMP_FILTER_ALL)

The next example shows the case where the MAC only supports software time stamping. The ptp4l program requires either the -S command line argument or the time_stamping software configuration option when using such interfaces.

Time stamping parameters for enp6s0:
Capabilities:
        software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
        software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
        software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
PTP Hardware Clock: none
Hardware Transmit Timestamp Modes: none
Hardware Receive Filter Modes: none

Note the software-transmit (SOF_TIMESTAMPING_TX_SOFTWARE) capability. If this is lacking, then the MAC cannot be used at all. However, adding this capability entails adding a single line of code to the device driver.

Installation

  1. Just type ‘make’
  2. If you compiled your own kernel (and the headers are not installed into the system path), then you should set the KBUILD_OUTPUT environment variable as in the example, above.
  3. In order to install the programs and man pages into /usr/local, run the ‘make install’ target. You can change the installation directories by setttings the variables prefix, sbindir, mandir, and man8dir on the make command line.

Getting Involved

The software development is hosted at Source Forge.

https://sourceforge.net/projects/linuxptp/

Reporting Bugs

Please report any bugs or other issues with the software to the linuxptp-users mailing list.

https://lists.sourceforge.net/lists/listinfo/linuxptp-users

Development

If you would like to get involved in improving the software, please join the linuxptp-devel mailing list.

https://lists.sourceforge.net/lists/listinfo/linuxptp-devel

Submitting Patches

  1. Before submitting patches, please make sure that you are starting your work on the current HEAD of the git repository.
  2. Please checkout the CODING_STYLE.org file for guidelines on how to properly format your code.
  3. Describe your changes. Each patch will be reviewed, and the reviewers need to understand why you did what you did.
  4. Sign-Off each commit, so the changes can be properly attributed to you and you explicitely give your agreement for distribution under linuxptp’s license. Signing-off is as simple as: 
    git commit -s

    or by adding the following line (replace your real name and email) to your patch:

    Signed-off-by: Random J Developer <random@developer.example.org>
  5. Finally, send your patches via email to the linuxptp-devel mailing list, where they will be reviewed, and eventually be included in the official code base. 
    git send-email --to linuxptp-devel@lists.sourceforge.net origin/master

Thanks

Thanks to AudioScience Inc for sponsoring the 8021.AS support.

Thanks to Exablaze for donating an ExaNIC X10

Thanks to Intel Corporation for donating four NICs, the 82574, 82580, 82599, and the i210.

Thanks to Meinberg Funkuhren for donating a LANTIME M1000.

Thanks to Moser Baer for sponsoring the Telecom Profiles and unicast support.

For testing I use an OTMC 100 grandmaster clock donated by OMICRON Lab.

Code reading

BC and OC, use the same events and dispatch. and in bc_event, we have “`c case FD_DELAY_TIMER: pr_debug(“%s: delay timeout”, p->log_name); port_set_delay_tmo(p); delay_req_prune(p); p->service_stats.delay_timeout++; if (port_delay_request(p)) { return EV_FAULT_DETECTED; } “`

State Transfer

In fsm.h, we have port_state, and fsm_event In fsm.c, we have ptp_fsm and ptp_slave_fsm, they are the stat machine for port

dispatch

bc_event (CLOCK_TYPE_BOUNDARY) -> process_follow_up -> port_syfunfsm -> different port functions (message_interval_request, port_synchronize, port_change_phc)->port_dispatch call p->dispatch, call e2e_dispatch

bc_dispatch will call port_e2e_transition too

enum fsm_event port_event(struct port *p, int fd_index)

任何 timer 或者 RTNL(netlink) 触发，进行对应处理, 同时会接受ptp报文，进行对应处理 “`c case FD_DELAY_TIMER: pr_debug(“%s: delay timeout”, p->log_name); port_set_delay_tmo(p); delay_req_prune(p); tc_prune(p); if (!clock_free_running(p->clock)) { switch (p->state) { case PS_UNCALIBRATED: case PS_SLAVE: if (port_delay_request(p)) { event = EV_FAULT_DETECTED; } break; default: break; }; } return event; …

switch (msg_type(msg)) { case SYNC: … case DELAY_RESP: if (tc_fwd_response(p, msg)) { event = EV_FAULT_DETECTED; } if (dup) { process_delay_resp(p, dup); } break; “` clock_poll (check all different timer, and new ptp frames) -> port_event

generate_clock_identity -> is this standard?

port_initialize will create all the fd (and all the file will be opened?).

port_open will create device poll event, for ptp packets