Permalink
Commits on Mar 26, 2017
  1. Merge pull request #88 from spaceinventor/rdpfixes

    johandc committed Mar 26, 2017
    Space Inventor: RDP fixes
Commits on Feb 13, 2017
  1. Merge pull request #91 from mndza/master

    johandc committed Feb 13, 2017
    Fix SocketCAN driver includes
Commits on Feb 7, 2017
  1. rdp: Always send ack on syn/ack

    johandc committed Oct 27, 2016
    We we relying on the userspace to send a packet in order to initiate
    the transfer. This will not work with services that does not use a
    request and relies on the fact that the connection is opened.
  2. bugfix: RDP do not close connection in CLOSE_WAIT

    johandc committed Oct 27, 2016
    When the first RST is receive the rdp state is set to CLOSE_WAIT, and
    the userspace is requested to close the connection by posting a null
    pointer to the RX queue.
    
    On the second RST message received, the connection is forcefully closed.
    If the userspace task was still working on the RX queue, it will
    therefore be denied the remainder of the received data.
    
    This fix means that a connection that was already closed by the
    userspace (on the first node to call csp_close()) and therefore put into
    CLOSE_WAIT, will remain in CLOSE_WAIT untill the rdp timeout happens and
    sliently sets the state to CLOSED.
    
    This resolution is more in line with the original RDP specification
    where connections were always left to just timeout in the CLOSE_WAIT
    state.
  3. rdp: Open connections before receiving first packet

    johandc committed Oct 24, 2016
    Previously we had to initiate a new RDP connection with a request containing a non-zero length packet. However, for streams this should not be necessary. So moving the conn-queue push to earlier will allow for a service to start transmitting as soon as somebody connects to the port.
    
    Also i have removed the call to srand() since i feel this is something that is best done outside of the RDP stack. Furthermore it resulted in using the same sequence number if two connections were made within the same millisecond
  4. Merge pull request #85 from libcsp/sam0

    johandc committed Feb 7, 2017
    Fixed warnings on newer arm compilers
Commits on Feb 6, 2017
  1. Merge pull request #79 from spaceinventor/upstreamthis

    johandc committed Feb 6, 2017
    Fix all warnings from -Wmissing-declarations
  2. Merge pull request #76 from pacheco017/less-cxx-warnings

    johandc committed Feb 6, 2017
    Avoid warnings when building with -std=c++11
  3. added missing baudrates + usart_init() now defaults to a baudrate of …

    sorennohr authored and jledet committed Feb 6, 2017
    …500000 instead of 0 (#83)
Commits on Jan 24, 2017
  1. Fix SocketCAN driver includes

    mndza committed Jan 24, 2017
    bits/socket.h should never be included directly, and only sys/socket.h must be used (which is already present).
    
    This change fixes build with musl libc.
Commits on Nov 8, 2016
  1. Fixed shadow warning.

    perwulff committed Nov 8, 2016
Commits on Nov 4, 2016
  1. Changed csp_set_hostname(), csp_set_model(), csp_set_revision() to use

    perwulff committed Nov 4, 2016
    const pointers - data is not modified, and prevents warnings.
Commits on Sep 9, 2016
  1. Fix all warnings from -Wmissing-declarations

    johandc committed Sep 9, 2016
    Warn if a global function is defined without a previous declaration.
    Do so even if the definition itself provides a prototype.
    Use this option to detect global functions that are not declared in
    header files.
  2. Added *.d to .gitignore

    johandc committed Sep 8, 2016
Commits on Aug 20, 2016
  1. csp/csp_debug: avoid warnings when building with -std=c++11 [-Wlitera…

    Matias Devenuta
    Matias Devenuta committed Jul 7, 2016
    …l-suffix]
    
    Avoid:
    
      warning: invalid suffix on literal; C++11 requires a space between literal
      and string macro [-Wliteral-suffix]
Commits on Jul 19, 2016
  1. csp: rdp: defer ACK transmission until two windows are available

    jledet committed Jul 19, 2016
    Also switch to use common csp_rdp_check_ack function at both ACK trigger
    points.
  2. csp: rdp: check packets in-flight before sending new packets

    jledet committed Jul 19, 2016
    Handle spurious wake-ups and sequence number wraparound.
Commits on Jun 24, 2016
  1. wscript: remove --with-freertos argument

    jledet committed Jun 24, 2016
    Use the existing --include=<path> option if you need to add additional
    FreeRTOS include paths.
  2. debug: remove verbose debug option

    jledet committed Jun 8, 2016
    The BASENAME macro is terribly slow, so it is unlikely that anyone would
    ever enable the verbose debug output (which adds file and lineno to
    debug messages). Kill the feature to simplify the code and remove an
    option.
    
    This patch also adds whitespace fixes to a couple of files.
  3. dedup: rename duplicate check function

    jledet committed Jun 7, 2016
    This should make it clearer what the return value means, without having
    to check the code. Also, add missing license header.
Commits on May 3, 2016
  1. csp_if_i2c.c: fix typo. (#71)

    silvawp authored and jledet committed May 3, 2016
Commits on Apr 14, 2016
  1. Merge pull request #69 from GomSpace/discard-unknown-rdp

    jledet committed Apr 14, 2016
    csp: route: drop packets with unsupported options early
Commits on Mar 30, 2016
  1. csp: route: drop packets with unsupported options early

    jledet committed Mar 17, 2016
    Incoming packets with unsupported options were freed, but if a new
    connection was allocated this was not. This patch splits the check for
    supported options into a new function and drops the packets early,
    before dealing with connections.
  2. Move defining ctx.env.OS earlier so it's defined when adding LIBS.

    Dolmio authored and jledet committed Feb 11, 2016
    Also rename undefined variable libs to ctx.env.LIBS in examples part.
    Fixes running examples on posix machine
Commits on Mar 29, 2016
  1. Added missing check on csp_malloc for CSP_PS in csp_service_handler()…

    kristianbay committed Mar 29, 2016
    … and added missing csp_buffer_free() in csp_ps()
Commits on Feb 26, 2016
  1. csp: do not expose global my_address variable

    jledet committed Feb 26, 2016
    This patch renames the global my_address variable and hides it behind
    two new access functions, csp_get_address and csp_set_address that
    should be used instead.
Commits on Feb 17, 2016
  1. io: add functions for setting the CMP ident revision

    jledet committed Feb 17, 2016
    The GIT_REV define is set to the git revision of the CSP repository, but
    usually you want to show a release number of the subsystem instead. This
    patch adds a csp_set_revision to override the GIT_REV value from the
    subsystem code.
Commits on Feb 16, 2016
  1. waf: update to version 1.8.19

    jledet committed Feb 16, 2016
Commits on Jan 20, 2016
  1. Merge pull request #62 from pacheco017/fix-can-tx-overflow

    jledet committed Jan 20, 2016
    csp: interfaces: can: fixed tx_count overflow in csp_can_tx()
Commits on Jan 19, 2016
  1. csp: interfaces: can: fixed tx_count overflow in csp_can_tx()

    Matias Devenuta
    Matias Devenuta committed Jan 19, 2016
    tx_count should be 16 bits long, same as CSP packet length; otherwise, overflow
    bugs will happen when sending packets with 256 or more bytes of data.
Commits on Dec 9, 2015
Commits on Dec 4, 2015
  1. Merge pull request #60 from GomSpace/can-refactor

    johandc committed Dec 4, 2015
    Refactor CAN
Commits on Dec 3, 2015
  1. csp: interfaces: can: rework CAN interface

    jledet committed Nov 25, 2015
    This patch restructures the CAN interface to do away with the strange
    callback architecture. Instead csp_can_tx simply calls the can_send
    function in a loop.
    
    Implementations of can_send *must* guarantee that messages are
    transmitted in the order that can_send is called. However, they are free
    to use any buffering/FIFOs available in hardware.
    
    The patch also removes all other CAN drivers than SocketCAN. Device
    specific drivers usually require hardware dependent header files for
    MMIO registers and detailed knowledge of the hardware configuration
    (e.g. clocks). Maintaining these drivers in CSP has never really been
    ideal, so lets just remove them. It should be a trivial fix to update
    them to the new CAN architecture.
    
    The Linux SocketCAN drivers is generic and useful as an example of the
    CAN driver interface, so it's allowed to stay.