]> git.ipfire.org Git - thirdparty/lldpd.git/log
thirdparty/lldpd.git
10 years agodebian: ship files in debian/etc/lldpd.d
Vincent Bernat [Wed, 17 Sep 2014 09:45:17 +0000 (11:45 +0200)] 
debian: ship files in debian/etc/lldpd.d

10 years agobuild: check if pkg-config is correctly installed in ./autogen.sh
Vincent Bernat [Fri, 5 Sep 2014 18:19:34 +0000 (20:19 +0200)] 
build: check if pkg-config is correctly installed in ./autogen.sh

Absence of pkg-config leads to some odd messages:

    configure.ac:27: error: possibly undefined macro: AC_SUBST
          If this token and others are legitimate, please use m4_pattern_allow.
          See the Autoconf documentation.
    configure.ac:163: error: possibly undefined macro: AC_MSG_FAILURE
    configure:21413: error: possibly undefined macro: AC_MSG_CHECKING
    configure:21418: error: possibly undefined macro: AC_TRY_LINK
    configure:21423: error: possibly undefined macro: AC_MSG_RESULT
    configure:21432: error: possibly undefined macro: AC_MSG_NOTICE
    configure:21894: error: possibly undefined macro: AC_PATH_TOOL
    configure:21896: error: possibly undefined macro: AC_MSG_ERROR
    configure:21975: error: possibly undefined macro: AC_DEFINE_UNQUOTED

We try to detect this situation by looking for `PKG_CHECK_MODULES` macro
in `aclocal.m4`. If not present, we issue some useful tips, like
installing the package or telling aclocal the location of
`pkg.m4` (mostly for OSX users).

10 years agobuild: check for needed commands in autogen.sh
Vincent Bernat [Fri, 5 Sep 2014 17:49:31 +0000 (19:49 +0200)] 
build: check for needed commands in autogen.sh

And display a more informative message.

10 years agobuild: provide a sysusers.d file
Vincent Bernat [Wed, 3 Sep 2014 23:02:39 +0000 (01:02 +0200)] 
build: provide a sysusers.d file

This allows systemd-sysusers to create _lldpd user/group. See:
 http://www.freedesktop.org/software/systemd/man/sysusers.d.html

This is currently untested.

10 years agolldpd: handle early cleanup in case interface cannot be created
Vincent Bernat [Sun, 31 Aug 2014 19:56:49 +0000 (21:56 +0200)] 
lldpd: handle early cleanup in case interface cannot be created

When an interface is about to be created but an external condition makes
it impossible, the cleanup may crash because the interface was
incomplete. Check things exist before cleanup.

10 years agolib: handle read failure in sync case
Vincent Bernat [Sun, 31 Aug 2014 19:53:54 +0000 (21:53 +0200)] 
lib: handle read failure in sync case

This amends the previous commit with a more concise condition.

10 years agolib: really correctly handle read failure
Vincent Bernat [Sun, 31 Aug 2014 19:53:17 +0000 (21:53 +0200)] 
lib: really correctly handle read failure

10 years agolldpcli: disable SIGPIPE
Vincent Bernat [Sun, 31 Aug 2014 05:35:06 +0000 (07:35 +0200)] 
lldpcli: disable SIGPIPE

We handle EOF inside the code.

10 years agolib: correctly handle read failure
Vincent Bernat [Sun, 31 Aug 2014 05:31:57 +0000 (07:31 +0200)] 
lib: correctly handle read failure

10 years agolib: stricter checks when resuming an operation in progress
Vincent Bernat [Sat, 30 Aug 2014 06:45:53 +0000 (08:45 +0200)] 
lib: stricter checks when resuming an operation in progress

When an operation is in progress, we used both state management and some
canary value to check if the user is resuming the right operation. The
canary was just a pointer and it was easy to get the same pointer as the
previous operation while not really resuming the previous operation. We
turn the check into a string comparison instead.

Moreover, add checks when setting configuration items too.

10 years agolldpcli: add editor local variables to bash completion
Vincent Bernat [Tue, 26 Aug 2014 15:58:53 +0000 (17:58 +0200)] 
lldpcli: add editor local variables to bash completion

10 years agolldpcli: add editor local variables to zsh completion
Vincent Bernat [Tue, 26 Aug 2014 15:58:00 +0000 (17:58 +0200)] 
lldpcli: add editor local variables to zsh completion

10 years agolldpcli: indent multiline texts (like system description)
Vincent Bernat [Tue, 12 Aug 2014 18:03:09 +0000 (20:03 +0200)] 
lldpcli: indent multiline texts (like system description)

This should fix #73 in an alternate way.

10 years agolldpcli: fix style convention to match other files
Vincent Bernat [Tue, 12 Aug 2014 17:45:14 +0000 (19:45 +0200)] 
lldpcli: fix style convention to match other files

There are still other files around with this convention that would need
to be fixed.

10 years agobash/zsh: ensure completions are present in dist tarball
Vincent Bernat [Sun, 3 Aug 2014 13:15:37 +0000 (15:15 +0200)] 
bash/zsh: ensure completions are present in dist tarball

10 years agoNEWS: add a word about bash/zsh completion
Vincent Bernat [Sun, 3 Aug 2014 10:33:52 +0000 (12:33 +0200)] 
NEWS: add a word about bash/zsh completion

10 years agozsh: add ZSH completion
Vincent Bernat [Sun, 3 Aug 2014 10:31:44 +0000 (12:31 +0200)] 
zsh: add ZSH completion

10 years agobash: move bash completion outside Debian packaging
Vincent Bernat [Sun, 3 Aug 2014 09:37:43 +0000 (11:37 +0200)] 
bash: move bash completion outside Debian packaging

10 years agorelease: new version 0.7.10
Vincent Bernat [Mon, 21 Jul 2014 19:53:22 +0000 (21:53 +0200)] 
release: new version 0.7.10

10 years agoNEWS: add a word about json-c library support 0.7.10
Vincent Bernat [Mon, 21 Jul 2014 16:06:54 +0000 (18:06 +0200)] 
NEWS: add a word about json-c library support

10 years agolldpcli: fix jansson implementation of the JSON output
Vincent Bernat [Mon, 21 Jul 2014 16:05:52 +0000 (18:05 +0200)] 
lldpcli: fix jansson implementation of the JSON output

10 years agoconfigure: if --with-json, default to jansson implementation
Vincent Bernat [Mon, 21 Jul 2014 16:01:49 +0000 (18:01 +0200)] 
configure: if --with-json, default to jansson implementation

10 years agocdp: complete manual page with CDP-related options
Vincent Bernat [Mon, 21 Jul 2014 16:00:51 +0000 (18:00 +0200)] 
cdp: complete manual page with CDP-related options

10 years agocdp: Make it possible to enable CDPv2 without enabling CDPv1 71/head
Michel Stam [Mon, 21 Jul 2014 15:30:28 +0000 (17:30 +0200)] 
cdp: Make it possible to enable CDPv2 without enabling CDPv1

10 years agolldpcli: Add json description to lldpcli usage
Michel Stam [Mon, 21 Jul 2014 15:10:21 +0000 (17:10 +0200)] 
lldpcli: Add json description to lldpcli usage

10 years agocdp: Add power requirement to CDPv2 frames
Michel Stam [Mon, 21 Jul 2014 14:40:41 +0000 (16:40 +0200)] 
cdp: Add power requirement to CDPv2 frames

LLDP already supports sending power requirements, for
IEEE 802.3at. Older Cisco PoE switches do not support
LLDP and need a custom CDPv2 frame. This patch enables
support if the LLDP MED PoE frame is also sent.

10 years agolldpcli: Add support for JSON-C
Michel Stam [Fri, 18 Jul 2014 15:00:32 +0000 (17:00 +0200)] 
lldpcli: Add support for JSON-C

OpenWRT is shipped with the JSON-C library, which is used for
several other packages. Having a separate JSON library (Jansson)
doing the exact same thing is wasting flash space.

10 years agolldpcli: make complete command work on privleged commands as well
Vincent Bernat [Sat, 12 Jul 2014 11:10:47 +0000 (13:10 +0200)] 
lldpcli: make complete command work on privleged commands as well

This avoids the use of sudo.

10 years agolldpcli: provide a hidden complete command for shell completion
Vincent Bernat [Sat, 12 Jul 2014 11:09:37 +0000 (13:09 +0200)] 
lldpcli: provide a hidden complete command for shell completion

10 years agolldpcli: change how privileged commands are declared
Vincent Bernat [Thu, 10 Jul 2014 20:55:58 +0000 (22:55 +0200)] 
lldpcli: change how privileged commands are declared

Even when not privileged, we declare privileged commands but tag them
accordingly. We need to be careful when executing the commands to check
the new bit.

This is a preliminary change to allow a completion command to complete
privileged commands even when not privileged.

10 years agolldpcli: reformat a bit bash completion to be more readable
Vincent Bernat [Thu, 10 Jul 2014 20:23:40 +0000 (22:23 +0200)] 
lldpcli: reformat a bit bash completion to be more readable

10 years agopriv: avoid a socket leak when interface is already in promisc
Vincent Bernat [Wed, 9 Jul 2014 06:52:43 +0000 (08:52 +0200)] 
priv: avoid a socket leak when interface is already in promisc

10 years agosnmp: avoid deferencing a pointer when it may be NULL
Vincent Bernat [Wed, 9 Jul 2014 06:50:04 +0000 (08:50 +0200)] 
snmp: avoid deferencing a pointer when it may be NULL

10 years agodcbx tlv recd are printed as log_debug
Sam Tannous [Tue, 8 Jul 2014 19:15:51 +0000 (15:15 -0400)] 
dcbx tlv recd are printed as log_debug

Changed dcbx tlv recd to be printed as log_debug
so that we dont flood daemon log (N3K sends dcbx
tlv by default.

Signed-off-by: Kanna Rajagopal <kanna@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agolldpcli: use protocol map from liblldpctl to select protocol
Vincent Bernat [Tue, 8 Jul 2014 20:57:54 +0000 (22:57 +0200)] 
lldpcli: use protocol map from liblldpctl to select protocol

10 years agolldpcli: document `-u` argument in synopsis
Vincent Bernat [Tue, 8 Jul 2014 20:57:03 +0000 (22:57 +0200)] 
lldpcli: document `-u` argument in synopsis

10 years agolldpcli: add display filter to show nbrs running specific protocols
Sam Tannous [Tue, 8 Jul 2014 19:15:50 +0000 (15:15 -0400)] 
lldpcli: add display filter to show nbrs running specific protocols

LLDPD allows a peer to run multiple protocols (LLDP, CDPv1, CDPv2).
lldpcli can be used to show the primary protocol for a nbr or all
the protocols using the "hidden" cmdline arg
This allows lldpcli to filter the nbr display based on the
protocol that is running on it.

Signed-off-by: Kanna Rajagopal <kanna@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agolib: update liblldpctl versioning
Vincent Bernat [Tue, 8 Jul 2014 20:41:29 +0000 (22:41 +0200)] 
lib: update liblldpctl versioning

Previous commit added a function to the API.

10 years agoAdd call to process more messages from data already read.
Sam Tannous [Tue, 8 Jul 2014 19:15:49 +0000 (15:15 -0400)] 
Add call to process more messages from data already read.

Currently, the calls within lldpctl are assumed to be all associated with
synchronous sockets. With async sockets, the data read can be varied and so
its useful to process as much of the data as possible. However, lldpctl_recv
only processes one message and there isn't a call to be made to process
data that's sitting in a control connection's buffer. This patch adds such a
call for external users to call. It simply calls check_for_notification and
returns whatever check_for_notification returns with 0 signalling more
processing is possible and non-zero indicating buffer is fully processed
(though there may still be data).

Signed-off-by: Dinesh G Dutt <ddutt@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoAdd error code to the multicast address add/delete failure msg
Sam Tannous [Tue, 8 Jul 2014 19:15:48 +0000 (15:15 -0400)] 
Add error code to the multicast address add/delete failure msg

Move mac filter delete info msg to debug and add error
to the multicast address add/delete failure msg.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoMove interface update msg to debug level
Sam Tannous [Tue, 8 Jul 2014 19:15:47 +0000 (15:15 -0400)] 
Move interface update msg to debug level

Move interface update msg to debug level so that we don't see
so many of them on large boxes.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoAdd ignore handler for SIGHUP in lldpcli
Sam Tannous [Tue, 8 Jul 2014 19:15:46 +0000 (15:15 -0400)] 
Add ignore handler for SIGHUP in lldpcli

lldpcli which is execed by lldpd is expected to resume lldpd operations.
But in some cases, lldpcli can exiting because of SIGHUP.  This fixes
these cases.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Kanna Rajagopal <kanna@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoThis patch adds bash completion for lldpcli.
Sam Tannous [Tue, 8 Jul 2014 20:34:07 +0000 (22:34 +0200)] 
This patch adds bash completion for lldpcli.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agosnmp: use poll() to wait for AgentX socket to be ready
Vincent Bernat [Tue, 8 Jul 2014 20:15:12 +0000 (22:15 +0200)] 
snmp: use poll() to wait for AgentX socket to be ready

Instead of sleeping for an arbitrary amount of time, query if the socket
is ready using poll() which should be portable enough (POSIX.1-2001) to
work everywhere.

10 years agosnmp: preserve previous flags when making AgentX socket non-blocking
Vincent Bernat [Tue, 8 Jul 2014 19:56:36 +0000 (21:56 +0200)] 
snmp: preserve previous flags when making AgentX socket non-blocking

10 years agomake agentx socket non-blocking
Sam Tannous [Tue, 8 Jul 2014 19:15:44 +0000 (15:15 -0400)] 
make agentx socket non-blocking

This patch makes the lldpd socket non-blocking. And also adds a brief
retry.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoMake "too many neighbors for port" msg appear less frequently
Sam Tannous [Tue, 8 Jul 2014 19:15:43 +0000 (15:15 -0400)] 
Make "too many neighbors for port" msg appear less frequently

There is a max neighbors setting (default value 4).
It prints a msg everytime it gets a new neighbor > 4.
This patch just makes it appear less often.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agolldpd: Fix netlink notification group for address changes
Sam Tannous [Tue, 8 Jul 2014 19:15:41 +0000 (15:15 -0400)] 
lldpd: Fix netlink notification group for address changes

The current netlink notification registration registers
for only link changes. This patch fixes it to include address
changes.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoIncrease event buffer
Sam Tannous [Tue, 8 Jul 2014 19:15:40 +0000 (15:15 -0400)] 
Increase event buffer

This buffer is just used to drain the event socket

During high event traffic saw a no buffer space error. This and other
buffers need to be revisited.

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agolog: info messages should be logged on syslog but not on first debug level
Vincent Bernat [Tue, 8 Jul 2014 19:41:48 +0000 (21:41 +0200)] 
log: info messages should be logged on syslog but not on first debug level

Without `-d`, syslog should log info messages. With `-d`, they should
not be displayed. With `-dd`, they should be displayed. This is an
amendment of the previous patch.

10 years agolldpd: fix log_info
Sam Tannous [Tue, 8 Jul 2014 19:15:39 +0000 (15:15 -0400)] 
lldpd: fix log_info

Currently none of the log_info msgs are getting
printed into syslog. vlog already does the right checks.
Remove the checks in log_info which is making all default log_info
prints no-ops

Also change severity of "interface reset timer" msg

Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoNEWS: credit seccomp fix
Vincent Bernat [Tue, 8 Jul 2014 19:30:18 +0000 (21:30 +0200)] 
NEWS: credit seccomp fix

10 years agolldpd: fix use of NULL in execl*
Vincent Bernat [Mon, 7 Jul 2014 07:32:12 +0000 (09:32 +0200)] 
lldpd: fix use of NULL in execl*

See:
 http://c-faq.com/null/nullreq.html

10 years agoMerge pull request #70 from chutz/seccomp-whitelist
Vincent Bernat [Tue, 8 Jul 2014 19:29:23 +0000 (21:29 +0200)] 
Merge pull request #70 from chutz/seccomp-whitelist

seccomp: whitelist sendto, poll, recvmsg and readv in seccomp sandbox

10 years agolldpd: whitelist sendto, poll, recvmsg and readv in seccomp sandbox 70/head
Patrick McLean [Tue, 8 Jul 2014 18:40:15 +0000 (11:40 -0700)] 
lldpd: whitelist sendto, poll, recvmsg and readv in seccomp sandbox

With privilege separation and seccomp enabled, the monitor thread was
dying due to seccomp errors, while leaving the child alive. This whitelists
some extra syscalls to allow the monitor process to stay alive.

10 years agoNEWS: add a word about ability to disable LLDP
Vincent Bernat [Tue, 17 Jun 2014 07:22:25 +0000 (09:22 +0200)] 
NEWS: add a word about ability to disable LLDP

10 years agolldpd: fix how LLDP can be disabled
Vincent Bernat [Sun, 15 Jun 2014 08:07:30 +0000 (10:07 +0200)] 
lldpd: fix how LLDP can be disabled

Commit 46baf62775eaca47075e1d9966a2103fd466e4b3 broke the ability to
disable LLDP. Instead of trying to be smart, just enumerate all the
possible cases.

10 years agocdp: don't expect and off-by-one checksum
Vincent Bernat [Sun, 15 Jun 2014 07:47:03 +0000 (09:47 +0200)] 
cdp: don't expect and off-by-one checksum

10 years agoframe: fix CDP checksum
udbxtd2008 [Sun, 15 Jun 2014 07:05:06 +0000 (15:05 +0800)] 
frame: fix CDP checksum

The checksum of CDP frame is calculated off-by-one error when the length
of frame is odd, and the highest bit of the last octet is 1, which is as
a known question mentioned in wireshark.

10 years agoREADME: document the new promisc interface in README as well
Vincent Bernat [Fri, 30 May 2014 13:26:03 +0000 (15:26 +0200)] 
README: document the new promisc interface in README as well

10 years agolldpcli: give more details in the manual page about promiscuous mode
Vincent Bernat [Fri, 30 May 2014 13:25:06 +0000 (15:25 +0200)] 
lldpcli: give more details in the manual page about promiscuous mode

10 years agolldpcli: add an option to enable promisc mode on managed interfaces
Vincent Bernat [Fri, 30 May 2014 13:21:34 +0000 (15:21 +0200)] 
lldpcli: add an option to enable promisc mode on managed interfaces

This allows LLDP frames to be received even when they are not in a
configured VLAN, like this may be the case with Cisco 2960.

10 years agopriv: don't output rc status twice when unable to open socket
Vincent Bernat [Fri, 30 May 2014 13:04:36 +0000 (15:04 +0200)] 
priv: don't output rc status twice when unable to open socket

10 years agoREADME: more about Cisco sending LLDP frames on VLAN 1
Vincent Bernat [Fri, 30 May 2014 10:07:39 +0000 (12:07 +0200)] 
README: more about Cisco sending LLDP frames on VLAN 1

10 years agoosx: update version to 0.7.9
Vincent Bernat [Wed, 28 May 2014 10:47:22 +0000 (12:47 +0200)] 
osx: update version to 0.7.9

10 years agodebian: sync a bit with official packaging 0.7.9
Vincent Bernat [Wed, 28 May 2014 10:08:51 +0000 (12:08 +0200)] 
debian: sync a bit with official packaging

10 years agoversion: prepare a new release
Vincent Bernat [Wed, 28 May 2014 10:06:32 +0000 (12:06 +0200)] 
version: prepare a new release

10 years agolldpd: trigger regular interface refresh when changing port ID
Vincent Bernat [Wed, 21 May 2014 16:40:02 +0000 (18:40 +0200)] 
lldpd: trigger regular interface refresh when changing port ID

We don't want to call OS-specific methods. Let's just trigger a regular
interface refresh instead.

10 years agolldpcli: document `configure lldp portidsubtype` command
Vincent Bernat [Wed, 21 May 2014 15:31:34 +0000 (17:31 +0200)] 
lldpcli: document `configure lldp portidsubtype` command

10 years agoNEWS: mention ability to configure port ID
Vincent Bernat [Wed, 21 May 2014 15:22:54 +0000 (17:22 +0200)] 
NEWS: mention ability to configure port ID

10 years agolldpcli: fix a format warning
Vincent Bernat [Wed, 21 May 2014 15:21:40 +0000 (17:21 +0200)] 
lldpcli: fix a format warning

10 years agolldpcli: remove unused variable
Vincent Bernat [Wed, 21 May 2014 15:21:03 +0000 (17:21 +0200)] 
lldpcli: remove unused variable

10 years agoSupport for modifying PortID TLV
Sam Tannous [Wed, 21 May 2014 14:11:24 +0000 (10:11 -0400)] 
Support for modifying PortID TLV

This patch adds the ability to change the behavior of what
lldpd sends in the PortID TLV.

By default, lldpd sets the MAC address in the PortID and the ifname
in the PortDescr field.  This patch allows the user to
configure lldpd via the CLI (lldpcli or /etc/lldpd.conf)
to send either the interface name
(ifname) or the MAC address in the PortID TLV.

Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
10 years agoREADME: point to the website for install instructions
Vincent Bernat [Fri, 16 May 2014 18:45:24 +0000 (20:45 +0200)] 
README: point to the website for install instructions

10 years agoREADME: update the intro to better match the reality
Vincent Bernat [Fri, 16 May 2014 18:41:48 +0000 (20:41 +0200)] 
README: update the intro to better match the reality

The inability to handle complex setup has been greatly diminished in 0.7
branch.

10 years agotravis: don't use regular options when on coverity branch
Vincent Bernat [Fri, 16 May 2014 09:06:04 +0000 (11:06 +0200)] 
travis: don't use regular options when on coverity branch

10 years agotravis: don't do anything if not on the coverity branch
Vincent Bernat [Fri, 16 May 2014 08:47:14 +0000 (10:47 +0200)] 
travis: don't do anything if not on the coverity branch

10 years agotravis: reduce matrix by not compiling everything with clang
Vincent Bernat [Fri, 16 May 2014 08:40:11 +0000 (10:40 +0200)] 
travis: reduce matrix by not compiling everything with clang

Only compile with all options when using clang.

10 years agotravis: integrate coverity script in regular run.sh
Vincent Bernat [Fri, 16 May 2014 08:34:16 +0000 (10:34 +0200)] 
travis: integrate coverity script in regular run.sh

10 years agotravis: when running coverity, skip regular build
Vincent Bernat [Fri, 16 May 2014 07:54:05 +0000 (09:54 +0200)] 
travis: when running coverity, skip regular build

10 years agotravis: don't try to install launchd/systemd stuff
Vincent Bernat [Fri, 16 May 2014 00:03:46 +0000 (02:03 +0200)] 
travis: don't try to install launchd/systemd stuff

This won't work as they are not installed inside prefix.

10 years agotravis: no need for `make install`
Vincent Bernat [Thu, 15 May 2014 23:52:52 +0000 (01:52 +0200)] 
travis: no need for `make install`

We already do a `make distcheck` for that purpose.

10 years agobsd: ensure we also have space for the padding after BPF header
Vincent Bernat [Thu, 15 May 2014 23:43:13 +0000 (01:43 +0200)] 
bsd: ensure we also have space for the padding after BPF header

10 years agobsd: ensure proper alignment of BPF header
Vincent Bernat [Thu, 15 May 2014 23:39:52 +0000 (01:39 +0200)] 
bsd: ensure proper alignment of BPF header

This is required as per bpf(4).

10 years agobsd: ensure we also have space for the BPF header
Vincent Bernat [Thu, 15 May 2014 23:18:07 +0000 (01:18 +0200)] 
bsd: ensure we also have space for the BPF header

10 years agobsd: fix clang warning about alignment problem with sockaddr
Vincent Bernat [Thu, 15 May 2014 23:08:27 +0000 (01:08 +0200)] 
bsd: fix clang warning about alignment problem with sockaddr

While this does not seem to be extensively documented, when casting a
`sockaddr *` to something else, like `sockaddr_dl *`, it is believed
that initially, it was a `sockaddr_dl *`, so there is no alignment
problem in this case. We use a macro to cast to `void *` to avoid the
warning.

10 years agotravis: run "make distcheck" with the appropriate configure arguments
Vincent Bernat [Thu, 15 May 2014 23:07:30 +0000 (01:07 +0200)] 
travis: run "make distcheck" with the appropriate configure arguments

Otherwise, we just use the default ones...

10 years agotravis: only run "make distcheck"
Vincent Bernat [Thu, 15 May 2014 13:47:42 +0000 (15:47 +0200)] 
travis: only run "make distcheck"

We don't need to compile stuff twice.

10 years agotravis: on OS X, try with/without embedded libevent and with/without SNMP
Vincent Bernat [Thu, 15 May 2014 13:26:14 +0000 (15:26 +0200)] 
travis: on OS X, try with/without embedded libevent and with/without SNMP

10 years agotravis: try to build only with clang on OS X
Vincent Bernat [Thu, 15 May 2014 13:23:05 +0000 (15:23 +0200)] 
travis: try to build only with clang on OS X

10 years agotravis: use devel version of net-snmp on OSX
Vincent Bernat [Thu, 15 May 2014 13:19:54 +0000 (15:19 +0200)] 
travis: use devel version of net-snmp on OSX

The current stable version does not work with Mavericks. See:
 https://github.com/Homebrew/homebrew/issues/25097

10 years agotravis: misc change to trigger a rebuild
Vincent Bernat [Thu, 15 May 2014 12:46:24 +0000 (14:46 +0200)] 
travis: misc change to trigger a rebuild

10 years agolldpcli: directive to configure IP management pattern
Vincent Bernat [Sun, 11 May 2014 14:21:33 +0000 (16:21 +0200)] 
lldpcli: directive to configure IP management pattern

This closes #54.

10 years agolldp: fix handling of LLDPDU expiration
Vincent Bernat [Sun, 11 May 2014 14:17:42 +0000 (16:17 +0200)] 
lldp: fix handling of LLDPDU expiration

Expiration logic for LLDPDU was totally flawed after the change to
support shutdown LLDPDU. Any LLDPDU would trigger infinite
expiration. Fix that.

10 years agoNEWS: mention support for shutdown LLDPDU
Vincent Bernat [Sat, 10 May 2014 16:02:40 +0000 (18:02 +0200)] 
NEWS: mention support for shutdown LLDPDU

10 years agolldp: support for shutdown LLDPU
Vincent Bernat [Sat, 10 May 2014 15:57:08 +0000 (17:57 +0200)] 
lldp: support for shutdown LLDPU

We just accept 0 as a valid TTL and let the expiration mechanism do its
work. We ensure that we can handle those 0 TTL and avoid underflow.

Closes #5.

10 years agointerfaces: default to use of a local administered MAC address for bonds
Vincent Bernat [Sat, 10 May 2014 15:32:10 +0000 (17:32 +0200)] 
interfaces: default to use of a local administered MAC address for bonds

Since in the default configuration, some swicthes (Cisco on IOS 15.x)
will complain about MAC addresses flapping when using `configure system
bond-slave-src-mac-type fixed`, and some other switches (still Cisco)
will complain when using `configure system bond-slave-src-mac-type
fixed`, we change the default to `local`. This is not the safest option
but the probability of having a collision is quite low.

Usually, locally administered addresses are used for VM but they are
taken from a known prefix and we are unlikely to have a collision with
this prefix. On VM, we will fallback to a fixed address but hopefully,
the switch will be a virtual one which won't have a problem with the
fixed address.

10 years agoNEWS: mention the fix for local address use
Vincent Bernat [Sat, 10 May 2014 08:29:42 +0000 (10:29 +0200)] 
NEWS: mention the fix for local address use

10 years agoman: more documentation for bond-slave-src-mac-type
Vincent Bernat [Sat, 10 May 2014 07:13:02 +0000 (09:13 +0200)] 
man: more documentation for bond-slave-src-mac-type

We document the MAC address used (to be able to find it through Google
or `git grep`). We also explain why the default is `fixed` and when it
should be changed to `local`.

10 years agointerfaces: fallback to fixed MAC if already a local MAC
Vincent Bernat [Sat, 10 May 2014 07:03:03 +0000 (09:03 +0200)] 
interfaces: fallback to fixed MAC if already a local MAC

This is what is documented in the manual page.