]> git.ipfire.org Git - thirdparty/lldpd.git/log
thirdparty/lldpd.git
8 years agobuild: check runtime support of constructor/destructor feature/test-runtime-ctors 139/head
Vincent Bernat [Tue, 6 Oct 2015 05:55:46 +0000 (07:55 +0200)] 
build: check runtime support of constructor/destructor

With some libc, we may have a compile-time support (through GCC), but
not a runtime support. For example, with uclibc, support for
constructors/destructors is optional. They can be enabled with
`CFLAGS=--uclibc-ctors` on some versions. On some others, they need to
be builtin by compiling uclibc with UCLIBC_CTOR_DTOR.

Unfortunately, cannot be tested when cross-compilation is in effect.

8 years agobuild: check for support for constructor/destructor function attributes
Vincent Bernat [Mon, 5 Oct 2015 13:47:59 +0000 (15:47 +0200)] 
build: check for support for constructor/destructor function attributes

8 years agoREADME: add a word about `--enable-oldies`
Vincent Bernat [Sun, 4 Oct 2015 06:58:53 +0000 (08:58 +0200)] 
README: add a word about `--enable-oldies`

8 years agotests: add missing include for `read()`
Vincent Bernat [Sun, 4 Oct 2015 06:52:12 +0000 (08:52 +0200)] 
tests: add missing include for `read()`

8 years agotests: fix SNMP tests
Vincent Bernat [Sun, 4 Oct 2015 00:32:19 +0000 (02:32 +0200)] 
tests: fix SNMP tests

Restore the `tohex()` function in `check_snmp.c`. Use a simpler version
in `decode.c`.

8 years agoprotocols: don't use assert on paths that can be reached
Vincent Bernat [Sun, 4 Oct 2015 00:24:29 +0000 (02:24 +0200)] 
protocols: don't use assert on paths that can be reached

Malformed packets should not make lldpd crash. Ensure we can handle them
by not using assert() in this part.

8 years agolldp: fix small indentation issue
Vincent Bernat [Sun, 4 Oct 2015 00:11:24 +0000 (02:11 +0200)] 
lldp: fix small indentation issue

8 years agoNEWS: add entry for last commit
Vincent Bernat [Sun, 4 Oct 2015 00:03:57 +0000 (02:03 +0200)] 
NEWS: add entry for last commit

8 years agolldp: fix a buffer overflow when handling management address TLV
Vincent Bernat [Sat, 3 Oct 2015 23:50:38 +0000 (01:50 +0200)] 
lldp: fix a buffer overflow when handling management address TLV

When a remote device was advertising a too large management address
while still respecting TLV boundaries, lldpd would crash due to a buffer
overflow. However, the buffer being a static one, this buffer overflow
is not exploitable if hardening was not disabled. This bug exists since
version 0.5.6.

8 years agotests: add instructions on how to use afl for fuzzing
Vincent Bernat [Sat, 3 Oct 2015 23:28:30 +0000 (01:28 +0200)] 
tests: add instructions on how to use afl for fuzzing

8 years agotests: add a simple program to test frame decoding
Vincent Bernat [Sat, 3 Oct 2015 23:21:45 +0000 (01:21 +0200)] 
tests: add a simple program to test frame decoding

8 years agolldpd: remove outdated comment about custom TLV
Vincent Bernat [Sat, 3 Oct 2015 23:11:03 +0000 (01:11 +0200)] 
lldpd: remove outdated comment about custom TLV

8 years agodebian: only enable systemd unit file on system recent enough
Vincent Bernat [Fri, 25 Sep 2015 11:52:36 +0000 (13:52 +0200)] 
debian: only enable systemd unit file on system recent enough

8 years agointerfaces: remove spurious #ifdef directive
Vincent Bernat [Fri, 25 Sep 2015 09:38:52 +0000 (11:38 +0200)] 
interfaces: remove spurious #ifdef directive

8 years agointerfaces: remove old bridge code
Vincent Bernat [Fri, 25 Sep 2015 09:04:02 +0000 (11:04 +0200)] 
interfaces: remove old bridge code

It has been ages since sysfs contains the interface for querying a
bridge (2.6.7). And nowadays, sysfs is always enabled.

8 years agoMerge pull request #138 from commodo/copy-paste-fix
Vincent Bernat [Thu, 24 Sep 2015 13:54:07 +0000 (15:54 +0200)] 
Merge pull request #138 from commodo/copy-paste-fix

client: fix copy+paste for some vars

8 years agoclient: fix copy+paste for some vars 138/head
Alexandru Ardelean [Thu, 24 Sep 2015 13:40:40 +0000 (16:40 +0300)] 
client: fix copy+paste for some vars

Only affects logging.

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
8 years agodebian: on old Debian releases, enable "oldies"
Vincent Bernat [Wed, 23 Sep 2015 21:24:05 +0000 (23:24 +0200)] 
debian: on old Debian releases, enable "oldies"

8 years agodoc: add a NEWS entry for latest commit
Vincent Bernat [Wed, 23 Sep 2015 18:08:22 +0000 (20:08 +0200)] 
doc: add a NEWS entry for latest commit

8 years agolldpd: add a warning if the kernel seems to be too old
Vincent Bernat [Wed, 23 Sep 2015 18:07:08 +0000 (20:07 +0200)] 
lldpd: add a warning if the kernel seems to be too old

8 years agoredhat: on RHEL 6, enable oldies support
Vincent Bernat [Wed, 23 Sep 2015 16:25:34 +0000 (18:25 +0200)] 
redhat: on RHEL 6, enable oldies support

8 years agobuild: declare kernels before 2.6.39 "too old"
Vincent Bernat [Wed, 23 Sep 2015 16:05:07 +0000 (18:05 +0200)] 
build: declare kernels before 2.6.39 "too old"

Unfortunately, netlink support for bond devices has been introduced in
2.6.33 (commit 88ead977109da926a03068e277869ea8fedd170d). For bridge
devices, this is an addition of Linux 2.6.39 (commit
bb900b27a2f49b37bc38c08e656ea13048fee13b).

8 years agocdp: fix log message spelling
Vincent Bernat [Wed, 23 Sep 2015 08:25:02 +0000 (10:25 +0200)] 
cdp: fix log message spelling

8 years agoredhat: for RHEL 7 and CentOS 7, install systemd unit file
Vincent Bernat [Sat, 19 Sep 2015 19:30:30 +0000 (21:30 +0200)] 
redhat: for RHEL 7 and CentOS 7, install systemd unit file

8 years agobuild: fool autoreconf to do a recursive reconfiguration
Vincent Bernat [Wed, 16 Sep 2015 20:15:25 +0000 (22:15 +0200)] 
build: fool autoreconf to do a recursive reconfiguration

We just redefine AC_CONFIG_SUBDIRS to our own empty version. autoreconf
will see this during the tracing step and will automatically reconfigure
libevent and libnl as well. This may unbreak people relying on tool
trying to regen configure without using the "autogen.sh" script.

Fix #133 (with a bit of luck)

8 years agorelease: 0.7.17 released
Vincent Bernat [Wed, 9 Sep 2015 15:15:51 +0000 (17:15 +0200)] 
release: 0.7.17 released

8 years agobuild: .git may be a file
Vincent Bernat [Wed, 9 Sep 2015 15:06:34 +0000 (17:06 +0200)] 
build: .git may be a file

It happens when using a worktree.

8 years agobuild: tell seccomp is experimental
Vincent Bernat [Wed, 9 Sep 2015 14:58:26 +0000 (16:58 +0200)] 
build: tell seccomp is experimental

I don't think many people are using it and the risk for lldpd to not
work is therefore increased.

8 years agoapparmor: fix profile so that it works on Ubuntu 15.04
Vincent Bernat [Sun, 30 Aug 2015 15:11:54 +0000 (17:11 +0200)] 
apparmor: fix profile so that it works on Ubuntu 15.04

8 years agoapparmor: provide an apparmor profile
Vincent Bernat [Sun, 30 Aug 2015 14:41:27 +0000 (16:41 +0200)] 
apparmor: provide an apparmor profile

Currently, lldpd has to be installed in /usr/sbin/lldpd. Will change
that later.

8 years agopriv: no need to access bridge files unless "oldies"
Vincent Bernat [Tue, 25 Aug 2015 12:20:39 +0000 (14:20 +0200)] 
priv: no need to access bridge files unless "oldies"

However, bonding stuff is still needed because we get the permanence MAC
address this way.

8 years agolib: map_reverse_lookup needs to be always available
Vincent Bernat [Sun, 23 Aug 2015 15:08:23 +0000 (17:08 +0200)] 
lib: map_reverse_lookup needs to be always available

8 years agoAbility to enable/disable RX/TX individually on each port
Vincent Bernat [Sun, 23 Aug 2015 14:01:48 +0000 (16:01 +0200)] 
Ability to enable/disable RX/TX individually on each port

8 years agolldpcli: fix per-port occurrences of "conf lldp"
Vincent Bernat [Sun, 23 Aug 2015 13:28:58 +0000 (15:28 +0200)] 
lldpcli: fix per-port occurrences of "conf lldp"

Some commands were documented as not accepting a per-port setting while
they accepted it (without any effect). On the other hand, some other
commands were accepting and used per-port setting while not
documented. Fix all that.

8 years agolldpd: ensure retransmit is triggered when a port goes up
Vincent Bernat [Sat, 22 Aug 2015 21:27:14 +0000 (23:27 +0200)] 
lldpd: ensure retransmit is triggered when a port goes up

8 years agolldpd: fix a memory leak in the way interfaces were cloned
Vincent Bernat [Sat, 22 Aug 2015 21:13:54 +0000 (23:13 +0200)] 
lldpd: fix a memory leak in the way interfaces were cloned

8 years agointerfaces: clean neighbors if an interface goes down
Vincent Bernat [Sat, 22 Aug 2015 21:09:42 +0000 (23:09 +0200)] 
interfaces: clean neighbors if an interface goes down

As we don't delete down interfaces anymore, we must at least clear
neighbors correctly.

8 years agointerfaces: don't discard down interfaces
Vincent Bernat [Sat, 22 Aug 2015 21:03:27 +0000 (23:03 +0200)] 
interfaces: don't discard down interfaces

This is an easy way to keep the configuration associated to those
interfaces. However, we still need to discard any neighbors in this
case.

8 years agolib/atom: only allow custom TLV modification for local ports
Vincent Bernat [Sat, 22 Aug 2015 14:39:13 +0000 (16:39 +0200)] 
lib/atom: only allow custom TLV modification for local ports

8 years agoAdd a default local port as a template for any future local port.
Vincent Bernat [Sat, 22 Aug 2015 14:34:24 +0000 (16:34 +0200)] 
Add a default local port as a template for any future local port.

This default local port can be modified through the library and any
future ports will use it as a template.

8 years agoclient: factor port modification in a dedicated function
Vincent Bernat [Sat, 22 Aug 2015 08:45:46 +0000 (10:45 +0200)] 
client: factor port modification in a dedicated function

8 years agomarshal: add a macro to repair TQ lists
Vincent Bernat [Sat, 22 Aug 2015 07:58:48 +0000 (09:58 +0200)] 
marshal: add a macro to repair TQ lists

8 years agoNEWS: add some entries for 0.7.17
Vincent Bernat [Wed, 19 Aug 2015 16:40:18 +0000 (18:40 +0200)] 
NEWS: add some entries for 0.7.17

8 years agolldpcli: display a warning when a configuration file is not found
Vincent Bernat [Wed, 19 Aug 2015 15:11:26 +0000 (17:11 +0200)] 
lldpcli: display a warning when a configuration file is not found

Otherwise, the output of lldpcli would look pretty empty.

8 years agonetlink: make error message more explicit
Vincent Bernat [Tue, 18 Aug 2015 08:58:15 +0000 (10:58 +0200)] 
netlink: make error message more explicit

But there is some other reason behind it. This needs to be investigated
at some point.

8 years agosnmp: make use of newer snmp_select_info2() function
Vincent Bernat [Tue, 18 Aug 2015 08:34:04 +0000 (10:34 +0200)] 
snmp: make use of newer snmp_select_info2() function

When available, it allows to not hit the 1024 FD limit of
snmp_select_info().

8 years agointerfaces: fix incomplete declaration of interfaces_cleanup
Vincent Bernat [Mon, 17 Aug 2015 12:22:06 +0000 (14:22 +0200)] 
interfaces: fix incomplete declaration of interfaces_cleanup

For BSD and Solaris, the declaration was incomplete.

8 years agobuild: ignore libnl documentation errors
Vincent Bernat [Mon, 17 Aug 2015 12:15:21 +0000 (14:15 +0200)] 
build: ignore libnl documentation errors

8 years agobuild: fix test building with libnl dependency
Vincent Bernat [Mon, 17 Aug 2015 12:07:34 +0000 (14:07 +0200)] 
build: fix test building with libnl dependency

8 years agobuild: better find system libnl3
Vincent Bernat [Mon, 17 Aug 2015 12:05:43 +0000 (14:05 +0200)] 
build: better find system libnl3

8 years agobuild: don't limit dist tarball files to 99 characters
Vincent Bernat [Mon, 17 Aug 2015 12:00:19 +0000 (14:00 +0200)] 
build: don't limit dist tarball files to 99 characters

Otherwise, some files from libnl will be missing.

8 years agobuild: don't forget to ship libnl directory too
Vincent Bernat [Mon, 17 Aug 2015 11:44:16 +0000 (13:44 +0200)] 
build: don't forget to ship libnl directory too

At least, on Linux. Therefore, dist tarballs should now be released on
Linux (otherwise, libnl won't be shipped).

8 years agobuild: ensure we can build libnl with -Wall -Werror
Vincent Bernat [Mon, 17 Aug 2015 11:43:44 +0000 (13:43 +0200)] 
build: ensure we can build libnl with -Wall -Werror

We need to silence out unused variables, in this case.

8 years agobuild: only test for libnl on Linux
Vincent Bernat [Mon, 17 Aug 2015 10:42:07 +0000 (12:42 +0200)] 
build: only test for libnl on Linux

8 years agotravis: install libnl-3-dev
Vincent Bernat [Mon, 17 Aug 2015 10:31:58 +0000 (12:31 +0200)] 
travis: install libnl-3-dev

Even if it will not be used due to being too old.

8 years agonetlink: workaround veth being peered together through IFLA_LINK
Vincent Bernat [Mon, 17 Aug 2015 06:51:46 +0000 (08:51 +0200)] 
netlink: workaround veth being peered together through IFLA_LINK

This is a change introduced in Linux 4.1. It's easy to detect, so,
workaround it. See:

    commit a45253bf32bf49cdb2807bad212b84f5ab51ac26
    Author: Nicolas Dichtel <nicolas.dichtel@6wind.com>
    Date:   Thu Apr 2 17:07:11 2015 +0200

        veth: set iflink to the peer veth

        Now that the peer netns is advertised in rtnl messages, we can set this property
        so that IFLA_LINK will advertise the peer ifindex. It allows the userland to get
        the full veth configuration.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonetlink: use libnl3 instead of custom netlink code
Vincent Bernat [Sun, 16 Aug 2015 23:19:52 +0000 (01:19 +0200)] 
netlink: use libnl3 instead of custom netlink code

The main goal of this change is to implement a caching system for
netlink and avoiding the full scan done when a change was
triggered. Implementing a netlink cache is not as funny as it seems and
therefore, it seems just better to use libnl3.

Licensing issues are explained in README. People concerned with that
should just use dynamic linking. An embedded copy of libnl3 is also
provided, just like libevent.

8 years agobuild: add a marker at the end of third-party configure
Vincent Bernat [Sun, 16 Aug 2015 13:12:50 +0000 (15:12 +0200)] 
build: add a marker at the end of third-party configure

8 years agobuild: make --with-embedded-libevent a tri-state switch
Vincent Bernat [Sun, 16 Aug 2015 13:02:36 +0000 (15:02 +0200)] 
build: make --with-embedded-libevent a tri-state switch

It defaults to auto (same as before) but not embedding libevent can be
enforced with `--without-embedded-libevent`.

8 years agobuild: don't let autoreconf rely on AC_CONFIG_SUBDIRS macro
Vincent Bernat [Sun, 16 Aug 2015 12:07:44 +0000 (14:07 +0200)] 
build: don't let autoreconf rely on AC_CONFIG_SUBDIRS macro

We don't use it anymore.

8 years agobuild: automatically reuse regular configure args for distcheck
Vincent Bernat [Sun, 16 Aug 2015 11:57:00 +0000 (13:57 +0200)] 
build: automatically reuse regular configure args for distcheck

8 years agobuild: run ./configure for libevent even when not needed
Vincent Bernat [Sun, 16 Aug 2015 10:31:45 +0000 (12:31 +0200)] 
build: run ./configure for libevent even when not needed

We need it for "make dist"...

8 years agobuild: only enable silent rules if they are not disabled
Vincent Bernat [Sun, 16 Aug 2015 10:05:03 +0000 (12:05 +0200)] 
build: only enable silent rules if they are not disabled

8 years agobuild: replace AC_CONFIG_SUBDIRS with custom version
Vincent Bernat [Sun, 16 Aug 2015 09:49:07 +0000 (11:49 +0200)] 
build: replace AC_CONFIG_SUBDIRS with custom version

Modifying `$ac_configure_args` breaks maintainer mode. Moreover,
./configure is run even when not needed. We use a custom version
allowing to pass custom parameters and not delaying ./configure
invocation.

8 years agocdp: compute default platform when updating the chassis
Vincent Bernat [Wed, 12 Aug 2015 15:01:46 +0000 (17:01 +0200)] 
cdp: compute default platform when updating the chassis

Like for other data, collect the default platform name for CDP in
`lldpd_update_localchassis()`. Code in `cdp.c` should be independant of
the environment. Also, fix the unittests.

8 years agocdp: make the default platform name really the kernel name
Baptiste Daroussin [Wed, 12 Aug 2015 12:40:45 +0000 (14:40 +0200)] 
cdp: make the default platform name really the kernel name

Instead of hardcoding 'Linux" directly query the kernel name of the user
is not overwriting it.

8 years agobuild: propagate appropriate build flags in src/compat
Vincent Bernat [Tue, 11 Aug 2015 09:40:10 +0000 (11:40 +0200)] 
build: propagate appropriate build flags in src/compat

8 years agoosx: remove unneeded patch for launchd
Vincent Bernat [Sat, 8 Aug 2015 19:24:06 +0000 (21:24 +0200)] 
osx: remove unneeded patch for launchd

It is not needed since 0.7.6.

8 years agoosx: fix configure command line in Homebrew
Vincent Bernat [Sat, 8 Aug 2015 19:16:52 +0000 (21:16 +0200)] 
osx: fix configure command line in Homebrew

The array concat operator has an higher precedence than the ternary
operator in Ruby.

8 years agoosx: restore use of _lldpd/_lldpd user/group in Homebrew
Vincent Bernat [Sat, 8 Aug 2015 18:16:39 +0000 (20:16 +0200)] 
osx: restore use of _lldpd/_lldpd user/group in Homebrew

This has been wrongly removed in a previous commit. Homebrew don't want
to do that, but I feel this is better to keep that here.

8 years agoosx: synchronize a bit with official HomeBrew formula
Vincent Bernat [Sat, 8 Aug 2015 18:14:57 +0000 (20:14 +0200)] 
osx: synchronize a bit with official HomeBrew formula

8 years agoREADME: incorrect instructions for OS X
Vincent Bernat [Sat, 8 Aug 2015 17:56:16 +0000 (19:56 +0200)] 
README: incorrect instructions for OS X

A backslash was missing.

8 years agodebian: update symbol versioning
Vincent Bernat [Sat, 8 Aug 2015 17:02:59 +0000 (19:02 +0200)] 
debian: update symbol versioning

Since all the work is now done in the `.map` file, we don't need to say
more here. As the symbols were previously unversioned, we have to start
both generations at 0.7.16.

8 years agoclient: parse options only once
Vincent Bernat [Sat, 8 Aug 2015 17:01:17 +0000 (19:01 +0200)] 
client: parse options only once

getopt() will parse them in order. If a user specifies `-d` after `-c`,
they will be ignored but this seems like a corner case.

8 years agoclient: make lldpctl also accepts `-u` flag
Vincent Bernat [Sat, 8 Aug 2015 16:24:40 +0000 (18:24 +0200)] 
client: make lldpctl also accepts `-u` flag

Close: #129

8 years agoMerge pull request #128 from baloo/patch-1
Vincent Bernat [Sat, 8 Aug 2015 09:53:01 +0000 (11:53 +0200)] 
Merge pull request #128 from baloo/patch-1

Typo in README.md

8 years agoTypo in README.md 128/head
Arthur Gautier [Sat, 8 Aug 2015 08:10:06 +0000 (08:10 +0000)] 
Typo in README.md

8 years agodebian: don't complain about too recent Standards-Version
Vincent Bernat [Fri, 7 Aug 2015 09:51:41 +0000 (11:51 +0200)] 
debian: don't complain about too recent Standards-Version

8 years agodebian: add a lintian override for debian-revision-should-not-be-zero
Vincent Bernat [Fri, 7 Aug 2015 09:35:37 +0000 (11:35 +0200)] 
debian: add a lintian override for debian-revision-should-not-be-zero

This is expected.

8 years agoredhat: on RHEL 5, use --enable-oldies
Vincent Bernat [Fri, 7 Aug 2015 09:22:38 +0000 (11:22 +0200)] 
redhat: on RHEL 5, use --enable-oldies

Kernel is too old to support everything we need (2.6.18).

8 years agotravis: also test when oldies are enabled
Vincent Bernat [Fri, 7 Aug 2015 09:19:25 +0000 (11:19 +0200)] 
travis: also test when oldies are enabled

8 years agotravis: don't add too many alternatives in configure
Vincent Bernat [Fri, 7 Aug 2015 09:19:00 +0000 (11:19 +0200)] 
travis: don't add too many alternatives in configure

8 years agorelease: prepare 0.7.16 0.7.16
Vincent Bernat [Fri, 7 Aug 2015 09:13:19 +0000 (11:13 +0200)] 
release: prepare 0.7.16

8 years agolib: accept "language" as a valid MED location type
Vincent Bernat [Fri, 7 Aug 2015 09:02:09 +0000 (11:02 +0200)] 
lib: accept "language" as a valid MED location type

Closes #127

8 years agolldpd: add version string to "starting" log entry
Vincent Bernat [Mon, 27 Jul 2015 06:55:39 +0000 (08:55 +0200)] 
lldpd: add version string to "starting" log entry

8 years agoMerge pull request #126 from idryzhov/master
Vincent Bernat [Wed, 5 Aug 2015 16:59:52 +0000 (18:59 +0200)] 
Merge pull request #126 from idryzhov/master

daemon: fix missing parenthesis

8 years agodaemon: fix missing parenthesis 126/head
Igor Ryzhov [Wed, 5 Aug 2015 15:28:38 +0000 (18:28 +0300)] 
daemon: fix missing parenthesis

Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
8 years agoMerge pull request #124 from commodo/rename-unconfig-custom-tlvs-cmd
Vincent Bernat [Thu, 30 Jul 2015 12:52:36 +0000 (14:52 +0200)] 
Merge pull request #124 from commodo/rename-unconfig-custom-tlvs-cmd

lldpcli: rename "unconfigure lldp custom-tlvs" to "unconfigure lldp custom-tlv"

8 years agolldpcli: rename "unconfigure lldp custom-tlvs" to "unconfigure lldp custom-tlv" 124/head
Alexandru Ardelean [Thu, 30 Jul 2015 11:26:13 +0000 (14:26 +0300)] 
lldpcli: rename "unconfigure lldp custom-tlvs" to "unconfigure lldp custom-tlv"

There was a bit of confusion regarding this command.
It's documented as 'unconfigure lldp custom-tlv', when in fact it's
'unconfigure lldp custom-tlvs' (plural).

That's also my fault, since I did not update much (or take care about)
documentation.
It seems simpler to rename the command as is in the doc than change
the doc.

Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
8 years agolldpcli: remove spurious space before port aggregation
Vincent Bernat [Sat, 25 Jul 2015 15:27:56 +0000 (17:27 +0200)] 
lldpcli: remove spurious space before port aggregation

8 years agonetlink: use netlink to retrieve bridge/bond/vlan information
Vincent Bernat [Sat, 25 Jul 2015 15:23:01 +0000 (17:23 +0200)] 
netlink: use netlink to retrieve bridge/bond/vlan information

On Linux, Netlink is more efficient to get this kind of information. Use
it all the time, unless we are configured to run on old kernels. In this
case, fallback to the previous functions.

An old kernel is now < 2.6.32 (RHEL 6, Ubuntu Lucid, Debian Squeeze).

8 years agobuild: add a note to remember why setproctitle may not work
Vincent Bernat [Fri, 24 Jul 2015 16:35:50 +0000 (18:35 +0200)] 
build: add a note to remember why setproctitle may not work

On libbsd < 0.5, this is only a stub function. So, it is expected to do
nothing on OS like Ubuntu Precise.

8 years agodoc: fix latest NEWS update
Vincent Bernat [Sat, 11 Jul 2015 21:12:01 +0000 (23:12 +0200)] 
doc: fix latest NEWS update

0.7.16 was not released yet.

8 years agolib: fix memory leak when handling MED location data
Vincent Bernat [Sat, 11 Jul 2015 21:10:04 +0000 (23:10 +0200)] 
lib: fix memory leak when handling MED location data

When formatting fixed precision data, the result was copied into the
atom but the original wasn't freed.

8 years agodoc: add a NEWS entry for the chassis-related stuff
Vincent Bernat [Sat, 11 Jul 2015 21:03:11 +0000 (23:03 +0200)] 
doc: add a NEWS entry for the chassis-related stuff

8 years agolldpcli: add "show chassis" command
Vincent Bernat [Sat, 11 Jul 2015 20:59:23 +0000 (22:59 +0200)] 
lldpcli: add "show chassis" command

This demonstrates the use of `lldpctl_get_local_chassis()` function.

8 years agolldpcli: get chassis atom to query chassis-related information
Vincent Bernat [Sat, 11 Jul 2015 20:39:13 +0000 (22:39 +0200)] 
lldpcli: get chassis atom to query chassis-related information

Don't rely on the previously introduced compatibility layer. There is no
plan to deprecate this layer, but as an example, we prefer to show how
stuff should be done.

8 years agoRevert "lib: bump ABI"
Vincent Bernat [Sat, 11 Jul 2015 20:30:40 +0000 (22:30 +0200)] 
Revert "lib: bump ABI"

This reverts commit 8ee9bc2cec66b080746372e0a7b5f3ec1363b694. Thanks to
the compatibility layer, we can only do a minor ABI bump.

8 years agolib: add compatibility layer with chassis-related information
Vincent Bernat [Sat, 11 Jul 2015 20:29:12 +0000 (22:29 +0200)] 
lib: add compatibility layer with chassis-related information

When querying a port, if we don't find the requested information,
check if the chassis has the information as well.

8 years agolib: bump ABI
Vincent Bernat [Sat, 11 Jul 2015 18:41:44 +0000 (20:41 +0200)] 
lib: bump ABI

Due to the absence of compatibility layer, all atom-related functions
are considered incompatible since they don't work on ports like they
used too. This is not strictly an ABI incompatibility, but we consider
the change important enough to do like it was.

This commit will be reverted once the compatibility layer is here.