Rename HAVE_POLL to HAVE_POLL_CALL, when compiling with
snmp and poll enabled this was causing issues.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This change adds three fields to thread_master and associated code to
use them. The fields are:
* long selectpoll_timeout
This is a millisecond value that, if nonzero, will override the
internally calculated timeout for select()/poll(). -1 indicates
nonblocking while a positive value indicates the desired timeout in
milliseconds.
* bool spin
This indicates whether a call to thread_fetch() should result in a loop
until work is available. By default this is set to true, in order to
keep the default behavior. In this case a return value of NULL indicates
that a fatal signal was received in select() or poll(). If it is set to
false, thread_fetch() will return immediately. NULL is then an
acceptable return value if there is no work to be done.
* bool handle_signals
This indicates whether or not the pthread that owns the thread master
is responsible for handling signals (since this is an MT-unsafe
operation, it is best to have just the root thread do it). It is set to
true by default. Non-root pthreads should set this to false.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Adds infrastructure for keeping track of pthreads.
The general idea is to maintain a daemon-wide table of all pthreads,
running or not. A pthread is associated with its own thread master that
can be used with existing thread.c code, which provides user-space
timers, an event loop, non-blocking I/O callbacks and other facilities.
Each frr_pthread has a unique identifier that can be used to fetch it
from the table. This is to allow naming threads using a macro, for
example:
#define WRITE_THREAD 0
#define READ_THREAD 1
#define WORK_THREAD 2
The idea here is to be relatively flexible with regard to how daemons
manage their collection of pthreads; the implementation could get away
with just some #define'd constants, or keep a dynamically allocated data
structure that provides organization, searching, prioritizing, etc.
Overall this interface should provide a way to maintain the familiar
thread.c userspace threading model while progressively introducing
pthreads.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Fixes a few insufficient critical sections. Adds back locking for
thread_cancel(), since while thread_cancel() is only safe to call from
the pthread which owns the thread master due to races involving
thread_fetch() modifying thread master's ready queue, we still need
mutual exclusion here for all of the other public thread.c functions to
maintain their MT-safety.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
This change introduces synchronization mechanisms to thread.c in order
to allow safe concurrent use.
Thread.c should now be threadstafe with respect to:
* struct thread
* struct thread_master
Calls into thread.c for operations upon data of this type should not
require external synchronization.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Remove the UNDEFINED_NODE as that it's implementation breaks
our ability in BGP to figure out where we are by allowing
default: in the switch statement.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Implement BGP Prefix-SID IETF draft to be able to signal a labeled-unicast
prefix with a label index (segment ID). This makes it easier to deploy
global MPLS labels with BGP, even without other aspects of Segment Routing
implemented.
This patch implements the handling of the BGP-Prefix-SID Label Index
attribute. When received from a peer and the index is acceptable, the local
label is picked up from the SRGB and is programmed as the incoming label as
well as advertised to peers. If the index is not acceptable, no local label
is assigned. The outgoing label will always be the one advertised by the
downstream neighbor.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Implement support for negotiating IPv4 or IPv6 labeled-unicast address
family, exchanging prefixes and installing them in the routing table, as
well as interactions with Zebra for FEC registration. This is the
implementation of RFC 3107.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Implement support for activating the labeled-unicast address family in
BGP and relevant configuration for this address family.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Internal and IANA definitions for labeled-unicast SAFI. Note that this SAFI
is specific to BGP and maps to the corresponding unicast SAFI in Zebra.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Support install of labeled-unicast routes by a client. This would be
BGP, in order to install routes corresponding to AFI/SAFI 1/4 (IPv4)
or 2/4 (IPv6). Convert labeled-unicast routes into label forwarding
entries (i.e., transit LSPs) when there is a static label binding.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Implement interface that allows a client to register a FEC for obtaining
a label binding (in-label). Update client whenever the label binding is
updated and cleanup when client goes away.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
Preface with line identifying which daemon it applies to.
[Also fixes a missed "plugin" -> "module" replace.]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
lib -> Add a bit of documentation about what units we are in.
zebra -> Fix failure case to be a bit better.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This is a prepatory commit for future improvements.
Add a change to the zapi to pass the interface speed up.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
On BSD systems, the getgrouplist() function returns 0 if successful and
-1 on error.
Linux in the other hand returns *ngroups (the number of groups of which
user is a member) on success and -1 on error.
Given this difference, the most portable way to use getgrouplist()
is use its return value only for checking if it succeeded or not.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The fact that I originally wrote this in Linux Kernel style and then
reindented it to GNU makes me want to gouge my eyes out every time I
look at it. Restore original indentation.
[This patch is whitespace-only.]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
the original version of this code already used _Atomic and atomic_*().
Restore this functionality for future multithreading.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Just adding -pthread to gcc options changes libc's behaviour, e.g.
making malloc() use proper locking. This means a SEGV inside malloc()
(e.g. because malloc bookkeeping structures have been damaged by writing
to a broken pointer) can lead to a lockup by the following chain:
- random_function()
- malloc()
--- SEGV
- core_handler()
- zlog_backtrace_sigsafe()
- backtrace()
- malloc()
This will hang forever waiting for the malloc() lock to be released.
Another failure mode is dynamic linking with lazy binding (-z lazy,
default). Since backtrace() is seldomly used, this means the call to
backtrace() in the core handler can in fact result in the dynamic linker
trying to resolve the "backtrace" symbol, which can also deadlock.
Add several safeguards to prevent any of this from happening.
(Unfortunately, these are not theoretical issues - I found them by
running into them headfirst.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This is very useful to check whether a command disappeared from a
specific daemon (by comparing against an earlier output of
"grammar find-ambiguous printall nodescan")
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The vtysh preprocessing stuff doesn't like the first argument to
install_element() being something other than a _NODE constant, and the
comment hack wasn't cutting it... just expand this.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
These have copies in vtysh that do the node-switch locally and are
listed in extract.pl's ignore list. The ignore list however is
redundant since DEFUN_NOSH does the same thing...
ldpd is a bit hacky, but Renato is reworking this anyway.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Only the parent process should handle the SIGHUP signal, but we need
to make sure that this signal is ignored in the child processes so a
command like "killall -SIGHUP ldpd" won't kill ldpd.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
This adds a "-M" option to each daemon, to load dynamic modules at
startup. Modules are by default located in /usr/lib/frr/modules (lib64
if appropriate). Unloading or loading at runtime is not supported at
this point to keep things simple.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Label Manager allows to share MPLS label space among different
daemons. Each daemon can request a chunk of consecutive labels and
release it if it doesn't need them anymore. Label Manager stores the
daemon protocol and instance to identify the owner client. It uses them
to perform garbage collection, releasing all label chunks from a client
when it gets disconnected or reconnected.
Additionally, every client can request that the chunk is never garbage
collected. In that case client has the responsibility to release
non-used labels.
Zebra can host the label manager itself (if no -l param is provided) or
connect to an external one using zserv/zclient (providing its address
with -l param).
Client code is in lib/zclient.c, but currently only LDP is using it.
TODO: Allow for custom ranges requests, i.e., specify the start label
besides the chunk.
TODO: Release labels from LDP.
Signed-off-by: Bingen Eguzkitza <bingen@voltanet.io>
The following changes do not apply on master because the code has
changed:
- "vtysh: fix completion"
reverts commit 09e61a383f.
- "Revert "lib: Fix tab completions memleak, memory stats corruption""
reverts commit 4dcee34bd6.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This reverts commit 039dc61292.
The patch actually made the situation worse since the return value from
cmd_complete_command_real() was now inconsistently allocated from
different memory stat pools.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Please Note, I will be redoing this commit message with
more information.
Additionally I will rework the lib/* changes into their
own commits.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
(Note: the allow_delete thing is called "zebra" on the commandline
because that's the clearest context there, while it is called "FRR" in
the CLI because that's considerably less confusing in a vtysh env.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
... no need to have struct zlog generally-exposed.
A few files get to include log_int.h because they use zlog/vzlog.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The protocols enum serves no purpose other than adding potential for
bugs and making it complicated to add a new protocol... nuke.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Contains the fetch-and-run-thread logic, and vty startup (which is the
last thing happening before entering the main loop).
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Centralise read_config/daemonize/dryrun/pidfile/vty_serv into libfrr.
This also makes multi-instance pid/config handling available as part of
the library. It's only wired up in ospfd, but the code is in lib/.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Start centralising startup & option parsing into the library.
FRR_DAEMON_INFO is a bit weird, but it will become useful later (e.g.
for killing the ZLOG_* enum, and having the daemon name available)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This provides DMVPN support and integrates to strongSwan. Please read
README.nhrpd and README.kernel for more details.
[DL: cherry-picked from dafa05e65fe4b3b3ed5525443f554215ba14f42c]
[DL: merge partially resolved, this commit will not build.]
Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Since `afi_t` and `struct vty` are used in plist.h, the appropriate
headers for them should be included.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
This define is used only to guard macros in lib/linklist.h which
themselves are not used anywhere in the codebase and have been marked
deprecated since anno domini 2005
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
This define is used only to guard macros in lib/linklist.h which
themselves are not used anywhere in the codebase and have been marked
deprecated since anno domini 2005
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
If an EVPN entry is detected, and type is not route type 5, displays the
Ethernet MAC configured, as it was before evpn is introduced.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
In the case, evpn prefix is requested to be transformed into string, and
if the evpn prefix is not an evpn route type 5 entry, then the prefix is
returning an initialised string that mentions the vpn prefix is unknown.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
This commit simplified the string to mac conversion, since it uses
sscanf, instead of depicting each incoming character one by one, and
doing self analysis. Also,this commit changes the internal usage of the
mac address representation in mac handling function.
Signed-off-by: Lou Berger <lberger@labn.net>
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
The introduction of AFI_L2VPN prefix makes usage of AFI_ETHER deprecated
and is of no usage currently. The latter define is linked to AFI_L2VPN.
For that, the prefix enumerate has the AFI_ETHER value removed.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
The case where no buffer is passed to the str2mac function is handled.
In that case, a buffer is allocated. Then the check against the buffer
length is not done.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
As mac-address structure is to be used as a prefix field, two new
functions permit handling mac address in order to convert it. either
from string to internal value, or the reverse operation.
Internal representation of a mac address is a 6 byte char value standing
for the 6 byte value the mac address has.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Because the prefix structure may include or not evpn sub structure, then
HAVE_EVPN compilation define is also used in prefix.c, because it
references the sub field evpn of prefix structure.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
This commit is also taking into account changes related to srcdes
feature introduction in zebra folder.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Ticket: CM-12262
Reviewed By: CCR-5065
Testing Done: Manual
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
'packed' keyword had no effect on packing the afi_ethernet structure.
The attribute keyword has been eppended in order to take into account
the packed feature.
Signed-off-by: Julien Courtat <julien.courtat@6wind.com>
The requirement from draft-ietf-bess-evpn-prefix-advertisement-03
mentions that the Eth-Tag ID, IP Prefix Length and IP Prefix will be
part of the route key used by BGP. The ip prefix length is then appended
to the evpn_addr. In addition to this, the ethernet tag ID is reused.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Extend the prefix data structure to allow for basic support for EVPN type-3
and type-2 routes.
Note: This may be revised in future.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Ticket: CM-11937
Reviewed By: CCR-5001
Testing Done: None
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
To support EVPN, a new AFI and SAFI value are defined here.
For internal processing, two other values are used. Those values will
be used to reach RIB entries by using internal afi and safi values
as indexes. This commit is using naming convention for using EVPN.
External value exchanged in BGP packets is called of with
IANA_ presence in macro, while internal value will not have _IANA_
presence.
Signed-off-by: Philippe Guibert <philippe.guibert@6wind.com>
Merge the parsed graph into the existing one as a separate step. This
makes it possible to merge identical subgraphs, which is used e.g. in
bgpd for <A.B.C.D|X:X::X:X|WORD> neighbor names.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Like config_write(), this should use rename(), even though atomicity is
not a real issue here.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
sync() has a HUGE impact on systems that perform actual I/O, i.e. real
servers...
Also, we were leaking a fd on each config write ever since
c5e69a0 "lib/vty: add separate output fd support to VTYs"
(by myself :( ...)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
sync() has a HUGE impact on systems that perform actual I/O, i.e. real
servers...
Also, we were leaking a fd on each config write ever since
c5e69a0 "lib/vty: add separate output fd support to VTYs"
(by myself :( ...)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The library libzebra that is installed with FRR will
conflict with Quagga. So let's rename it to libfrr.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The code was making the correct assumption
that the v4 and v6 addresses start in the
same spot in memory and since we were looking
at a v6 prefix it would just work. This
causes distress in SA systems, so let's just
make it happy.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
It is possible if the hostname is > 32
characters that we would just overrun the
client_name data structure. Truncate
the hostname string to 31 characters (to allow for NULL)
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
was_stdio was never set up with a 'correct'
initial value, leading to cases where
we would choose what to do based upon
what was in the stack.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
In the case where we are using select as
the operator *and* we call
funcname_thread_add_read_write *and* the
fd is already set, we would overwrite
the read/write direction to always be READ.
Clearly this was a bad idea.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
The cli could be reduced for v4 and v6 code
paths into 1 function. Additionally the v6
code path had a SA issue found where it
"theoratically" could have caused a null
de-reference. This issue has been removed
as well.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
We were checking for non-null of 'struct stream *s'
after we did a stream_getl, which would have crashed
the program.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This patch introduces several new configuration commands to ldpd. These
commands should allow the operator to define advanced filtering policies
for things like label advertisement, label allocation, etc.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
prefix_list_delete wasn't releasing chained trie entries, only the main
one. Just call the proper trie_del.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Since the at_close call for the stdio vty might exit() the process, move
it to the end of the function, after freeing all memory.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
To make it possible for testcli to get a clean memory management bill.
(Note: XFREE() is NULL-safe, just like free().)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
command_lex.l was allocating as MTYPE_TMP, while command_parse.y would
just call free(). Make both use MTYPE_LEX.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Move the data structure used to have knowledge about
the zapi message types to zclient.h where it belongs.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
These error codes have ended up only being used
for socket type interfaces to the kernel(*bsd),
yet we were exposing the #defines to the entirety
of the project.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
IPv6 srcdest routes need to be keyed by both destination and source
prefix. Since the lookup order is destination first, the simplest thing
to do here is to add a second route_table to destination entries, which
then contain source entries. Sadly, the result is somewhat confusing
since a route_node might now be either a source node or a destination
node.
There are helper functions to get source and destination prefix from a
given route node (which can be either a destination or a source route).
The following bits have been added by Christian Franke
<chris@opensourcerouting.org>:
- make srcdest routing table reusable by moving it into lib
- make the srcdest routing table structure more opaque
- implement a srcdest routing table iterator
- fix a refcounting issue in src_node_lookup
- match route_node_lookup behavior with srcdest_rnode_lookup
- add accessor for the route_node table and table_info
- add string formatter srcdest_rnode2str
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Signed-off-by: Christian Franke <chris@opensourcerouting.org>
[v3: adapted for cmaster-next as of 2016-12-05]
The sourcedest code needs to get the route_node even if its info pointer
is NULL (which occurs when there are srcdest routes, but no general
destination route.)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This introduces ZAPI_MESSAGE_SRCPFX, and if set adds a source prefix
field to ZAPI IPv6 route messages sent from daemons to zebra. The
function calls all have a new prefix_ipv6 * argument specifying the
source, or NULL. All daemons currently supply NULL.
Zebra support for processing the field was added in the previous patch,
however, zebra does not do anything useful with the value yet.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Check and read the IPv6 source prefix on ZAPI messages, and pass it down
to the RIB functions (which do nothing with it yet.) Since the RIB
functions now all have a new extra argument, this also updates the
kernel route read functions to supply NULL.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
vector_remove would corrupt the data in the following sequence:
1. assume vector v = [a, b], active = 2
2. vector_unset(v, 0) => v = [NULL, b], active = 2
3. vector_remove(v, 1)
vector_remove calls vector_unset(v, 1), vector_unset notices index #0 is
also NULL and thus sets active to 0.
The equality test in vector_remove() now fails, leading it to decrement
v->active *again*, leading to an underflow that will likely crash the
daemon (and might even be exploitable).
This call sequence does not happen in existing code since vector_unset()
is not used on graph from/to lists. Nonetheless this is a buried land
mine in the code at best.
Rewrite the function - while we're at it, there's no reason to move the
entire array around, just fill the hole with the last element.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Cc: Quentin Young <qlyoung@cumulusnetworks.com>
We don't need to copy the from/to arrays, we can just iterate backwards.
NB: this makes graph_remove_edge delete only one edge (which is more
consistent with graph_add_edge allowing parallel edges).
Iterating graph->nodes backwards also makes graph_delete_graph faster
since that also iterates backwards.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Iterating over an array while deleting items needs to consider
interactions between the iteration position and deletion. The previous
code completely ignored that problem, leading to memleaks (graph_delete
skipping half of the nodes) and dangling pointers (if parallel edges
exist in graph_remove_edge).
Iterating backwards is safe and reduces "move to fill hole" overhead in
deletion.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Cc: Quentin Young <qlyoung@cumulusnetworks.com>
command.c had:
DEFINE_MTYPE_STATIC(LIB, CMD_TOKENS, "Command desc")
while command_match.c had:
DEFINE_MTYPE_STATIC(LIB, CMD_TOKENS, "Command Tokens")
... which means that there are 2 distinct MTYPE_CMD_TOKENS.
(The description text being different does not matter, even with the
same text it'd be 2 distinct types.)
command_match.c allocates token->arg in command_match_r() while
command.c frees it in del_cmd_token(). Therefore with each command
being executed, the allocation count goes up on one, down on the other.
=> clean up parser allocation counting. Also, use separate MTYPEs for
the different fields in struct cmd_token.
Fixes: #108 / ee9216cf ("lib, ripngd: clean up merge leftovers")
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Cc: Quentin Young <qlyoung@cumulusnetworks.com>
