automake file lists haven't quite kept up with recent changes, time to
fix them up so the dist tarball actually works...
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
If enabled with --with-pkg-gitversion on ./configure, this will append
git version strings and branch information at the following places:
- overall version number: 0.99.21-g0123456
- login motd and show version: tag information + git id + branches
Sample output:
Hello, this is Quagga (version 0.99.21-g14b49ad-dirty).
Copyright 1996-2005 Kunihiro Ishiguro, et al.
This is a git build of quagga_0_99_21_release-106-g14b49ad-dirty
Associated branch(es):
local:master
[v2]: fix build without gitinfo (add "else" branch)
[v2]: fix for repos without any tags (different git describe output)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Bring in sys/queue.h from the FreeBSD tree as lib/queue.h.
This header implements lists of various flavors using inline
linkages. The imported file corresponds to SVN revision 221843 (url
below) and is available under the terms of the New BSD license
(3-clause).
http://svnweb.freebsd.org/base/head/sys/sys/queue.h?revision=221843
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* lib/zebra.h
Add macro ZEBRA_NUM_OF, which returns the number of elements in a
static array.
* zebra/rib.h
Add the rib_tables_iter_t structure and associated functions,
which allow one to walk all tables in the rib.
* zebra/zebra_rib.c
- Add vrf_id_get_next() to retrieve the first VRF id (if any) that
is greater than a given VRF id.
- Add rib_tables_iter_next().
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Add some code that allows us to determine which VRF and AFI/SAFI a
given RIB table corresponds to.
* zebra/rib.h
Add rib_table_info_t structure, which contains information about
the VRF, AFI and SAFI that a table is for.
* zebra/zebra_rib.c
- Add the vrf_table_create() function, which creates a table and
sets its 'info' pointer to a newly created rib_table_info_t.
The 'info' pointer allows us to go from a route_node or a table
to the associated vrf.
- vrf_alloc(): Use vrf_create_table() to create tables.
* lib/memtypes.c
Add memory type for rib_table_info_t.
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Add the rib_dest_t structure to hold per-prefix state in the routing
information base. This gives us an appropriate place to maintain the
queueing state of a route_node. Queuing state was previously being
stored on the first rib in the list of ribs hanging off the
route_node.
* zebra/rib.h
- Add new structure rib_dest_t.
- Remove the rn_status field from 'struct rib', it is no longer
required.
- Add macros (RNODE_FOREACH_RIB, RNODE_FOREACH_RIB_SAFE) for
walking all 'struct ribs' corresponding to a route_node. These
hide the fact that there is an intermediate rib_dest_t
structure.
- Add a few utility inlines to go between a rib_dest_t and
associated structures.
* zebra/zebra_rib.c
- rib_link()/rib_unlink()
Tweak for new behavior, where the 'info' pointer of a route_node
points to a rib_dest_t. The list of ribs for a prefix now hangs
off of the dest.
Change the way we ref count route_nodes. We now hold a single
ref count on a route_node if there is a corresponding
rib_dest_t.
- Maintain the queuing state of a route_node on the flags field of
the rib_dest_t.
- Add the rib_gc_dest() function, which deletes a rib_dest_t if it
is no longer required. A rib_dest_t can be deleted iff there are
no struct ribs hanging off of it.
- Call rib_gc_dest() any time we unlink a rib from the
rib_dest_t. Currently we only need to call it once, just before
we return from rib_process().
* zebra/{redistribute,zebra_rib,zebra_snmp,zebra_vty}.c
Use new macros to walk over route_node ribs.
* lib/memtypes.c
Add memory type for rib_dest_t.
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
There is no Posix CLOCK_MONOTONIC in Darwin, but monotonically
increasing clock can be implemented using mach_absolute_time().
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Mac OS X needs HAVE_IP_HDRINCL_BSD_ORDER defined like BSD. If it's not
defined, it'll fail like this:
*** sendmsg in ospf_write failed to 224.0.0.5, id 0, off 0, len 64,
interface en0, mtu 1500: Invalid argument
Which is caused by reordering iph->ip_len in
sockopt_iphdrincl_swab_htosys.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Use the array_size() helper macro. Replaces several instances of local
macros with the same definition.
Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* lib/stream.c: (stream_set_endp) Add checks to make sure that the
supplied 'endp' is within the 'size' of the stream, and that the
current read pointer 'getp' is not beyond the specified 'endp'.
OVERVIEW
The checksum used in OSPF (rfc2328) is specified in rc905 annex B. There is an
sample implementation in rfc1008 which forms the basis of the quagga
implementation. This algorithm works perfectly when generating a checksum;
however, validation is a bit problematic.
The following LSA (generated by a non-quagga implementation) is received by
quagga and marked with an invalid checksum; however, it passes both the rfc905
and rfc1008 validation checks.
static uint8_t lsa_10_121_233_29[] = {
0x0e, 0x10, 0x02, 0x03,
0x09, 0x00, 0x35, 0x40,
0x0a, 0x79, 0xe9, 0x1d,
0x80, 0x00, 0x00, 0x03,
0x00, 0x8a, 0x00, 0x1c,
0xff, 0xff, 0xff, 0xe0,
0x00, 0x00, 0x36, 0xb0
};
LS Type: Summary-LSA (IP network)
LS Age: 3600 seconds
Do Not Age: False
Options: 0x02 (E)
Link-State Advertisement Type: Summary-LSA (IP network) (3)
Link State ID: 9.0.53.64
Advertising Router: 10.121.233.29 (10.121.233.29)
LS Sequence Number: 0x80000003
LS Checksum: 0x008a
Length: 28
Netmask: 255.255.255.224
Metric: 14000
You'll note that one byte of the checksum is 0x00; quagga would calculate the
checksum as 0xff8a.
It can be argued that the sourcing implementation generates an incorrect
checksum; however, rfc905 indicates that, for 1's complement arithmetic, the
value 255 shall be regarded as 0, thus either values are valid.
EXPLANATION
The quagga ospfd and ospf6d implementations operate by copying the PDU's
existing checksum in a holding variable, calculating the checksum, and comparing
the resulting checksum to the original. As a note, this implementation has the
side effect of modifying the contents of the PDU.
Evaluation of both rfc905 and rfc1008 shows that checksum validation should
involve calculating the sum over the PDU and checking that both resulting C0 and
C1 values are zero. This behavior is enacted in the rfc1008 implementation by
calling encodecc with k = 0 (checksum offset); however, this functionality had
been omitted from the quagga implementation.
PATCH
This patch adds the ability to call the quagga's fletcher_checksum() with a
checksum offset value of 0xffff (aka FLETCHER_CHECKSUM_VALIDATE) which returns
the sum over the buffer (a value of 0 indicates a valid checksum). This is
similar to the mechanism in rfc1008 when called with k = 0. The patch also
introduces ospf_lsa_checksum_valid().
ospf6d had it's own implementation of the fletcher checksum in
ospf6_lsa_checksum(); it's the same algorithm as in fletcher_checksum(). This
patch removes the local implementation in favor of the library's as well as creates
and uses ospf6_lsa_checksum_valid().
quagga's ISIS implementation suffers from the same problem; however, I do not
have the facilities to validate a fix to ISIS, thus this change has been left to
the ISIS maintainers. The function iso_csum_verify() should be reduced to
running the fletcher checksum over the buffer using an offset of 0.
Signed-off-by: JR Rivers <jrrivers@cumulusnetworks.com>
Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com>
Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com>
Reviewed-by: Ayan Banerjee <ayan@cumulusnetworks.com>
Reviewed-by: Shrijeet Mukherjee <shm@cumulusnetworks.com>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* lib/table.[ch]
- Add a function (route_table_get_next()) to get the route_node in
a tree that succeeds a given prefix in iteration order.
This allows one to reliably walk nodes in a tree while allowing
modifications, and is useful for achieving scale and
performance. Other approaches are also possible -- the main plus
point of this one is that it does not require any state about
the walk to be maintained in the table data structures.
- Add an iterator for walking the nodes in a tree. This introduces
a new structure (route_table_iter_t) and the following main
functions.
route_table_iter_init()
route_table_iter_pause()
route_table_iter_next()
route_table_iter_cleanup()
The iterator normally uses node pointers and the existing
route_next() function to walk nodes efficiently. When an
iteration is 'paused' with route_table_iter_pause(), it stores
the last prefix processed. The next call to
route_table_iter_next() transparently invokes
route_table_get_next() with the prefix to resume iteration.
* bgpd/bgp_table.[ch]
Add wrappers for the new table features described above.
* tests/table_test.c
Add tests for the new table code.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* lib/table.[ch]
- Add a macro (ROUTE_NODE_FIELDS) that expands to all the fields
of a route_node structure.
- Add the route_table_delegate_t structure, a function vector
which allows clients to customize the behavior of one or more
tables.
The delegate currently contains the 'create_node' and
'destroy_node' functions, and hence enables a table to use an
alternative node structure. The alternative node is expected to
embed the fields of a route_node using ROUTE_NODE_FIELDS.
- Add route_table_init_with_delegate() to create a new table with
a given delegate.
- Make route_table_init() a thin wrapper around
route_table_init_with_delegate(). The delegate it passes in
simply creates/destroys route_node structures as before.
- Add a user data pointer (info) to the route_table
structure. This can be used by a client to keep per-table state.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* lib/table.c
- Maintain table node count. Expose it via the route_table_count()
function (from revision cbdfbaa5).
- route_unlock_node(): Add assertion (from revision 228da428).
- route_table_free(): Make static and fix up cleanup code (from
revision 228da428).
- route_node_delete(): Change to be static.
- Add 'const' qualifier in a couple places.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This makes it possible to have both "show babel route A.B.C.D/M"
and "show babel route X:X::X:X/M" commands at the same time without
the parser complaining about ambiguity.
* lib/command.c: only accept STATE_DOT after : was seen.
Reported-by: Juliusz Chroboczek <jch@pps.jussieu.fr>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Most table handlers do not expect to be given an OID whose prefix is
outside what they can handle. This is not a problem with the SMUX
implementation since it always correct the OID such that the prefix
matches. However, this is not the case for the AgentX
implementation. A new function, smux_header_table() is used to do this
normalization.
smux_trap() signature has been changed to provide appropriate level
information to send SNMPv2 notifications. This includes the addition
of the enterprise OID to use (from which is derived the SNMP trap OID)
and the MIB registry to locate the appropriate function for variable
bindings provided by the trap.
The SMUX implementation has been updated but ignore the provided
enterprise OID. Instead, it still uses the SMUX peer OID to keep
compatibility with previous versions of Quagga. The SMUX
implementation also ignores the provided MIB registry since it uses
smux_get() function to grab the appropriate values. This is not
possible with the AgentX implementation since there is no such
function provided by NetSNMP.
smux_trap() contains an argument whose use appears to be to set
sysUpTime.0/timestamp field in SNMP trap. However, this value is not
used in smux_trap(). Moreover, it is expected that this field is the
value of sysUpTime.0 when the trap was sent and not any other time
related to the trap. To avoid any confusion, we remove this field from
the signature of the function.
--enable-snmp will enable AgentX support in Quagga. SMUX is still here
and can be enabled with --enable-snmp=smux. AgentX support can be
enabled with "agentx" in configuration file. As for SMUX, this command
is not understood by vtysh. It can be disabled with "no agentx",
though there is no real use of this since this command cannot be used
with vtysh.
If "agentx" and "no agentx" command were added to vtysh, it would not
be possible to disable agentx support after enabling it because
NetSNMP does not expose the appropriate methods for this.
The internals of AgentX are hidden by NetSNMP. Therefore, we don't
have a file descriptor to add to the threading system. We do not have
the timers to set either. Therefore, the event loop is modified to
make use of snmp_select_info() from NetSNMP.
Traps are not supported yet.
lib/snmp.c gets OID related helper functions that can be used with
another SNMP interface. smux.h is cleaned of SMUX specific bits to
only expose functions that may be used by an alternative
implementation. We also do not redefine functions already present in
NetSNMP. Just use the appropriate headers.
The correct method to link to NetSNMP is to use net-snmp-config (which
is like pkg-config). Explicit link to libcrypto is also dropped
(NetSNMP libs are linked to libcrypto, no need to link Quagga to
it). Moreover, @SNMP_INCLUDES@ is dropped because useless. Due to a
bug in configure.ac, it was properly populated.
Some .h files in lib/ are autogenerated. The search path should
include the build directory and the source directory. They usually
match but sometimes, they may be different. For example:
$ mkdir build
$ cd build
$ ../configure
$ make
* sockunion.h: Add back sockunion_str2su, its removal breaks things needlessly
(e.g. our own unit tests).
* sockunion.c: (sockunion_str2su) implement on top of str2sockunion.
This avoids memory heap fragmentation and imposses less load on the
system memory allocator.
* thread.h: FUNCNAME_LEN defined to 64 (ISO C99 says max 63)
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
[changed FUNCNAME_LEN to a less arbitrary value]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
thread_trim_head() already checks that the list is not empty.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* thread.c: It's only temporarily used in thread_call() to calculate the
diffs. Saves 80 bytes per copy.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Reduce indirection for values that doesn't change in the loop.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
[adjusted after dropping previous patch]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This function scores 2nd, profiling a full internet table load. It's called
for every prefix received.
Instead of looping in the interface lists comparing addresses use a hash
to mantain them.
* bgpd.c: Init the own address hash.
* bgp_nexthop.c: Introduce methods to maintain an own address hash.
(bgp_connected_add) add addresses to the hash.
(bgp_connected_delete) delete addresses from the hash.
(bgp_nexthop_self) lookup addresses in the hash. Removed the unused afi_t
parameter.
* bgp_route.c: (bgp_update_main) Micro-optimization, rearranged condition to
not lookup the hash for bogus nexthops (0.0.0.0 or a class D/E address)
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
ospfd was crashing some times on neighbour going down. The cause was that
ospf_nsm_event() was accessing already freed memory in ospf_nbr_delete()
call from ospf_nsm_event().
What happens is that since commit b5043aab (lib: fix incorrect thread
list...) now a thread can be on the event and ready lists but
thread_cancel_event() doesn't account for that.
* thread.c: (thread_cancel_event) loop on the ready list too to cancel
pending events.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
* command.c: (cmd_ipv6_match) Drop IPv6 address recognition
heuristics and solely rely on inet_pton, because strings
like "abcd" were mistaken for IPv6 addresses.
This affects e.g. the command "neighbour WORD peer-group",
which won't work with words consisting of up to 4 characters
between 'a' and 'f' and digits.
From: Roman Hoog Antink <rha@open.ch>
[full delete instead of #if 0]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
sockunion_str2su() use is prone to memory leaks. Remove it's use all over
the code.
At least these commands leaked a sockunion union:
- show ip bgp vpnv4 ... routes
- show ip bgp ... received prefix-filter
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@diac24.net>
Use of this function is prone to memory leaks.
This fixes a memory accounting bug for vty denied connections.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@diac24.net>
A route-map with the mentioned statement causes a memory leak for every
prefix that matches.
Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David Lamparter <equinox@diac24.net>
continually changing the zserv protocol without bumping up the version
number has made it impossible to talk to zebra without knowing the exact
version. in reality, increasing the version number more often guards
against inadvertedly running incompatible versions of a daemon and zebra
as well as aids external development.
* lib/zclient.h: #define ZSERV_VERSION 2
Signed-off-by: David Lamparter <equinox@diac24.net>
The steps were:
$ git checkout google-is-is
$ git merge quagga
$ git checkout google-is-is -- isisd
# Resolve conflicts in the following:
lib/md5.h
zebra/rt_netlink.c
zebra/zebra_rib.c
zebra/zserv.c
Note that the content in the isisd directory is left unchanged in the
merge. As a result, changes made to isisd as part of the following
commits on the quagga mainline are dropped.
# 8ced4e82 is the merge base, e96b3121 is the current quagga master
$ git log --oneline --reverse 8ced4e82..e96b3121 -- isisd
5574999 isisd: fix crash on "no router isis" (BZ#536)
8998075 isisd: raise hello rate for DIS (BZ#539)
306ca83 isisd: include hash.h, not hash.c
b82cdeb delete CVS keywords
2f65867 isisd: indent longopts array
b511468 quagga: option "-z" ("--socket <path>") added
05e54ee build: delete .cvsignore files
b4e45f6 fix zebra protocol after MP-BGP changes
7fd6cd8 isisd: fix circuit state machine
907fd95 isisd: send proper LSP after DIS election
d034aa0 isisd: fix wrong next-hops from SPF
c25eaff isisd: unexpected kernel routing table (BZ#544)
e6b03b7 isisd: implement MD5 circuit authentication
Change interface up/down notification messages to also include the
hardware address of the interface. The format of these messages is now
identical to the interface add message -- move the serialization code
to common functions.
* lib/zclient.c: Modify zebra_interface_if_set_value() to also parse
the hardware address. Invoke it from zebra_interface_add_read()
and and zebra_interface_state_read().
* zebra/zserv.c: Add zserv_encode_interface(). Invoke it from
zserv_interface_add(), zserv_interface_delete() and
zserv_interface_update().
* lib/if.c: Change if_delete_retain() to delete all connected
addresses, but to not free the list that holds them. Free the list
just before the interface structure itself is freed, in
if_delete().
* lib/stream.[ch]:
- Add stream_set_endp(). This can be used to trim data (for
example, padding) at the end of a stream.
- Fix swapped 'getp' and 'endp' parameters in STREAM_WARN_OFFSETS.
From: Subbaiah Venkata <svenkata@google.com>
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@diac24.net>
* lib/linklist.h
- Change the listnextnode, listhead and listtail macros so that
they allow the list pointer to be NULL.
- Modify the ALL_LIST_ELEMENTS* macros such that they clear the
data pointer at the beginning and end of the loop.
From: Josh Bailey <joshb@google.com>
Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org>
Signed-off-by: David Lamparter <equinox@diac24.net>
nonwithstanding any desire for optimisation, these versions are shorter
and more concise. reading the comments, they might even be easier to
understand.
I've tested them on i686 and x86_64, and checked that correct assembler
code is emitted for ARM, MIPS and PowerPC.
IPv6 is left as an exercise for another day, none of the ideas I had led
to a "yes, this is the one to go with" solution.
Signed-off-by: David Lamparter <equinox@diac24.net>
This reverts commit d171bf58ef.
There are multiple reasons for this revert, including bug #720, but also
quite simply the unreadability of having 2000 lines of an autogenerated
table in the middle of a source file.
Signed-off-by: David Lamparter <equinox@diac24.net>
Reported-by: Martin Winter <mwinter@opensourcerouting.org>
* Initial import of the Babel routing protocol, ported to Quagga.
* LICENCE: Update the original LICENCE file to include all known potentially
applicable copyright claims. Ask that any future contributors to babeld/
grant MIT/X11 licence to their work.
* *.{c,h}: Add GPL headers, in according with the SFLC guidance on
dealing with potentially mixed GPL/other licensed work, at:
https://www.softwarefreedom.org/resources/2007/gpl-non-gpl-collaboration.html
While defining two "maskbytes" arrays for the respective endiannesses,
the code was unconditionally using one that only worked on little endian
systems.
Use preprocessor macros to avoid the somewhat expensive
htonl/ntohl macros.
From: Renato Westphal <renatowestphal@gmail.com>
Signed-off-by: David Lamparter <equinox@diac24.net>
* thread.c: (thread_timer_process,thread_process) thread_list_delete nulls
thread->next. Loops need to save next first, or will only process the head.
Problem noted by Lou Berger <lberger@labn.net>.
* lib/thread.h: new timer macro, millisecond precision
(this was cherrypicked from pimd to isolate all non-contained changes)
From: Everton Marques <everton.marques@gmail.com>
Signed-off-by: David Lamparter <equinox@diac24.net>
This patch contains the following:
1. Addition of IPv6 SAFI_MULTICAST BGP routes into the BGP Multicast RIB.
2. Deletion of IPv6 SAFI_MULTICAST BGP routes from the BGP Multicast RIB.
This patch contains the following:
1. Addition of IPv4 SAFI_MULTICAST BGP routes into the BGP Multicast RIB.
2. Deletion of IPv4 SAFI_MULTICAST BGP routes from the BGP Multicast RIB.
* if.c: (if_dump) loop that doesn't do anything, wants to be
before the zlog of what it's meant to print out so all the connected
addresses get printed out. Trival: just a debug function
* (general) Move functions in headers into files, to be compiled into
shared object files. Remove inline qualifier from functions. Let the
compiler do the work.
IPV4_ADDR_SAME() wasn't passed the right union member, this could cause
a bug due to strict-aliasing. IPV6_ADDR_SAME() case got its fix before
the error could be created by macro upgrade.
* lib/prefix.h
* IPV4_CLASS_DE(): make consistent with counterpart macros
* bgp_packet.c
* bgp_open_receive(): test using macro instead of ">="
* bgp_route.c
* bgp_update_rsclient(): idem
* bgp_update_main(): idem
ZEBRA_HELLO message is used by routing daemons to inform zebra
what type of routes daemon will be announcing to zebra. Also
zebra uses route_type_oaths array to track which daemon announces
which protocol. Zebra mops up routes if daemon didn't for some
reason.
The new implementation makes use of a 64KB mapping table, which makes it
possible to compute masklen faster and with constant execution time. The
map also allows for additional version of the function, which can detect
errors in input argument.
The previous implementation had a variable cost of execution, which
depended on masklen in a non-linear manner, and at its worst (/31) was
4 times slower, than the new implementation. The only case of old
function just slightly outperforming the new one is /0, which is of
little practical interest.
getaddrinfo returns a list of socket parameters for listening. it
will contain both IPv4 and IPv6 listening sockets. unless we use
IPV6_V6ONLY on the IPv6 ones, only the socket listed first will
work. if the IPv4 one came first, the IPv6 one would get an
"Address in use" error.
this functionality was already present for bgpd and its listening
sockets. as it is needed for vtys as well, make it a common helper.
Conflicts:
lib/sockunion.c
this replaces most occurences of routing protocol lists by preprocessor
defines from route_types.h. the latter is autogenerated from
route_types.txt by a perl script (previously awk). adding a routing
protocol now is mostly a matter of changing route_types.txt and log.c.
Conflicts:
lib/route_types.awk
All daemons modified to support custom path to zserv
socket.
lib: generalize a zclient connection
zclient_socket_connect added. zclient_socket and
zclient_socket_un were hidden under static expression.
"zclient_serv_path_set" modified.
The delete_hook was being run after calling access list delete function.
This would cause ospf to dereference a NULL, in ospf_filter_update
because 'access->name' was already freed.
See also:
https://bugzilla.vyatta.com/show_bug.cgi?id=7654
IPv6 supports the same concept of differentiated service for routing
protocols as IPv4, but like too many things, the standards committee
decided that having two names for the same thing wasn't good enough and
introduced a third more generic term transport class.
The socket option to set transport class works the same as IPv4, but the
arguments are different.
* lib/sockopt.[ch]
* setsockopt_ipv6_tclass(): new function
* bgpd/bgp_network.c
* bgp_connect(): set socket option
* bgp_listener(): set socket option
* ospf6d/ospf6_network.c
* ospf6_set_transport_class(): new function
* ospf6_serv_sock(): set socket option
* ripngd/ripngd.c
* ripng_make_socket(): set socket option
* lib/prefix.h
* IPV4_CLASS_DE(): new helper macro
* bgp_attr.c
* bgp_attr_nexthop(): add check for "partial" bit, refresh flag error
reporting, explain meaning of RFC4271 section 6.3 and implement it
Conflicts:
bgpd/bgp_attr.c
- SAFI value 3 is reserved. It was assigned by RFC 2858 for a use
that was never fully implemented, so it is deprecated by this
document.
* zebra.h: rename macro
* bgp_fsm.c: (bgp_graceful_restart_timer_expire,
bgp_graceful_stale_timer_expire, bgp_stop, bgp_establish): update
* bgpd.c: (peer_nsf_stop): update
* bgp_open.c: (bgp_capability_vty_out): SAFI 3 isn't a recognized case
any more
* log.[ch]
* mes_lookup: add a parameter with the name of the message list, print
the name in case of failure.
* LOOKUP macro: pass the name of the message list.
(This commit is based on the patch from BZ#420, and should fix that bug.)
* configure.ac: detect availability of that API
* sockopt.c (setsockopt_ipv4_multicast): use it for join/leave IPv4
multicast groups
* sockopt.[ch] (setsockopt_ipv4_multicast): ifindex is now mandatory (all
non-ancient OSes can use it anyway), and if_addr parameter (the address
of the interface) is now gone. (setsockopt_ipv4_multicast_if):
IP_MULTICAST_IF processing moved to this new function
* ospf_network.c (ospf_if_add_allspfrouters, ospf_if_drop_allspfrouters,
ospf_if_add_alldrouters, ospf_if_drop_alldrouters, ospf_if_ipmulticast),
rip_interface.c (ipv4_multicast_join, ipv4_multicast_leave,
rip_interface_new): adapt to the new interface
information based on the multipath list (mp_list) generated during
the best path calculation. Display "multipath" for paths that are
multipath and also on bestpath if the route is multipath. Flag a
best path with the BGP_INFO_MULTIPATH_CHG if the multipath
set has changed since the last update. This can be used to trigger
updates to zebra and peers.
The multipath markup is a lazily allocated bgp_info_mpath structure
that is added to the best path and any multipaths. The mpath structures
are linked together with the best path element at the head and the
other elements ordered by nexthop and then by peer address. This
markup scheme is updated by calling bgp_info_mpath_update() and passing
in a new mp_list the the current multipath set. There are additional
API's for walking the multipath set, querying the count of multipaths,
and for cleaning up the multipath markup information when freeing path
information.
* bgpd/bgp_mpath.c
* bgp_info_mpath_new(): Allocation of new mpath element
* bgp_info_mpath_free(): Release memory for mpath element
* bgp_info_mpath_get(): Access mpath element of path. Allocate memory
on-demand
* bgp_info_mpath_enqueue(): Enqueue a path onto the multipath list
* bgp_info_mpath_dequeue(): Remove a path from the multipath list
* bgp_info_mpath_first(): Return first path on the multipath list
* bgp_info_mpath_next(): Return next path on the multipath list
* bgp_info_mpath_count(): Return the number of paths on the multipath list
* bgp_info_mpath_count_set(): Set the number of paths on the multipath list
* bgp_info_mpath_update(): Update multipath markup on bgp route table entry
and flag any changes. Emit 'debug bgp event' output on any multipath
change.
* bgpd/bgp_mpath.h
* struct bgp_info_mpath: Information added to a bgp_info path to record
multipath information
* External declarations for new functions in bgp_mpath.c
* bgpd/bgp_route.c
* bgp_info_free(): Free mpath memory when freeing path information
* bgp_info_reap(): Dequeue path from multipath queue before deleting it
* bgp_best_selection(): Calls bgp_info_mpath_update() with latest
mp_list to mark-up rib table entry
* bgp_vty_out_detail(): Add display of multipath flag for a path. Also
display 'multipath' for bestpath if it is a multipath route
* bgpd/bgp_route.h
* struct bgp_info: Add pointer to bgp_info_mpath information
* Add flags to mark a path as multipath (BGP_INFO_MULTIPATH) and
to mark bestpath if multipath information has changed
(BGP_INFO_MULTIPATH_CHG)
* lib/memtypes.c
* Add MTYPE_BGP_MPATH_INFO for allocating memory for bgp_info_mpath
* tests/bgp_mpath_test.c
* Add test case for bgp_info_mpath_update() and supporting functions
* bgp_vty.c: (peer_ebgp_multihop_{un,}set_vty) tail-call cleanup.
({no_,}neighbor_ttl_security) ditto.
* bgpd.c: (peer_ttl_security_hops_set) Peer group checks and TTL set only
need to be done on transition.
* sockunion.c: (sockopt_minttl) remove always-on debug and improve readability.
* bgpd: Add support for RFC 5082 GTSM, which allows the TTL field to be used
to verify that incoming packets have been sent from neighbours no more
than X IP hops away. In other words, this allows packets that were sent from
further away (i.e. not by the neighbour with known distance, and so possibly
a miscreant) to be filtered out.
* lib/sockunion.{c,h}: (sockopt_minttl) new function, to set a minimum TTL
using the IP_MINTTL socket opt.
* bgpd.h: (BGP_ERR_NO_EBGP_MULTIHOP_WITH_TTLHACK) define for command
error for minttl.
(struct peer) add a config variable, to store the configured minttl.
(peer_ttl_security_hops_{set,unset}) configuration handlers
* bgpd.c: (peer_group_get) init gtsm_hops
(peer_ebgp_multihop_{un,}set) check for conflicts with GTSM. Multihop and
GTSM can't both be active for a peer at the same time.
(peer_ttl_security_hops_set) set minttl, taking care to avoid conflicts with
ebgp_multihop.
(bgp_config_write_peer) write out minttl as "neighbor .. ttl-security hops X".
* bgp_vty.c: (bgp_vty_return) message for
BGP_ERR_NO_EBGP_MULTIHOP_WITH_TTLHACK
(peer_ebgp_multihop_{un,}set_vty)
* bgp_network.c: (bgp_accept) set minttl on accepted sockets if appropriate.
(bgp_connect) ditto for outbound.
* hash.{h,c}: (string_hash_make) Hash optimised for strings, current
implementation using Bernstein hash, which offers a good compromise
between distribution and performance.
* distribute.c: (distribute_hash_make) use previous instead of additive
string hash.
* if_rmap.c: (if_rmap_hash_make) ditto
* bgpd: Connected table locks were being locked but not unlocked, such that
eventually a lock would exceed 2^31 and become negative, thus triggering
an assert later on.
* bgp_main.c: (bgp_exit) delete connected elements along with ifp's.
* bgp_nexthop.c: (bgp_nexthop_lookup{,_ipv6}) add missing unlocks
(bgp_multiaccess_check_v4) ditto
(bgp_connected_{add,delete}) Use a distinct memtype for bgp_connected_ref.
(bgp_scan_finish) reset the nexthop cache to clean it up when bgpd exits
* bgp_route.c: fix missing bgp_node unlocks
* lib/memtype.c: (memory_list_bgp) add MTYPE_BGP_CONN
* testing: has been tested for almost 2 months now.
* lib/memory.c: (zrealloc) If is called with NULL pointer then it should
increment allocations because it behaves the same as zmalloc.
(zfree) is called with NULL pointer, it does nothing therefore allocation
count should not change.
* lib/table.c: (route_node_match) fix overshoot that was causing this
function to go 1 bit too far and thus reading past end of prefix.
(route_node_lookup) be defensive - don't assume others will clean up
leaves when removing info.
* workqueue.c: (work_queue_run) Err more on the side of keeping granularity
down, by being more conservative about increasing it.
Also, fix mispelling.
* (general) this can be useful when investigating thread latency problems,
when you don't want to have to restart a daemon between tests.
* thread.c: (cpu_record_(hash_)clear) wipe the stored thread cpu history
data, according to the filter, similar to the vty print code.
(clear_thread_cpu_cmd) new command to clear data.
* thread.h: export new command
* command.c: install it
* thread.c: (thread_fetch) the current scheduler will service events
indefinitely, ignoring I/O and timers, so long as there are events.
In other words, events can crowd out I/O and timers. In theory this
shouldn't be a huge problem as events are generated only by timers
and I/O, however in practice it means normal-load behaviour is not as
useful a predictor of high-load behaviour as it should be.
Fix this by considering all the kinds of threads, in every run of the
scheduler. For any given run, we prioritise events, however across
runs the scheduler should be fair.
This has been observed to give more stable inter-packet times in
testing of ospfd (i.e. lower std-dev).
(thread_process) new heler to queue all the given threads onto the ready
list
* thread.h: (struct cpu_thread_history.funcname) malloc gets called on this
so it should not be const
* thread.c: (cpu_record_print) cast the (const char *) to (char *), this
function knows what it's doing (free wont be called on it).
If router-id is not specified in ospf6d.conf, ospf6d will get it from
the zebra daemon. But ospf6d originates Link LSAs before the router-id
is returned by zebra, thus this router's Link LSAs will be flooded
with AdvRouter set to 0.
* zclient.c: zclient_start(): send ZEBRA_INTERFACE_ADD message after
ZEBRA_ROUTER_ID_ADD, not before
* bgpd/bgp_attr.c, bgpd/bgp_open.h, bgpd/bgp_route.c, lib/prefix.c,
lib/prefix.h: Various integer types were being used where, if we
had strict type checking, afi_t and safi_t would be required.
Signed-off-by: G.Balaji <balajig81@gmail.com>
(cherry picked from commit c8af35ffa2dc79ff7d7ff00b1b61f1f50a100ab6)
Stephen:
A recent change breaks build Quagga on Debian Lenny with amd64 (but not
i386). Not sure whether this is cause by ld, gcc, or libtool; but
whatever it just won't work for a large number of users.
Mathias:
I would guess it's a problem of libtool because it adds "-fPIC -DPIC" to
the compiler flags but doesn't remove -fPIE. That wouldn't be a problem
if the compiler would ignore the -fPIE in favor to the latter -fPIC, but
obviously it does not. [...] Those objects were actually compiled with
-fPIC but also -fPIE which seems to confuse gcc.
Because the final executables are built as position independent
executables (PIE) -- when configure has detected compiler supported for
PIE -- the objects in the library archive must be built in that way,
too. Otherwise the runtime linker has to do unneccesary relocation for
each start of the program. Even worse, the programs won't even be able
to start on a (hardened) kernel that doesn't allow those relocations to
happen by preventing making the .text segment writable (PaX's MPROTECT
feature comes to mind). The attached patch fixes this issue by adding
the appropriate flags to the Makefile.am.
* lib/memory.h
* mtype_zcalloc(): correct function prototype
* XFREE(): make both definitions consistent in setting
the pointer to NULL after freeing the memory
These changes will only have an effect if MEMORY_LOG is defined
(it is not by default).
* isisd/isis_pdu.c
* maskbit: this pre-initialized array is not modified
* lib/prefix.c
* maskbit: idem
* lib/command.c
* facility_map: idem
* itoa64: idem
* default_motd: make local var static
* facility_name(): update local var accordingly
* facility_match(): idem
Make one version of check prefix bit, and put it inline
with proper prototype. This gets rid of some macro's and also some
assert() that can never happen on a non-broken compiler.
* bgpd/bgp_table.c
* CHECK_BIT(): sayonara
* check_bit(): sayonara
* SET_LINK(): sayonara
* set_link(): make use of prefix_bit() instead of check_bit()
* bgp_node_match(): idem
* bgp_node_lookup(): idem
* bgp_node_get(): idem
* lib/prefix.h
* prefix_bit(): new inline version of check_bit()
* lib/table.c
* CHECK_BIT(): sayonara
* check_bit(): sayonara
* SET_LINK(): sayonara
* set_link(): make use of prefix_bit() instead of check_bit()
* route_node_match(): idem
* route_node_lookup(): idem
* route_node_get(): idem
* ospf6d/ospf6_lsdb.c
* CHECK_BIT(): sayonara
* ospf6_lsdb_lookup_next(): make use of prefix_bit() instead of
CHECK_BIT()
* ospf6_lsdb_type_router_head(): idem
* ospf6_lsdb_type_head(): idem
* ospf6d/ospf6_route.c
* CHECK_BIT(): sayonara
* ospf6_route_match_head() make use of prefix_bit() instead of
* CHECK_BIT()
* md5.c
* md5_calc(): If building on little endian then X
will be 'const uint32_t' which causes compiler
warning in code that never gets used. Just move
the endif to make sure code is not checked.
* table.c: general type safety and compiler help:
* maskbit[]: become const
* route_node_match(): take const args
* route_node_match_ipv4(): idem
* route_node_match_ipv6(): idem
* check_bit(): idem, plus adjust local vars typing
Start BGP listener only after first instance is started. This helps the
security if BGP is not used but daemon is started. It also addresses some
issues like MD5 not working on listener unless IPV6 configured (because
listener was not in list); as well as compiler warnings.
* bgp_network.c: (bgp_listener) listen socket creation consolidated here
(bgp_socket) Use bgp_listener
* bgpd.c: (bgp_get) call bgp_socket on creation of first struct bgp.
(bgp_init) remove bgp_socket call.
* memtypes.c: Add MTYPE_BGP_LISTENER
I've spent the last several weeks working on stability fixes to bgpd.
These patches fix all of the numerous crashes, assertion failures, memory
leaks and memory stomping I could find. Valgrind was used extensively.
Added new function bgp_exit() to help catch problems. If "debug bgp" is
configured and bgpd exits with status of 0, statistics on remaining
lib/memory.c allocations are printed to stderr. It is my hope that other
developers will use this to stay on top of memory issues.
Example questionable exit:
bgpd: memstats: Current memory utilization in module LIB:
bgpd: memstats: Link List : 6
bgpd: memstats: Link Node : 5
bgpd: memstats: Hash : 8
bgpd: memstats: Hash Bucket : 2
bgpd: memstats: Hash Index : 8
bgpd: memstats: Work queue : 3
bgpd: memstats: Work queue item : 2
bgpd: memstats: Work queue name string : 3
bgpd: memstats: Current memory utilization in module BGP:
bgpd: memstats: BGP instance : 1
bgpd: memstats: BGP peer : 1
bgpd: memstats: BGP peer hostname : 1
bgpd: memstats: BGP attribute : 1
bgpd: memstats: BGP extra attributes : 1
bgpd: memstats: BGP aspath : 1
bgpd: memstats: BGP aspath str : 1
bgpd: memstats: BGP table : 24
bgpd: memstats: BGP node : 1
bgpd: memstats: BGP route : 1
bgpd: memstats: BGP synchronise : 8
bgpd: memstats: BGP Process queue : 1
bgpd: memstats: BGP node clear queue : 1
bgpd: memstats: NOTE: If configuration exists, utilization may be expected.
Example clean exit:
bgpd: memstats: No remaining tracked memory utilization.
This patch fixes bug #397: "Invalid free in bgp_announce_check()".
This patch fixes bug #492: "SIGBUS in bgpd/bgp_route.c:
bgp_clear_route_node()".
My apologies for not separating out these changes into individual patches.
The complexity of doing so boggled what is left of my brain. I hope this
is all still useful to the community.
This code has been production tested, in non-route-server-client mode, on
a linux 32-bit box and a 64-bit box.
Release/reset functions, used by bgp_exit(), added to:
bgpd/bgp_attr.c,h
bgpd/bgp_community.c,h
bgpd/bgp_dump.c,h
bgpd/bgp_ecommunity.c,h
bgpd/bgp_filter.c,h
bgpd/bgp_nexthop.c,h
bgpd/bgp_route.c,h
lib/routemap.c,h
File by file analysis:
* bgpd/bgp_aspath.c: Prevent re-use of ashash after it is released.
* bgpd/bgp_attr.c: #if removed uncalled cluster_dup().
* bgpd/bgp_clist.c,h: Allow community_list_terminate() to be called from
bgp_exit().
* bgpd/bgp_filter.c: Fix aslist->name use without allocation check, and
also fix memory leak.
* bgpd/bgp_main.c: Created bgp_exit() exit routine. This function frees
allocations made as part of bgpd initialization and, to some extent,
configuration. If "debug bgp" is configured, memory stats are printed
as described above.
* bgpd/bgp_nexthop.c: zclient_new() already allocates stream for
ibuf/obuf, so bgp_scan_init() shouldn't do it too. Also, made it so
zlookup is global so bgp_exit() can use it.
* bgpd/bgp_packet.c: bgp_capability_msg_parse() call to bgp_clear_route()
adjusted to use new BGP_CLEAR_ROUTE_NORMAL flag.
* bgpd/bgp_route.h: Correct reference counter "lock" to be signed.
bgp_clear_route() now accepts a bgp_clear_route_type of either
BGP_CLEAR_ROUTE_NORMAL or BGP_CLEAR_ROUTE_MY_RSCLIENT.
* bgpd/bgp_route.c:
- bgp_process_rsclient(): attr was being zero'ed and then
bgp_attr_extra_free() was being called with it, even though it was
never filled with valid data.
- bgp_process_rsclient(): Make sure rsclient->group is not NULL before
use.
- bgp_processq_del(): Add call to bgp_table_unlock().
- bgp_process(): Add call to bgp_table_lock().
- bgp_update_rsclient(): memset clearing of new_attr not needed since
declarationw with "= { 0 }" does it. memset was already commented
out.
- bgp_update_rsclient(): Fix screwed up misleading indentation.
- bgp_withdraw_rsclient(): Fix screwed up misleading indentation.
- bgp_clear_route_node(): Support BGP_CLEAR_ROUTE_MY_RSCLIENT.
- bgp_clear_node_queue_del(): Add call to bgp_table_unlock() and also
free struct bgp_clear_node_queue used for work item.
- bgp_clear_node_complete(): Do peer_unlock() after BGP_EVENT_ADD() in
case peer is released by peer_unlock() call.
- bgp_clear_route_table(): Support BGP_CLEAR_ROUTE_MY_RSCLIENT. Use
struct bgp_clear_node_queue to supply data to worker. Add call to
bgp_table_lock().
- bgp_clear_route(): Add support for BGP_CLEAR_ROUTE_NORMAL or
BGP_CLEAR_ROUTE_MY_RSCLIENT.
- bgp_clear_route_all(): Use BGP_CLEAR_ROUTE_NORMAL.
Bug 397 fixes:
- bgp_default_originate()
- bgp_announce_table()
* bgpd/bgp_table.h:
- struct bgp_table: Added reference count. Changed type of owner to be
"struct peer *" rather than "void *".
- struct bgp_node: Correct reference counter "lock" to be signed.
* bgpd/bgp_table.c:
- Added bgp_table reference counting.
- bgp_table_free(): Fixed cleanup code. Call peer_unlock() on owner if
set.
- bgp_unlock_node(): Added assertion.
- bgp_node_get(): Added call to bgp_lock_node() to code path that it was
missing from.
* bgpd/bgp_vty.c:
- peer_rsclient_set_vty(): Call peer_lock() as part of peer assignment
to owner. Handle failure gracefully.
- peer_rsclient_unset_vty(): Add call to bgp_clear_route() with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
* bgpd/bgp_zebra.c: Made it so zclient is global so bgp_exit() can use it.
* bgpd/bgpd.c:
- peer_lock(): Allow to be called when status is "Deleted".
- peer_deactivate(): Supply BGP_CLEAR_ROUTE_NORMAL purpose to
bgp_clear_route() call.
- peer_delete(): Common variable listnode pn. Fix bug in which rsclient
was only dealt with if not part of a peer group. Call
bgp_clear_route() for rsclient, if appropriate, and do so with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
- peer_group_get(): Use XSTRDUP() instead of strdup() for conf->host.
- peer_group_bind(): Call bgp_clear_route() for rsclient, and do so with
BGP_CLEAR_ROUTE_MY_RSCLIENT purpose.
- bgp_create(): Use XSTRDUP() instead of strdup() for peer_self->host.
- bgp_delete(): Delete peers before groups, rather than after. And then
rather than deleting rsclients, verify that there are none at this
point.
- bgp_unlock(): Add assertion.
- bgp_free(): Call bgp_table_finish() rather than doing XFREE() itself.
* lib/command.c,h: Compiler warning fixes. Add cmd_terminate(). Fixed
massive leak in install_element() in which cmd_make_descvec() was being
called more than once for the same cmd->strvec/string/doc.
* lib/log.c: Make closezlog() check fp before calling fclose().
* lib/memory.c: Catch when alloc count goes negative by using signed
counts. Correct #endif comment. Add log_memstats_stderr().
* lib/memory.h: Add log_memstats_stderr().
* lib/thread.c: thread->funcname was being accessed in thread_call() after
it had been freed. Rearranged things so that thread_call() frees
funcname. Also made it so thread_master_free() cleans up cpu_record.
* lib/vty.c,h: Use global command_cr. Add vty_terminate().
* lib/zclient.c,h: Re-enable zclient_free().
* lib/sockunion.c: (sockunion_normalise_mapped) The code to normalize
address was not copying port value - probably reason why IPV4 in IPV6
never worked right.
* bgpd/bgp_network.c: Fix MD5 listen in IPv4 version of bgp_socket() by
adding listen socket to listen_sockets list so that MD5 passwords can
get set.
* lib/sockopt.c: (sockopt_tcp_signature) Fix bogus "% Error while applying
TCP-Sig to session(s)" / "can't set TCP_MD5SIG option" startup error
messages by not returning error when there isn't one.
* configure.ac: Check for the actual CLOCK_MONOTONIC symbol, as clock_gettime
need not imply that clock type is supported
* lib/zebra.h: use the method given in autoconf docs for sys/time.h
inclusion
* configure.ac: Add a --enable-linux24-tcp-md5 argument, to compile in
support for the old TCP-MD5 patch for Linux 2.4 systems. This overrides
auto-detection of TCP-MD5 supported by the target system.
* lib/sockopt.c: (sockopt_tcp_signature) add in a variant for the old
Linux 2.4, TCP_MD5_AUTH variant of TCP-MD5 support, conditional
on the previous configure arg.
