summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2013-01-07ospfd: compile warning cleanupsAndrew Certain
A set of patches to clarify some comments as well as cleanup code that was causing warnings. After these patches, the code can be compiled with -Wall -Wsign-compare -Wpointer-arith -Wbad-function-cast -Wwrite-strings -Wmissing-prototypes -Wmissing-declarations -Wchar-subscripts -Wcast-qual -Wextra -Wno-unused-parameter -Wno-missing-field-initializers (what is current in trunk plus -Wextra -Wno-unused-parameter -Wno-missing-field-initializers). Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: Update nsm_change_state to static scope, as it is not called from ↵Andrew Certain
elsewhere Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: Changed TE instance check to remove -Wtype-limits warningAndrew Certain
Since LEGAL_TE_INSTANCE_RANGE() was being passed an unsigned int, a warning was being thrown due to the compare against >= 0. Since this macro was used only in one place, I removed the macro for an explict compare against a constant for the MAX. Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07lib: Changes to VTY string-parsing macros to remove warningsAndrew Certain
The VTY_GET_INTEGER_RANGE macro was being used also just to check the range on a variable that wasn't used (for the "no" version of a VTY command), so I split the macro into two. Also, since the variable is unsigned, if MIN is zero, you get a warning about comparing an unsigned number against 0, giving rise to slightly convoluted logic. Note that the previous two patches were found by the -Wtype-limits and -Wunused-variables warnings. Without the changes to these macros, these warnings are triggered erroneously, making it harder to find the real problems. Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: Update comments to be more clear in packet processingAndrew Certain
Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: Fixed signed/unsigned masking of negative metricsAndrew Certain
In the original code, negative metrics would be converted successfully by atoi() and then converted to an unsigned int that would always compare successfully against >= 0, leaving a large positive metric in the route map. Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: Fixed typo bug in ospf_vty.h:ospf_neighborAndrew Certain
Typo bug. ospf_nbr_nbma_poll_interval_set() was being sent priority instead of interval. Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospfd: fix argument parsing for distribute-listChristian Franke
Use the correct argument for the protocol lookup in ospf distribute-list commands. Signed-off-by: Christian Franke <chris@opensourcerouting.org> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: fix apiserver enableDavid Lamparter
The ospf_apiserver_enable flag was being cleared _after_ the "-a" command-line option set it to 1. Move up the initialisation, so enabling the OSPF API is actually possible. Reported-by: Rosario Mattera <rosmattera@gmail.com> Signed-off-by: David Lamparter <equinox@opensourcerouting.org> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: Fix type-4 network mask to 0 per RFCLeonard Tracy
The OSPF RFC (2328) states that the network mask field of a type 4 LSA "is not meaningful and must be zero". OSPFD has been setting the mask as /32. This patch changes OSPFD to set the mask to 0 per the RFC Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: Reduce MaxAge log levelAyan Banerjee
Reduce the log level for the MaxAge LSA reception when such an LSA does not exist in the database. Signed-off-by: Ayan Banerjee <ayan@cumulusnetworks.com> Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com> Reviewed-by: Nolan Leake <nolan@cumulusnetworks.com> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: Convert MAX_AGE LSA list to treeDinesh Dutt
Store the MaxAge LSA list in a tree instead of a linked list for efficient access. Walking the list can be quite inefficient in some large systems and under certain tests. ospfd maintains the list of LSA's that have been MaxAge'd out in a separate linked list for removal by a remover/walker thread. When a new LSA is to be installed, the old LSA is ejected and when it is ejected, the MaxAge LSA list is traversed to ensure that the old LSA is also removed from this list if it exists on this list. When a large number (> 5K) MaxAge LSAs are bombarding the system, walking this list takes a significant time causing timers to fire and actions to be taken such as expiring neighbors due to expiry of DeadInterval (especially when timer is really low, <= 12s), creating a spiral of instability. By making this MaxAge LSA list be a tree, this problem is mitigated. Signed-off-by: Dinesh Dutt <ddutt@cumulusnetworks.com> Reviewed-by: Ayan Banerjee <ayan@cumulusnetworks.com> Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com> Reviewed-by: Shrijeet Mukherjee <shm@cumulusnetworks.com> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: forward ref. of areas for "max-metric router-lsa administrative" cmdAyan Banerjee
In the event areas are created at a later point of time with respect to the playback of the "max-metric router-lsa administrative" command, those areas do not get into indefinite max-metric mode. This patch is inteneded to store the configuration and apply it to all future areas that may be created. In the process, some other bugs that were there with respect to restart etc are fixed up. Tested locally to see that the fix works across multiple areas and across multiple restarts. Signed-off-by: Ayan Banerjee <ayan@cumulusnetworks.com> Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com> Reviewed-by: JR Rivers <jrrivers@cumulusnetworks.com> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2013-01-07ospf: suppress delete using replacementAyan Banerjee
After a SPF run, OSPF deletes routes that have changed in terms of any metric, type, and/or next-hops and re-adds them. Given that the Zebra-RIB already support replacement semantics, we suppress deletes for routes that will be added back again. This has the following advantages. It reduces the number of IPC messages between OSPF/Zebra. Also, in the current flow, a batch of route deletes were followed by a batch of adds even for say a metric change. With the change, routes are sent as "add" when they are modified. Zebra already implicitly deletes older routes. Signed-off-by: Ayan Banerjee <ayan@cumulusnetworks.com> Reviewed-by: Scott Feldman <sfeldma@cumulusnetworks.com> Reviewed-by: Dinesh Dutt <ddutt@cumulusnetworks.com> Signed-off-by: Scott Feldman <sfeldma@cumulusnetworks.com>
2012-12-13build: check for .git in configure.acDavid Lamparter
Don't error out when someone tries using --with-pkg-git-version on something that isn't actually a git checkout (like a dist tarball). Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-13build: fix dist tarballDavid Lamparter
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>
2012-12-13build: update ICC warning CFLAGSDavid Lamparter
Intel's icc doesn't accept "-wd <number>" anymore, it's "-wd<number>" these days. But, anyhow, the warnings disabled in Quagga's configure.ac don't seem to appear anywhere at all, so let's just remove the option completely. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: fix ipv6 metric endiannessNick Hilliard
the isis ipv6 reachability metric is transmitted in big endian / network format, but isis_spf_process_lsp() does not convert this into host endian format when mucking around with local cost + received metric. This patch fixes this problem and makes received ipv6 metrics work properly on little-endian machines. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: verify metrics on metric-style transitionChristian Franke
When switching to metric-style transition, circuit metrics should also be verified to be in the narrow range 0..63. Signed-off-by: Christian Franke <chris@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: fix metrics check for metric-style narrowChristian Franke
When switching to narrow metric style, all configured circuits are verified to have a valid narrow style metric. Check te_metric instead of metric_default as the latter is only 8bit wide and may overflow for wide style metrics. Signed-off-by: Christian Franke <chris@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: address Coverity warningsDavid Lamparter
this fixes a bunch of issues found by Coverity SCAN and flagged as "high" impact -- although, they're all rather minute issues. * isisd/isis_adjacency.c: one superfluous check, one possible NULL deref * isisd/isis_circuit.c: two prefix memory leaks * isisd/isis_csm.c: one missing break * isisd/isis_lsp.c: one possible NULL deref * isisd/isis_pfpacket.c: one error-case fd leak * isisd/isis_route.c: one isis_route_info memory leak * isisd/isis_routemap.c: one... fnord * isisd/isis_tlv.c: one infinite loop Reported-by: Coverity SCAN Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: always join all IS-IS multicast groupsDavid Lamparter
The socket is only created once when an interface is brought up, and the multicast groups were joined according to configuration at that point. This breaks when later switching an interface to another IS-IS level. Since, for a separate conformance issue (ANVL ISIS-6.4), we should be inspecting the destination address anyway, the simplest fix here is to just join all groups unconditionally. There shouldn't be much traffic on these anyway, worst case we might be picking up some unrelated multicast groups due to NIC filter aliasing though... Signed-off-by: David Lamparter <equinox@opensourcerouting.org> Tested-by: Martin Winter <mwinter@opensourcerouting.org>
2012-12-12isisd: save metric-style narrowChristian Franke
isisd defaults to wide metric style. So if narrow metric style is configured, a matching setting should be written to the configuration, allowing a narrow metric-style setting to be saved. Signed-off-by: Christian Franke <chris@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: fix spftree_area_del typo causing SEGVDavid Lamparter
spftree_area_del didn't clear the IPv6 L2 spftree due to a simple typo, leading to a SEGV on shutdown when the still-armed timer would try to run an IPv6 L2 SPF calculation with its data free'd already. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: drop hellos without supported protocol listDavid Lamparter
isisd should not form adjacencies on receiving an IS-IS Hello without a list of supported protocols (cf. RFC 1195 s4.4 p32 "Maintaining Router Adjacencies") Also fixes memleaks in these error cases. * isisd/isis_pdu.c: improve TLVFLAG_NLPID handling Signed-off-by: David Lamparter <equinox@opensourcerouting.org> Tested-by: Martin Winter <mwinter@opensourcerouting.org>
2012-12-12isisd: refuse adjacencies with our own system IDDavid Lamparter
isisd would form an adjacency with another router despite the system IDs being identical. This would later cause an assertion failure like this: assertion=0x555555596db8 "isis_find_vertex (spftree->paths, id, vtype) == ((void *)0)", file=0x555555596c60 "isis_spf.c", line=515, function=0x555555597900 "isis_spf_add2tent") at log.c:619 which is caused by trying to add a path expected to not exist, but suddenly colliding due to the duplicate system ID. * isis_pdu.c: check for system ID collision on receiving Hello Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-12-12isisd: do not add >63 IP addresses to helloDavid Lamparter
RFC1195 s4.2 "Multiple IP Addresses per Interface" explicitly forbids us from adding multiple tuples of IP addresses, putting a hard cutoff at 63 IP addresses. * isisd/isis_tlv.c: cut off (and return success) at 63 addrs. Signed-off-by: David Lamparter <equinox@opensourcerouting.org> Tested-by: Martin Winter <mwinter@opensourcerouting.org>
2012-12-12build: include git infoDavid Lamparter
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>
2012-11-30zebra: include intf when sending IPv4 nexthop to FPMAvneesh Sachdev
* zebra/zebra_fpm_netlink.c Change the zebra FPM code to include an interface index when encoding a nexthop even if the protocol only provided a gateway address (e.g, NEXTHOP_TYPE_IPV4). Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30doc: add blurbs on zebra FPM interface and commandsAvneesh Sachdev
Update documentation with some text on the zebra interface to the optional Forwarding Path Manager component, and the related cli commands. Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30zebra: add module to communicate routes to FPMAvneesh Sachdev
Enhance zebra to send routes to the (optional) Forwarding Path Manager component using the interface defined by fpm/fpm.h. * configure.ac - Add --enable-fpm flag. The FPM-related code in zebra is activated only if the build is configured with '--enable-fpm'. - Add HAVE_NETLINK automake conditional. This allows us to conditionally build netlink-dependent C code. * zebra/{rib.h,zebra_rib.c} - Add the 'fpm_q_entries' field to the rib_dest_t structure. This allows dests to be placed on the fpm queue. - Define a couple new rib_dest_t flags that hold FPM-related state. - Invoke the zfpm_trigger_update() function for a route_node whenever the information to be sent to the FPM changes. - rib_can_delete_dest(): Return FALSE if we have to update the FPM about the given dest. This ensures that the dest is not deleted even if there are no ribs hanging off of it. * zebra/zebra_fpm.c This file holds most of the code for interacting with the FPM. - If quagga was configured with '--enable-fpm', periodically try to connect to the FPM. - When the connection comes up, enqueue all relevent dests to the FPM queue. - When the FPM socket is readable, dequeue the next rib_dest_t from the FPM queue, encode it in to a message and send the message to the FPM. - When the connection to the FPM goes down, remove all dests from the FPM queue, and then start trying to connect to the FPM again. - Expose the following new operational commands: show zebra fpm stats clear zebra fpm stats * zebra/zebra_fpm_netlink.c - zfpm_netlink_encode_route(): Function to encode information about a rib_dest_t in netlink format. * zebra/zebra_fpm_private.h Private header file for the zebra FPM module. * zebra/zebra_fpm.h Header file exported by zebra FPM module to the rest of zebra. * zebra/debug.c Add the 'debug zebra fpm' command. * zebra/main.c Initialize the zebra-FPM code on startup. * zebra/misc_null.c Add stub for zfpm_trigger_update(). * zebra/Makefile.am - Include new file zebra_fpm.c in build. - Include zebra_fpm_netlink.c in build if HAVE_NETLINK is defined. * vtysh/Makefile.am Include zebra_fpm.c in list of files that define cli commands. Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30fpm: Add public header for Forwarding Plane ManagerAvneesh Sachdev
The Forwarding Plane Manager (FPM) is an optional component that may be used in scenarios where the router has a forwarding path that is distinct from the kernel, commonly a hardware-based fast path. It is responsible for programming forwarding information (such as routes and nexthops) in the fast path. In Quagga, the Routing Information Base is maintained in the 'zebra' infrastructure daemon. Routing protocols communicate their best routes to zebra, and zebra computes the best route across protocols for each prefix. This latter information comprises the bulk of the Forwarding Information Base. The new header file added by this patch, 'fpm/fpm.h', defines a point-to-point interface using which zebra can update the FPM about changes in routes. The communication takes place over a stream socket. The FPM listens on a well-known TCP port, and zebra initiates the connection. All messages sent over the connection start with a short 'FPM header'. In the case of route add/delete messages, the header is followed by a netlink message. Zebra should send a complete copy of the forwarding table(s) to the FPM, including routes that it may have picked up from the kernel. The FPM interface uses replace semantics. That is, if a 'route add' message for a prefix is followed by another 'route add' message, the information in the second message is complete by itself, and replaces the information sent in the first message. If the connection to the FPM goes down for some reason, the client (zebra) should send the FPM a complete copy of the forwarding table(s) when it reconnects. Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30lib: bring in sys/queue.h from FreeBSD treeAvneesh Sachdev
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>
2012-11-30zebra: extern/extract some functions from rt_netlink.cAvneesh Sachdev
* zebra/{rib.h,zebra_rib.c} Add nexthop_type_to_str(), which returns a human-readable string corresponding to a nexthop type. * zebra/rt_netlink.[hc] - Add new header file that exposes some existing and new netlink-related functions from rt_netlink.c to the rest of zebra. addattr32 addattr_l rta_addattr_l nl_msg_type_to_str (new) nl_rtproto_to_str (new) - Use nexthop_type_to_str() instead of the static array 'nexthop_types_desc'. Signed-off-by: Avneesh Sachdev <avneesh@opensourcerouting.org> Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30zebra: add iterator for walking all tables in RIBAvneesh Sachdev
* 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>
2012-11-30zebra: add way to determine VRF/AFI/SAFI of tableAvneesh Sachdev
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>
2012-11-30zebra: add structure to hold per-prefix state in RIBAvneesh Sachdev
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>
2012-11-30bgpd: document bgp neighbor local-as peer commandAndrew Certain
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30bgpd: Fixed out-of-date commentAndrew Certain
When going through the code to write the documentation for local-as, I discovered that one of the comments was out-of-date. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-30bgpd: add replace-as modifier for BGP neighborAndrew Certain
Added replace-as modifier for BGP neighbors when using local-as. If the replace-as modifier is specified, only the replacement AS as specified by the local-as modifier is prepended to the AS_PATH, not the process's AS. In bgp_attr.c, I decided that if (peer->change_local_as) { /* If replace-as is specified, we only use the change_local_as when advertising routes. */ if( ! CHECK_FLAG (peer->flags, PEER_FLAG_LOCAL_AS_REPLACE_AS) ) { aspath = aspath_add_seq (aspath, peer->local_as); } aspath = aspath_add_seq (aspath, peer->change_local_as); } else { aspath = aspath_add_seq (aspath, peer->local_as); } was clearer than the alternative that didn't duplicate the prepending of the process's AS: /* First, append the process local AS unless we have an alternate local_as * and we're replacing it (as opposed to just prepending it). */ if (! (peer->change_local_as && CHECK_FLAG (peer->flags, PEER_FLAG_LOCAL_AS_REPLACE_AS) ) ) { aspath = aspath_add_seq (aspath, peer->local_as); } if (peer->change_local_as) aspath = aspath_add_seq (aspath, peer->change_local_as); } But I could be convinced otherwise. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-05lib: Implement monotonically increasing clock for Darwin.Hasso Tepper
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>
2012-11-05isisd: clock_gettime() -> quagga_gettime() conversion.Hasso Tepper
* isisd/isis_spf.c: Use portable quagga_gettime() like the rest of the Quagga code. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-05ospfd: remove some buggy extra ';' symbols.Hasso Tepper
* ospfd/ospf_apiserver.c: extra ; causing lookup to fail always * ospfd/ospf_lsa.c: extra ; causing debug output even when disabled Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-05build: check actually-used BSD link state fieldsDoug VanLeuven
ifi_link_state missing in OS X. There could be other *BSD's that haven't implemented it and possibly affects older implementations. The existing HAVE_BSD_LINK_DETECT configure.ac check is only confirming the link state detection using ifmediareq.ifm_status found in <net/if_media.h>. This is the link state detection used in zebra/ioctl.c. Later, *BSD redefined struct if_data in <net/if.h> and included link state detection. This is the method used in zebra/kernel_socket.c Additional test defined in config.ac to test for member struct if_data.ifi_link_state defined in <net/if.h> separate from test for <net/if_media.h> ifmediareq.ifm_status Fixed #ifdef's in zebra/kernel_socket.c to use the new #define No impact on older function calls in zebra/ioctl.c Tested on 64bit OS X 10.7, FreeBSD 9.0 amd64 & i386 (32bit) using gcc & clang. Tested on linux 64bit. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-05zebra: kernel_socket: fix overflow in RTA_ADDR & RTA_ATTRDoug VanLeuven
In zebra/kernel_socket.c, copying sockaddr from *_msghdr: There are really 2 different lengths that need to be determined. 1) the length required to point to the next sockaddr in the mesg buffer which might include any required padding and 2) the actual length of the sockaddr data that needs to be copied into the destination field. They may or may not be the same value. Sizeof sockaddr_in6 is 28, which to pad for alignment purposes on 32 bit systems with a long of 4 bytes is evenly divided and requires no padding. On 64 bit systems, with a long of 8 it is padded with 4 extra bytes.So the current RTA_* macros are copying 32 bytes into a 28 byte field on 64 bitsystems, where the field overflow did not occur on the 32 bit systems. Since using sa_len required the use of an #ifdef which couldn't be used directly inside a #define, it made sense to move the copy into the function to allow typdef checking throughout and eliminate the hack to suppress compiler warnings. Fixed declaration of cp in ifm_read after compiler noticed type mismatch. Tested on 64bit OS X 10.7, FreeBSD 9.0 amd64 & i386 (32bit) using gcc & clang Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-05zebra: kernel_socket: fix 64bit MacOS X alignmentDoug VanLeuven
In OS X 10.7 zebra crashed on invalid execution address. sockaddr padding in *_msghdr is observed to be 4 bytes in 64bit OS X. The ROUNDUP macro assumed alignment on sizeof(long) which allocates 8 bytes on 64bit systems, 4 bytes on 32bit systems which is true for BSD generally. Test for Apple and use sizeof(int) which allocates 4 bytes on 32 & 64bit systems. Tested on 64bit OS X 10.7, FreeBSD 9.0 amd64 & i386 (32bit) using gcc & clang Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-03doc: update installation instructionsDavid Lamparter
configure parameters have changed quite a bit, several options are enabled by default now and there's --disable-tests. Update documentation to match. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-03build: compile tests/ by defaultDavid Lamparter
Broke the tests again... let's just build them by default so it's easier to notice. If anyone doesn't want to build tests, there's --disable-tests. NB: tests will be neither run nor installed. Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-03tests: fix missing array_size() includeDavid Lamparter
* tests/test-sig.c: add #include "lib/memory.h" to get array_size() Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
2012-11-01bgpd: CVE-2012-1820, DoS in bgp_capability_orf()Denis Ovsienko
An ORF (code 3) capability TLV is defined to contain exactly one AFI/SAFI block. Function bgp_capability_orf(), which parses ORF capability TLV, uses do-while cycle to call its helper function bgp_capability_orf_entry(), which actually processes the AFI/SAFI data block. The call is made at least once and repeated as long as the input buffer has enough data for the next call. The helper function, bgp_capability_orf_entry(), uses "Number of ORFs" field of the provided AFI/SAFI block to verify, if it fits the input buffer. However, the check is made based on the total length of the ORF TLV regardless of the data already consumed by the previous helper function call(s). This way, the check condition is only valid for the first AFI/SAFI block inside an ORF capability TLV. For the subsequent calls of the helper function, if any are made, the check condition may erroneously tell, that the current "Number of ORFs" field fits the buffer boundary, where in fact it does not. This makes it possible to trigger an assertion by feeding an OPEN message with a specially-crafted malformed ORF capability TLV. This commit fixes the vulnerability by making the implementation follow the spec.