Memory unmapping causes slow address space fragmentation, leading in
extreme cases to failing to allocate pages at all. Removing this problem
by keeping all the pages allocated to us, yet calling madvise() to let
kernel dispose of them.
This adds a little complexity and overhead as we have to keep the
pointers to the free pages, therefore to hold e.g. 1 GB of 4K pages with
8B pointers, we have to store 2 MB of data.
While onlink flag is meaningful only with explicit next hops, it can be
defined also on direct routes. Parse it also in this case to avoid
periodic updates of the same route.
Thanks to Marcin Saklak for the bugreport.
Changes in commit eb937358 broke setting of channel preference for alien
routes learned during scan. The preference was set only for async routes.
Move common attribute processing part of functions krt_learn_async() and
krt_learn_async() to a separate function to have only one place for such
changes.
Seems like the previous patch was too optimistic, as route replace is
still broken even in Linux 4.19 LTS (but fixed in Linux 5.10 LTS) for:
ip route add 2001:db8::/32 via fe80::1 dev eth0
ip route replace 2001:db8::/32 dev eth0
It ends with two routes instead of just the second.
The issue is limited to direct and special type (e.g. unreachable)
routes, the patch restricts route replace for cases when the new route
is a regular route (with a next hop address).
When IPv6 ECMP support first appeared in Linux kernel, it used different
API than IPv4 ECMP. Individual next hops were updated and announced
separately, instead of using RTA_MULTIPATH as in IPv4. This has several
drawbacks and requires complex code to merge received notifications to
one multipath route.
When Linux came with IPv6 RTA_MULTIPATH support, the initial versions
were somewhat buggy, so we kept using the old API for updates (splitting
multipath routes to sequences of route updates), while accepting both
old-style routes and RTA_MULTIPATH routes in scans / notifications.
As IPv6 RTA_MULTIPATH support is here for a long time, this patch fully
switches Netlink to the IPv6 RTA_MULTIPATH API and removes old complex
code for handling individual next hop announces.
The required Linux version is at least 4.11 for reliable operation.
Thanks to Daniel Gröber for the original patch.
Remove compile-time sysdep option CONFIG_ALL_TABLES_AT_ONCE, replace it
with runtime ability to run either separate table scans or shared scan.
On Linux, use separate table scans by default when the netlink socket
option NETLINK_GET_STRICT_CHK is available, but retreat to shared scan
when it fails.
Running separate table scans has advantages where some routing tables are
managed independently, e.g. when multiple routing daemons are running on
the same machine, as kernel routing table modification performance is
significantly reduced when the table is modified while it is being
scanned.
Thanks Daniel Gröber for the original patch and Toke Høiland-Jørgensen
for suggestions.
Passing protocol to preexport was in fact a historical relic from the
old times when channels weren't a thing. Refactoring that to match
current extensibility needs.
When BIRD was munmapping too many pages, it sometimes aborted, saying
that munmap failed with "Not enough memory" as the address space was
getting more and more fragmented.
There is a workaround in place, simply keeping that page for future use,
yet it has never been compiled in because I somehow forgot to include
errno.h. And because I also thought that somebody may have ENOMEM not
defined (why?!), there was a check which quietly omitted that
workaround.
Anyway, ENOMEM is POSIX. It's an utter nonsense to check for its
existence. If it doesn't exist, something is broken.
Add option 'netlink rx buffer' to specify netlink socket receive buffer
size. Uses SO_RCVBUFFORCE, so it can override rmem_max limit.
Thanks to Trisha Biswas and Michal for the original patches.
Add strict checking for netlink KRT dumps to avoid PMTU cache records
from FNHE table dump along with KRT.
Linux Kernel added FNHE table dump to the netlink API in patch:
8d3b68cd37.1561131177.git.sbrivio@redhat.com/
Therefore, since Linux 5.3 these route cache entries are dumped together
with regular routes during periodic KRT scans, which in some cases may be
huge amount of useless data. This can be avoided by using strict checking
for netlink dumps:
https://lore.kernel.org/netdev/20181008031644.15989-1-dsahern@kernel.org/
The patch mitigates the risk of receiving unknown and potentially large
number of FNHE records that would block BIRD I/O in each sync. There is a
known issue caused by the GRE tunnels on Linux that seems to be creating
one FNHE record for each destination IP address that is routed through
the tunnel, even when the PMTU equals to GRE interface MTU.
Thanks to Tomas Hlavacek for the original patch.
Kernel uses cloned routes to keep route cache entries, but reports them
together with regular routes. They were skipped implicitly as they
do not have rtm_protocol filled. Add explicit check for cloned flag
and skip such routes explicitly.
Also, improve debug logs of skipped routes.
Add option to socket interface for nonlocal binding, i.e. binding to an
IP address that is not present on interfaces. This behaviour is enabled
when SKF_FREEBIND socket flag is set. For Linux systems, it is
implemented by IP_FREEBIND socket flag.
Minor changes done by commiter.
Currently, BIRD ignores dead routes to consider them absent. But it also
ignores its own routes and thus it can not correctly manage such routes
in some cases. This patch makes an exception for routes with proto bird
when ignoring dead routes, so they can be properly updated or removed.
Thanks to Alexander Zubkov for the original patch.
The BSD kernel does not support the onlink flag and BIRD does not use
direct routes for next hop validation, instead depends on interface
address ranges. We would like to handle PtMP cases with only host
addresses configured, like:
ifconfig wg0 192.168.0.10/32
route add 192.168.0.4 -iface wg0
route add 192.168.0.8 -iface wg0
To accept BIRD routes with onlink next-hop, like:
route 192.168.42.0/24 via 192.168.0.4%wg0 onlink
BIRD would dismiss the route when receiving from the kernel, as the
next-hop 192.168.0.4 is not part of any interface subnet and onlink
flag is not kept by the BSD kernel.
The commit fixes this by assuming that for routes received from the
kernel, any next-hop is onlink on ifaces with only host addresses.
Thanks to Stefan Haller for the original patch.
We can also quite simply allocate bigger blocks. Anyway, we need these
blocks to be aligned to their size which needs one mmap() two times
bigger and then two munmap()s returning the unaligned parts.
The user can specify -B <N> on startup when <N> is the exponent of 2,
setting the block size to 2^N. On most systems, N is 12, anyway if you
know that your configuration is going to eat gigabytes of RAM, you are
almost forced to raise your block size as you may easily get into memory
fragmentation issues or you have to raise your maximum mapping count,
e.g. "sysctl vm.max_map_count=(number)".
Add a wrapper function in sysdep to get random bytes, and required checks
in configure.ac to select how to do it. The configure script tries, in
order, getrandom(), getentropy() and reading from /dev/urandom.
The BSD code did not propagate the OS-level IFF_MULTICAST flag to the
Bird-internal IF_MULTICAST flag, which causes problems with Wireguard
interfaces on FreeBSD. The Linux sysdep code does propagate the flag
already, so just copy over the same check and flag update.
For logging purposes a stack allocated net_addr struct was passed by
value as vararg (instead of the expected pointer). This resulted in
a segfault when the specific error condition got logged.
From now, there are no auxiliary pointers stored in the free slab nodes.
This led to strange debugging problems if use-after-free happened in
slab-allocated structures, especially if the structure's first member is
a next pointer.
This also reduces the memory needed by 1 pointer per allocated object.
OTOH, we now rely on pages being aligned to their size's multiple, which
is quite common anyway.
In general, events are code handling some some condition, which is
scheduled when such condition happened and executed independently from
I/O loop. Work-events are a subgroup of events that are scheduled
repeatedly until some (often significant) work is done (e.g. feeding
routes to protocol). All scheduled events are executed during each
I/O loop iteration.
Separate work-events from regular events to a separate queue and
rate limit their execution to a fixed number per I/O loop iteration.
That should prevent excess latency when many work-events are
scheduled at one time (e.g. simultaneous reload of many BGP sessions).
This is an implementation of draft-walton-bgp-hostname-capability-02.
It is implemented since quite some time for FRR and in datacenter, this
gives a nice output to avoid using IP addresses.
It is disabled by default. The hostname is retrieved from uname(2) and
can be overriden with "hostname" option. The domain name is never set
nor displayed.
Minor changes by committer.
With net.ipv4.conf.XXX.ignore_routes_with_linkdown sysctl, a user can
ensure the kernel does not use a route whose target interface is down.
Such route is marked with a 'dead' / RTNH_F_DEAD flag.
Ignore these routes or multipath nexthops during scan.
Thanks to Vincent Bernat for the original patch.
So one can define kernel protocol template without channels.
For other protocols, it is either irrelevant or already done.
Thanks to Clemens Schrimpe for the bugreport.
For ECMP routes, RTA_FLOW attribute must be set per-nexthop, not
per-route. Our corresponding krt_realm attribute is per-route.
Thanks to Mikhail Petrov for the bugreport.
The log subsystem should be locked earlier, as default_log_list() may
internally manipulate with the current_log_list (if it is also a default
log list).
The static logging structures are reused, we need to reinitialize them
otherwise add_tail() would fail in debug build. Reinitializing these
structures should be fine as the list they belong to is being
reinitialized on entry to the very same function.
Thanks to Andreas Rammhold and Mikael Magnusson for patches.
Add support for RTA_MULTIPATH attribute parsing for AF_MPLS routes.
BIRD is capable of installing a multipath route into kernel on Linux,
but it would not be seen because parsing fails. This made BIRD attempt
to install the same route repeatedly.
(The patch minorly updated by committer)
This is a quick workaround for an issue where configured logfiles are
opened/created during parsing of a config file even when parse-and-exit
option is active. We should later refactor the logging code to avoid
opening log during parsing altogether.
When dynamic BGP with remote range is configured, MD5SIG needs to use
newer socket option (TCP_MD5SIG_EXT) to specify remote addres range for
listening socket.
Thanks to Adam Kułagowski for the suggestion.
The old code stored route verdicts and temporary routes directly in
rtable. The new code do not store received routes (it immediately
compares them with exported routes and resolves conflicts) and uses
internal bitmap to keep track of which routes were received and which
needs to be reinstalled.
By not putting 'invalid' temporary routes to rtable, we keep rtable
in consistent state, therefore scan no longer needs to be atomic
operation and could be splitted to multiple events.
Use a hierarchical bitmap in a routing table to assign ids to routes, and
then use bitmaps (indexed by route id) in channels to keep track whether
routes were exported. This avoids unreliable and inefficient re-evaluation
of filters for old routes in order to determine whether they were exported.
Accept RTA_VIA attribute in all cases. The old code always used
RTA_GATEWAY for IPv4 / IPv6 and RTA_VIA for MPLS. The new code uses
RTA_VIA in cases where AF of network and AF of nexthop differs.
Names read from texfiles in /etc/iproute2/* are normalized by replacing
non-alphanumeric chars with underscore. The patch fixes handling of
uppercase letters, which were handled as non-alphanumberic.
Thanks to Igor Gavrilov for the bugreport.
The C11 specification allows only sig_atomic_t and _Atomic variable
access. All other accesses to global variables are undefined behavior.
Using int was probably OK on x86 and x86_64; yet there were some reports
from other architectures (especially some MIPS) that in rare cases,
after issuing SIGHUP, BIRD did strange things.
The command initiating planned graceful restart including bird shutdown
should be called 'graceful restart' instead of 'graceful down', as the
later should be reserved for graceful shutdown in style of RFC 8326.
