From discussion we decided that we should use ETH_ALEN instead
of ETHER_ADDR_LEN. Add ETH_ALEN to prefix.h and make
ETHER_ADDR_LEN generate a warning when used( but still work ).
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Since we were only setting vty->wfd in config_write, vty->fd would
remain 0 and vty_close() wouldn't close vty->wfd.
Clean up the entire fd closing and make it more explicit. We were even
trying to write to stdin...
Reported-by: Jorge Boncompte <jbonor@gmail.com>
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
BUILT_SOURCES doesn't do what the name suggests. What it actually means
is "these files should be built first when doing a 'make' without
explicit target" (or "make all").
It's pretty much almost always wrong to use BUILT_SOURCES, the only
correct use is when a file is needed by an unspecified / large set of
files.
Also remove version.h and route_types.h from dist tarball while we're at
it. configure will create them anyway.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Oops, forgot this path... in the --terminal case, stdio is closed when
the user ends the terminal session, but without terminal it was left
open.
(This caused a ssh session hang in the CentOS6 CI because the file
descriptors were still open, so ssh would keep the session alive...)
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Now that the logging hole is plugged, we can just print config-loading
errors to the log. This has 2 hidden advantages:
- vty_read_config calls in SIGHUP don't print errors to /dev/null
- errors are consistently printed to syslog on --enable-cumulus
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
zlog_* doesn't work in startup before we've loaded the real logging
configuration. Add some code to log to stderr for that window of time.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
If the paths for pid or vty don't exist, try creating them. Failure is
ignored (on EEXIST) or prints a non-fatal warning (other errors).
Fixes: #507
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This splits off privs_preinit(), which does the lookups for user and
group IDs. This is so the init code can create state directories while
still running as root.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The OpenBSD's cp(1) command doesn't support the -v option. This will
change in the next releases (starting from v6.2) but this patch fixes
the problem for v6.1 and older releases.
Fixes Issue #875.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
- SIGTSTP appropriately suspends the foreground terminal
- SIGINT causes the daemon to exit, regardless of -d
- SIGQUIT causes the daemon to daemonize, regardless of -d
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Block the parent process until the child has reached the main loop, e.g.
full service is available.
This means it's no longer neccessary to add a "safety sleep" for daemon
cross-dependencies, when using the -d startup option. This doesn't help
if -d isn't used.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Linux -> TCP_CORK is enabled by default
*BSD -> The equivalent is TCP_NOPUSH
As such sockopt_cork is effectively useless
especially since the two places it is used
do it right before TCP writes then disable
it right after the writes are over.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Switch statements are more elegant (and potentially faster... but that's
not the main motivation).
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
SAFI values have been a major source of confusion over the last few
years. That's because each SAFI needs to be represented in two different
ways:
* IANA's value used to send/receive packets over the network;
* Internal value used for array indexing.
In the second case, defining reserved values makes no sense because we
don't want to index SAFIs that simply don't exist. The sole purpose of
the internal SAFI values is to remove the gaps we have among the IANA
values, which would represent wasted memory in C arrays. With that said,
remove these reserved SAFIs to avoid further confusion in the future.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
swpX peers all start out with the same sockunion so initially they all
go into the same hash bucket. Once IPv6 ND has worked its magic they
will have different sockunions and will go in different buckets...life
is good.
Until then though, we are in a phase where all swpX peers have the same
socknunion. Once we have HASH_THRESHOLD (10) swpX peers and call
hash_get for a new swpX peer the hash code calls hash_expand(). This
happens because there are more than HASH_THRESHOLD entries in a single
bucket so the logic is "expand the hash to spread things out"...in our
case expanding doesn't spread out the swpX peers because all of their
sockunions are the same.
I looked at having peer_hash_make and peer_hash_same consider the ifname
of the swpX peer but that is a large change that we don't want to make
at the moment. So the fix is to put a cap on how large we are
willing to let the hash table get. By default there is no limit but if
max_size is set we will not allow the hash to expand above that.
Adds an array of descriptive names for each CLI node, plus a runtime
check to make sure folks don't forget to update it.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Commit 8f942af90 introduced a bug while silencing a clang warning. Silence
the warning in a different way to fix our red-black tree implementation.
Fixes#841.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
We should preserve the original indentation to make it easier to keep
these files in sync with the upstream.
Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
The pimregX devices when created by the kernel are put into
the default vrf. When pim gets the callback that the device
exists, check to see if it is a pimregX device and if so
move it into the appropriate vrf.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
This reverts commit c14777c6bf.
clang 5 is not widely available enough for people to indent with. This
is particularly problematic when rebasing/adjusting branches.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Before the fix NHT had each path resolving via swp1
cel-redxp-10# show ip route 20.0.11.253
Routing entry for 20.0.11.0/24
Known via "bgp", distance 20, metric 0, best
Last update 00:00:20 ago
* 169.254.0.1, via swp1
* 169.254.0.17, via swp2
cel-redxp-10#
cel-redxp-10# show ip nht
[snip]
20.0.11.253
resolved via bgp
via 169.254.0.1, swp1
via 169.254.0.1, swp1
Client list: pim(fd 19)
After the fix
cel-redxp-10# show ip nht
[snip]
20.0.11.253
resolved via bgp
via 169.254.0.1, swp1
via 169.254.0.17, swp2
Client list: pim(fd 19)
A {foo|bar|baz} graph node will have more than 1/2 incoming links even
if no other references are left to it (which is what the assert was
previously trying to ensure.)
I don't see a good way to "fix" the assert so I'm just removing it.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
The label initializer & nhrpd variable are just to shut up GCC 7,
the other two are actual bugs.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
This allows frr-reload.py (or anything else that scripts via vtysh)
to know if the vtysh command worked or hit an error.
Instead of having an ?: expression embedded in every single caller of
vty_out, just expand \n to \r\n in the vty code if neccessary.
(Deprecation warnings will be enabled in the next commits which will do
the search-and-replace over the codebase.)
[This reverts commit 4d5f445 "lib: add vty_outln()"]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Pretty-prints variable autocompletions by breaking them up into multiple
lines, indenting them consistently and respecting the column width of
the terminal.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Implement configuration options for EVPN. The configuration options include
VNI configuration with RD and Import and Export Route Targets. Also, display
the EVPN configuration.
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Implement handling of MACs and Neighbors (ARP/ND entries) in zebra:
- MAC and Neighbor database handlers
- Read MACs and Neighbors from the kernel, when needed and create
entries in zebra's MAC and Neighbor databases.
- Handle add/update/delete notifications from the kernel for MACs and
Neighbors and update zebra's database appropriately
- Inform locally learnt MACs and Neighbors to client
- Handle MACIP add/delete from client and install appriporiate entries
into the kernel
- Since Neighbor entries will be installed on an SVI, implement the
needed mappings
NOTE: kernel interface is only implemented for Linux/netlink
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Implement fundamental handling for VNIs and VTEPs:
- Handle EVPN enable/disable by client (advertise-all-vni)
- Create/update/delete VNIs based on VxLAN interface events and inform
client
- Handle VTEP add/delete from client and install into kernel
- New debug command for VxLAN/EVPN
- kernel interface (Linux/netlink only)
Signed-off-by: Vivek Venkatraman <vivek@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Some platforms don't support 64-bit atomics, missed converting a
floating point pow() to an integral mul when changing SD algo.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Signed-off-by: Daniel Walton <dwalton@cumulusnetworks.com>
Reviewed-by: Donald Sharp <sharpd@cumulusnetworks.com>
Restoring some code that was unintentionally removed when we were
migrating to Quentin's parser.
This shows up as still reachable allocation on exit, which is somewhat
misleading. Note this only affects --enable-dev-build.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
"$(top_srcdir)" is not on the include path, but "$(top_srcdir)/lib" is.
This is relevant when building with a separate build directory.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
hash_cmd_init will overwrite _hashes with a new list, while _hashes
already has been initialised from cmd_init(), thread_master_create(), or
any other function that may have created a hash.
Found while valgrind'ing ospf6d/test_lsdb.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Most read accesses of route_table are actually exact matches where
walking down the tree is wildly inefficient. Use a parallel hash
structure instead.
This significantly speeds up processes that are performance-bound by
table accesses, e.g. BGP withdraw processing. In other locations, the
improvement is not seen as strongly, e.g. when filter processing is the
limiting factor.
[includes fix to ignore prefix host bits in hash comparison]
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
route_node->parent and route_node->link shouldn't be touched by user
code since that is a recipe for trouble once we have a hash table in
parallel.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
This allows passing struct prefix_{ipv4,ipv6,evpn} * in addition to
struct prefix * without an extra cast (since the union uses the gcc
transparent-union extension present in all compilers that we support.)
Also applies some "const" while we're at it.
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
Consuming va_args modifies its internal bits, hence the need to copy
it... but the copying wasn't quite right just yet.
Fixes: 4d5f445 ("lib: add vty_outln()")
Signed-off-by: David Lamparter <equinox@opensourcerouting.org>
NetBSD can't take the square root of a long double
and we should be fine just using a double here anyway
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Adds the ability to name hash tables, and a new cli command that will
show various summary statistics for named hash tables.
Statistics computed are
- load factor
- full load factor (see comments)
- stddev of full load factor
Standard deviation is computed by storing the sum of squares of bucket
lengths. This is somewhat susceptible to overflow. On platforms where a
double is 32 bits, placing 65535 or more elements into a hash table
opens up the potential for overflow, depending on how they are arranged
in buckets (which depends on the hash function). For example, placing
65535 elements into one hash bucket would cause ssq overflow, but
distributing 40000000 elements evenly among 400000 buckets (100 elements
per bucket) would not.
These cases are extremely degenerate, so the vague possibility of
overflow in an informational command is deemed an acceptable tradeoff
for constant time calculation of variance without locks or compromising
efficiency of actual table operations.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Adds a function that calculates various statistics on our implementation
of a hash table. These are useful for evaluating performance.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
* Fix bad format specifier in thread.[ch]
* Move PRINTF_ATTRIBUTE macro to zebra.h
* Use PRINTF_ATTRIBUTE on termtable printers
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
Add support for naming pthreads. Also, note that we don't have any
records yet if that's the case.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
This patch fixes up show thread commands so that they know about
and operate on all extant thread_masters, since we can now have multiple
running in any given application.
This change also eliminates a heap use after free that appears when
using a single cpu_record shared among multiple threads. Since struct
thread's have pointers to bits of memory that are freed when the global
statistics hash table is freed, later accesses are invalid. By moving
the stats hash to be unique to each thread_master this problem is
sidestepped.
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
json-c does not (yet) offer support for unsigned integer types, and
furthermore, the docs state that all integers are stored internally as
64-bit. So there's never a case in which we would want to limit,
implicitly or otherwise, the range of an integer when adding it to a
json object.
Among other things this fixes the display of ASN values greater than
(1/2) * (2^32 - 1)
Signed-off-by: Quentin Young <qlyoung@cumulusnetworks.com>
The if_update function was taking the interface name as
input and reapplying it, using strncpy to reapply the name.
This has several issues. strncpy should not be used
to copy memory in place. The second issue is that
the interface name is not actually changing when we
update interface to be in the new vrf.
Since every usage of if_update was just reapplying the same
name the interface actually had, just remove that part of
the function and rename it to if_update_to_new_vrf
to represent what it is actually doing.
Signed-off-by: Donald Sharp <sharpd@cumulusnetworks.com>
Issue reported that a configuration commonly used on other routing implementations
fails in frr. If under ospf, "network 172.16.1.1/32 area 0" or under eigrp, "network
172.16.1.1/32" is entered, the appropriate interfaces are not included in the routing
protocol. This was because the code was calling prefix_match, which did not match if
the network statement had a longer mask than the interface being matched. This fix
takes away that restriction by creating a "lib/prefix_match_network_statement" function
which doesn't care about the mask of the interface. Manual testing shows both ospf and
eigrp now can be defined with more specific network statements.
Signed-off-by: Don Slice <dslice@cumulusnetworks.com>
