This code provided inter-VM (2 cores per VM) throughput of
22Gbps using iperf through VPP (1 core) with 9k frames.
With the same setup and pktgen running on both sides, it
reached 5Mpps with no packets drop (Equivalent to before the patch).
During the tests the average vector length was about 1, which
likely means that VPP is not the bottleneck.
The patch also includes some generic functions for vlib buffers
allowing for chained buffer construction whether or not DPDK is enabled.
Change-Id: Icfd1803e84b2b4578f305ab730576211f6242d6a
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
It also includes check to ensure that number of
per-cpu mheaps is not lower than number of cpus.
Change-Id: Ibc68b34dda130f922243f9ea15b03e44bbcac269
Signed-off-by: Damjan Marion <damarion@cisco.com>
vec_sort macro was using gcc proprietary nested functions that
require a executable stack and they are considered as unsafe.
Also, nested functions are not supported by other compilers.
vec_sort_with_function() should be used instead.
Change-Id: I05959da63d222ec71c090ba63420b427ce10c79b
Signed-off-by: Damjan Marion <damarion@cisco.com>
This patch alters behavior for dpdk rx packets. Depending on test
results, it may be necessary to extend the scheme to packets received
on non-dpdk paravirtualized interfaces, and packets originating in the
vpp stack itself.
Change-Id: I8444232a90ff176e7d6a688e36801174575251a1
Signed-off-by: Dave Barach <dave@barachs.net>
Maximum number of TX queues ca be defined by NIC driver
or configured manualy with 'dpdk { max-tx-queues X }'.
If system have more worker threads than TX queues they will
be shared between them. Before this change only one tx
queue was used in such cases.
Change-Id: Iab68170ab45fa6b9925fc4e79ccab9222f095e7e
Signed-off-by: Damjan Marion <damarion@cisco.com>
The fix tests if the ARP table exist before attempting to
interate over the entries.
Change-Id: I22cad5ba3a01c3fba1feac33745a29ff1f7f8fe1
Signed-off-by: Keith Wiles <keith.wiles@intel.com>
Ensure that some additional rte_* symbols (from the dpdk library) and
pulled into the vpp image so that they will be available to plugins.
Also, provide vlib's dpdk configuration function so that plugins that
need to make dpdk calls during initialization/configuration may ensure
that dpdk itself is initialized first.
Change-Id: Iff8b518c6fc23e9e2f24a85944e5b246c80e5f67
Signed-off-by: Josh Gahm <jgahm@cisco.com>
GetStringUTFChars() can fail, returning NULL. Make sure we do not trip
over it.
Change-Id: I2d6b9c72c353c2423042fd035087f0d57fdc08ca
Signed-off-by: Robert Varga <nite@hq.sk>
Allocation of the integer array may fail, guard against NPE.
Change-Id: I3e860153597638392c9b669a6ad586fd2e03a64d
Signed-off-by: Robert Varga <nite@hq.sk>
Pure cleanup of indentation and brace placement, so the file ends up
being consisteny.
Change-Id: Idd2f20deda486d16b455f3e13cacbc2f3baa50f1
Signed-off-by: Robert Varga <nite@hq.sk>
The symbol is not referenced outside of the source file, hide it from
the world.
Change-Id: Id5db73efff0767657ebc5a0b25dcccd2b85e354d
Signed-off-by: Robert Varga <nite@hq.sk>
Adding hidden attribute results in compilation failure if the symbol is
not found when linking the .so. It should also lead to better
performance, as it side-steps GOT.
Change-Id: I7b6f39e663ca2b3e432669a3e2b36d7395f555b6
Signed-off-by: Robert Varga <nite@hq.sk>
There is skew between symbol names of vppjni_env.[ch], causing the
library to fail to load. This patch fixes that up.
Change-Id: I972e6d0ce72fd05ee3518e7161e20946ff3426ca
Signed-off-by: Robert Varga <nite@hq.sk>
This can cause packets to be mishandled in systems with more than
one type of driver in use.
Change-Id: I73ae91ff5d7a22fa542c3b39f6ccfe50437ec71b
Signed-off-by: Todd Foggoa <tfoggoa@cisco.com>
Some applications may wish to define their own implentation of
usec delay other than the existing blocking one. The default
behavior remains unchanged.
Change-Id: I699b2ed460763492502d22b7e8532514e0886ecd
Signed-off-by: Todd Foggoa <tfoggoa@cisco.com>
does not close the socket. Resulting in the main thread being stuck
in a tight infinite loop polling on the erronous socket.
Change-Id: I630b84b97c059acce117d56e41cd201131db4cab
Signed-off-by: Ole Troan <ot@cisco.com>
The DPDK glue did not support cloned packets which do not
have a freelist handler. Add support for this case.
Change-Id: I8f17cd4952df97989d90d3f3e39792bc3739705c
Signed-off-by: Kevin Paul Herbert <kph@cisco.com>
This symbol is defined in the math library, and is a compiler
built-in. Plug-ins that need it should link with -lm.
Change-Id: I967c87ccea0e762b68ca59cfaa460f3d6e534a59
Signed-off-by: Kevin Paul Herbert <kph@cisco.com>
build-root/vagrant/Vagrantfile
was always mounting the vpp into /vpp
Now rather than cloning it and building,
we just use it as mounted.
In order to let folks know what happened,
a README.moved is copied into the ~/git/vpp
so folks know what happened.
In addition to make it easier for folks
to do commits from withing the vagrant,
we install git-review, and copy in the
users .gitconfig and .gnupg directory.
A couple of notes about this. VMWare goes much
much faster in all cases. Virtualbox is a
bit slower in the very first run (without ccache).
One of the benefits of using the mounted /vpp though
is that after your first vagrant up, you always
have access to the .ccache, as it lives
outside the vagrant, and so in steady state
everything is faster.
Change-Id: I2cd2c28181b3d7e664240dfe2249b5be3f1b9241
Signed-off-by: Ed Warnicke <eaw@cisco.com>
Limit buffer tracing to 50 in order to limit large output, unless
the user over rides the max "sh trace max <number>".
Add trace filtering, to be able to only trace packets that were
processed by a specific node or exclude packets processed by a node.
Example, only include packets processed by error-drop:
# trace filter include error-drop 1
# trace add dpdk-input 1000000
<wait for packets, to come in>
# show trace
Change-Id: I5d9e15d2268ea55e6ef87b2b8756049c49b2791b
Signed-off-by: Todd Foggoa <tfoggoa@cisco.com>
