In vnet/vnet/devices/dpdk/device.c, post 2 event data instead of
merging them into 1 u64.
Change-Id: I8b1d61b894279fb6eb57bb82a05affc14360e6b8
Signed-off-by: Christophe Fontaine <christophe.fontaine@qosmos.com>
On non numa aware platforms, the path "/sys/devices/system/node/node%u"
does not exist: the "free_hugepages" count is available in "/sys/kernel/mm/".
Change-Id: Id62018a2bd0937d267663699c620e89d02c26a72
Signed-off-by: Christophe Fontaine <christophe.fontaine@qosmos.com>
To better handle lines longer than the terminal width, re-work how
the pager collates lines. We still store each output line in a vector
or lines; additionally we index the start and end of displayable lines.
The index then becomes the data to work out what to display.
If the terminal is resized this index is rebuilt and the current page
redisplayed and an attempt made to keep the current top-of-screen line
in place.
Change-Id: Icc6f4f72dabad89682cd82be88adb03eadb7811d
Signed-off-by: Chris Luke <chrisy@flirble.org>
- Had issues with original patch and CSIT, breaking into
-- Makefile and API changes
-- Follow up patch removing NSH directory
Change-Id: I6c2f49a74d59b97ce4ea43799cc6f01f5cb1d5a2
Signed-off-by: Keith Burns (alagalah) <alagalah@gmail.com>
The classifying capabilities are pretty limited if only a single table can be used.
This patch adds support for chained classify tables to L2 Rewrite feature.
Change-Id: Ib678fe2028f17b441adb75d85bf3e31185e42c59
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
It appeared to me that it might be usefull
to allow users to create fibs or bridge domains without
a complete knowledge of the current used IDs.
These changes define fib and bridge domain constructors
when the provided ID is ~0. In such a case, an unused ID is
used to create a new fib or bridge domain.
Change-Id: Iaba69a023296e6d17bdde45980f9db84832a3995
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
This makes Doxygen think the API definitions are structs which is close
enough to be able to document the API methods.
It also has logic to create an indexed API page but that's disabled for
now because it duplicates the "brief" text twice in the struct doc.
Fixes a minor line numbering issue in filter_c.py.
Change-Id: If380160b73e7c10d999b35a76f55d0e27cbc91cc
Signed-off-by: Chris Luke <chrisy@flirble.org>
We now try to find the first map-resolver IP we have a route to and
decide the ip lookup node according to map-resolver IP.
Change-Id: I68fc12861dc1e9a203ebf6caefd4f053a741a785
Signed-off-by: Florin Coras <fcoras@cisco.com>
- Configures Doxygen.
- Adds a source filter to do magic on our use of the preprocessor to do
constructor stuff to make Doxygen grok it better.
- Adds a convenience helper to the root Makefile.
- Adds a README.md to the root directory (and which Doxygem uses as its
"mainpage".
- Add several other documentative files.
- Currently using SVG for call graphs, though this may have a
load-time performance impact in browsers.
Change-Id: I25fc6fb5bf634319dcb36a7f0e32031921c125ac
Signed-off-by: Chris Luke <chrisy@flirble.org>
- removed vnet/vnet/nsh-gre
- removed all nsh from vnet/vnet/nsh_vxlan_gpe to
vnet/vnet/nsh
- moved vnet/vnet/nsh_vxlan_gpe to vnet/vnet/vxlan_gpe
- added cli and binary api for VXLAN GPE tunnels
- plan to move vnet/vnet/nsh to new repo (sfc_nsh) and make plugin
- added cli for NSH (binary API will be done in sfc_nsh)
- vnet/vnet/gre will be extended in VPP-54
Change-Id: I1d27def916532321577ccd68cb982ae0d0a07e6f
Signed-off-by: Keith Burns (alagalah) <alagalah@gmail.com>
- check input variables
- in locator_set dump, add support for remote locator_set
Change-Id: Ib10028e83fead358f820ae45c71b6ca4dfbe2f1e
Signed-off-by: Andrej Kozemcak <akozemca@cisco.com>
When inserting routes into ip4/6 fibs, we first added a dummy adjacency
and afterwards manually updated its rewrite header to enable src/dst
forwarding. The downside to this is that the adj signature is changed
and therefore when deleting a route the adjacency signature is not
removed from adj_index_by_signature hash resulting in crash if the same
adjacency is re-inserted.
This patch avoids the issue by enforcing the insertion of 'complete'
adjacencies thereby obviating the need to update the rewrite header.
Change-Id: Ib43bfe72a65e2cf9ef7685a99596eb1d7723e543
Signed-off-by: Florin Coras <fcoras@cisco.com>
Since the move to line-mode the debug CLI was eating everything in the
input_vector but only processing upto the first newline. Cut-and-paste
type operations generally send a large block of input with multiple
newlines and thus all but the very first line were simply ignored.
This patch fixes that and also cleans up the difference between
input_vector and current_command which in turn removes a lot of cruft
from the keystroke parser.
Previously current_command was just the character accumulator inside the
char-by-char keystroke parser; complete commands were copied back to
input_vector (overwriting anything already in there).
Now, in char-by-char mode:
- input_vector is the stream of incoming bytes yet to be processed
- current_command is the accumulated characters of the next command to
be executed; once newline is found, it is the complete command to be
executed.
In line mode:
- input_vector and current_command are the same thing.
Change-Id: I72d21f0f3508b413879071ab186a71cef1124a2b
Signed-off-by: Chris Luke <chrisy@flirble.org>
A problem is easily reproducible by taking the test harness code from the commit,
and launching it in two terminals with some time overlap - the outputs will
be sent to the wrong session. This commit moves the output_function and argument
from a global structure into the process structure, thus the output_function
is not clobbered anymore and each session gets only its own output.
To ensure the callers can redirect the outputs to different destinations
(e.g. the API calls via shared memory, etc.) the existing logic
for vlib_cli_input() was retained.
To avoid the magic numbers usage in the logic that does the page-alignment
of the process stack, there are changes around the stack[] member
of vlib_process_t. Also added a compile-time assert to ensure that
the stack does indeed start on the page size multiple boundary.
Change-Id: I128680ac480735e5f214f81a884e414268e5d652
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
This patch adds an enable/disable API/CLI for control plane
which calls similar functions for data plane. When re-enabling
it also re-populates dataplane with tunnels and interfaces.
Change-Id: Id8c3d6af90ecc0be331d502756914b1f62824046
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
(Also checked other API calls for same error without finding any.)
Change-Id: I1062ecf8eff004efb3fad4582a00c6b512c36999
Signed-off-by: Ole Troan <ot@cisco.com>
By default, jdks bring a bunch of UI related things
we don't need, so switch to headeless.
Also, use default-jdk-headless for Ubuntu after 14.04.
Use openjdk-8-jdk-headless for Ubuntu 14.04.
Change-Id: I3cf14c39c9f59dc2f1beba8dfb19971f4b67f5a6
Signed-off-by: Ed Warnicke <eaw@cisco.com>
If a GRE tunnel is created, no other interface added and the GRE tunnel
is not set "up" then a crash occurs on the first packet for this tunnel
because fib_index_by_sw_if_index[] does not yet have a mapping to the fib
the new interface is in. The code to set this is missing from
gre/interface.c
Change-Id: I567ad74a2af3ea5afe4a40ed39a1d4395642f77c
Signed-off-by: Chris Luke <chrisy@flirble.org>
The GRE code is missing the logic to indicate the "hardware" interface
is up. The fix is to listen for admin up/down events on the "software"
interface and reflect that into the hardware interface state.
Change-Id: If06e4f03989b2c52c32f50c11e1943e42bb2609f
Signed-off-by: Chris Luke <chrisy@flirble.org>
