vpp/src/vnet/udp, branch v17.10 Vector Packet Processing http://git.fd.io/vpp/atom?h=v17.10 2017-09-20T17:50:48+00:00 session: store tep port in net order 2017-09-20T17:50:48+00:00 Florin Coras fcoras@cisco.com 2017-09-20T05:27:18+00:00 urn:sha1:0e4956869b7b919957433f04cc8fe52a2f391251 Change-Id: Ie3a99f09f44ec081d9b88a213bdb8d987fb462de Signed-off-by: Florin Coras <fcoras@cisco.com> Make tcp active open data structures thread safe 2017-07-30T15:04:26+00:00 Florin Coras fcoras@cisco.com 2017-07-25T00:40:28+00:00 urn:sha1:68810624f84467503482b82662c980e8f0e36deb - Cleanup half-open connections and timers on the right thread - Ensure half-open connection and transport endpoint pools are thread safe - Enqueue TX events to the correct vpp thread in the builtin client - Use transport proto in transport connections instead of session type Change-Id: Id13239a206afbff6f34a38afa510fe014e4b2049 Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net> Cleanup/refactor session layer code 2017-07-25T11:17:20+00:00 Florin Coras fcoras@cisco.com 2017-07-17T00:12:15+00:00 urn:sha1:04e5344a358a9ad42d896486d2d226149fd326f4 Change-Id: Ica99e8cb919fca6b069c37c969d60e8ccc2c6bf9 Signed-off-by: Florin Coras <fcoras@cisco.com> Horizontal (nSessions) scaling draft 2017-07-11T15:40:19+00:00 Dave Barach dbarach@cisco.com 2017-06-26T15:35:07+00:00 urn:sha1:2c25a62cc1cc4937165de740a3b32d78429c72d6 - Data structure preallocation. - Input state machine fixes for mid-stream 3-way handshake retries. - Batch connections in the builtin_client - Multiple private fifo segment support - Fix elog simultaneous event type registration - Fix sacks when segment hole is added after highest sacked - Add "accepting" session state for sessions pending accept - Add ssvm non-recursive locking - Estimate RTT for syn-ack - Don't init fifo pointers. We're using relative offsets for ooo segments - CLI to dump individual session Change-Id: Ie0598563fd246537bafba4feed7985478ea1d415 Signed-off-by: Dave Barach <dbarach@cisco.com> Signed-off-by: Florin Coras <fcoras@cisco.com> Rework vxlan-gpe to support FIB 2.0 and bypass mode 2017-06-06T07:26:00+00:00 Hongjun Ni hongjun.ni@intel.com 2017-05-27T12:23:09+00:00 urn:sha1:8a0a0ae60b8dd9da7cf53c895e85dc6daf67143d Change-Id: I0324f945bdb4dd3b19151be6f3ce24a47a000104 Signed-off-by: Hongjun Ni <hongjun.ni@intel.com> Improve fifo allocator performance 2017-06-01T16:02:55+00:00 Dave Barach dave@barachs.net 2017-05-30T13:30:07+00:00 urn:sha1:10d8cc6bf92851fcaec4a6b4c6d3554dc1eb2386 - add option to preallocate fifos in a segment - track active fifos with doubly linked list instead of vector - update udp redirect test code to read fifo pointers from API call instead of digging them up from fifo segment header - input-node based active-open session generator Change-Id: I804b81e99d95f8690d17e12660c6645995e28a9a Signed-off-by: Dave Barach <dave@barachs.net> Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dbarach@cisco.com> Add GTP-U plugin. VPP-694 2017-05-15T19:34:00+00:00 Hongjun Ni hongjun.ni@intel.com 2017-04-12T11:21:16+00:00 urn:sha1:ef486b1545d892f9f0e0d35e7e57cb0ca04d7ff7 Basic GTP-U feature Change-Id: I31226f890a92c5303ac06e112ed7820cae52d9bd Signed-off-by: Hongjun Ni <hongjun.ni@intel.com> Session layer improvements 2017-04-24T12:02:14+00:00 Florin Coras fcoras@cisco.com 2017-04-19T20:00:05+00:00 urn:sha1:a5464817522c7a7dc760af4612f1d6a68ed0afc8 Among others: - Moved app event queue to shared memory segment - Use private memory segment for builtin apps - Remove pid from svm fifo - Protect session fifo (de)allocation - Use fifo event for session disconnects - Have session queue node poll in all wk threads Change-Id: I89dbf7fdfebef12f5ef2b34ba3ef3c2c07f49ff2 Signed-off-by: Florin Coras <fcoras@cisco.com> Session layer refactoring 2017-04-14T01:35:50+00:00 Florin Coras fcoras@cisco.com 2017-04-05T06:08:23+00:00 urn:sha1:6cf30adc2cd3aa818e5d97cf71ea8b2fc2aaefa7 Major refactoring of the session layer api - Add attatch api for application binding to the the session layer - Simplify listen/connect calls - Update application CLI - Add transport endpoint to accept callback - Associate segment manager to application and allow for multiple binds/connects per app Additional: - svm fifo cleanup - add fifo free, format fns - add fifo offset enqueue unit test Change-Id: Id93a65047de61afc2bf3d58c9b544339c02065af Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net> Use thread local storage for thread index 2017-04-06T11:31:39+00:00 Damjan Marion damarion@cisco.com 2017-04-05T17:18:20+00:00 urn:sha1:586afd762bfa149f5ca167bd5fd5a0cd59ce94fe This patch deprecates stack-based thread identification, Also removes requirement that thread stacks are adjacent. Finally, possibly annoying for some folks, it renames all occurences of cpu_index and cpu_number with thread index. Using word "cpu" is misleading here as thread can be migrated ti different CPU, and also it is not related to linux cpu index. Change-Id: I68cdaf661e701d2336fc953dcb9978d10a70f7c1 Signed-off-by: Damjan Marion <damarion@cisco.com>
This XML file does not appear to have any style information associated with it. The document tree is shown below.
<feed xmlns="http://www.w3.org/2005/Atom">
<title>vpp/src/vnet/udp, branch v17.10</title>
<subtitle>Vector Packet Processing</subtitle>
<id>http://git.fd.io/vpp/atom?h=v17.10</id>
<link rel="self" href="http://git.fd.io/vpp/atom?h=v17.10"/>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/"/>
<updated>2017-09-20T17:50:48+00:00</updated>
<entry>
<title>session: store tep port in net order</title>
<updated>2017-09-20T17:50:48+00:00</updated>
<author>
<name>Florin Coras</name>
<email>fcoras@cisco.com</email>
</author>
<published>2017-09-20T05:27:18+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=0e4956869b7b919957433f04cc8fe52a2f391251"/>
<id>urn:sha1:0e4956869b7b919957433f04cc8fe52a2f391251</id>
<content type="text"> Change-Id: Ie3a99f09f44ec081d9b88a213bdb8d987fb462de Signed-off-by: Florin Coras <fcoras@cisco.com> </content>
</entry>
<entry>
<title>Make tcp active open data structures thread safe</title>
<updated>2017-07-30T15:04:26+00:00</updated>
<author>
<name>Florin Coras</name>
<email>fcoras@cisco.com</email>
</author>
<published>2017-07-25T00:40:28+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=68810624f84467503482b82662c980e8f0e36deb"/>
<id>urn:sha1:68810624f84467503482b82662c980e8f0e36deb</id>
<content type="text"> - Cleanup half-open connections and timers on the right thread - Ensure half-open connection and transport endpoint pools are thread safe - Enqueue TX events to the correct vpp thread in the builtin client - Use transport proto in transport connections instead of session type Change-Id: Id13239a206afbff6f34a38afa510fe014e4b2049 Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net> </content>
</entry>
<entry>
<title>Cleanup/refactor session layer code</title>
<updated>2017-07-25T11:17:20+00:00</updated>
<author>
<name>Florin Coras</name>
<email>fcoras@cisco.com</email>
</author>
<published>2017-07-17T00:12:15+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=04e5344a358a9ad42d896486d2d226149fd326f4"/>
<id>urn:sha1:04e5344a358a9ad42d896486d2d226149fd326f4</id>
<content type="text"> Change-Id: Ica99e8cb919fca6b069c37c969d60e8ccc2c6bf9 Signed-off-by: Florin Coras <fcoras@cisco.com> </content>
</entry>
<entry>
<title>Horizontal (nSessions) scaling draft</title>
<updated>2017-07-11T15:40:19+00:00</updated>
<author>
<name>Dave Barach</name>
<email>dbarach@cisco.com</email>
</author>
<published>2017-06-26T15:35:07+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=2c25a62cc1cc4937165de740a3b32d78429c72d6"/>
<id>urn:sha1:2c25a62cc1cc4937165de740a3b32d78429c72d6</id>
<content type="text"> - Data structure preallocation. - Input state machine fixes for mid-stream 3-way handshake retries. - Batch connections in the builtin_client - Multiple private fifo segment support - Fix elog simultaneous event type registration - Fix sacks when segment hole is added after highest sacked - Add "accepting" session state for sessions pending accept - Add ssvm non-recursive locking - Estimate RTT for syn-ack - Don't init fifo pointers. We're using relative offsets for ooo segments - CLI to dump individual session Change-Id: Ie0598563fd246537bafba4feed7985478ea1d415 Signed-off-by: Dave Barach <dbarach@cisco.com> Signed-off-by: Florin Coras <fcoras@cisco.com> </content>
</entry>
<entry>
<title>Rework vxlan-gpe to support FIB 2.0 and bypass mode</title>
<updated>2017-06-06T07:26:00+00:00</updated>
<author>
<name>Hongjun Ni</name>
<email>hongjun.ni@intel.com</email>
</author>
<published>2017-05-27T12:23:09+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=8a0a0ae60b8dd9da7cf53c895e85dc6daf67143d"/>
<id>urn:sha1:8a0a0ae60b8dd9da7cf53c895e85dc6daf67143d</id>
<content type="text"> Change-Id: I0324f945bdb4dd3b19151be6f3ce24a47a000104 Signed-off-by: Hongjun Ni <hongjun.ni@intel.com> </content>
</entry>
<entry>
<title>Improve fifo allocator performance</title>
<updated>2017-06-01T16:02:55+00:00</updated>
<author>
<name>Dave Barach</name>
<email>dave@barachs.net</email>
</author>
<published>2017-05-30T13:30:07+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=10d8cc6bf92851fcaec4a6b4c6d3554dc1eb2386"/>
<id>urn:sha1:10d8cc6bf92851fcaec4a6b4c6d3554dc1eb2386</id>
<content type="text"> - add option to preallocate fifos in a segment - track active fifos with doubly linked list instead of vector - update udp redirect test code to read fifo pointers from API call instead of digging them up from fifo segment header - input-node based active-open session generator Change-Id: I804b81e99d95f8690d17e12660c6645995e28a9a Signed-off-by: Dave Barach <dave@barachs.net> Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dbarach@cisco.com> </content>
</entry>
<entry>
<title>Add GTP-U plugin. VPP-694</title>
<updated>2017-05-15T19:34:00+00:00</updated>
<author>
<name>Hongjun Ni</name>
<email>hongjun.ni@intel.com</email>
</author>
<published>2017-04-12T11:21:16+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=ef486b1545d892f9f0e0d35e7e57cb0ca04d7ff7"/>
<id>urn:sha1:ef486b1545d892f9f0e0d35e7e57cb0ca04d7ff7</id>
<content type="text"> Basic GTP-U feature Change-Id: I31226f890a92c5303ac06e112ed7820cae52d9bd Signed-off-by: Hongjun Ni <hongjun.ni@intel.com> </content>
</entry>
<entry>
<title>Session layer improvements</title>
<updated>2017-04-24T12:02:14+00:00</updated>
<author>
<name>Florin Coras</name>
<email>fcoras@cisco.com</email>
</author>
<published>2017-04-19T20:00:05+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=a5464817522c7a7dc760af4612f1d6a68ed0afc8"/>
<id>urn:sha1:a5464817522c7a7dc760af4612f1d6a68ed0afc8</id>
<content type="text"> Among others: - Moved app event queue to shared memory segment - Use private memory segment for builtin apps - Remove pid from svm fifo - Protect session fifo (de)allocation - Use fifo event for session disconnects - Have session queue node poll in all wk threads Change-Id: I89dbf7fdfebef12f5ef2b34ba3ef3c2c07f49ff2 Signed-off-by: Florin Coras <fcoras@cisco.com> </content>
</entry>
<entry>
<title>Session layer refactoring</title>
<updated>2017-04-14T01:35:50+00:00</updated>
<author>
<name>Florin Coras</name>
<email>fcoras@cisco.com</email>
</author>
<published>2017-04-05T06:08:23+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=6cf30adc2cd3aa818e5d97cf71ea8b2fc2aaefa7"/>
<id>urn:sha1:6cf30adc2cd3aa818e5d97cf71ea8b2fc2aaefa7</id>
<content type="text"> Major refactoring of the session layer api - Add attatch api for application binding to the the session layer - Simplify listen/connect calls - Update application CLI - Add transport endpoint to accept callback - Associate segment manager to application and allow for multiple binds/connects per app Additional: - svm fifo cleanup - add fifo free, format fns - add fifo offset enqueue unit test Change-Id: Id93a65047de61afc2bf3d58c9b544339c02065af Signed-off-by: Florin Coras <fcoras@cisco.com> Signed-off-by: Dave Barach <dave@barachs.net> </content>
</entry>
<entry>
<title>Use thread local storage for thread index</title>
<updated>2017-04-06T11:31:39+00:00</updated>
<author>
<name>Damjan Marion</name>
<email>damarion@cisco.com</email>
</author>
<published>2017-04-05T17:18:20+00:00</published>
<link rel="alternate" type="text/html" href="http://git.fd.io/vpp/commit/?id=586afd762bfa149f5ca167bd5fd5a0cd59ce94fe"/>
<id>urn:sha1:586afd762bfa149f5ca167bd5fd5a0cd59ce94fe</id>
<content type="text"> This patch deprecates stack-based thread identification, Also removes requirement that thread stacks are adjacent. Finally, possibly annoying for some folks, it renames all occurences of cpu_index and cpu_number with thread index. Using word "cpu" is misleading here as thread can be migrated ti different CPU, and also it is not related to linux cpu index. Change-Id: I68cdaf661e701d2336fc953dcb9978d10a70f7c1 Signed-off-by: Damjan Marion <damarion@cisco.com> </content>
</entry>
</feed>