[lng-odp] [PATCH API-NEXT v5 1/6] validation: crypto: stop declaring test functions as public

From: Dmitry Eremin-Solenikov There is no use to declare test functions as public, they are defined and used in the same file. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 352

[lng-odp] [PATCH API-NEXT v5 0/6] Add separate auth IV

I was thinking again about GMAC auth case. Currently GMAC just reuses cipher IV data fields, which is not future proof and not quite obvious. So let's add separate set of IV-related fields for authentication algorithms. github /** Email created from pull

[lng-odp] [PATCH API-NEXT v5 2/6] linux-gen: crypto: move session type to odp_crypto module

From: Dmitry Eremin-Solenikov There is no point in having odp_crypto_generic_session_t definition in global include file. Move it to odp_crypto module. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull

[lng-odp] [PATCH API-NEXT v5 5/6] validation: crypto: use auth IV for GMAC test case

From: Dmitry Eremin-Solenikov Change GMAC testing code to use new auth IV fields. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 352 (lumag:crypto_gmac_iv) **

[lng-odp] [PATCH API-NEXT v5 6/6] linux-gen: crypto, ipsec: use auth_iv.

From: Dmitry Eremin-Solenikov Separate handling of authentication IV data. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 352 (lumag:crypto_gmac_iv) ** https://github.com/Linaro/odp/pull/352

Re: [lng-odp] [PATCH v11] Misc fixes

Dmitry Eremin-Solenikov(lumag) replied on github web page: test/validation/api/time/time.c @@ -436,12 +436,12 @@ static void time_test_accuracy(time_cb time_cur, time_from_ns_cb time_from_ns) wait = odp_time_sum(wait, sec); } + c2 = time(NULL); t2 =

[lng-odp] [PATCH v1 1/4] validation: crypto: print untested capabilities

From: Dmitry Eremin-Solenikov Print all capability variants that were not covered by reference tests. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 379 (lumag:crypto-untested) **

[lng-odp] [PATCH v1 2/4] validation: crypto: add test vectors for full HMAC length

From: Dmitry Eremin-Solenikov for HMAC tests add test vectors covering full HMAC length in addition to truncated tests. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 379

[lng-odp] [PATCH API-NEXT v5 3/6] api: crypto: use cipher_iv instead of iv in session params

From: Dmitry Eremin-Solenikov In preparation to add auth-specific IV, rename iv field to ciper_iv. Provide deprecated compatibility field iv. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 352

[lng-odp] [PATCH API-NEXT v5 4/6] api: crypto: add separate auth IV

From: Dmitry Eremin-Solenikov GMAC auth algorithm requires IV to work. Instead of hacking the ODP_CIPHER_ALG_NULL iv to include value for ODP_AUTH_ALG_GMAC, provide separate iv (in auth capability, session params and operation params). Signed-off-by: Dmitry

[lng-odp] [PATCH v1 3/4] validation: crypto: fail if no tests were executed

From: Dmitry Eremin-Solenikov Fail the test if we were not able to match any test vectors to capabilities. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 379 (lumag:crypto-untested) **

[lng-odp] [PATCH v11 0/13] Misc fixes

Several small fixes produced while working on updating ODP packaging. github /** Email created from pull request 377 (lumag:misc-fixes) ** https://github.com/Linaro/odp/pull/377 ** Patch: https://github.com/Linaro/odp/pull/377.patch ** Base sha:

[lng-odp] [PATCH v11 4/13] performance: odp_crypto: correct typo

From: Dmitry Eremin-Solenikov Receieved -> Received. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) ** https://github.com/Linaro/odp/pull/377 ** Patch:

[lng-odp] [PATCH v11 13/13] example: ipsec_api: remove file conflict with ipsec example

From: Dmitry Eremin-Solenikov ipsec_api example installs program with the same name, as ipsec example does. Remove file conflict between two examples. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull

[lng-odp] [PATCH v1 4/4] validation: crypto: don't output untested cipher for NULL cipher

From: Dmitry Eremin-Solenikov AES-GMAC adds synthetic NULL cipher capability with IV length = 12. Skip CIPHER_ALG_NULL entries when printing untested messages (as untested entry does not really mean a thing for NULL cipher algo). Signed-off-by: Dmitry

[lng-odp] [PATCH v11 7/13] helper/tests: use common Makefile.inc for tests

From: Dmitry Eremin-Solenikov Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) ** https://github.com/Linaro/odp/pull/377 ** Patch:

[lng-odp] [PATCH v11 10/13] validation: time: use gettimeofday() instead of clock()

From: Dmitry Eremin-Solenikov ODP test for time API uses clock() to compare time against. However clock() returns processor time used by program, which can differ between runs. Use gettimeofday() as a time source to compare against. Signed-off-by: Dmitry

[lng-odp] [PATCH v11 8/13] test, linux-gen: use common test/Makefile.inc

From: Dmitry Eremin-Solenikov Populate test/Makefile.inc with more variables, allowing us to drop platform test/Makefile.inc. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377

[lng-odp] [PATCH v11 12/13] tests: don't use ODP_SYSCONFIG_FILE

From: Dmitry Eremin-Solenikov libconfig is not used in TigerMoth, so let's drop the variable in test's Makefile.inc. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) **

[lng-odp] [PATCH v11 1/13] helper: link against libpthread and libodp-linux

From: Dmitry Eremin-Solenikov libodphelper.so uses symbols from libpthread and libodp-linux.so, link it aganst those libraries. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377

[lng-odp] [PATCH v11 2/13] doc: add missing user-guide images to Makefile.am

From: Dmitry Eremin-Solenikov Omitting those images in Makefile.am made them disappear from distribution tarball resulting in an improperly-built documentation files. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email

[lng-odp] [PATCH v11 11/13] linux-gen: don't use SYSCONFDIR

From: Dmitry Eremin-Solenikov sysconfig is not used in TigerMoth, so let's drop the variable in linux-gen's Makefile. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) **

[lng-odp] [PATCH v11 9/13] validation: stop overriding LDADD variable

From: Dmitry Eremin-Solenikov LDADD variable can be used to pass other libraries (like DPDK). Stop overriding LDADD entirely. Instead use PRELDADD variable or appending to LDADD. Signed-off-by: Dmitry Eremin-Solenikov ---

[lng-odp] [PATCH v11 5/13] linux-gen: don't pass DPKD_LIBS for dynamic linking with libodp-linux

From: Dmitry Eremin-Solenikov Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) ** https://github.com/Linaro/odp/pull/377 ** Patch:

[lng-odp] [PATCH v11 6/13] example: add missing libraries to LDADD variables

From: Dmitry Eremin-Solenikov Signed-off-by: Dmitry Eremin-Solenikov --- /** Email created from pull request 377 (lumag:misc-fixes) ** https://github.com/Linaro/odp/pull/377 ** Patch:

[lng-odp] [PATCH v11 3/13] doc: add missing implementers-guide images to Makefile.am

From: Dmitry Eremin-Solenikov Omitting those images in Makefile.am made them disappear from distribution tarball resulting in an improperly-built documentation files. Signed-off-by: Dmitry Eremin-Solenikov --- /** Email

[lng-odp] [Linaro/odp] 9eebbe: linux-gen: add memory-mapped I/O access API

Branch: refs/heads/caterpillar Home: https://github.com/Linaro/odp Commit: 9eebbe7d21f0b933190fd910c54c872eecfe180e https://github.com/Linaro/odp/commit/9eebbe7d21f0b933190fd910c54c872eecfe180e Author: Mykyta Iziumtsev Date: 2018-01-04 (Thu, 04

[lng-odp] [PATCH CATERPILLAR v4 4/4] linux-gen: packet: calculate the start of packet pointer

From: Brian Brooks The start of packet pointer was adjusted on every operation involving headroom manipulation. This meant that this pointer had to be restored to its pristine (aligned) offset when the buffer was recycled for reuse. Instead of maintaining a pristine offset

[lng-odp] [PATCH CATERPILLAR v4 1/4] linux-gen: packet: remove duplicated odp_packet_hdr() function

From: Brian Brooks odp_packet_hdr() and packet_hdr() are both internal-only functions that do the same exact thing. Remove odp_packet_hdr() version since "odp_" prefix is reserved for top-level API identifiers. This is also consistent with other packet_xxx() internal-only

[lng-odp] [PATCH CATERPILLAR v4 2/4] linux-gen: packet: remove duplicated _odp_packet_from_buf_hdr() function

From: Brian Brooks _odp_packet_from_buf_hdr() and packet_from_buf_hdr() are both internal-only functions that do the exact same thing. Remove _odp_packet_from_buf_hdr() version since internal-only identifiers should not use any form of an "odp_" prefix. This is also

[lng-odp] [PATCH CATERPILLAR v4 3/4] linux-gen: packet: remove excess segmentation metadata

From: Brian Brooks Use a single pointer to next segment instead of an array of pointers. This reduces buffer metadata by ~140 bytes. Signed-off-by: Brian Brooks Signed-off-by: Honnappa Nagarahalli --- /** Email created

[lng-odp] RSS in ODP

Hello and happy new year, In our company we are looking into scaling the odp_schedule() calls. Currently we are manually doing Receive Side Scaling, which requires one CPU to receive all the packets and pass them to other worker CPU in a flow-deterministic way (i.e., not spreading the packets

Re: [lng-odp] [PATCH v9] Misc fixes

Petri Savolainen(psavol) replied on github web page: test/validation/api/time/time.c line 25 @@ -436,12 +436,12 @@ static void time_test_accuracy(time_cb time_cur, time_from_ns_cb time_from_ns) wait = odp_time_sum(wait, sec); } + c2 = time(NULL); t2 =

Re: [lng-odp] RSS in ODP

ODP has several mechanisms for controlling workload distribution. The hash functions can distribute packets from a given PktIO to a group of queues, however the classifier is the most flexible means of controlling this. A CoS defines a target queue for the flow. This is the queue that receives