Re: [lng-odp] [API-NEXT RFC] api: packet: add define for max segmentation

On 07/23/2015 07:43 AM, Bala Manoharan wrote: > > > On 21 July 2015 at 13:05, Nicolas Morey-Chaisemartin > wrote: > > > > On 07/20/2015 07:24 PM, Bala Manoharan wrote: >> Hi, >> >> Few comments inline >> >> On 20 July 2015 at 22:38, Nicolas Morey-Chaisema

Re: [lng-odp] [API-NEXT RFC] api: packet: add define for max segmentation

On 21 July 2015 at 13:05, Nicolas Morey-Chaisemartin wrote: > > > On 07/20/2015 07:24 PM, Bala Manoharan wrote: > > Hi, > > Few comments inline > > On 20 July 2015 at 22:38, Nicolas Morey-Chaisemartin < > nmo...@kalray.eu> wrote: > >> Replace current segmentation with an explicit define. >> Thi

Re: [lng-odp] [API-NEXT RFC] api: packet: add define for max segmentation

Hi, Even with this current pool parameters if the application provides "seg_len" pool parameter equal to the expected packet length then the implementation will have to store the entire packet in a single segment but it will not be performant in platforms which support segmentation. I believe this

Re: [lng-odp] [PATCH] validation: packet: do not account for head or tail room in packet length computations

On Tue, Jul 21, 2015 at 3:05 AM, Nicolas Morey-Chaisemartin < nmo...@kalray.eu> wrote: > It is the pool responsibility to ensure that there is enough space > available for headroom and tailroom in the segment(s) > Also fixes the issue where ODP_CONFIG_PACKET_SEG_MIN_LEN < > ODP_CONFIG_PACKET_HEADR

Re: [lng-odp] [API-NEXT RFC] api: packet: add define for max segmentation

In the original packet API design it was proposed that one of the pool options should be UNSEGMENTED, which says the application does not wish to see segments at all. The implementation would then either comply or else fail the pool create if it is unable to support unsegmented pools. However, that

[lng-odp] [PATCH] linux-generic: Makefile.am: fix basename conflicts

Building generated this warning message: "copying selected object files to avoid basename conflicts..." Reported-by: Mile Holmes Signed-off-by: Anders Roxell --- platform/linux-generic/Makefile.am | 2 +- .../linux-generic/arch/linux/{odp_time.c => odp_time_

[lng-odp] [PATCH] update version number from v1.1.0.0 to v1.2

Signed-off-by: Maxim Uvarov --- Changelog notes for next release. thread mask api will also added before release. Maxim. debian/changelog | 50 +++ include/odp/api/version.h | 2 +- 2 files changed, 51 insertions(+), 1 deletion(-) diff

[lng-odp] [Bug 1706] New: odp_term_local() not being called from main thread

https://bugs.linaro.org/show_bug.cgi?id=1706 Bug ID: 1706 Summary: odp_term_local() not being called from main thread Product: OpenDataPlane Version: 1.1 Hardware: Other OS: Linux Status: UNCONFIRMED Sever

[lng-odp] [API-NEXT PATCHv4 10/10] validation: thrmask: test worker and control APIs

Add tests for odp_thrmask_worker() and odp_thrmask_control() Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard --- test/validation/thread/thread.c | 70 + 1 file changed, 70 insertions(+) diff --git a/test/validation/thread/thread.c b/test/vali

[lng-odp] [API-NEXT PATCHv4 08/10] validation: thrmask: introduce thrmask tests

Signed-off-by: Stuart Haslam --- test/validation/Makefile.inc | 1 + test/validation/common/.gitignore| 1 + test/validation/common/Makefile.am | 5 - test/validation/common/mask_common.h | 12 test/validation/thread/Makefile.am | 2 ++ test/validation/thread/

[lng-odp] [API-NEXT PATCHv4 09/10] validation: common: return number of threads created

odph_linux_pthread_create() now returns the number of threads created, so modify odp_cunit_thread_create() to do the same. Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard --- test/validation/common/odp_cunit_common.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff -

[lng-odp] [API-NEXT PATCHv4 06/10] validation: split out cpumask tests into separate file

Split the majority of the test cases out into a separate file in preparation for them being reused to perform thrmask tests. Signed-off-by: Stuart Haslam --- test/validation/Makefile.inc | 3 +- test/validation/common/Makefile.am | 4 +- test/validation/common/mask_common.c | 472 +

[lng-odp] [API-NEXT PATCHv4 07/10] validation: refactor cpumask tests

Refactor the cpumask tests a bit so that they may be reused to perform the thrmask testing. Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard --- test/validation/common/mask_common.c | 478 +-- test/validation/common/mask_common.h | 51 ++-- test/valid

[lng-odp] [API-NEXT PATCHv4 05/10] validation: use convenience libs rather than static libs

Convert libraries used under test/validation to libtool convenience libraries (.la files) rather than regular static libraries. These are used elsewhere in the tree and allow for dependencies to be tracked automatically. Signed-off-by: Stuart Haslam --- .gitignore

[lng-odp] [API-NEXT PATCHv4 04/10] validation: cpumask: def_worker and def_control fixes

The odp_cpumask_def_worker() and odp_cpumask_def_control() APIs are used to populate a mask with the CPUs available for running threads of the same type. This does not necessarily mean those CPUs are dedicated to running threads of that type, on a particular platform the same CPU may appear in both

[lng-odp] [API-NEXT PATCHv4 03/10] linux-generic: cpumask: zero mask before setting

odp_cpumask_def_control() was neglecting to zero out the provided mask. Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard --- platform/linux-generic/odp_cpumask.c | 1 + 1 file changed, 1 insertion(+) diff --git a/platform/linux-generic/odp_cpumask.c b/platform/linux-generic/odp_cpu

[lng-odp] [API-NEXT PATCHv4 02/10] validation: cpumask: use odp_cpumask_setall() to determine mask capacity

The existing test uses odp_cpu_count() to attempt to determine the capacity of odp_cpumask_t. This was a workaround as there was previously no reliable method to determine this but the odp_cpumask_setall() API was added to address this, so use it. Signed-off-by: Stuart Haslam Reviewed-by: Christo

[lng-odp] [API-NEXT PATCHv4 01/10] validation: add missing include to odp_cunit_common.h

stdint.h is required for uint32_t Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard --- test/validation/common/odp_cunit_common.h | 1 + 1 file changed, 1 insertion(+) diff --git a/test/validation/common/odp_cunit_common.h b/test/validation/common/odp_cunit_common.h index 98bccfb..7

[lng-odp] [API-NEXT PATCHv4 00/10] validation: add thrmask tests

This series adds tests for the new odp_thrmask_* APIs. As many of the APIs are the same as the odp_cpumask_* APIs but operating on a different type, the existing cpumask tests are initially refactored so that they can be reused. Changes v3 -> v4; - added symbols to libthread and libcpumask libra

Re: [lng-odp] [API-NEXT PATCH 1/3] merge fix: remove files accidentally added when resolving git merge

series Reviewed-by: Mike Holmes On 20 July 2015 at 11:51, Maxim Uvarov wrote: > That commit: > fd855d7 Merge branch 'master' into api-next > > Signed-off-by: Maxim Uvarov > --- > platform/linux-generic/pktio/.socket.c.swp | Bin 16384 -> 0 bytes > test/performance/.libs/lt-odp_l2fwd

Re: [lng-odp] [PATCH v2] configure: use stricter warnings

Merged, Maxim. On 07/22/15 17:34, Anders Roxell wrote: On 2015-07-20 16:32, Mike Holmes wrote: Signed-off-by: Mike Holmes Reviewed-by: Anders Roxell --- Makefile.inc | 1 - bootstrap | 2 +- configure.ac | 1 + examp

Re: [lng-odp] [PATCH] linux-generic: buffer: removed unused _odp_buffer_type_set

I will cherry-pick original patch from api-next to master. Maxim. On 07/22/15 17:46, Bill Fischofer wrote: This is a duplicate of Patch http://patches.opendataplane.org/patch/2215/ that I submitted on July 10th. On Wed, Jul 22, 2015 at 9:35 AM, Mike Holmes > wr

Re: [lng-odp] [PATCH] linux-generic: buffer: removed unused _odp_buffer_type_set

This is a duplicate of Patch http://patches.opendataplane.org/patch/2215/ that I submitted on July 10th. On Wed, Jul 22, 2015 at 9:35 AM, Mike Holmes wrote: > Signed-off-by: Mike Holmes > --- > platform/linux-generic/include/odp_buffer_internal.h | 9 - > platform/linux-generic/odp_buf

Re: [lng-odp] [PATCH v2] configure: use stricter warnings

On 2015-07-20 16:32, Mike Holmes wrote: > Signed-off-by: Mike Holmes Reviewed-by: Anders Roxell > --- > Makefile.inc | 1 - > bootstrap | 2 +- > configure.ac | 1 + > example/Makefile.inc | 1 - > platform/linu

[lng-odp] [PATCH] linux-generic: buffer: removed unused _odp_buffer_type_set

Signed-off-by: Mike Holmes --- platform/linux-generic/include/odp_buffer_internal.h | 9 - platform/linux-generic/odp_buffer.c | 8 2 files changed, 17 deletions(-) diff --git a/platform/linux-generic/include/odp_buffer_internal.h b/platform/linux-generic/inclu

[lng-odp] [PATCHv2] linux-generic: pktio: remove _pkt suffix for open, close, recv and send pktio methods

Unifies the format for all pktio_if_ops_t functions Signed-off-by: Nicolas Morey-Chaisemartin --- v2: Drop suffix for recv/send functionx too platform/linux-generic/pktio/loop.c| 12 +-- platform/linux-generic/pktio/socket.c | 32 +++--- platform/lin

Re: [lng-odp] [API-NEXT PATCH] api: crypto: add crypto IPSec extension

Minor fixes On 22 July 2015 at 04:26, wrote: > From: Alexandru Badicioiu > > This patch adds IPSec protocol processing capabilities to crypto > sesssions. Implementations which have these capabilities in hardware > crypto engines can use the extension to offload the application from > IPSec pro

Re: [lng-odp] [PATCH 2/2] linux-generic: pktio/sock: remove _pkt suffix for open and close functions

On 07/22/15 15:02, Nicolas Morey-Chaisemartin wrote: On 07/22/2015 01:23 PM, Maxim Uvarov wrote: On 07/22/15 13:57, Nicolas Morey-Chaisemartin wrote: On 07/22/2015 12:00 PM, Maxim Uvarov wrote: On 07/17/15 12:15, Nicolas Morey-Chaisemartin wrote: .open = sock_mmsg_open, -.close =

Re: [lng-odp] [PATCH 2/2] linux-generic: pktio/sock: remove _pkt suffix for open and close functions

On 07/22/2015 01:23 PM, Maxim Uvarov wrote: > On 07/22/15 13:57, Nicolas Morey-Chaisemartin wrote: >> >> On 07/22/2015 12:00 PM, Maxim Uvarov wrote: >>> On 07/17/15 12:15, Nicolas Morey-Chaisemartin wrote: .open = sock_mmsg_open, -.close = sock_close_pkt, +.close = s

Re: [lng-odp] [PATCH] api: config: replace defines with apis

Lets bring it up in the call, looks like there are issues of grouping, naming and how many APIs there are to retrieve the data. On 22 July 2015 at 05:31, Maxim Uvarov wrote: > I would add only one odp api call to fetch platform config. > > Like > > int odp_get_platform_config(struct platform_con

Re: [lng-odp] [API-NEXT PATCHv3 5/9] validation: split out cpumask tests into separate file

On Wed, Jul 22, 2015 at 11:06:00AM +0200, Christophe Milard wrote: > On 2015-07-21 14:32, Stuart Haslam wrote: > > Split the majority of the test cases out into a separate file in > > preparation for them being reused to perform thrmask tests. > > > > Signed-off-by: Stuart Haslam > > --- > > tes

Re: [lng-odp] [PATCH 2/2] linux-generic: pktio/sock: remove _pkt suffix for open and close functions

On 07/22/15 13:57, Nicolas Morey-Chaisemartin wrote: On 07/22/2015 12:00 PM, Maxim Uvarov wrote: On 07/17/15 12:15, Nicolas Morey-Chaisemartin wrote: .open = sock_mmsg_open, -.close = sock_close_pkt, +.close = sock_close, .recv = sock_mmsg_recv_pkt, .send = sock_mm

Re: [lng-odp] [API-NEXT PATCHv3 9/9] validation: thrmask: test worker and control APIs

On 2015-07-21 14:32, Stuart Haslam wrote: > Add tests for odp_thrmask_worker() and odp_thrmask_control() > > Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard > --- > test/validation/thread/thread.c | 70 > + > 1 file changed, 70 insertions(+

Re: [lng-odp] [API-NEXT PATCHv3 6/9] validation: refactor cpumask tests

On 2015-07-21 14:32, Stuart Haslam wrote: > Refactor the cpumask tests a bit so that they may be reused to perform > the thrmask testing. > > Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard > --- > test/validation/common/mask_common.c | 478 > +-- >

Re: [lng-odp] [API-NEXT PATCHv3 4/9] validation: cpumask: def_worker and def_control fixes

On 2015-07-21 14:32, Stuart Haslam wrote: > The odp_cpumask_def_worker() and odp_cpumask_def_control() APIs are used > to populate a mask with the CPUs available for running threads of the > same type. This does not necessarily mean those CPUs are dedicated to > running threads of that type, on a p

Re: [lng-odp] [API-NEXT PATCHv3 3/9] linux-generic: cpumask: zero mask before setting

On 2015-07-21 14:32, Stuart Haslam wrote: > odp_cpumask_def_control() was neglecting to zero out the provided mask. > > Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard > --- > platform/linux-generic/odp_cpumask.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/platform/l

Re: [lng-odp] [API-NEXT PATCHv3 8/9] validation: common: return number of threads created

On 2015-07-21 14:32, Stuart Haslam wrote: > odph_linux_pthread_create() now returns the number of threads created, > so modify odp_cunit_thread_create() to do the same. > > Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard > --- > test/validation/common/odp_cunit_common.c | 5 ++---

Re: [lng-odp] [API-NEXT PATCHv3 2/9] validation: cpumask: use odp_cpumask_setall() to determine mask capacity

On 2015-07-21 14:32, Stuart Haslam wrote: > The existing test uses odp_cpu_count() to attempt to determine the > capacity of odp_cpumask_t. This was a workaround as there was previously > no reliable method to determine this but the odp_cpumask_setall() API was > added to address this, so use it. >

Re: [lng-odp] [API-NEXT PATCHv3 1/9] validation: add missing include to odp_cunit_common.h

On 2015-07-21 14:32, Stuart Haslam wrote: > stdint.h is required for uint32_t > > Signed-off-by: Stuart Haslam Reviewed-by: Christophe Milard > --- > test/validation/common/odp_cunit_common.h | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/test/validation/common/odp_cunit_common.h

Re: [lng-odp] [PATCH 2/2] linux-generic: pktio/sock: remove _pkt suffix for open and close functions

On 07/22/2015 12:00 PM, Maxim Uvarov wrote: > On 07/17/15 12:15, Nicolas Morey-Chaisemartin wrote: >> .open = sock_mmsg_open, >> -.close = sock_close_pkt, >> +.close = sock_close, >> .recv = sock_mmsg_recv_pkt, >> .send = sock_mmsg_send_pkt, > Nicolas, I'm going to apply

Re: [lng-odp] [PATCH 1/2] linux-generic: pktio: rename loop and sock _init functions to _open

That patch 1/1 is merged. Maxim. On 07/17/15 12:14, Nicolas Morey-Chaisemartin wrote: Function names are now coherent with pktio_if_opts names Signed-off-by: Nicolas Morey-Chaisemartin --- platform/linux-generic/pktio/loop.c | 4 ++-- platform/linux-generic/pktio/socket.c | 8

Re: [lng-odp] [API-NEXT PATCHv3 7/9] validation: thrmask: introduce thrmask tests

On 2015-07-21 14:32, Stuart Haslam wrote: > Signed-off-by: Stuart Haslam > --- > test/validation/Makefile.inc | 1 + > test/validation/common/.gitignore| 1 + > test/validation/common/Makefile.am | 5 - > test/validation/common/mask_common.h | 12 > test/validati

Re: [lng-odp] [PATCH 2/2] linux-generic: pktio/sock: remove _pkt suffix for open and close functions

On 07/17/15 12:15, Nicolas Morey-Chaisemartin wrote: .open = sock_mmsg_open, - .close = sock_close_pkt, + .close = sock_close, .recv = sock_mmsg_recv_pkt, .send = sock_mmsg_send_pkt, Nicolas, I'm going to apply first patch. But in that patch it's not clear why

Re: [lng-odp] [PATCH] api: config: replace defines with apis

I would add only one odp api call to fetch platform config. Like int odp_get_platform_config(struct platform_config *config); where: struct platform_config { struct { } queue; struct { } pool; } One function instead 20 others for each layer it better

Re: [lng-odp] [API-NEXT PATCHv3 5/9] validation: split out cpumask tests into separate file

On 2015-07-21 14:32, Stuart Haslam wrote: > Split the majority of the test cases out into a separate file in > preparation for them being reused to perform thrmask tests. > > Signed-off-by: Stuart Haslam > --- > test/validation/Makefile.inc | 7 +- > test/validation/common/.gitignore

[lng-odp] [API-NEXT PATCH] api: crypto: add crypto IPSec extension

From: Alexandru Badicioiu This patch adds IPSec protocol processing capabilities to crypto sesssions. Implementations which have these capabilities in hardware crypto engines can use the extension to offload the application from IPSec protocol processing. Signed-off-by: Alexandru Badicioiu ---