Re: [lng-odp] [API-NEXT PATCH 1/2] api: version: added implementation name str

This patch introduces odp_version_impl_name(), which is used to identify platform e.g. if (strcmp(odp_version_impl_name(), "odp-dpdk") == 0) { // odp-dpdk specific worker thread creation, etc } Otherwise it's just documenting what we already have. Implementation specific defines would

Re: [lng-odp] [API-NEXT PATCHv2] linux-generic: queue: use locking hierarchy for ordered queueing

On 11/02/2015 06:11 PM, Bill Fischofer wrote: > Enqueueing to ordered queues requires two locks (source and target > queues). Since any queue can be either source or target, queues do not > have a natural locking hierarchy. Create one by using the address of > the qentry as the lock order. > >

[lng-odp] [PATCH] validation: pktio: initialize queue parameters correctly

A call to odp_queue_param_init() is missing so uninitialized parameters are passed to odp_queue_create() leading to unpredictable results. Signed-off-by: Stuart Haslam --- pktio tests fail fairly regularly when built with clang. test/validation/pktio/pktio.c | 1 + 1

Re: [lng-odp] [API-NEXT PATCHv2] linux-generic: queue: use locking hierarchy for ordered queueing

On 11/03/2015 10:21 AM, Nicolas Morey-Chaisemartin wrote: > > On 11/02/2015 06:11 PM, Bill Fischofer wrote: >> Enqueueing to ordered queues requires two locks (source and target >> queues). Since any queue can be either source or target, queues do not >> have a natural locking hierarchy. Create

Re: [lng-odp] [API-NEXT PATCHv2] linux-generic: queue: use locking hierarchy for ordered queueing

It's fixing the deadlock issue reported by Carl where A is queueing to B and B is queuing to A. The previous scheme relied on a statistical locking strategy that was prone to stalls, especially on uniprocessor systems. Your suggestion solves this issue and is also simpler and faster. Concurrent

Re: [lng-odp] [API-NEXT PATCHv2] linux-generic: queue: use locking hierarchy for ordered queueing

Also, why is this only applied to API-NEXT ? Doesn't master have the same "deadlock" issue? On 11/02/2015 06:11 PM, Bill Fischofer wrote: > Enqueueing to ordered queues requires two locks (source and target > queues). Since any queue can be either source or target, queues do not > have a natural

Re: [lng-odp] [PATCH v2 0/6] l2fwd optimizations

Ping. L2fwd is used to benchmark pktio development. The app should be optimized so that it does not hide optimizations on the implementation side. -Petri > -Original Message- > From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of > EXT Petri Savolainen > Sent: Friday,

Re: [lng-odp] [API-NEXT PATCHv2] linux-generic: queue: use locking hierarchy for ordered queueing

On Tue, Nov 3, 2015 at 3:21 AM, Nicolas Morey-Chaisemartin wrote: > > > On 11/02/2015 06:11 PM, Bill Fischofer wrote: > > Enqueueing to ordered queues requires two locks (source and target > > queues). Since any queue can be either source or target, queues do not > > have a

Re: [lng-odp] [PATCH v2 1/6] test: l2fwd: added option to disable error check

On 11/03/2015 12:52 PM, Nicolas Morey-Chaisemartin wrote: > > On 10/30/2015 10:24 AM, Petri Savolainen wrote: >> Added command line option to disable packet error check. Error >> check requires full packet parse. Max packet rate measurements >> should be done with minimal feature set. >> >> This

[lng-odp] [Bug 1879] Ordered queues occasionally not restoring order properly

https://bugs.linaro.org/show_bug.cgi?id=1879 --- Comment #10 from Carl Wallen --- (In reply to Bill Fischofer from comment #9) > v2 of patch posted at http://patches.opendataplane.org/patch/3583/ This patch also works for me, however as pointed out by Nicolas, I would

Re: [lng-odp] [PATCH v2 1/6] test: l2fwd: added option to disable error check

On 10/30/2015 10:24 AM, Petri Savolainen wrote: > Added command line option to disable packet error check. Error > check requires full packet parse. Max packet rate measurements > should be done with minimal feature set. > > This change gives +10% packet rate increase in direct recv mode > with

Re: [lng-odp] [PATCH v2 1/6] test: l2fwd: added option to disable error check

For the series: Reviewed-by: Nicolas Morey-Chaisemartin On 10/30/2015 10:24 AM, Petri Savolainen wrote: > Added command line option to disable packet error check. Error > check requires full packet parse. Max packet rate measurements > should be done with minimal feature set. >

[lng-odp] [PATCH v1] test: performance: add crypto test

From: Alexandru Badicioiu Test program to measure crypto operation performance. Measures the time required to launch and get the result of ODP crypto operations in async and sync API mode with configurable payload size, crypto algorithms and iteration number. Both

Re: [lng-odp] [PATCH v1] test: performance: add crypto test

On 11/03/2015 18:20, alexandru.badici...@linaro.org wrote: memset(, 0, sizeof(params)); From quick look 2 notes: 1. alphabetical order for Makefile.am 2. change memset to appropriate init function. Maxim. ___ lng-odp mailing list

Re: [lng-odp] [API-NEXT RFC 2/2] api: pktio: added direct queue receive

On Tue, Nov 3, 2015 at 8:29 AM, Petri Savolainen wrote: > Added odp_pktio_recv_queue for direct packet receive from pktio input > hash queues. > > Signed-off-by: Petri Savolainen > --- > include/odp/api/packet_io.h | 25

Re: [lng-odp] [API-NEXT RFC 1/2] api: pktio: added input hash parameters

On Tue, Nov 3, 2015 at 8:29 AM, Petri Savolainen wrote: > Added parameters for setting up input hashing during pktio_open > and odp_pktio_input_hash_queues to query queue handles. > > Signed-off-by: Petri Savolainen > --- >

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

Ok, turns out that at the time the set function was removed odp_timer.c was still using this function, which is why it wasn't removed as part of that patch. Since then odp_timer.c has been changed to use official odp_event_type() API, so this function is no longer used anywhere. Patch submitted

[lng-odp] [PATCH] linux-generic: buffer: remove unused internal _odp_buffer_type() function

Signed-off-by: Bill Fischofer --- platform/linux-generic/include/odp_buffer_internal.h | 10 -- platform/linux-generic/odp_buffer.c | 8 2 files changed, 18 deletions(-) diff --git

Re: [lng-odp] [PATCHv3 1/5] helper: add table api

Merged to branch next, with is staging branch for next 1.5 release. Maxim. On 10/29/2015 10:02, huanggaoyang wrote: Signed-off-by: huanggaoyang --- helper/include/odp/helper/table.h | 222 ++ 1 file changed, 222 insertions(+)

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

Let me investigate and I'll post either a patch to do the delete or an explanation of why we shouldn't do that. On Tue, Nov 3, 2015 at 8:55 AM, Maxim Uvarov wrote: > That commit is for _odp_buffer_type_set but we need to remove > _odp_buffer_type. > > Maxim. > > On

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

ping On 22 July 2015 at 10:49, Maxim Uvarov wrote: > 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

Re: [lng-odp] [API-NEXT PATCH v2] validation: pktio: test odp_pktio_print()

Merged, Maxim. On 11/02/2015 16:10, Elo, Matias (Nokia - FI/Espoo) wrote: -Original Message- From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of EXT Maxim Uvarov Sent: Monday, November 02, 2015 2:44 PM To: lng-odp@lists.linaro.org Subject: Re: [lng-odp] [API-NEXT

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

That commit is for _odp_buffer_type_set but we need to remove _odp_buffer_type. Maxim. On 11/03/2015 17:46, Mike Holmes wrote: ping On 22 July 2015 at 10:49, Maxim Uvarov > wrote: I will cherry-pick original patch from api-next

Re: [lng-odp] [PATCH v2 0/6] l2fwd optimizations

Merged, candidate to 1.4.1 Maxim. On 11/03/2015 13:57, Savolainen, Petri (Nokia - FI/Espoo) wrote: Ping. L2fwd is used to benchmark pktio development. The app should be optimized so that it does not hide optimizations on the implementation side. -Petri -Original Message- From:

[lng-odp] [API-NEXT RFC 1/2] api: pktio: added input hash parameters

Added parameters for setting up input hashing during pktio_open and odp_pktio_input_hash_queues to query queue handles. Signed-off-by: Petri Savolainen --- include/odp/api/packet_io.h | 60 + 1 file changed, 60

[lng-odp] [API-NEXT RFC 2/2] api: pktio: added direct queue receive

Added odp_pktio_recv_queue for direct packet receive from pktio input hash queues. Signed-off-by: Petri Savolainen --- include/odp/api/packet_io.h | 25 - 1 file changed, 20 insertions(+), 5 deletions(-) diff --git

[lng-odp] [PATCH] example: ipsec: check push_tail return code

Signed-off-by: Nicolas Morey-Chaisemartin --- example/ipsec/odp_ipsec.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/example/ipsec/odp_ipsec.c b/example/ipsec/odp_ipsec.c index 2aaae1d..9950ccf 100644 --- a/example/ipsec/odp_ipsec.c +++

Re: [lng-odp] [PATCH] validation: pktio: initialize queue parameters correctly

Reviewed-by: Nicolas Morey-Chaisemartion On 11/03/2015 04:00 PM, Stuart Haslam wrote: > A call to odp_queue_param_init() is missing so uninitialized parameters > are passed to odp_queue_create() leading to unpredictable results. > > Signed-off-by: Stuart Haslam

Re: [lng-odp] [PATCH] linux-generic: check return codes in odp_pktio_term_global

On 10/30/2015 10:27 AM, Maxim Uvarov wrote: > According to API odp_pktio_close() can be called only for stopped > pktio. So in odp_pktio_term_global try to stop it first then > call close. Also check all returns codes. > https://bugs.linaro.org/show_bug.cgi?id=1854 > > Signed-off-by: Maxim