[lng-odp] [API_NEXT RFC V2 00/01] linux-generic: add support for initial cpumasks

Oops! Multi-tasking error on my part - forgot to run 'make check' after my last code refinement and had an unused variable. This update removes that variable and allows make check to complete. Sorry about the noise! Gary S. Robertson (1): linux-generic: add support for initial cpumasks

Re: [lng-odp] [PATCH 0/5] validation: cls: correct tests a little

On 14.01.16 07:51, Bala Manoharan wrote: Hi Ivan, It would be better if you can send this patch on API-NEXT branch as I said before I am working on API proposal to modify odp_pmr_create() function syntax and it would be easy if we could modify your changes to API-NEXT and then we can move it

Re: [lng-odp] [PATCH 5/5] validation: cls: split pmr chain test

Agreed. Error CoS and L2 should be added as separate suites and I would preferable add them in odp_classification_test_pmr.c so that each are tested separately. Maybe rename the file also :) As I pointed out enhancements can be done once API freeze is accepted. Regards, Bala On 14 January 2016

Re: [lng-odp] [PATCH 3/5] validation: cls: test_pmr: don't create default input queue

The reason for default input queue is coz thats how linux generic pktio works and as you said when running the same tests on HW the default input queue will be ignored and the packet will be processed correctly. Regards, Bala Regards, Bala On 14 January 2016 at 13:15, Ivan Khoronzhuk

Re: [lng-odp] [PATCH 4/5] validation: cls: use correct MAC addresses

In odp_classification_tests.c a common pktio_loop is used coz in that suite the objective was to test the behaviour of a classification system with all different possible configurations existing concurrently. ie Configure L2 PMR, L3 PMR, default_cos, error_cos and send an error packet and check

[lng-odp] ODP 226: need for shmem->refresh()?

This is regarding ODP 226 (Tests assuming pthreads) Kalray is facing a problem actually larger than this thread vs process problem: The basic question is: when N processors share the same memory (shmem object), is it acceptable to force a cache update for the whole shmem'd area for N-1

[lng-odp] [Bug 1834] Untested internal API verify_pmr_dmac

https://bugs.linaro.org/show_bug.cgi?id=1834 Bala Manoharan changed: What|Removed |Added Resolution|--- |FIXED

[lng-odp] [Bug 1541] Untested API verify_pmr_packet_len

https://bugs.linaro.org/show_bug.cgi?id=1541 Bala Manoharan changed: What|Removed |Added Resolution|--- |FIXED

[lng-odp] [PATCH 2/2] tests: harmonize posix extensions level defines

Allways use _GNU_SOURCE instead of various _POSIX_C_SOURCE defines. Signed-off-by: Petri Savolainen --- example/generator/odp_generator.c | 4 +++- example/ipsec/odp_ipsec.c | 6 -- example/ipsec/odp_ipsec_fwd_db.c | 4 +++-

[lng-odp] [PATCH] validation: pktio: assign MAC address if one loop pktio is used

In case of one loop pktio the MAC address is not set in the packets but should be. Signed-off-by: Ivan Khoronzhuk --- test/validation/pktio/pktio.c | 18 ++ 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/test/validation/pktio/pktio.c

Re: [lng-odp] [API-NEXT PATCH 1/2] doc: users: add TM example

Merged, Maxim. On 01/14/2016 02:15, Bill Fischofer wrote: For this series: Reviewed-by: Bill Fischofer > On Wed, Jan 13, 2016 at 4:07 PM, Mike Holmes > wrote:

Re: [lng-odp] odp_pool_param_t::num versus per-thread object caches

Num is exact number so that application can limit the number of in-flight packets (number of per packet contexts and latency due to buffered packets). Optimal level of per thread stashing is implementation specific, so water marks (if any) should remain internal to the implementation. It’s OK

[lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

Signed-off-by: Bill Fischofer --- CONTRIBUTING | 5 ++--- DEPENDENCIES | 7 ++- 2 files changed, 4 insertions(+), 8 deletions(-) diff --git a/CONTRIBUTING b/CONTRIBUTING index 4ad964e..f6e3fc6 100644 --- a/CONTRIBUTING +++ b/CONTRIBUTING @@ -126,9 +126,8 @@ Code

Re: [lng-odp] [RFC PATCH] api: packet_io: return headroom when set new one

Once we add the ability to attach additional segments to a packet (as is being discussed in other threads) we may wish to revisit the concept of "extensible" headroom/tailroom. As long as packets can be extended there's no logical reason why headroom/tailroom need be limited to a single segment.

Re: [lng-odp] [RFC PATCH] api: packet_io: return headroom when set new one

Maximum headroom is defined in ODP_PACKET_MAX_HEADROOM and is an implementation specific value and will not be greater than uint32/2. Having said that the current definition of headroom in ODP is that it will not exceed a single segment meaning the headroom will NOT overflow by adding an empty

Re: [lng-odp] [RFC PATCH] api: packet_io: return headroom when set new one

On 14.01.16 14:03, Bala Manoharan wrote: Maximum headroom is defined in ODP_PACKET_MAX_HEADROOM and is an implementation specific value and will not be greater than uint32/2. Having said that the current definition of headroom in ODP is that it will not exceed a single segment meaning the

Re: [lng-odp] [API-NEXT PATCHv3 1/2] doc: remove imagemagic dependency and use svg diagrams in all asciidocs

I think you deleted the rule for the msg files as well - resource_management.svg is no longer built asciidoc -a data-uri -b html5 -a icons -a toc2 -a max-width=55em --out-file=users-guide.html users-guide.adoc sh: 1: cannot open

[lng-odp] [API-NEXT PATCHv4 1/2] doc: remove imagemagic dependency and use svg diagrams in all asciidocs

Simplify ODP documentation builds by using .svg files directly for all images. This eliminates the need to create .png and .eps files via the imagemagic package. Note that this removes easy PDF output for ODP user documents, however there is no real demand for such capability. Signed-off-by: Bill

Re: [lng-odp] [API-NEXT PATCHv3 1/2] doc: remove imagemagic dependency and use svg diagrams in all asciidocs

v4 sent to correct this. On Thu, Jan 14, 2016 at 6:51 AM, Mike Holmes wrote: > I think you deleted the rule for the msg files as well > - resource_management.svg is no longer built > > asciidoc -a data-uri -b html5 -a icons -a toc2 -a max-width=55em >

Re: [lng-odp] [API_NEXT RFC 01/01] ODP API: add control/worker cpumasks to init data

> -Original Message- > From: EXT Gary S. Robertson [mailto:gary.robert...@linaro.org] > Sent: Wednesday, January 13, 2016 10:01 PM > To: mike.hol...@linaro.org; bill.fischo...@linaro.org; > maxim.uva...@linaro.org; anders.rox...@linaro.org; > stuart.has...@linaro.org;

Re: [lng-odp] ODP 226: need for shmem->refresh()?

In general, this issue is about - how to support non cache coherent systems. Additional calls specified for non-coherent systems could be added (to shm and potentially elsewhere) but should be optional for applications, since it’s quite tricky to (efficiently) ensure cache coherency in SW.

Re: [lng-odp] Maintain code copy-pasted from linux-generic in ODP-DPDK

On 13/01/16 13:37, Mike Holmes wrote: On 13 January 2016 at 03:49, Maxim Uvarov > wrote: On 01/12/2016 21:57, Zoltan Kiss wrote: Hi, We have a couple of places where the entire source file couldn't be

Re: [lng-odp] Maintain code copy-pasted from linux-generic in ODP-DPDK

On 13/01/16 13:46, Maxim Uvarov wrote: On 01/13/2016 00:15, Nicolas Morey-Chaisemartin wrote: Hi, I have the same issue with our implementation. I usually either work from the git diff of linux generic before and post update and backport changed or to a diff of the diff of linux-generic vs

Re: [lng-odp] Maintain code copy-pasted from linux-generic in ODP-DPDK

On 14/01/16 16:25, Stuart Haslam wrote: On Thu, Jan 14, 2016 at 04:13:06PM +, Zoltan Kiss wrote: On 13/01/16 13:37, Mike Holmes wrote: On 13 January 2016 at 03:49, Maxim Uvarov > wrote: On 01/12/2016 21:57, Zoltan Kiss

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

https://bugs.linaro.org/show_bug.cgi?id=1706 Mike Holmes changed: What|Removed |Added Assignee|kapil.h...@linaro.org |lng-odp@lists.linaro.org

Re: [lng-odp] Maintain code copy-pasted from linux-generic in ODP-DPDK

On 01/13/2016 02:46 PM, Maxim Uvarov wrote: > On 01/13/2016 00:15, Nicolas Morey-Chaisemartin wrote: >> Hi, >> >> I have the same issue with our implementation. >> I usually either work from the git diff of linux generic before and post >> update and backport changed or to a diff of the diff of

[lng-odp] [PATCH] test: change l2fwd pool size

The current code looks awkward, it defines pkt.num as 564. Clean that up and set it to 8k to make sure even several big descriptor rings could be accomodated. Signed-off-by: Zoltan Kiss diff --git a/test/performance/odp_l2fwd.c b/test/performance/odp_l2fwd.c index

[lng-odp] 答复: [PATCH] pool:support to create pool on user's own memory

On 01/14/2016 04:28, Xu Zaibo wrote: > Hi, > > Firstly, I don't want to bring any new type of pool into ODP. > > I just want to give user more chance to use more areas of memory. For > example, referring to your example of various HW offload engines, if there > two or more types of engines > >

[lng-odp] [PATCH] queue: queue_deq: delete atomic fetch and inc

As all the queue_deq is locked by queue lock, I think it doesn't need to atomic opertion here any more. Signed-off-by: Zaibo Xu --- platform/linux-generic/odp_queue.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/platform/linux-generic/odp_queue.c

[lng-odp] [PATCH] DEPENDENCIES: add librsvg2-bin to HTML

Signed-off-by: Anders Roxell --- DEPENDENCIES | 7 ++- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/DEPENDENCIES b/DEPENDENCIES index fec65d2..3563b36 100644 --- a/DEPENDENCIES +++ b/DEPENDENCIES @@ -215,8 +215,5 @@ The tested version of doxygen is

[lng-odp] [API-NEXT PATCHv5 1/2] doc: remove imagemagic dependency and use svg diagrams in all asciidocs

Simplify ODP documentation builds by using .svg files directly for all images. This eliminates the need to create .png and .eps files via the imagemagic package. Note that this removes easy PDF output for ODP user documents, however there is no real demand for such capability. Signed-off-by: Bill

[lng-odp] [API-NEXT PATCHv5 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

Signed-off-by: Bill Fischofer --- CONTRIBUTING | 5 ++--- DEPENDENCIES | 11 ++- 2 files changed, 4 insertions(+), 12 deletions(-) diff --git a/CONTRIBUTING b/CONTRIBUTING index 4ad964e..f6e3fc6 100644 --- a/CONTRIBUTING +++ b/CONTRIBUTING @@ -126,9 +126,8 @@

[lng-odp] 答复: [PATCH] pool:support to create pool on user's own memory

Hi, So, it is sure that implementation may need different kinds of memory. After adding user's own memory operations to the pool param, all the functions of odp_pool_create before are not changed. An implementation is still asking memory from ODP shm or some other internal memory as before. I

Re: [lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

If it is growing in scope I agree to follow ons, let's focus on what we broke with this change then. We only ever supported html and pdf before so this should just clean the remaining pdf text so people dont read about it. On 14 January 2016 at 17:08, Bill Fischofer

Re: [lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

OK v5 sent. Only Delta is removing the PDF stuff in DEPENDENCIES. On Thu, Jan 14, 2016 at 4:32 PM, Mike Holmes wrote: > If it is growing in scope I agree to follow ons, let's focus on what we > broke with this change then. > We only ever supported html and pdf before so

Re: [lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

I think all at once with new subject like - remove all pdf support On 14 January 2016 at 15:35, Bill Fischofer wrote: > Sounds good. Do you want that as a v5 or should that be a separate patch? > > On Thu, Jan 14, 2016 at 1:10 PM, Mike Holmes

[lng-odp] [PATCH] linux-generic: init: handle local/global init/term cleanly

Restructure odp_init_global() and odp_init_local() so that they recover cleanly if initialization fails. At exit any partial initialization is reversed so that system is in the same state as before the failing call was made. This includes adding dummy termination calls for functions that do not

[lng-odp] [PATCH] validation: system: add validation tests for odp_cpu_cycles_* calls

https://bugs.linaro.org/show_bug.cgi?id=1906 Signed-off-by: Ivan Khoronzhuk --- test/validation/system/system.c | 111 test/validation/system/system.h | 4 ++ 2 files changed, 115 insertions(+) diff --git

Re: [lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

Sounds good. Do you want that as a v5 or should that be a separate patch? On Thu, Jan 14, 2016 at 1:10 PM, Mike Holmes wrote: > The dependencies still talks about pdfs which we dont really support now > for doxygen. > > We can opt at this point to completely remove all

Re: [lng-odp] [PATCH] doc: move api docs to application-api-guide

ping On 7 January 2016 at 16:48, Bill Fischofer wrote: > > > On Wed, Jan 6, 2016 at 6:22 AM, Mike Holmes > wrote: > >> The other documentation is contained in its own directory, the API guide >> for the application should be no different. >>

Re: [lng-odp] ODP 226: need for shmem->refresh()?

A full memory barrier would actually work for us but the cost is unacceptable most of the time. To get all the performances out of our core, it will be necessary for our ODP user to be aware the the non-cache coherency and deal with it manually. However what I'm looking for (as a first step at

Re: [lng-odp] [PATCH] pool:support to create pool on user's own memory

Hi, So, it would be memory for any pool type. If an implementation asks memory from the application, portability becomes a problem. How application would know which kind of memory each implementation would need for each pool type? Linux-generic uses ODP shm to reserve memory for pools, but

Re: [lng-odp] odp_pool_param_t::num versus per-thread object caches

Ok, then the next question comes: how this timer unit test timer_test_odp_timer_all should be changed? I've put Ola in the To field as he created it originally. It creates the pool with (num_workers * X) timeouts, and each workers then tries to allocate X timeouts from the pool. It fails

Re: [lng-odp] Maintain code copy-pasted from linux-generic in ODP-DPDK

On Thu, Jan 14, 2016 at 04:13:06PM +, Zoltan Kiss wrote: > > > On 13/01/16 13:37, Mike Holmes wrote: > > > > > >On 13 January 2016 at 03:49, Maxim Uvarov >> wrote: > > > >On 01/12/2016 21:57, Zoltan Kiss wrote: > > > >Hi,

Re: [lng-odp] 答复: [PATCH] pool:support to create pool on user's own memory

On 01/14/2016 04:28, Xu Zaibo wrote: Hi, Firstly, I don't want to bring any new type of pool into ODP. I just want to give user more chance to use more areas of memory. For example, referring to your example of various HW offload engines, if there two or more types of engines which have

Re: [lng-odp] [API-NEXT PATCH 1/2] validation: synchro tests split into 3 groups

ping. On 12/17/2015 18:37, Christophe Milard wrote: No functionnal changes: just code reordering to match the ODP modules. Signed-off-by: Christophe Milard --- configure.ac|3 + platform/linux-generic/test/Makefile.am |3 +

[lng-odp] [Bug 1976] example/generator should use odp time instead of system sys/time.h

https://bugs.linaro.org/show_bug.cgi?id=1976 Mike Holmes changed: What|Removed |Added Priority|--- |Lowest

[lng-odp] [Bug 1881] linux-generic/arch does not contain architectures

https://bugs.linaro.org/show_bug.cgi?id=1881 Mike Holmes changed: What|Removed |Added Status|UNCONFIRMED |CONFIRMED Ever

Re: [lng-odp] [API-NEXT PATCHv4 2/2] doc: update CONTRIBUTING and DEPENDENCIES files for images

Taking a look at this, we also have make rules for rtf, chm, xml, ps, and pdf output so this removal patch is somewhat bigger. I think I'd prefer to see that as a separate patch. This one really just cleaning up the image files and standardizing on svgs. I think a separate one to standardize on