Re: [lng-odp] Mailing test again

It is still: Why is this message in Spam? It has a from address in kalray.eu but has failed kalray.eu's required tests for authentication. Learn more The learn more link points here: https://support.google.com/mail/answer/1366858?hl=en-GB&expand=5 Here are your headers: Delivered-To: zolta

Re: [lng-odp] [PATCH] configure.ac: fix have_pcap AM_CONDITIONAL

Ping On 18/07/16 15:40, Zoltan Kiss wrote: This causes "./configure: line 20048: test: =: unary operator expected" error messages, while the generated code was like this: if test $pktio_dpdk_support = yes ; then PKTIO_DPDK_TRUE= PKTIO_DPDK_FALSE='#' else ... Signed

Re: [lng-odp] [PATCH] linux-dpdk/m4/configure.m4: move m4_include above ac_check_*

Applied to ODP-DPDK (with subject line formatted a bit) On 21/07/16 15:42, Anders Roxell wrote: Signed-off-by: Anders Roxell --- platform/linux-dpdk/m4/configure.m4 | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/platform/linux-dpdk/m4/configure.m4 b/platform/

Re: [lng-odp] [PATCH] linux-dpdk/m4/configure.m4: move m4_include above ac_check_*

Reviewed-by: Zoltan Kiss On 21/07/16 15:42, Anders Roxell wrote: Signed-off-by: Anders Roxell --- platform/linux-dpdk/m4/configure.m4 | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/platform/linux-dpdk/m4/configure.m4 b/platform/linux-dpdk/m4/configure.m4

[lng-odp] [PATCH] configure.ac: fix have_pcap AM_CONDITIONAL

This causes "./configure: line 20048: test: =: unary operator expected" error messages, while the generated code was like this: if test $pktio_dpdk_support = yes ; then PKTIO_DPDK_TRUE= PKTIO_DPDK_FALSE='#' else ... Signed-off-by: Zoltan Kiss --- diff --git a/configu

Re: [lng-odp] Test mail again

It ended up in the spam folder again, but I forget to check why before moving it from them, and gmail forgot that. You could try sending an email to me to my address privately. On 08/07/16 06:59, Nicolas Morey-Chaisemartin wrote: Me again and again. Sorry for the spam. As for the previous one

Re: [lng-odp] Yet Another Test Mail

Hi, Yes, both your mails landed in my spam folder. Gmail said: Why is this message in Spam? It has a from address in kalray.eu but has failed kalray.eu's required tests for authentication. Learn more Then there is a link: https://support.google.com/mail/answer/1366858?hl=en-GB&expand=5 Zol

Re: [lng-odp] [lng-odp-ovs] Error while installing ODP-v1.8

t 11:05 AM, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: Hi, ODP_CACHE_LINE_SIZE is defined in platform/linux-generic/arch/[your arch]/odp/api/cpu_arch.h, or in platform/linux-generic/include/odp/api/align.h, I guess there is some screwup there. You should ask about

Re: [lng-odp] Error while installing ODP-v1.8

Hi, ODP_CACHE_LINE_SIZE is defined in platform/linux-generic/arch/[your arch]/odp/api/cpu_arch.h, or in platform/linux-generic/include/odp/api/align.h, I guess there is some screwup there. You should ask about this on the lng-odp list, I've added it to this mail. Regards, Zoltan On 21/06/

Re: [lng-odp] [PATCH] linux-generic: timer: correct definition of ODP_TIMEOUT_INVALID

Hi, ODP-DPDK uses 0 or NULL for all _INVALID constants, because that's what rte_mbuf's using for that purpose. How could this fix work there? I mean, simply just applying this patch doesn't work, because then this value will be in conflict with ODP_EVENT_INVALID et all. Zoli On 10/06/16 01:

[lng-odp] [PATCH] linux-generic: classification: use proper accessor to set packet length

And repurpose packet_set_len() for this, as it is no longer used. This was introduced by the following commit, and break compatibility with ODP-DPDK: bd18047a "linux-gen: pktio: don't allocate new packets in classifier" Signed-off-by: Zoltan Kiss --- platform/linux-

Re: [lng-odp] [PATCH] autotools: define test extensions to skip on valgrind test

Hi, Now that I tried to port this to ODP-DPDK, the pktio tests started to fail. It turned out Valgrind overwrites the LOG_COMPILER, which is not used in odp-linux, but odp-dpdk has that. TEST_EXTENSIONS makes the testcase run without the wrapper, even if Valgrind is not used. Moreover, do we n

Re: [lng-odp] [PATCHv2 1/2] linux-generic: packet: copy user area as part of odp_packet_copy()

I think that last hunk for test/validation/packet/packet.c should go to the second patch, otherwise, for the series: Reviewed-by: Zoltan Kiss On 09/06/16 00:11, Bill Fischofer wrote: Resolve Bug https://bugs.linaro.org/show_bug.cgi?id=2310 by copying the user area as part of odp_packet_copy

Re: [lng-odp] [API-NEXT PATCHv1 0/4] implement mechanism to allow inline and no inline use of ODP API

Hi, ODP-DPDK already has a simpler solution for this problem, the first patch was posted here as well: https://git.linaro.org/lng/odp-dpdk.git/commitdiff/968237b592a8162041a4edf0091575fd1390d647 It can get simpler as it doesn't need a separate config option, but it automatically turns on whe

Re: [lng-odp] [PATCH 2/2] validation: packet: ensure user area is copied correctly

Reviewed-by: Zoltan Kiss On 07/06/16 15:53, Bill Fischofer wrote: As part of resolution of Bug https://bugs.linaro.org/show_bug.cgi?id=2310 make sure that odp_packet_copy() handles user area copies correctly. The copy should fail if the target pool's user area size is not large enou

Re: [lng-odp] [PATCH 1/2] linux-generic: packet: copy user area as part of odp_packet_copy()

Hi, On 07/06/16 15:53, Bill Fischofer wrote: Resolve Bug https://bugs.linaro.org/show_bug.cgi?id=2310 by copying the user area as part of odp_packet_copy(). The copy fails if the user area size of the destination pool is not large enough to hold the source packet user area. Signed-off-by: Bill

Re: [lng-odp] packet copy questions

@linaro.org <mailto:bill.fischo...@linaro.org>] Sent: Monday, June 06, 2016 10:53 PM To: Zoltan Kiss mailto:zoltan.k...@linaro.org>> Cc: lng-odp mailto:lng-odp@lists.linaro.org>>; Savolainen, Petri (Nokia - FI/Espoo) mailto:petri.savolai...@nokia.com>>

Re: [lng-odp] [lng-odp-dpdk] [RFC PATCH] linux-dpdk: Enable inlines if dynamic library is not used

I've merged this patch as it was required for 1.10 to work on ODP-DPDK: https://git.linaro.org/lng/odp-dpdk.git/commitdiff/968237b592a8162041a4edf0091575fd1390d647 The changes I made: - packet_flags also handled - now it called _ODP_INLINES On 30/05/16 11:03, Zoltan Kiss wrote: Hi, On

[lng-odp] packet copy questions

Hi, During the ODP-DPDK 1.10 upgrade I found these two things: - Should _odp_packet_copy_md_to_packet() copy timestamp and op_result? - odp_packet_copy() doesn't copy the user area. I think it would be better if it just calls _odp_packet_copy_md_to_packet() Regards, Zoltan __

Re: [lng-odp] [lng-odp-dpdk] [RFC PATCH] linux-dpdk: Enable inlines if dynamic library is not used

Hi, On 30/05/16 02:46, Bill Fischofer wrote: diff --git a/configure.ac b/configure.ac index 99772a1..199a6e2 100644 --- a/configure.ac +++ b/configure.ac @@ -141,6 +141,13 @@ AM_CONDI

[lng-odp] [RFC PATCH] linux-dpdk: Enable inlines if dynamic library is not used

functions. If INLINES defined, the accessors appear as static inline functions in the header, and the application can directly inline them. Signed-off-by: Zoltan Kiss --- .gitignore | 1 + configure.ac | 7 ++ plat

[lng-odp] opendataplane.org "Downloads" page changes

Hi, The odp-dpdk part of http://www.opendataplane.org/downloads/ needs updating: - The patchworks links should be deleted, as nobody uses it, given that I'm the only maintainer and the only regular commiter. - The "latest stable release link" is also very old, and doesn't fit for odp-dpdk: I on

Re: [lng-odp] lng-odp mailman settings

+1 for mandatory plain text. HTML mail is for newsletters, but as soon as you try to reply to a section of someone else's mail, it breaks down to chaos. My observation is that mail clients tend to handle it differently, and screwing up each other (and their users) in the process. And setting fon

Re: [lng-odp] [PATCH 1/2] linux-generic: includes: typo correction

Reviewed-by: Zoltan Kiss On 12/05/16 17:29, Bill Fischofer wrote: Correct typo in spelling of _LARGEFILE64_SOURCE. This also removes an extraneous non-ascii character from the source file. Suggested-by: Zoltan Kiss Signed-off-by: Bill Fischofer --- platform/linux-generic/include

Re: [lng-odp] OpenDataPlane is now on Wikipedia

Hi, The link to Wind River is broken, it should point here: https://en.wikipedia.org/wiki/Wind_River_Systems On 12/05/16 13:27, Raj Murali wrote: Dear All, I am happy to inform you all that, we are now on Wikipedia - at last. https://en.wikipedia.org/wiki/Open_Data_Plane_(ODP) Thanks to all

[lng-odp] Non-ASCII characters in the repo

Hi, I've came accross a lot in the doc folder, I don't know if we want to do anything about them, but here is an easy way to find them: git grep --color='auto' -P -n "[\x80-\xFF]" There is only one place outside docs: diff --git a/platform/linux-generic/include/odp_posix_extensions.h b/platfo

[lng-odp] run checkpatch automatically

Hi, As Mike raised my attention, I'm sloppy in that, so I decided to integrate it with my git wrapper. It is a shell script in my home directory which precedes git so I can enforce stuff not configurable for git. git hooks can do similar stuff, but they have to be created for each repository, w

[lng-odp] [PATCH v2 3/3] linux-generic: pktio: don't return packet after failed enqueue

igned-off-by: Zoltan Kiss --- v2: - rebase - don't report discards as failure - add comment .../linux-generic/include/odp_packet_io_internal.h | 3 +- platform/linux-generic/pktio/dpdk.c| 10 +++--- platform/linux-generic/pktio/netmap.c | 8 ++--- platform/lin

[lng-odp] [PATCH v2 2/3] linux-generic: pktio: classification error handling fixes for loop

not count these packets - as a side effect, separate this code path from normal handling completely Signed-off-by: Zoltan Kiss --- v2: - rebase - fixs discards as well platform/linux-generic/pktio/loop.c | 22 +++--- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git

[lng-odp] [PATCH v2 1/3] linux-generic: classification: release the packet as soon as an error happens

-by: Zoltan Kiss --- v2: - add comment for this function - use errno.h values - let caller handle EINVAL and ENOMEM - loopback reenques them - fix in stats that CoS drop means actually discard platform/linux-generic/odp_classification.c | 41 ++--- platform/linux-generic

[lng-odp] [PATCH v2 0/3] Fix pktio classification error handling

Fixing up things in error handling, and harmonizing different codebases. Zoltan Kiss (3): linux-generic: classification: release the packet as soon as an error happens linux-generic: pktio: classification error handling fixes for loop linux-generic: pktio: don't return packet

Re: [lng-odp] [PATCH 1/3] linux-generic: classification: release the packet as soon as an error happens

On 02/05/16 09:56, Bala Manoharan wrote: On 29 April 2016 at 18:28, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: On 28/04/16 18:08, Bala Manoharan wrote: On 28 April 2016 at 21:50, Zoltan Kiss mailto:zoltan.k...@linaro.org> <mailto:zoltan.k..

Re: [lng-odp] [PATCH 1/3] linux-generic: classification: release the packet as soon as an error happens

On 28/04/16 18:08, Bala Manoharan wrote: On 28 April 2016 at 21:50, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: On 28/04/16 10:29, Bala Manoharan wrote: On 27 April 2016 at 21:43, Zoltan Kiss mailto:zoltan.k...@linaro.org> <mailto:zoltan.k..

Re: [lng-odp] [PATCH 1/3] linux-generic: classification: release the packet as soon as an error happens

On 28/04/16 10:29, Bala Manoharan wrote: On 27 April 2016 at 21:43, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: Move release to _odp_packet_classifier(), because caller has no way to know if new_pkt were allocated. In that case there would be a double free on pk

[lng-odp] [PATCH 3/3] linux-generic: pktio: don't return packet after failed enqueue

If queue_enq() fails, there is a serious internal error, the packet should be released instead of passing it back to pktin_poll() for enqueueing on an another one. Also return the error if there is any, not just the queue_enq() ones. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include

[lng-odp] [PATCH 1/3] linux-generic: classification: release the packet as soon as an error happens

Move release to _odp_packet_classifier(), because caller has no way to know if new_pkt were allocated. In that case there would be a double free on pkt, and new_pkt would be leaked. Signed-off-by: Zoltan Kiss --- platform/linux-generic/odp_classification.c | 21 - platform

[lng-odp] [PATCH 2/3] linux-generic: pktio: classification error handling fixes for loop

this code path from normal handling completely Signed-off-by: Zoltan Kiss --- platform/linux-generic/pktio/loop.c | 18 +- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/platform/linux-generic/pktio/loop.c b/platform/linux-generic/pktio/loop.c index 676e98b..

[lng-odp] [PATCH 0/3] Fix pktio classification error handling

Hopefully we covered every case now. Zoltan Kiss (3): linux-generic: classification: release the packet as soon as an error happens linux-generic: pktio: classification error handling fixes for loop linux-generic: pktio: don't return packet after failed enqueue .../linux-ge

Re: [lng-odp] classification questions

Bala, any opinions? On 25/04/16 18:51, Zoltan Kiss wrote: Hi, This is what we talked about at the arch call, I summarized it up what I've found, so hopefully it'll be easier for you to comment: Socket, socket_mmap and dpdk pktio uses _odp_packet_cls_enq(), it can have 3 differe

Re: [lng-odp] [RFC PATCH] linux-dpdk: add --enable-inlines configure option

s how do you use that setting in your source files ... On 22/04/16 19:11, Zoltan Kiss wrote: In order to have a fixed ABI, we need each function to be in the library file. But that hurts performance a bit, as small accessor functions couldn't be inlined. To have the ability to use both,

[lng-odp] classification questions

Hi, This is what we talked about at the arch call, I summarized it up what I've found, so hopefully it'll be easier for you to comment: Socket, socket_mmap and dpdk pktio uses _odp_packet_cls_enq(), it can have 3 different returns: - pktio_select_cos(), odp_packet_alloc() or odp_packet_copydata

Re: [lng-odp] [RFC PATCH] linux-dpdk: add --enable-inlines configure option

nt an umbrella --enable-abi option that turns on all sub-options (whatever they may be) to achieve ABI mode? On Mon, Apr 25, 2016 at 6:39 AM, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: Originally I planned to have something like "-DINLINES=0/1", but that would for

Re: [lng-odp] [RFC PATCH] linux-dpdk: add --enable-inlines configure option

Originally I planned to have something like "-DINLINES=0/1", but that would force the application to define this macro in any case. Creating a config.h would be a way to overcome this maybe. On 22/04/16 19:11, Zoltan

Re: [lng-odp] [PATCH v2] linux-generic: classification: fix error checking _odp_packet_classifier()

On 22/04/16 17:22, Bala Manoharan wrote: On 22 April 2016 at 20:28, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: Hi Bala, While looking at this, I realized I don't understand something: we pass the received packets to classification, and it enqueues them to

[lng-odp] [RFC PATCH] linux-dpdk: add --enable-inlines configure option

#define INLINES" before including odp.h Signed-off-by: Zoltan Kiss --- configure.ac | 9 +++ platform/linux-dpdk/Makefile.am| 1 + platform/linux-dpdk/include/odp/api/packet.h | 59 +-- .../linux-dpdk/include/odp/

Re: [lng-odp] [PATCH v2] linux-generic: classification: fix error checking _odp_packet_classifier()

from the caller of odp_pktin_recv(): it can't actually use those handles, as those packets are on a queue already. Am I missing something? Or should we mention this in the function description? Regards, Zoltan On 13/04/16 15:20, Zoltan Kiss wrote: diff --git a/platform/linux-generic/pk

[lng-odp] [PATCH v2] validation: pktio: remove checks from stats test

c out on the newly created interfaces. It won't be visible for ODP when going out, but coming in it will increase the counters. This breaks the test on ODP-DPDK, unnecessarily. On ODP-Linux it does not, because it checks the system level statistics, not the ODP level ones. Signed-off-by: Z

Re: [lng-odp] [PATCH] validation: pktio: remove octet check from stats test

On 15/04/16 14:45, Maxim Uvarov wrote: On 04/05/16 19:16, Zoltan Kiss wrote: This test sets up two interface and connect them to each other, so in theory these two numbers should be the same. However when you use a pktio which doesn't have full control of the interface, it could happen

[lng-odp] [API-NEXT PATCH] api: queue: remove const qualifier of passed down event array

As the description says, normally this function consumes the events, which means there is no point to expect the array to stay intact. Other functions like odp_pktout_send() doesn't do that as well. Signed-off-by: Zoltan Kiss --- diff --git a/include/odp/api/spec/queue.h b/include/odp/api

Re: [lng-odp] [PATCH v2] linux-generic: classification: fix error checking _odp_packet_classifier()

Maxim, what's the planned time to apply this? On 15/04/16 06:31, Bala Manoharan wrote: Reviewed-by: Balasubramanian Manoharan mailto:bala.manoha...@linaro.org>> On 13 April 2016 at 19:50, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: In case of error the 'pkt&

[lng-odp] odp_cpumask_default_worker() and availability of the returned CPU's

Hi, A quick question: my understanding is that odp_cpumask_default_worker() returns the available CPU's, but it doesn't guarantee that when you start your thread's/processes they will be still available, right? It should be the applications responsibility to not start more threads than it wan

Re: [lng-odp] odp_queue_enq_multi() question

On 18/04/16 15:47, Bill Fischofer wrote: On Mon, Apr 18, 2016 at 9:25 AM, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: Hi, The second parameter of this function is an array pointer for the events we want to send: int odp_queue_enq_multi(odp_queue_t queue,

[lng-odp] odp_queue_enq_multi() question

Hi, The second parameter of this function is an array pointer for the events we want to send: int odp_queue_enq_multi(odp_queue_t queue, const odp_event_t events[], int num); I wonder if that const qualifier makes sense there. As the description says, normally the events are consumed by th

Re: [lng-odp] [lng-odp-ovs] error while configuring ODP-OVS for v1.6 and v1.8 ./configure --with-odp, cannot find ODP headers

4/16 14:18, Maxim Uvarov wrote: On 04/18/16 15:50, Zoltan Kiss wrote: Hi, I've tried this out as well, and it fails for me too. I've used the following configure command: ./configure--with-odp=[installdir] --with-odp-platform=linux-generic LDFLAGS="$(LDFLAGS) -L$(REPOS)/[install

Re: [lng-odp] [lng-odp-ovs] error while configuring ODP-OVS for v1.6 and v1.8 ./configure --with-odp, cannot find ODP headers

Hi, I've tried this out as well, and it fails for me too. I've used the following configure command: ./configure--with-odp=[installdir] --with-odp-platform=linux-generic LDFLAGS="$(LDFLAGS) -L$(REPOS)/[installdir]/lib/ -lpcap -lm The problem is that libodp.a is in /lib/.libs, not in a direc

Re: [lng-odp] [PATCH] validation: pktio: remove octet check from stats test

Ping On 06/04/16 11:43, Zoltan Kiss wrote: On 05/04/16 17:16, Zoltan Kiss wrote: This test sets up two interface and connect them to each other, so in s/interface/interfaces/ Maxim, could you fix that when commit? (assuming the patch is OK) And an another note: Maxim told me when I

Re: [lng-odp] [PATCH 1/2] linux-generic: classification: fix error checking _odp_packet_classifier()

Hi, I've sent a new patch, I just wanted to mention that I couldn't find this behavior described in user-guide-cls.adoc Zoli On 13/04/16 13:40, Bala Manoharan wrote: The classification module drops a packet only on specific cases either on error CoS or when the packet

[lng-odp] [PATCH v2] linux-generic: classification: fix error checking _odp_packet_classifier()

In case of error the 'pkt' should be released by the calling function. Currently loopback gives it back to the receiver and report it as success in the stats. Signed-off-by: Zoltan Kiss --- v2: handle release in caller instead, and adjust stats. platform/linux-generic/odp_classi

Re: [lng-odp] [PATCH 1/2] linux-generic: classification: fix error checking _odp_packet_classifier()

On 13/04/16 13:40, Bala Manoharan wrote: Regards, Bala On 12 April 2016 at 17:34, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: On 11/04/16 05:01, Bala Manoharan wrote: Hi, Comments inline... Regards, Bala On 8 April 2016 at

Re: [lng-odp] [PATCH 1/2] linux-generic: classification: fix error checking _odp_packet_classifier()

The original behavior of loopback_recv() is incorrect then, because it gives back the failed packets to the caller in pkts[], some of the might had been already released. On 11/04/16 05:01, Bala Manoharan wrote: int packet_classifier(odp_pktio_t pktio, odp_packet_t pkt) diff --git a/platform

Re: [lng-odp] [PATCH 1/2] linux-generic: classification: fix error checking _odp_packet_classifier()

On 11/04/16 05:01, Bala Manoharan wrote: Hi, Comments inline... Regards, Bala On 8 April 2016 at 00:23, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: The callers are only interested whether there is a CoS or not. It is not an error if there isn't, so it has a posi

[lng-odp] [PATCH 2/2] linux-generic: classification: remove unused code

This function is not referred. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_classification_internal.h | 9 ++--- platform/linux-generic/odp_classification.c | 10 -- 2 files changed, 2 insertions(+), 17 deletions(-) diff --git a/platform/linux

[lng-odp] [PATCH 1/2] linux-generic: classification: fix error checking _odp_packet_classifier()

The callers are only interested whether there is a CoS or not. It is not an error if there isn't, so it has a positive return value. Any other case needs to release the packet, which is not handled after calling queue_enq(). Signed-off-by: Zoltan Kiss --- platform/linux-ge

[lng-odp] [PATCH 0/2] Minor classification fixes

The first one prevents memory leak/double-free situations though. Zoltan Kiss (2): linux-generic: classification: fix error checking _odp_packet_classifier() linux-generic: classification: remove unused code .../linux-generic/include/odp_classification_internal.h | 9

Re: [lng-odp] [PATCH] linux-generic: packet_io: fix state handling

Ping, I think it only waits for merge On 03/04/16 20:41, Bill Fischofer wrote: On Fri, Apr 1, 2016 at 11:44 AM, Zoltan Kiss mailto:zoltan.k...@linaro.org>> wrote: The current two state doesn't tell odp_pktio_close() whether the device has ever been started, which is i

Re: [lng-odp] [PATCH] validation: pktio: remove octet check from stats test

On 05/04/16 17:16, Zoltan Kiss wrote: This test sets up two interface and connect them to each other, so in s/interface/interfaces/ Maxim, could you fix that when commit? (assuming the patch is OK) And an another note: Maxim told me when I brought this up first to disable stuff, and I

[lng-odp] [PATCH] validation: pktio: remove octet check from stats test

affic out on the newly created interfaces. It won't be visible for ODP when going out, but coming in it will increase the counters. This breaks the test on ODP-DPDK, unnecessarily. On ODP-Linux it does not, because it checks the system level statistics, not the ODP level ones. Signed-off-by: Z

Re: [lng-odp] flush_in_queues() on devices never started

e: // Flush if state STOPPED, don’t flush if only OPENED // Error check if state is STARTED (interface is active). if (entry->s.state == STATE_STOPPED) flush_in_queues(entry); -Petri -Original Message- From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro

[lng-odp] [PATCH] linux-generic: packet_io: fix state handling

alid for configuration, so instead of "== STATE_STOPPED", check "!= STATE_STARTED". Signed-off-by: Zoltan Kiss --- diff --git a/platform/linux-generic/include/odp_packet_io_internal.h b/platform/linux-generic/include/odp_packet_io_internal.h index 7636768..cca5c39 100644

Re: [lng-odp] flush_in_queues() on devices never started

I forgot to Cc the list On 31/03/16 17:03, Zoltan Kiss wrote: Hi, I ran into a problem when pktio_test_pktin_queue_config_direct() plays with queue config then call close. That calls flush_in_queues(), which tries to receive on a device which was never started, therefore run into a segfault

Re: [lng-odp] [PATCH] validation: timer: don't access non-existing timers

Ping On 18/03/16 17:25, Zoltan Kiss wrote: Since e5c85d3f "validation: timer: handle early exhaustion of pool" the workers can handle if object caches retain packets, but with enough threads it can happen that a late starting thread won't be able to allocate any. This for loop s

Re: [lng-odp] [PATCH v2 0/2] packet_flags changes

Ping, this patch is a prereq for a fix in ODP-DPDK On 18/03/16 17:20, Zoltan Kiss wrote: These have to be applied in order, otherwise the code is not dependent on each other. v2: remove "odp_" from the internal function names Zoltan Kiss (2): linux-generic: packet_flags: use ac

[lng-odp] [PATCH v2 0/2] packet_flags changes

These have to be applied in order, otherwise the code is not dependent on each other. v2: remove "odp_" from the internal function names Zoltan Kiss (2): linux-generic: packet_flags: use accessors to modify eth and l2 flag linux-generic: packet: don't look for L2 header if

[lng-odp] [PATCH v2 2/2] linux-generic: packet: don't look for L2 header if there isn't any

The L2 offset functions should consider the L2 flag: return negative answer if there isn't any, and implicitly set it when offset is set. E.g. user created packets don't have L2 headers immediately. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_packet_inte

Re: [lng-odp] [PATCH 2/2] linux-generic: packet: don't look for L2 header if there isn't any

My bad, I read those comments made for the other patch but I haven't applied them on this one. On 18/03/16 12:07, Maxim Uvarov wrote: On 03/18/16 14:39, Zoltan Kiss wrote: Ping, added Bill ping for what? There are comments to rename functions. Maxim. On 11/03/16 07:03, Zoltan

[lng-odp] [PATCH v2 1/2] linux-generic: packet_flags: use accessors to modify eth and l2 flag

This makes it possible for other implementations like ODP-DPDK to reuse classification code while using a different packet API. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_classification_inlines.h | 2 +- platform/linux-generic/include/odp_packet_internal.h| 10

Re: [lng-odp] [PATCH] validation: timer: don't access non-existing timers

Note, this is just a resend with a different set of CC recipients On 18/03/16 17:25, Zoltan Kiss wrote: Since e5c85d3f "validation: timer: handle early exhaustion of pool" the workers can handle if object caches retain packets, but with enough threads it can happen that a late start

Re: [lng-odp] [PATCH 2/2] linux-generic: packet: don't look for L2 header if there isn't any

Ping, added Bill On 11/03/16 07:03, Zoltan Kiss wrote: Ping On 03/03/16 03:05, Zoltan Kiss wrote: The L2 offset functions should consider the L2 flag: return negative answer if there isn't any, and implicitly set it when offset is set. E.g. user created packets don't have

[lng-odp] [PATCH] validation: timer: don't access non-existing timers

access tt[0].ev. Signed-off-by: Zoltan Kiss --- test/validation/timer/timer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/validation/timer/timer.c b/test/validation/timer/timer.c index 5854b32..7b76dbf 100644 --- a/test/validation/timer/timer.c +++ b/test/validatio

[lng-odp] [PATCH] validation: pktio: add more debug log to pktio_pkt_seq()

Signed-off-by: Zoltan Kiss --- test/validation/pktio/pktio.c | 26 +- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/test/validation/pktio/pktio.c b/test/validation/pktio/pktio.c index 9443cd2..cb403a6 100644 --- a/test/validation/pktio/pktio.c +++ b/test

Re: [lng-odp] [PATCH 2/2] linux-generic: packet: don't look for L2 header if there isn't any

Ping On 03/03/16 03:05, Zoltan Kiss wrote: The L2 offset functions should consider the L2 flag: return negative answer if there isn't any, and implicitly set it when offset is set. E.g. user created packets don't have L2 headers immediately. Signed-off-by: Zoltan Kiss --- plat

Re: [lng-odp] [PATCH 1/2] linux-generic: packet_flags: use accessors to modify eth and l2 flag

Ping On 03/03/16 03:05, Zoltan Kiss wrote: This makes it possible for other implementations like ODP-DPDK to reuse classification code while using a different packet API. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_classification_inlines.h | 2 +- platform/linux

Re: [lng-odp] [PATCH 1/2] linux-generic: packet_flags: use accessors to modify eth and l2 flag

Ping On 03/03/16 03:05, Zoltan Kiss wrote: This makes it possible for other implementations like ODP-DPDK to reuse classification code while using a different packet API. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_classification_inlines.h | 2 +- platform/linux

[lng-odp] [PATCH] validation: timer: don't access non-existing timers

access tt[0].ev. Signed-off-by: Zoltan Kiss --- test/validation/timer/timer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/validation/timer/timer.c b/test/validation/timer/timer.c index 004670a..85a5061 100644 --- a/test/validation/timer/timer.c +++ b/test/validatio

[lng-odp] [PATCH] validation: pktio: add more debug log to pktio_pkt_seq()

Signed-off-by: Zoltan Kiss --- test/validation/pktio/pktio.c | 26 +- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/test/validation/pktio/pktio.c b/test/validation/pktio/pktio.c index a5f25da..084e5cd 100644 --- a/test/validation/pktio/pktio.c +++ b/test

[lng-odp] [PATCH 2/2] linux-generic: packet: don't look for L2 header if there isn't any

The L2 offset functions should consider the L2 flag: return negative answer if there isn't any, and implicitly set it when offset is set. E.g. user created packets don't have L2 headers immediately. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_packet_inte

[lng-odp] [PATCH 0/2] packet_flags changes

These have to be applied in order, otherwise the code is not dependent on each other. Zoltan Kiss (2) linux-generic: packet_flags: use accessors to modify eth and l2 flag linux-generic: packet: don't look for L2 header if there isn't any

[lng-odp] [PATCH 1/2] linux-generic: packet_flags: use accessors to modify eth and l2 flag

This makes it possible for other implementations like ODP-DPDK to reuse classification code while using a different packet API. Signed-off-by: Zoltan Kiss --- platform/linux-generic/include/odp_classification_inlines.h | 2 +- platform/linux-generic/include/odp_packet_internal.h| 10

Re: [lng-odp] pktio stats counters

2/16 12:40, Zoltan Kiss wrote: Yes, I've already have that, and some more switches, like "-multicast", but it's impossible to predict all the services which will run on that interface. Zoli On 19/02/16 08:45, Maxim Uvarov wrote: In pktio_env for linux-generic we use disable

Re: [lng-odp] pktio stats counters

In that case there is not dhcp and other traffic on veth. Will it work for you? Maxim. On 02/18/16 21:06, Zoltan Kiss wrote: Hi, On ODP-DPDK this assert fails every now and then: https://git.linaro.org/lng/odp.git/blob/refs/heads/api-next:/test/validation/pktio/pktio.c#l1193 The reason is t

Re: [lng-odp] [API-NEXT PATCH v4 13/13] linux-generic: dpdk: close resources in odp_pktio_close()

On 25/02/16 15:53, Zoltan Kiss wrote: pktio_entry already has a 'state' field, why not using that? Sorry for the slow reply. The state field in pktio_entry is modified outside dpdk pktio code, so it cannot be used to track if rte_eth_dev_start() has been actually

Re: [lng-odp] [API-NEXT PATCH v4 13/13] linux-generic: dpdk: close resources in odp_pktio_close()

On 25/02/16 07:22, Elo, Matias (Nokia - FI/Espoo) wrote: -Original Message- From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro.org] Sent: Wednesday, February 24, 2016 9:10 PM To: Elo, Matias (Nokia - FI/Espoo) ; lng- o...@lists.linaro.org Subject: Re: [lng-odp] [API-NEXT PATCH v4 13

Re: [lng-odp] [API-NEXT PATCH v4 13/13] linux-generic: dpdk: close resources in odp_pktio_close()

On 24/02/16 07:36, Elo, Matias (Nokia - FI/Espoo) wrote: -Original Message- From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro.org] Sent: Friday, February 19, 2016 6:15 PM To: Elo, Matias (Nokia - FI/Espoo) ; lng- o...@lists.linaro.org Subject: Re: [lng-odp] [API-NEXT PATCH v4 13/13

Re: [lng-odp] [API-NEXT PATCH v4 10/13] linux-generic: dpdk: add odp_pktio_input_queues_config()

On 22/02/16 09:07, Elo, Matias (Nokia - FI/Espoo) wrote: -Original Message- From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro.org] Sent: Friday, February 19, 2016 6:13 PM To: Elo, Matias (Nokia - FI/Espoo) ; lng- o...@lists.linaro.org Cc: Savolainen, Petri (Nokia - FI/Espoo) Subject: Re

[lng-odp] locking in queue recv/send

Hi, I've noticed that "struct pktio_entry" locks 'rxl' and 'txl' are not grabbed when you call odp_pktio_recv/send_queue() functions, only when you have a single queue, which goes to the old codepath. This is particularly dangerous when doing close, but other cases might have trouble with it

Re: [lng-odp] [API-NEXT PATCH v4 13/13] linux-generic: dpdk: close resources in odp_pktio_close()

pktio_entry already has a 'state' field, why not using that? On 15/02/16 10:50, Matias Elo wrote: Free/close open resources in odp_pktio_close(). Reviewed-by: Petri Savolainen Signed-off-by: Matias Elo --- platform/linux-generic/include/odp_packet_dpdk.h | 1 + platform/linux-generic/pkti

Re: [lng-odp] [API-NEXT PATCH v4 10/13] linux-generic: dpdk: add odp_pktio_input_queues_config()

Hi, I'm implementing multiqueue for ODP-DPDK as well, where I'm taking a lot of your code, so finally I'm doing a bigger review of this code: On 15/02/16 10:49, Matias Elo wrote: @@ -309,6 +335,29 @@ static int odp_dpdk_pktio_init_local(void) return 0; } +static int dpdk_input_queu

[lng-odp] pktio stats counters

, which seems not to be able to catch all the outgoing packets. I guess on linux-generic we use the kernel stats in the unit tests which can do that. Any idea how to solve this problem? Regards, Zoltan Kiss ___ lng-odp mailing list lng-odp

Re: [lng-odp] scheduler_test_wait_time() tolerance

er platform tuning is required but we dont mention it ? It is not really a platform dependent stuff, this test is inherently unreliable, but we don't have a better idea, and adjusting the tolerance is just good enough. On 17 February 2016 at 10:51, Zoltan Kiss

Re: [lng-odp] [API-NEXT PATCH] api: time: fix typo for cmp function

Signed-off-by: Zoltan Kiss On 17/02/16 15:34, Ivan Khoronzhuk wrote: Signed-off-by: Ivan Khoronzhuk --- include/odp/api/time.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/include/odp/api/time.h b/include/odp/api/time.h index efc5478..85692ec 100644 --- a/include

  1   2   3   4   5   >