Re: [lng-odp] [PATCH 00/10] Pktio checks

So, critical bug fixes from master side are not visible to api-next until a month or so? Most bug fixes should apply directly to api-next as well, since those are implementation/test code/build system fixes (not api changing fixes). Some fixes may touch the same lines that api-next have changed

[lng-odp] [PATCHv5] helper : Fix UDP checksum computation

From: Grigore Ion This patch fixes the following problems: - checksum computation for LE platforms - checksum is computed in the CPU endianness. The returned result must be converted to the BE ordering when it is used to update the UDP checksum in a packet. - checksum computation for packets havi

Re: [lng-odp] [PATCHv4] helper : Fix UDP checksum computation

Resent with the fix inside. Thanks, Grig -Original Message- From: Ilya Maximets [mailto:i.maxim...@samsung.com] Sent: Wednesday, December 09, 2015 8:02 AM To: Grigore Ion-B17953 ; lng-odp@lists.linaro.org Subject: Re: [PATCHv4] helper : Fix UDP checksum computation It's a completely the

Re: [lng-odp] [PATCH 00/10] Pktio checks

On 12/09/2015 11:54, Savolainen, Petri (Nokia - FI/Espoo) wrote: So, critical bug fixes from master side are not visible to api-next until a month or so? Exactly. api-next was broken with TM. Until issue if fixed and branch operates as usual, there is no pull of critical bug and etc. http

Re: [lng-odp] api-next broken?

So the problem appears only in libc6-dev 2.21. No problem with libc6-dev 2.19 Diff is: /usr/include/time.h @@ -108,7 +108,7 @@ typedef __timer_t timer_t; #if (!defined __timespec_defined\ && ((defined _TIME_H\ - && (defined __USE_POSIX19

Re: [lng-odp] [PATCHv5] helper : Fix UDP checksum computation

Some comments below. On 09.12.2015 11:59, ion.grig...@freescale.com wrote: > From: Grigore Ion > > This patch fixes the following problems: > - checksum computation for LE platforms > - checksum is computed in the CPU endianness. The returned result > must be converted to the BE ordering when it

Re: [lng-odp] [API-NEXT PATCH v2] helper: add cuckoo hash implementation

On 12/09/2015 08:54, HePeng wrote: Let me further explain this. Cuckoo hash is a hash table, and as a hash table, it needs to store the key-value paris. For any input, e.g. a key ( a byte string with some size), a hash table needs to calculate the hash value from the key, find the right bucke

Re: [lng-odp] [API-NEXT PATCH v2] helper: add cuckoo hash implementation

On 9 December 2015 at 06:31, HePeng wrote: > > 在 2015年12月8日，下午9:34，Ola Liljedahl 写道： > > On 8 December 2015 at 12:42, Bill Fischofer > wrote: > >> This is an interesting topic. I'd like to discuss this a bit during >> today's ODP public call. >> >> I think the issue is that while a ring is a v

Re: [lng-odp] [API-NEXT PATCH v2] helper: add cuckoo hash implementation

On 9 December 2015 at 11:20, Maxim Uvarov wrote: > On 12/09/2015 08:54, HePeng wrote: > >> Let me further explain this. >> >> Cuckoo hash is a hash table, and as a hash table, it needs to store the >> key-value paris. >> For any input, e.g. a key ( a byte string with some size), a hash table >> n

Re: [lng-odp] [API-NEXT PATCH v3] api: atomic: added atomic_lock_free_u64

It patches API-NEXT branch. Did you try to apply it to master? Just tried it out, and it applies nicely. “[API-NEXT,v3,7/7] api: atomic: added 32 bit acquire and release” is in api-next. The idea is to add those operation-mem_model combinations, that are commonly used and make sense (from our

Re: [lng-odp] api-next broken?

I'm running vanilla Ubuntu 15.04 at current service. gcc shows: bill@Ubuntu15:~/linaro/timefix$ gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.

Re: [lng-odp] api-next broken?

yes, I'm installing 15.10 to VM to check that odp works there. Also it looks like in CI we need different containers to check most of common distributive versions. Maxim. On 12/09/2015 15:00, Bill Fischofer wrote: I'm running vanilla Ubuntu 15.04 at current service. gcc shows: bill@Ubuntu1

Re: [lng-odp] [PATCHv5] helper : Fix UDP checksum computation

See inline comments. -Original Message- From: Ilya Maximets [mailto:i.maxim...@samsung.com] Sent: Wednesday, December 09, 2015 12:04 PM To: Grigore Ion-B17953 ; lng-odp@lists.linaro.org Subject: Re: [PATCHv5] helper : Fix UDP checksum computation Some comments below. On 09.12.2015 11:59

Re: [lng-odp] [API-NEXT PATCH v3] api: atomic: added atomic_lock_free_u64

On 9 December 2015 at 11:58, Savolainen, Petri (Nokia - FI/Espoo) < petri.savolai...@nokia.com> wrote: > It patches API-NEXT branch. Did you try to apply it to master? Just tried > it out, and it applies nicely. > I get it now. The branch is actually called "api-next" (lower case), not API-NEXT. Y

Re: [lng-odp] [PATCHv5] helper : Fix UDP checksum computation

On 09.12.2015 15:26, Ion Grigore wrote: > See inline comments. > > -Original Message- > From: Ilya Maximets [mailto:i.maxim...@samsung.com] > Sent: Wednesday, December 09, 2015 12:04 PM > To: Grigore Ion-B17953 ; lng-odp@lists.linaro.org > Subject: Re: [PATCHv5] helper : Fix UDP checksu

Re: [lng-odp] [API-NEXT PATCH v2] helper: add cuckoo hash implementation

> 在 2015年12月9日，下午6:20，Maxim Uvarov 写道： > > On 12/09/2015 08:54, HePeng wrote: >> Let me further explain this. >> >> Cuckoo hash is a hash table, and as a hash table, it needs to store the >> key-value paris. >> For any input, e.g. a key ( a byte string with some size), a hash table >> needs

Re: [lng-odp] [PATCH] linux-generic: validation: run config tests

On 2015-11-13 15:35, Stuart Haslam wrote: > Add the new config tests to the TESTS list so that they get run via > "make check". > > Signed-off-by: Stuart Haslam Reviewed-and-Tested-by: Anders Roxell > --- > platform/linux-generic/test/Makefile.am | 1 + > 1 file changed, 1 insertion(+) > > d

Re: [lng-odp] [PATCH] linux-generic: validation: run config tests

Merged, Maxim. On 12/09/2015 17:58, Anders Roxell wrote: On 2015-11-13 15:35, Stuart Haslam wrote: Add the new config tests to the TESTS list so that they get run via "make check". Signed-off-by: Stuart Haslam Reviewed-and-Tested-by: Anders Roxell --- platform/linux-generic/test/Makefil

[lng-odp] [API-NEXT PATCHv3] api: pktio link

Signed-off-by: Maxim Uvarov --- include/odp/api/packet_io.h | 11 +++ 1 file changed, 11 insertions(+) diff --git a/include/odp/api/packet_io.h b/include/odp/api/packet_io.h index 443841e..0df46ba 100644 --- a/include/odp/api/packet_io.h +++ b/include/odp/api/packet_io.h @@ -673,6 +673,1

Re: [lng-odp] [API-NEXT PATCHv3] api: pktio link

On Wed, Dec 9, 2015 at 9:21 AM, Maxim Uvarov wrote: > Signed-off-by: Maxim Uvarov > --- > include/odp/api/packet_io.h | 11 +++ > 1 file changed, 11 insertions(+) > > diff --git a/include/odp/api/packet_io.h b/include/odp/api/packet_io.h > index 443841e..0df46ba 100644 > --- a/include/o

[lng-odp] [API-NEXT PATCH] linux-generic: time: remove posix bleed through on odp_time_t

The linux-generic implementation of odp_time_t makes use of POSIX APIs that are sensitive to the _POSIX_C_SOURCE level. Use an indirection mechanism so that these dependencies do not "bleed through" the ODP API. This means that ODP applications can be independent of _POSIX_C_SOURCE level. Signed-o

Re: [lng-odp] [API-NEXT PATCHv3] api: pktio link

On 12/09/2015 18:26, Bill Fischofer wrote: On Wed, Dec 9, 2015 at 9:21 AM, Maxim Uvarov > wrote: Signed-off-by: Maxim Uvarov mailto:maxim.uva...@linaro.org>> --- include/odp/api/packet_io.h | 11 +++ 1 file changed, 11 insertions(+)

[lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

No need to define odp_time_null() as it only initialize to 0 time. Remove it completely with needed to add doxygen comment for it. Signed-off-by: Maxim Uvarov --- platform/linux-generic/include/odp/plat/time_types.h | 4 +--- platform/linux-generic/odp_time.c| 7 +-- 2 fi

[lng-odp] [API-NEXT PATCH 2/2] linux-generic: time: use same type as returned

Signed-off-by: Maxim Uvarov --- platform/linux-generic/odp_time.c | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/platform/linux-generic/odp_time.c b/platform/linux-generic/odp_time.c index 1d374ca..500ba4b 100644 --- a/platform/linux-generic/odp_time.c +++ b/platform

Re: [lng-odp] [API-NEXT PATCHv3] api: pktio link

On 9 December 2015 at 17:09, Maxim Uvarov wrote: > On 12/09/2015 18:26, Bill Fischofer wrote: > >> >> >> On Wed, Dec 9, 2015 at 9:21 AM, Maxim Uvarov > > wrote: >> >> Signed-off-by: Maxim Uvarov > > >> --- >> include

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

This is proposing an API change. It's OK to change how odp_time_null() is implemented in linux-generic, but we can't arbitrarily remove this public API from the linux-generic implementation. On Wed, Dec 9, 2015 at 10:08 AM, Maxim Uvarov wrote: > No need to define odp_time_null() as it only init

Re: [lng-odp] [API-NEXT PATCH] linux-generic: time: remove posix bleed through on odp_time_t

On 9 December 2015 at 16:53, Bill Fischofer wrote: > The linux-generic implementation of odp_time_t makes use of POSIX > APIs that are sensitive to the _POSIX_C_SOURCE level. Use an indirection > mechanism so that these dependencies do not "bleed through" the ODP API. > This means that ODP applic

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

On 12/09/2015 19:26, Bill Fischofer wrote: This is proposing an API change. It's OK to change how odp_time_null() is implemented in linux-generic, but we can't arbitrarily remove this public API from the linux-generic implementation. But it's not defined in API headers: fgrep -r odp_time_null

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

Sorry, my mistake. I had assumed that using odp_time_null() rather than _odp_time_null() was because it was implementing an external API. However that does raise the question of whether odp_time_null() or ODP_TIME_NULL should be part of the external API. On Wed, Dec 9, 2015 at 10:29 AM, Maxim Uv

Re: [lng-odp] [API-NEXT PATCH 2/2] linux-generic: time: use same type as returned

My earlier patch http://patches.opendataplane.org/patch/4080/ deals with this in a more general manner. On Wed, Dec 9, 2015 at 10:08 AM, Maxim Uvarov wrote: > Signed-off-by: Maxim Uvarov > --- > platform/linux-generic/odp_time.c | 8 > 1 file changed, 4 insertions(+), 4 deletions(-) >

Re: [lng-odp] [PATCH] queue: fix memory corruption in reorder_enq()

Hi, On 08/12/15 02:43, Bill Fischofer wrote: As to your other question, reorder_tail exists because the reorder queue is a queue, however elements on it are sorted by order rather than arrival time. The tail is used because dequeues can dequeue multiple elements at a time. My problem is that t

Re: [lng-odp] [PATCH] queue: fix memory corruption in reorder_enq()

It's there for consistency with other queue structures. Are you looking to remove it because you want to save 8 bytes? That should be doable if needed. On Wed, Dec 9, 2015 at 10:54 AM, Zoltan Kiss wrote: > Hi, > > On 08/12/15 02:43, Bill Fischofer wrote: > >> As to your other question, reorder_

[lng-odp] [Bug 1937] New: CID 155811: Error handling issues odp_classification_common.c

https://bugs.linaro.org/show_bug.cgi?id=1937 Bug ID: 1937 Summary: CID 155811: Error handling issues odp_classification_common.c Product: OpenDataPlane - linux- generic reference Version: 1.5 Hardware: Other

Re: [lng-odp] [PATCH] queue: fix memory corruption in reorder_enq()

On 09/12/15 17:00, Bill Fischofer wrote: It's there for consistency with other queue structures. Are you looking to remove it because you want to save 8 bytes? That should be doable if needed. I don't mind the 8 bytes too much, though if it's absolutely unused then we should rather remove i

Re: [lng-odp] [API-NEXT PATCH 2/2] linux-generic: time: use same type as returned

On 12/09/2015 19:30, Bill Fischofer wrote: My earlier patch http://patches.opendataplane.org/patch/4080/ deals with this in a more general manner. yes, that patch is not needed now. Maxim. On Wed, Dec 9, 2015 at 10:08 AM, Maxim Uvarov > wrote: Signed-off

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

On 12/09/2015 19:33, Bill Fischofer wrote: Sorry, my mistake. I had assumed that using odp_time_null() rather than _odp_time_null() was because it was implementing an external API. However that does raise the question of whether odp_time_null() or ODP_TIME_NULL should be part of the external

Re: [lng-odp] [PATCH 00/10] Pktio checks

Can we merge [1] that was ready and reviewed last week with no objections Maxim Petri - can you review that patch and if there are changes needed I can update. [1] [PATCH] doc: process-guide: add release process On 9 December 2015 at 04:35, Maxim Uvarov wrote: > On 12/09/2015 11:54, Savolainen

Re: [lng-odp] [API-NEXT PATCH 2/2] linux-generic: time: use same type as returned

On 9 December 2015 at 19:22, Maxim Uvarov wrote: > On 12/09/2015 19:30, Bill Fischofer wrote: > >> My earlier patch http://patches.opendataplane.org/patch/4080/ deals with >> this in a more general manner. >> > > yes, that patch is not needed now. > I assume Bill meant that a more general solutio

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

On 09.12.15 20:24, Maxim Uvarov wrote: On 12/09/2015 19:33, Bill Fischofer wrote: Sorry, my mistake. I had assumed that using odp_time_null() rather than _odp_time_null() was because it was implementing an external API. However that does raise the question of whether odp_time_null() or ODP

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

On 09.12.15 18:33, Bill Fischofer wrote: Sorry, my mistake. I had assumed that using odp_time_null() rather than _odp_time_null() was because it was implementing an external API However that does raise the question of whether odp_time_null() or ODP_TIME_NULL should be part of the external A

Re: [lng-odp] [API-NEXT PATCH 1/2] api: linux-generic: no need to define odp_time_null()

On 09.12.15 18:29, Maxim Uvarov wrote: On 12/09/2015 19:26, Bill Fischofer wrote: This is proposing an API change. It's OK to change how odp_time_null() is implemented in linux-generic, but we can't arbitrarily remove this public API from the linux-generic implementation. But it's not def

Re: [lng-odp] [API-NEXT PATCH v2] helper: add cuckoo hash implementation

> 在 2015年12月9日，下午6:49，Ola Liljedahl 写道： > > On 9 December 2015 at 06:31, HePeng > wrote: > >> 在 2015年12月8日，下午9:34，Ola Liljedahl > > 写道： >> >> On 8 December 2015 at 12:42, Bill Fischofer > > wrote: >

Re: [lng-odp] [API-NEXT PATCH] linux-generic: time: remove posix bleed through on odp_time_t

Going to apply this patch to unblock builds in new distros. Any objections? Maxim. On 12/09/2015 19:25, Ola Liljedahl wrote: On 9 December 2015 at 16:53, Bill Fischofer > wrote: The linux-generic implementation of odp_time_t makes use of POSIX APIs tha

Re: [lng-odp] [PATCH] validation/config/Makefile: change install location

Merged, Thanks, Maxim. On 12/09/2015 02:18, Anders Roxell wrote: All tests should be installed under testdir variable and not in the bin directory. Signed-off-by: Anders Roxell --- test/validation/config/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/va