v5 was sent. It's better to continue discussion there.
That changes will break pdf creation and everybody agree with it?
Maxim.
On 01/26/2016 00:03, Bill Fischofer wrote:
This was targeted at api-next, however it can be applied there and
promoted to master as part of ODP v1.7 or before as
Avoid seg. fault if invalid handle provided to queue destroy.
Reviewed-by: Zoltan Kiss
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/odp_queue.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/platform/linux-generic/odp_queue.c
Signed-off-by: Nicolas Morey-Chaisemartin
---
test/validation/init/init.c | 16
test/validation/init/init.h | 1 +
2 files changed, 17 insertions(+)
diff --git a/test/validation/init/init.c b/test/validation/init/init.c
index 62bd75c..f7424a9 100644
---
Signed-off-by: Nicolas Morey-Chaisemartin
---
I'm not very happy with a odp_platform_init_init but odp_platform_init would be
misleading for users
and odp_platform_param_init wouldn't fit ODP semantics.
If someone has a better idea, I'm all ears.
include/odp/api/init.h
Hm, yes, that is reasonable patch.
Will send it to mainline list.
Maxim.
On 01/25/2016 18:03, Zoltan Kiss wrote:
Hi Maxim,
I've spotted this patch which went in to ODP-DPDK instead of mainline
ODP. I think it would be better to push it there as well. You can add my
Reviewed-by: Zoltan
I think those are just early definitions and nobody paid attention ever since.
I think you are right, those should be prefixed also to avoid name space
clashes.
The __bitwise definition is needed by sparse tool, but likely the type name can
be anything. Maybe something like this:
odp_u16le_t,
On 20 January 2016 at 10:19, Petri Savolainen
wrote:
> Completed additions to atomic API. Added exchange operation in relaxed
> memory
> model. Added CAS operations in rel and acq_rel memory order. Added 64 bit
> versions of all non-relaxed operations defined so far.
On 01/26/2016 11:45, Savolainen, Petri (Nokia - FI/Espoo) wrote:
Yes it is. I'm asking for the next time. Where tags should be copy-pasted into
the cover letter?
-Petri
It's better to add to patches. But in cover letter it's also ok.
Maxim.
-Original Message-
From: EXT Maxim
On 26 January 2016 at 22:45, Mike Holmes wrote:
> Inline comment on include structure
>
> On 26 January 2016 at 13:04, Christophe Milard <
> christophe.mil...@linaro.org> wrote:
>
>> This patch series implements a new structure in the ODP repo so that new
>> interfaces
The tag was there as soon as possible. Stuart reviewed v3 and I added the
review-by tag to v4 comments. It was about 10 lines down on the cover letter
comments. Should it be copied as the first line? Or in to the latest patch
version comment (v4 in this case)?
-Petri
> -Original
On 01/26/2016 11:08, Savolainen, Petri (Nokia - FI/Espoo) wrote:
The tag was there as soon as possible. Stuart reviewed v3 and I added the
review-by tag to v4 comments. It was about 10 lines down on the cover letter
comments. Should it be copied as the first line? Or in to the latest patch
Yes it is. I'm asking for the next time. Where tags should be copy-pasted into
the cover letter?
-Petri
> -Original Message-
> From: EXT Maxim Uvarov [mailto:maxim.uva...@linaro.org]
> Sent: Tuesday, January 26, 2016 10:19 AM
> To: Savolainen, Petri (Nokia - FI/Espoo)
> Cc:
Hi Ola,
Would you mind taking a look at this?
Zoli
On 22/01/16 18:24, Zoltan Kiss wrote:
As per-thread caches might retain some elements, no particular thread
should assume that a certain amount of elements are available at any
time.
Also, we can't reliable check the high watermark anymore.
On Fri, Jan 22, 2016 at 03:45:40PM +0300, Maxim Uvarov wrote:
> Signed-off-by: Maxim Uvarov
> ---
> test/validation/pktio/pktio.c | 166
> ++
> test/validation/pktio/pktio.h | 2 +
> 2 files changed, 168 insertions(+)
>
> diff
Reviewed-by: Petri Savolainen
> -Original Message-
> From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro.org]
> Sent: Friday, January 22, 2016 8:24 PM
> To: lng-odp@lists.linaro.org
> Cc: ola.liljed...@arm.com; Savolainen, Petri (Nokia - FI/Espoo)
> Subject:
On 26 January 2016 at 14:59, Savolainen, Petri (Nokia - FI/Espoo) <
petri.savolai...@nokia.com> wrote:
> Reviewed-by: Petri Savolainen
>
>
> > -Original Message-
> > From: EXT Zoltan Kiss [mailto:zoltan.k...@linaro.org]
> > Sent: Friday, January 22, 2016 8:24
This would also allow HW blocks (like packet input) to cache some free buffers
(for incoming packets). It would be hard to guarantee that any given time a
single thread would be able to allocate (steal) all pre-allocated buffers from
all pool users. For example, could you steal all packet
Signed-off-by: Maxim Uvarov
---
test/validation/timer/timer.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/test/validation/timer/timer.c b/test/validation/timer/timer.c
index 5d89700..0bd67fb 100644
--- a/test/validation/timer/timer.c
+++
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/Makefile.am | 2 +
.../linux-generic/include/odp_packet_io_internal.h | 15 ++
platform/linux-generic/include/odp_packet_socket.h | 5 +
platform/linux-generic/odp_packet_io.c | 53
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/pktio/pcap.c | 28 +++-
1 file changed, 27 insertions(+), 1 deletion(-)
diff --git a/platform/linux-generic/pktio/pcap.c
b/platform/linux-generic/pktio/pcap.c
index 7da36aa..c22cce0 100644
v8: - fix validation test to support unsupported counters :)
v7: - fix stats struction init if stats not supported in stats() call.
v6: - reorganize stats code to support different sockets and not fail on
missing ethtool
strings.
v5: - update to the latest api-next. Fix logic to not fail if
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/pktio/netmap.c | 44 +++
1 file changed, 44 insertions(+)
diff --git a/platform/linux-generic/pktio/netmap.c
b/platform/linux-generic/pktio/netmap.c
index 2a14861..16898ec 100644
---
Signed-off-by: Maxim Uvarov
---
test/validation/pktio/pktio.c | 170 ++
test/validation/pktio/pktio.h | 2 +
2 files changed, 172 insertions(+)
diff --git a/test/validation/pktio/pktio.c b/test/validation/pktio/pktio.c
index
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/pktio/loop.c | 42 ++---
1 file changed, 39 insertions(+), 3 deletions(-)
diff --git a/platform/linux-generic/pktio/loop.c
b/platform/linux-generic/pktio/loop.c
index
On 26.01.16 17:46, Maxim Uvarov wrote:
Signed-off-by: Maxim Uvarov
Reviewed-by: Ivan Khoronzhuk
---
test/validation/timer/timer.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git
On 26 January 2016 at 14:32, Zoltan Kiss wrote:
> Hi Ola,
>
> Would you mind taking a look at this?
>
> Zoli
>
> On 22/01/16 18:24, Zoltan Kiss wrote:
>
>> As per-thread caches might retain some elements, no particular thread
>> should assume that a certain amount of
On Tue, Jan 26, 2016 at 07:00:37PM +0300, Maxim Uvarov wrote:
> v8: - fix validation test to support unsupported counters :)
> v7: - fix stats struction init if stats not supported in stats() call.
> v6: - reorganize stats code to support different sockets and not fail on
> missing ethtool
>
On 26 January 2016 at 06:25, Maxim Uvarov wrote:
> v5 was sent. It's better to continue discussion there.
>
> That changes will break pdf creation and everybody agree with it?
>
Yes, I agree
>
> Maxim.
>
>
> On 01/26/2016 00:03, Bill Fischofer wrote:
>
>> This was
The tests-validation.env file lists the test binaries/scripts to be run
during post-install testing, but the Makefile dependencies are wrong so
it gets created once and never updated. This means the following
sequence doesn't actually peform testing as expected -
./configure
make
./configure
As discussed on the ODP architecture call. It could be useful to be able to
know how much you need to over-provision a pool in order to be guaranteed
that you can allocate N elements from that pool (allowing for an
implementation dependent amount of per-CPU caching etc), regardless of how
that
This patch series implements a new structure in the ODP repo so that new
interfaces can be more easily added to ODP
Saddly changing the current structure seems needed as just adding to it
results in a quite fuzzy file organisation:
Today:
include/odp/api is the reference for the ODP API
odp.h actually describes the applicatiion interface of ODP and is
logicaly moved in the ODP include space to api.h.
Applications now includes .
This simplify the addition of other ODP interfaces (e.g. drivers may
include odp/drv.h in the future.)
This patch has been done by a script which is
Signed-off-by: Christophe Milard
---
doc/implementers-guide/implementers-guide.adoc | 74 +++---
doc/users-guide/users-guide.adoc | 15 +++---
2 files changed, 53 insertions(+), 36 deletions(-)
diff --git
Before sending my reviewed-by I started working on implementing these
(along with the tests). The flags themselves are straightforward, however
the parse definitions are somewhat complicated and we may want to simplify
things a bit to capture what applications are really looking for.
First, we
On 26 January 2016 at 11:39, Stuart Haslam wrote:
> The tests-validation.env file lists the test binaries/scripts to be run
> during post-install testing, but the Makefile dependencies are wrong so
> it gets created once and never updated. This means the following
>
Signed-off-by: Mike Holmes
---
doc/application-api-guide/doxygen.cfg | 2 +-
doc/images/ODP-Logo-HQ.png| Bin 5736 -> 0 bytes
doc/images/ODP-Logo-HQ.svg| 209 ++
3 files changed, 210 insertions(+), 1 deletion(-)
Signed-off-by: Mike Holmes
---
doc/application-api-guide/doxygen.cfg | 1 +
1 file changed, 1 insertion(+)
diff --git a/doc/application-api-guide/doxygen.cfg
b/doc/application-api-guide/doxygen.cfg
index 909f0ce..75006e2 100644
---
On 20 January 2016 at 19:52, Mike Holmes wrote:
> Currently docs will build the images regardless of the setting for
> enable-user-docs at configure time.
> The image generation should be dependent on actually building docs so
> that systems without the image tools
Inline comment on include structure
On 26 January 2016 at 13:04, Christophe Milard wrote:
> This patch series implements a new structure in the ODP repo so that new
> interfaces can be more easily added to ODP
>
> Saddly changing the current structure seems needed
Signed-off-by: Bill Fischofer
---
.../linux-generic/include/odp_packet_internal.h| 4 +++
platform/linux-generic/odp_packet_flags.c | 40 ++
2 files changed, 44 insertions(+)
diff --git
This patch series adds the implementation, CUnit validation tests, and
documentation for the addition of the broadcast and multicast packet metadata
APIs.
This series must be applied following patch
http://patches.opendataplane.org/patch/4653/
Bill Fischofer (5):
linux-generic: packet:
Signed-off-by: Bill Fischofer
---
platform/linux-generic/odp_packet.c | 41 +
1 file changed, 28 insertions(+), 13 deletions(-)
diff --git a/platform/linux-generic/odp_packet.c
b/platform/linux-generic/odp_packet.c
index
Signed-off-by: Bill Fischofer
---
test/validation/packet/packet.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/test/validation/packet/packet.c b/test/validation/packet/packet.c
index 2467dec..c5e2356 100644
---
Signed-off-by: Bill Fischofer
---
doc/users-guide/users-guide.adoc | 121 +++
1 file changed, 121 insertions(+)
diff --git a/doc/users-guide/users-guide.adoc b/doc/users-guide/users-guide.adoc
index b3c08a1..513f383 100644
---
the file:
odp/platform/linux-generic/include/odp/plat/byteorder_types.h defines:
typedef uint16_t __odp_bitwise uint16le_t;
typedef uint16_t __odp_bitwise uint16be_t;
typedef uint32_t __odp_bitwise uint32le_t;
typedef uint32_t __odp_bitwise uint32be_t;
typedef uint64_t __odp_bitwise
Petri, I think it's a quick one to review, as we already discussed this
on an arch call. If you agree, please provide a quick Reviewed-by.
Zoli
On 22/01/16 18:24, Zoltan Kiss wrote:
This addition explicitly loses the requirement that a single thread should
be able to allocate the whole pool
Thanks. Maxim could you first try to merge and resolve conflicts. Those should
be quite simple to resolve: additional atomic validation tests into
validation/atomic/atomic.c instead of old
validation/synchronizers/synchronizers.c.
-Petri
From: EXT Ola Liljedahl
47 matches
Mail list logo