Hi all,
My name is Genís Riera Pérez and I use the ODP framework to develop some
apps, and I would want to ask you some questions about the output queues
and the priority level of ODP queues. I'm based on v1.1.0 API documentation.
So, basically my doubt is if there's an approach (by using the
Hi Genís,
There is a draft comprehensive set of ODP APIs for Egress Traffic
Management that has been worked on over the past few months that addresses
this area. It was the main focus are of the ODP Design Sprint we held a
couple of weeks ago. The APIs for this should be hitting this list very
On 06/30/15 15:52, Stuart Haslam wrote:
On Tue, Jun 30, 2015 at 03:20:34PM +0300, Maxim Uvarov wrote:
On 06/29/15 19:32, Nicolas Morey-Chaisemartin wrote:
I'm fine with this but why should it only be for this test and not all of them?
Nicolas, can you do required changes for your platform and
Ping.
On 06/18/2015 08:53 PM, Benoît Ganne wrote:
The application can now specify a packet offset
in PMR rules. This offset is absolute from the
frame start. It is used to extract the PMR value.
This is useful to support arbitrary backplane
protocols and extensions.
Signed-off-by: Benoît
Merged,
thanks,
Maxim.
On 07/01/15 11:19, Christophe Milard wrote:
thread tests now go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of validation: scheduler tests moved to platform side
Christophe Milard (5):
validation: own
From: Hongbo Zhang hongbo.zh...@linaro.org
v1 - v2 changes:
- separate original 1/3 patch to two
- add validation patch to test new APIs
- other minor updates upon review comments
v1 notes:
The current API of getting CPU frequency is really ambiguous.
CPU has its max frequency, and if enabled,
From: Hongbo Zhang hongbo.zh...@linaro.org
This patch adds test for the newly introduced per-CPU system APIs:
odp_sys_cpu_model_str_amp()
odp_sys_cpu_hz_amp()
odp_sys_cpu_hz_current()
Signed-off-by: Hongbo Zhang hongbo.zh...@linaro.org
---
test/validation/ver_abt_log_dbg/odp_system.c | 28
From: Hongbo Zhang hongbo.zh...@linaro.org
This patch add API to return the current frequency of the CPU on which
the thread is running.
Only x86 platform is now implemented, others are to be done.
Signed-off-by: Hongbo Zhang hongbo.zh...@linaro.org
---
include/odp/api/system_info.h
On 07/01/15 16:40, Nicolas Morey-Chaisemartin wrote:
Offset wrongly used num_tx_workers instead of num_rx_workers which caused
threads to be badly joined.
Without this patch:
$./test/performance/odp_pktio_perf -c 3
Segmentation fault (core dumped)
Nikolas, I can not reproduce that bug.
On Thu, Jul 02, 2015 at 11:21:33AM +0200, Nicolas Morey-Chaisemartin wrote:
It may not segfault depending on what is on your stack at this point and in
the pthread_t.
Depending on #RX #Tx or the opposit, it eithers join a thread twice, or
join an unitiliazed pthread_t.
On 07/02/2015
Merged!
Thanks,
Maxim.
On 07/02/15 12:32, Stuart Haslam wrote:
On Thu, Jul 02, 2015 at 11:21:33AM +0200, Nicolas Morey-Chaisemartin wrote:
It may not segfault depending on what is on your stack at this point and in the
pthread_t.
Depending on #RX #Tx or the opposit, it eithers join a thread
On Thu, Jul 02, 2015 at 10:47:44AM +0200, Nicolas Morey-Chaisemartin wrote:
Both linux/if.h and net/if.h were used but they are not necesseraly compatible
and it is not possible to include both in the same file.
Trying to understand the problem.. is it actually causing issues how it
is or are
On Wed, Jul 01, 2015 at 06:17:03PM +0200, Christophe Milard wrote:
Synchronizers testsnow go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of: cpumask tests moved to platform side
Christophe Milard (5):
validation: own main in
On 07/01/15 20:17, Mike Holmes wrote:
odp_timer fails periodically with segfaults in CI
https://bugs.linaro.org/show_bug.cgi?id=1615
You can replicate on my machine by running a load and then running
this a few times until it pops, I assume sometimes the CI build
machines are also loaded and
On 07/02/2015 12:23 PM, Stuart Haslam wrote:
On Thu, Jul 02, 2015 at 10:47:44AM +0200, Nicolas Morey-Chaisemartin wrote:
Both linux/if.h and net/if.h were used but they are not necesseraly
compatible
and it is not possible to include both in the same file.
Trying to understand the
From: Hongbo Zhang hongbo.zh...@linaro.org
For AMP system such as ARM big.LITTLE, cores are heterogeneous, and cpu_hz
for each core may be different too, so this patch changes the cpu_hz to
data array cpu_hz[] to contain data for each different core, while for the
common SMP system, we can simply
From: Hongbo Zhang hongbo.zh...@linaro.org
For AMP system such as ARM big.LITTLE, cores are heterogeneous, and the
model_str for each core may be different too, so this patch changes the
model_str to data array model_str[] to contain data for each different
core, while for the common SMP system,
From: Hongbo Zhang hongbo.zh...@linaro.org
In the previous x86 platform code, the variable is acquired from segment
cpu MHz of /proc/cpuinfo, this is the current CPU frequency when
acquired, and it may be scaled from time to time. But most of the use
cases when odp_sys_cpu_hz() is called, users
Both linux/if.h and net/if.h were used but they are not necesseraly compatible
and it is not possible to include both in the same file.
Signed-off-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
---
platform/linux-generic/include/odp_packet_io_internal.h | 4 ++--
That is the change which we discussed on Sprint.
Bala, Petri - no objection to include it?
Thank you,
Maxim.
On 07/02/15 13:40, Benoît Ganne wrote:
Ping.
On 06/18/2015 08:53 PM, Benoît Ganne wrote:
The application can now specify a packet offset
in PMR rules. This offset is absolute from
Merged,
Maxim.
On 07/02/15 13:22, Stuart Haslam wrote:
On Wed, Jul 01, 2015 at 06:17:03PM +0200, Christophe Milard wrote:
Synchronizers testsnow go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of: cpumask tests moved to platform
Signed-off-by: Mike Holmes mike.hol...@linaro.org
---
v2
Dont put a period at the end of lists
doc/odp.dox | 7 ---
doc/release.dox | 65 +
2 files changed, 65 insertions(+), 7 deletions(-)
create mode 100644 doc/release.dox
diff
It may not segfault depending on what is on your stack at this point and in the
pthread_t.
Depending on #RX #Tx or the opposit, it eithers join a thread twice, or join
an unitiliazed pthread_t.
On 07/02/2015 11:18 AM, Maxim Uvarov wrote:
On 07/01/15 16:40, Nicolas Morey-Chaisemartin wrote:
Merged,
Maxim.
On 07/01/15 21:31, Mike Holmes wrote:
On 1 July 2015 at 09:40, Nicolas Morey-Chaisemartin nmo...@kalray.eu
mailto:nmo...@kalray.eu wrote:
Signed-off-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
mailto:nmo...@kalray.eu
Reviewed-by: Mike Holmes
User should pick a synchronisation method and use SYNC_NONE if
does not need synchronization.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/queue.h | 5 -
platform/linux-generic/include/odp/plat/queue_types.h | 2 --
Default prio is used when priority scheduling is not important.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/queue.h | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/include/odp/api/queue.h b/include/odp/api/queue.h
index bdda845..10bac16
On 2 July 2015 at 19:03, Savolainen, Petri (Nokia - FI/Espoo)
petri.savolai...@nokia.com wrote:
-Original Message-
From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of ext
Benoît Ganne
Sent: Thursday, June 18, 2015 9:53 PM
To: lng-odp@lists.linaro.org
This is a hint to the scheduler for prefetching more events.
Performance is improved when scheduler may prefetch events in
parallel to application finishing processing of previous events.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule.h | 14
For performance optimization user can allow ordered context
release as soon as the last enqueue which needs ordering support
is finished.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule.h | 17 +
1 file changed, 17 insertions(+)
diff --git
Hi,
I like to start the discussion thread here to gather the the requirements
to test the odp scheduling capabilities. That would be the input for
updating odp generator for demo purpose. Some of the proposed requirements
I heard are,
1) using multiple flows of simple traffic
2) IPSec traffic
Hi,
I have a few concerns with this API proposal,
1. The term ODP_PMR_OFFSET_ABS could be renamed as ODP_PMR_OFFSET_L2 as
this offset is from L2.
This would help some platforms which can support additional offset from
custom layer other than L2 and they can add additional Enum in their
private
Merged,
Maxim.
On 07/01/15 13:41, Stuart Haslam wrote:
On Wed, Jul 01, 2015 at 12:05:36PM +0200, Christophe Milard wrote:
Cpumask tests now go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of validation: thread tests moved to
Hi Ben,
Pls find my comments inline.
On 2 July 2015 at 18:48, Benoît Ganne bga...@kalray.eu wrote:
Hi Bala,
Thanks for your feedback. My comments inline.
1. The term ODP_PMR_OFFSET_ABS could be renamed as ODP_PMR_OFFSET_L2 as
this offset is from L2.
This would help some platforms which
Extracted scheduler parameters into a new type and harmonized
documentation of the types.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/queue.h | 34 --
1 file changed, 20 insertions(+), 14 deletions(-)
diff --git
User should care which group of thread will process events
from a queue. Removed default group and use all threads
group instead.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
example/classifier/odp_classifier.c | 2 +-
example/generator/odp_generator.c
Hi Petri,
related to your scheduling API patches - how many contexts and of what type
one thread can hold at a given moment? If there's only one , atomic or
ordered, it does make sense to have only one _release_context_ function as
the type of context is somehow implied.
Different functions for
New file is used to avoid include loop between queue and
scheduler api. Both include the new file.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/queue.h| 93 +---
include/odp/api/schedule.h | 3 +-
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule.h | 41 ++
.../include/odp/plat/schedule_types.h | 2 ++
.../linux-generic/include/odp/plat/thrmask_types.h | 2 +-
3 files changed, 44 insertions(+),
Hi Bill Fischofer,
Thank you so much for provide this information. I will keep track of it.
Again, thanks for your help.
Best,
Genís Riera Pérez.
Genís Riera Pérez
E-mail: genis.riera.pe...@gmail.com
2015-07-02 13:33 GMT+02:00 Bill Fischofer bill.fischo...@linaro.org:
Hi Genís,
There is
Lock protects a critical section in ordered context processing.
Threads enter the section in order. Added dummy typedef for
not breaking the build.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule.h | 42 ++
Hi Bala,
Thanks for your feedback. My comments inline.
1. The term ODP_PMR_OFFSET_ABS could be renamed as ODP_PMR_OFFSET_L2 as
this offset is from L2.
This would help some platforms which can support additional offset from
custom layer other than L2 and they can add additional Enum in their
-Original Message-
From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of ext
Benoît Ganne
Sent: Thursday, June 18, 2015 9:53 PM
To: lng-odp@lists.linaro.org
Subject: [lng-odp] [API-NEXT PATCH 1/3] api: classification: add
ODP_PMR_OFFSET_ABS
The application can
Clarified that a thread holds synchronization context between
schedule calls. Next schedule call implicitly release the context.
Clafified atomic and ordered context definitions.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule.h | 18
The predefined group includes all worker threads and
is automanically updated.
Signed-off-by: Petri Savolainen petri.savolai...@nokia.com
---
include/odp/api/schedule_types.h | 7 +++
platform/linux-generic/include/odp/plat/schedule_types.h | 1 +
2 files changed, 8
To please check-patch (called by check-odp) when the file will be moved
(next patch)
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/ver_abt_log_dbg/odp_errno.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/test/validation/ver_abt_log_dbg/odp_errno.c
Errno (earlyer part of the odp_ver_abt_log_dbg module)
now gets its own module. Test contents remains unchanged.
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
configure.ac | 1 +
platform/linux-generic/test/Makefile.am
Renaming of things which may be, one day, exported in a lib.
This renaming is important, as it creates consistency between test
symbols, which is needed if things get eventually exported in the lib.
Also, tests are often created from other tests: Fixing the first exemples
will help geting future
Hi Petri,
--- a/include/odp/api/classification.h
+++ b/include/odp/api/classification.h
@@ -215,6 +215,8 @@ typedef enum odp_pmr_term {
ODP_PMR_IPSEC_SPI, /** IPsec session
identifier(*val=uint32_t)*/
ODP_PMR_LD_VNI, /** NVGRE/VXLAN network identifier
From: ext Alexandru Badicioiu [mailto:alexandru.badici...@linaro.org]
Sent: Thursday, July 02, 2015 4:35 PM
To: Petri Savolainen; LNG ODP Mailman List
Subject: api-next scheduler patches - context questions
Hi Petri,
related to your scheduling API patches - how many contexts and of what type
Module system (which just contained time tests) is now logicaly renamed
time. The renaming includes files and function names, to match
test conventions.
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
configure.ac | 2 +-
Former system module just contained time tests, and is
therefore renamed.
This also prepare for the proper system module (coming soon...)
To be applied on top of validation: new module errno
Christophe Milard (1):
validation: renaming system to time
configure.ac
Hi Bala,
But what if you create a PMR on ODP_PMR_DMAC and then a
ODP_PMR_OFFSET_ABS where the offset is greater than that? I believe
this is an implementation restriction. I think it should be possible
to reorder the rules in the implementation before configuring the HW.
This will not be
The tests for errno are removed from module odp_ver_abt_log_dbg.
Errno tests will get their own module in next patches
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/Makefile.am | 1 -
test/validation/ver_abt_log_dbg/odp_errno.h
https://bugs.linaro.org/show_bug.cgi?id=1657
Maxim Uvarov maxim.uva...@linaro.org changed:
What|Removed |Added
CC||maxim.uva...@linaro.org
https://bugs.linaro.org/show_bug.cgi?id=1657
Maxim Uvarov maxim.uva...@linaro.org changed:
What|Removed |Added
Assignee|lng-odp@lists.linaro.org
Hi Bala,
I'd define it like this,
odp_pmr_t odp_pmr_create_custom(uint32_t offset, const void *val,
const void *mask, uint32_t val_sz);
It would fail if the requested custom rule is not supported or too
many custom rules are used.
This signature should work fine. Can we additionally add the
Errno, previously part of module ver_abt_log_dbg is now separated
as its own module, and gets through the normal changes to conform
to the new test structure (lib...) and naming convention.
Christophe Milard (4):
validation: removed errno from odp_ver_abt_log_dbg
validation: cosmetic change
On 2 July 2015 at 21:17, Benoît Ganne bga...@kalray.eu wrote:
Hi Bala,
I'd define it like this,
odp_pmr_t odp_pmr_create_custom(uint32_t offset, const void *val,
const void *mask, uint32_t val_sz);
It would fail if the requested custom rule is not supported or too
many custom rules are
The original design for the user metadata feature had persistence as an
odp_pool_param option at pool creation. We should revisit this in that
context as introducing more hardcoded #defines into odp_config.h doesn't
seem the right way to go for these things.
On Thu, Jul 2, 2015 at 11:42 AM,
Yes - at the latest lets use Tuesdays ODP planning meeting to find a pick
what makes sense in preparation for ODP 1.2 at the end of August.
On 2 July 2015 at 04:51, Maxim Uvarov maxim.uva...@linaro.org wrote:
On 06/30/15 15:52, Stuart Haslam wrote:
On Tue, Jun 30, 2015 at 03:20:34PM +0300,
I will try, I did get one ages ago and some people took a look, the short
time I spent with it did not help.
On 2 July 2015 at 05:05, Maxim Uvarov maxim.uva...@linaro.org wrote:
On 07/01/15 20:17, Mike Holmes wrote:
odp_timer fails periodically with segfaults in CI
If the requested seg_len exceeds ODP_CONFIG_PACKET_SEG_LEN_MAX the
odp_pool_create() call will fail, so applications need to be mindful that
implementations may have maximum segment size limits, so requesting large
segment sizes that aren't actually needed is not a good idea.
If an application
That's a good start. To this I'd add:
1. Support for multiple concurrent I/O interfaces
2. Traffic replay capability
The latter is to be able to take things like Wireshark sample captures
https://wiki.wireshark.org/SampleCaptures and send them to a DUT for
validation testing.
On Thu, Jul 2,
On Thu, Jul 2, 2015 at 6:29 AM, Mike Holmes mike.hol...@linaro.org wrote:
Signed-off-by: Mike Holmes mike.hol...@linaro.org
Reviewed-by: Bill Fischofer bill.fischo...@linaro.org
---
v2
Dont put a period at the end of lists
doc/odp.dox | 7 ---
doc/release.dox | 65
This requirement came from ODP-OVS: it handles buffers through 'struct ofpbuf'
(in upstream it became 'struct dp_packet'), and they can be non-ODP buffers as
well. Therefore after receiving a packet currently we have to reset three
fields in it: 'source' (to ODP), 'allocated' (segment size, OVS
On 2 July 2015 at 06:45, hongbo.zh...@freescale.com wrote:
From: Hongbo Zhang hongbo.zh...@linaro.org
This patch adds test for the newly introduced per-CPU system APIs:
odp_sys_cpu_model_str_amp()
odp_sys_cpu_hz_amp()
odp_sys_cpu_hz_current()
Signed-off-by: Hongbo Zhang
66 matches
Mail list logo