Needed to please check-patch called via check-odp when the file will
be moved, at next patch.
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/odp_thread.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/test/validation/odp_thread.c
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/Makefile.am | 2 ++
test/validation/odp_thread.c | 15 +--
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/test/validation/Makefile.am b/test/validation/Makefile.am
index
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 main in odp_thread.c
validation: renaming in odp_thread.c
validation:
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
On Tue, Jun 30, 2015 at 03:31:03PM +0300, Petri Savolainen wrote:
Added max number of worker/control threads as global init
parameter. Implementation can e.g. optimize it's per worker
thread resource reservation or configuration accordingly.
Signed-off-by: Petri Savolainen
On 1 July 2015 at 09:40, Nicolas Morey-Chaisemartin nmo...@kalray.eu
wrote:
Signed-off-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
Reviewed-by: Mike Holmes mike.hol...@linaro.org
---
helper/linux.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git
Merged,
Thanks,
Maxim.
On 07/01/15 13:36, Stuart Haslam wrote:
On Tue, Jun 30, 2015 at 06:14:51PM +0200, Christophe Milard wrote:
Scheduler tests now go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of validation: call to
The odp-dpdk.git repo now relies on this DPDK version.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
diff --git a/helper/platform/dpdk b/helper/platform/dpdk
index 05c46e9..120da1d 100644
--- a/helper/platform/dpdk
+++ b/helper/platform/dpdk
@@ -25,7 +25,7 @@ dpdk_get(){
fi
cd
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
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
platform/linux-generic/odp_event.c | 24
1 file changed, 24 insertions(+)
diff --git a/platform/linux-generic/odp_event.c
b/platform/linux-generic/odp_event.c
index d9e86a9..c50103c 100644
---
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 it surfaces as an issue. This is seen on
-Original Message-
From: ext Stuart Haslam [mailto:stuart.has...@linaro.org]
Sent: Wednesday, July 01, 2015 12:03 PM
To: Savolainen, Petri (Nokia - FI/Espoo)
Cc: lng-odp@lists.linaro.org
Subject: Re: [lng-odp] [API-NEXT PATCH] api: init: added thread count
params
On Tue, Jun
On Wed, Jul 01, 2015 at 12:15:54PM +, Prashant Upadhyaya wrote:
Hi guys,
I object too. Why would you want to 'remove' a functionality.
The 'basic' socket is a very good lowest common denominator of implementation
across all the machines/linux
And it was anyway a fallback mechanism.
The intent is that a pktio has a single classification context at a time.
These contexts can contain any number of PMRs however the application
wishes to arrange them. So a context is a complete description of what
should be done for packets arriving on a single pktio.
If a PMR is associated
Mike, you a little bit clean up text. Few notes are bellow:
Thanks,
Maxim.
On 06/03/15 20:23, Mike Holmes wrote:
Signed-off-by: Mike Holmes mike.hol...@linaro.org
---
doc/odp.dox | 7 ---
doc/release.dox | 65 +
2 files
Signed-off-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
---
helper/linux.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/helper/linux.c b/helper/linux.c
index be1ea13..2aa6835 100644
--- a/helper/linux.c
+++ b/helper/linux.c
@@ -139,10 +139,15 @@ int
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)
Signed-off-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
---
Hi, Bill
On 01.07.15 05:12, Bill Fischofer wrote:
Signed-off-by: Bill Fischofer bill.fischo...@linaro.org
---
include/odp/api/classification.h | 106 +++
1 file changed, 106 insertions(+)
diff --git a/include/odp/api/classification.h
On Wed, Jul 01, 2015 at 10:19:28AM +0200, 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
On 30 June 2015 at 04:15, Mike Holmes mike.hol...@linaro.org wrote:
Need to send this with API-NEXT in the subject because it touches
include/odp/api
OK, I see.
AND in the subject, make it a patch series for the two separate ideas AND is
a warning that you tried to do more than one thing in
On Tue, Jun 30, 2015 at 06:14:51PM +0200, Christophe Milard wrote:
Scheduler tests now go through the same modification as pktio,
and gets finally called from the platform side.
To be applied on top of validation: call to buffer_main from platform
Christophe Milard (5):
validation: own
On 06/30/15 19:06, Nicolas Morey-Chaisemartin wrote:
I got a little bit mixed up between the msg and mmsg version.
After a quick look, we should be able to fall back to recvmsg/sendmsg for a
basic socket implementation
But yes, this patch removed a lot of code and a new version would be very
Are people seeing a measurable performance benefit from using more advanced
socket APIs? Last time I compared basic socket to the default (a year ago?)
there was no performance benefit. If the benefit is neglect or small (
20-30%), I'd simplify the packet IO socket code by removing everything
Merged,
Thanks,
Maxim.
On 06/26/15 00:48, Bill Fischofer wrote:
Fix for Bug https://bugs.linaro.org/show_bug.cgi?id=1661
Signed-off-by: Bill Fischofer bill.fischo...@linaro.org
---
platform/linux-generic/include/odp_buffer_inlines.h | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
https://bugs.linaro.org/show_bug.cgi?id=1670
Bug 1670 depends on bug 1661, which changed state.
Bug 1661 Summary: Invalid data in segmented packet when copying data with
odp_packet_copydata_in()
https://bugs.linaro.org/show_bug.cgi?id=1661
What|Removed |Added
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
configure.ac | 1 +
test/validation/.gitignore | 1 -
test/validation/Makefile.am | 10 --
test/validation/cpumask/.gitignore
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/Makefile.am | 2 ++
test/validation/odp_cpumask.c | 15 +--
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/test/validation/Makefile.am b/test/validation/Makefile.am
index
We could simply factor basic and mmsg code by doing something like this in the
mmsg implementation
if ( /* FIXME BASIC */) {
ssize_t recv_bytes;
for(i = 0; i msgvec_len; ++i){
recv_bytes = recvmsg(sockfd, msgvec[i], MSG_DONTWAIT);
if
To please check-patch as much as possible before the
file gets moved (and rechecked by check-odp)
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/odp_cpumask.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git
Hi guys,
I object too. Why would you want to 'remove' a functionality.
The 'basic' socket is a very good lowest common denominator of implementation
across all the machines/linux
And it was anyway a fallback mechanism.
Currently atleast my application depends on it, so guys, please don't remove
On 30 June 2015 at 04:40, Mike Holmes mike.hol...@linaro.org wrote:
This touches include/odp/api so the subject must include API-NEXT
On 25 June 2015 at 05:05, hongbo.zh...@freescale.com wrote:
From: Hongbo Zhang hongbo.zh...@linaro.org
This patch add API to return the current frequency of
https://bugs.linaro.org/show_bug.cgi?id=1661
Bill Fischofer bill.fischo...@linaro.org changed:
What|Removed |Added
Status|IN_PROGRESS |RESOLVED
https://bugs.linaro.org/show_bug.cgi?id=1661
--- Comment #8 from Bill Fischofer bill.fischo...@linaro.org ---
Fix merged. Commit ID 828f2e54799dd6f0f277aa125fbb325f48da9362
--
You are receiving this mail because:
You are on the CC list for the
The intent behind the odp_structname_init() routines (following the POSIX
model) is simply to initialize struct values to default values to handle
things like reserved fields that the application might not know about.
Extending that to provide fill in for the various subfields is really
beyond the
After reading this patch it looks like buffer size is good parameter to
init function. Also if we anyway need to set up
seg_len, len and num, why not to pass that values to init?
Maxim.
On 07/01/15 04:39, Bill Fischofer wrote:
Signed-off-by: Bill Fischofer bill.fischo...@linaro.org
---
On 30 June 2015 at 12:16, Ivan Khoronzhuk ivan.khoronz...@linaro.org
wrote:
It's more generic implementation simplify reusing timer API
for other platforms.
Signed-off-by: Ivan Khoronzhuk ivan.khoronz...@linaro.org
Reviewed-by: Ola Liljedahl ola.liljed...@linaro.org
-
I got a (spurious)
Ola,
I need this change.
Could you please comment on this if you have some objections.
or maybe send your version.
On 30.06.15 13:16, Ivan Khoronzhuk wrote:
It's more generic implementation simplify reusing timer API
for other platforms.
Signed-off-by: Ivan Khoronzhuk
https://bugs.linaro.org/show_bug.cgi?id=1675
Bill Fischofer bill.fischo...@linaro.org changed:
What|Removed |Added
CC|
The odp_pool_param_t is a struct that is part of the public API.
odp_pool_param_init() simply sets that struct to a default value (memset's
it to 0 in linux-generic) meaning that the application need not be
concerned with whether zeros or other values represent the default values
for fields
The ODP API does not mandate that an implementation uses segments, only
that applications can discover and work with any segments that may be
present.
So it is not possible to force an implementation to use segments since, as
Stuart points out, seg_len is simply an application minimum value that
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 odp_synchronizers.c
validation: renaming in odp_synchronizers.c
synchronizers_main is now ran from the platform side.
As synchronizers tests do not have
any platform dependency, the call to the platform agnostic tests is
performed directly in the Makefele.am (TESTS).
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
Sorry, I've forgot the check-odp string from the subject, I've sent it again
On 01/07/15 17:12, Zoltan Kiss wrote:
The odp-dpdk.git repo now relies on this DPDK version.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
diff --git a/helper/platform/dpdk b/helper/platform/dpdk
index
Clarify that the user owns all events not sent/enqueued successfully.
Currently our examples are not handling this situation as well.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
v4:
- improve commit message
- move changes into the body of the description
include/odp/api/packet_io.h |
It happens with ODP_PKTIO_TYPE_SOCKET_BASIC.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
platform/linux-generic/odp_packet_socket.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/platform/linux-generic/odp_packet_socket.c
Unsent packet has to be released. If the event type is obvious from the
context, use directly the relevant release functions, otherwise
odp_event(free).
Wider error handling is attempted, but this patch can't fix all the flaws
in the many calling functions of odp_queue_enq()
Signed-off-by: Zoltan
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
example/timer/odp_timer_test.c | 26 +-
1 file changed, 1 insertion(+), 25 deletions(-)
diff --git a/example/timer/odp_timer_test.c b/example/timer/odp_timer_test.c
index f4af0e2..ccd07fd 100644
---
The odp-dpdk.git repo now relies on this DPDK version.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
diff --git a/helper/platform/dpdk b/helper/platform/dpdk
index 05c46e9..120da1d 100644
--- a/helper/platform/dpdk
+++ b/helper/platform/dpdk
@@ -25,7 +25,7 @@ dpdk_get(){
fi
cd
On 07/01/15 18:41, Bill Fischofer wrote:
The intent behind the odp_structname_init() routines (following the
POSIX model) is simply to initialize struct values to default values
to handle things like reserved fields that the application might not
know about.
In that case init() should
To please check-patch as much as possible before the
file gets moved (and rechecked by check-odp)
Signed-off-by: Christophe Milard christophe.mil...@linaro.org
---
test/validation/odp_synchronizers.c | 62 -
1 file changed, 34 insertions(+), 28 deletions(-)
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
example/packet/odp_pktio.c | 10 +-
test/performance/odp_l2fwd.c | 12 ++--
2 files changed, 19 insertions(+), 3 deletions(-)
diff --git a/example/packet/odp_pktio.c b/example/packet/odp_pktio.c
index f08d9f4..0b4a8f1 100644
We need a convenient function which calls the right type-depending free()
function.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
v4: incorporate Petri's suggestions into formatting
include/odp/api/event.h | 11 +++
1 file changed, 11 insertions(+)
diff --git
Instead of converting types let odp_event_free() to do the right thing.
Signed-off-by: Zoltan Kiss zoltan.k...@linaro.org
---
v3: rebase to latest master
example/timer/odp_timer_test.c| 2 +-
platform/linux-generic/odp_schedule.c | 5 +
Unsent packet has to be released. If the event type is obvious from the
context, use directly the relevant release functions, otherwise
odp_event(free).
Wider error handling is attempted, but this patch can't fix all the flaws
in the many calling functions of odp_queue_enq()
Signed-off-by: Zoltan
If the idea of this patch is to test segmented packets it can be
accomplished by allocating packets of size greater than seg_len in an
additional test case rather than modifying the segment length in pool
create function.
Regards,
Bala
On 30 June 2015 at 22:26, Stuart Haslam
It does not apply anymore due to Christophe patches which moved odp_packet.c.
Otherwise Reviewed and tested.
On 06/25/2015 11:48 PM, Bill Fischofer wrote:
Signed-off-by: Bill Fischofer bill.fischo...@linaro.org
Extend segment tests to catch Bug https://bugs.linaro.org/show_bug.cgi?id=1661
Also, Since seg_len gives information to the implementation as to how many
bytes are required to be contiguous on the first segment for faster access.
I would like to have the value of seg_len as something in the lines of 256
or 512 bytes rather than 3K or 9K.
Regards,
Bala
On 2 July 2015 at
This should be on the lng-odp-dpdk list. Should also mention that
tag v1.1.0.0_DPDK_1.7.1 marks the code versioned for DPDK v1.7.1
On Wed, Jul 1, 2015 at 11:21 AM, Zoltan Kiss zoltan.k...@linaro.org wrote:
Sorry, I've forgot the check-odp string from the subject, I've sent it
again
On
On 06/25/2015 11:48 PM, Bill Fischofer wrote:
Fix for Bug https://bugs.linaro.org/show_bug.cgi?id=1661
Signed-off-by: Bill Fischofer bill.fischo...@linaro.org
Review-and-tested-by: Nicolas Morey-Chaisemartin nmo...@kalray.eu
---
platform/linux-generic/include/odp_buffer_inlines.h | 4 +++-
59 matches
Mail list logo