On Fri, Aug 26, 2016 at 11:14 AM, Mike Holmes
wrote:
> A recent bug showed the limitations of using powers of two, add a prime
> number of threads as a new test case.
>
> Signed-off-by: Mike Holmes
>
Reviewed-by: Bill Fischofer
> ---
> v2
>Test with max cores (Maxim)
>
> DEPENDENCIES
On 20 June 2016 at 10:17, Maxim Uvarov wrote:
> Add initial version for more verbose stacktrace for linux-
> generic odp platform.
> Usage:
>
> export ODP_STACKTRACE_ENABLE=y
>
> Thread 0 terminated with signal=11 (SIGSEGV)
> ./test/performance/odp_scheduling[0x4250b3]
> /lib/x86_64-linux-gnu/lib
ping
On 16 August 2016 at 17:10, Mike Holmes wrote:
> Add process mode run for odp_crypto_run, odp_l2fwd_run,
> odp_pktio_perf_run.
>
> Currently failing tests retun SKIP (77) until process mode is fixed
>
> Signed-off-by: Mike Holmes
> ---
> test/linux-generic/performance/Makefile.am
A recent bug showed the limitations of using powers of two, add a prime
number of threads as a new test case.
Signed-off-by: Mike Holmes
---
v2
Test with max cores (Maxim)
DEPENDENCIES | 4
test/common_plat/performance/odp_scheduling_run.sh
ping
On 06/20/16 17:17, Maxim Uvarov wrote:
Add initial version for more verbose stacktrace for linux-
generic odp platform.
Usage:
export ODP_STACKTRACE_ENABLE=y
Thread 0 terminated with signal=11 (SIGSEGV)
./test/performance/odp_scheduling[0x4250b3]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x1
On 25 August 2016 at 10:14, Maxim Uvarov wrote:
> On 08/25/16 00:34, Anders Roxell wrote:
>
> +#include
> >why do we include an internal header file?
> >
>>>
>>> >
>>> >because we need access to classifier bits, that is why this test is
>>> > under
>>> >linux-generic.
>>
>>
https://bugs.linaro.org/show_bug.cgi?id=2309
--- Comment #6 from Mike Holmes ---
This has also failed in CI
https://ci.linaro.org/view/odp-ci/job/odp-tool-check/build_type=run_valid_from_install,label=docker-jessie-amd64/546/consoleFull
../../helper/linux.c:273:odpthread_run_start_routine():hel
https://bugs.linaro.org/show_bug.cgi?id=2490
Bug ID: 2490
Summary: linux-generic: does not remove static inline when
building .so
Product: OpenDataPlane - linux- generic reference
Version: v1.11.0.0
Hardware: Other
From: Xuelin Shi
increase the hash size.
pre-compute the hash keys of possible ip destinations in a subnet.
Signed-off-by: Xuelin Shi
---
example/l3fwd/odp_l3fwd.c| 3 -
example/l3fwd/odp_l3fwd_db.c | 160 ++-
example/l3fwd/odp_l3fwd_db.h | 35 +++
From: Xuelin Shi
multi-thread, multi-queues and bi-directional forwarding.
support (port, queue, thread) arguments in cmdline which specify how
the threads handle which rx queue at which port, if no this argument,
default specification used.
both hash and lpm based lookup methods are supported,
make check is still missing. You need to add script similar to:
./test/common_plat/performance/odp_l2fwd_run.sh
On 08/25/16 09:02, forrest@linaro.org wrote:
From: Xuelin Shi
multi-thread, multi-queues and bi-directional forwarding.
support (port, queue, thread) arguments in cmdline which
On 08/26/16 04:37, forrest.shi wrote:
Hi Maxim,
I will send another patch to resolve the hash collision issue.
BTW: did you received my resend v8, this time is TO: lng-odp@ lists.linaro.org?
I received it but not see on patch list.
Thanks,
Forrest
Now I see it with RESEND tag. W
12 matches
Mail list logo