On 26 July 2016 at 23:09, Brian Brooks wrote:
> On 07/26 22:38:20, Bala Manoharan wrote:
>> On 26 July 2016 at 22:26, Brian Brooks wrote:
>> > On 07/26 10:44:17, Maxim Uvarov wrote:
>> >> On 07/26/16 05:56, Brian Brooks wrote:
>> >> > --- a/platform/linux-generic/arch/mips64/cpu_arch.h
>> >> > ++
Hi Maxim,
Many termination code fragments may like below.
if (odp_pool_destroy(pool)) {
LOG_ERR("Error: pool destroy\n");
exit(EXIT_FAILURE);
}
if (odp_term_local()) {
LOG_ERR("Error: term local\n");
exit(EXI
Signed-off-by: Bill Fischofer
---
CHANGELOG | 99 +++
1 file changed, 99 insertions(+)
diff --git a/CHANGELOG b/CHANGELOG
index d8230cd..5c028d0 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -1,3 +1,102 @@
+== OpenDataPlane (1.11.0.0)
+ODP
During last week's Wednesday ODP call the topic of service threads was brought
up and there was an involved discussion on how to cope with the resource impact
of having extra internal threads.Some work I am doing internally at ARM has
been to research into the benefits of modifying ODP to be
On 07/26 10:13:09, Christophe Milard wrote:
> I am confused about patch 3 which seems to remove arch specific for
> arm and patch 6 which reintroduce it. Probably something I don't
> understand... can you explain? Is this series working fine on ARM?
Currently
platform/linux-generic/arch/arm/
i
On 07/26 22:38:20, Bala Manoharan wrote:
> On 26 July 2016 at 22:26, Brian Brooks wrote:
> > On 07/26 10:44:17, Maxim Uvarov wrote:
> >> On 07/26/16 05:56, Brian Brooks wrote:
> >> > --- a/platform/linux-generic/arch/mips64/cpu_arch.h
> >> > +++ b/platform/linux-generic/arch/mips64/cpu_arch.h
> >>
On 07/26 10:51:56, Maxim Uvarov wrote:
> On 07/26/16 05:56, Brian Brooks wrote:
> > Signed-off-by: Brian Brooks
> > ---
> > configure.ac | 12
> > ++--
> > platform/Makefile.inc| 6 +++---
> > pla
On 26 July 2016 at 22:26, Brian Brooks wrote:
> On 07/26 10:44:17, Maxim Uvarov wrote:
>> On 07/26/16 05:56, Brian Brooks wrote:
>> > --- a/platform/linux-generic/arch/mips64/cpu_arch.h
>> > +++ b/platform/linux-generic/arch/mips64/cpu_arch.h
>> > @@ -7,6 +7,8 @@
>> > #ifndef ODP_PLAT_CPU_ARCH_H
On 07/26 10:47:44, Maxim Uvarov wrote:
> On 07/26/16 05:56, Brian Brooks wrote:
> > Signed-off-by: Brian Brooks
> > ---
> > platform/linux-generic/arch/powerpc/odp_cpu_arch.c | 26
> > --
> > 1 file changed, 9 insertions(+), 17 deletions(-)
> >
> > diff --git a/platform/li
On 07/26 10:46:08, Maxim Uvarov wrote:
> On 07/26/16 05:56, Brian Brooks wrote:
> > Signed-off-by: Brian Brooks
> > ---
> > platform/linux-generic/arch/x86/odp_cpu_arch.c | 19 ++-
> > 1 file changed, 6 insertions(+), 13 deletions(-)
> >
> > diff --git a/platform/linux-generic/
On 07/26 10:44:17, Maxim Uvarov wrote:
> On 07/26/16 05:56, Brian Brooks wrote:
> > --- a/platform/linux-generic/arch/mips64/cpu_arch.h
> > +++ b/platform/linux-generic/arch/mips64/cpu_arch.h
> > @@ -7,6 +7,8 @@
> > #ifndef ODP_PLAT_CPU_ARCH_H_
> > #define ODP_PLAT_CPU_ARCH_H_
> > +#include
>
Define the ODP API for cache line size to the cache line size defined
in the internal architecture directories. Prefix internal cache line
size identifier with '_odp'.
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/default/odp/api/cpu_arch.h | 10 +-
platform/linux-generic/a
On 07/26 13:49:15, Christophe Milard wrote:
> On 26 July 2016 at 04:56, Brian Brooks wrote:
> > Define the ODP API for cache line size to the cache line size defined
> > in the internal architecture specific directories. Prefix internal
> > cache line size identifier with '_'.
> >
> > Signed-off-b
On 07/26 17:39:31, Maxim Uvarov wrote:
> On 07/26/16 05:56, Brian Brooks wrote:
> > #if defined __OCTEON__
> > #define ODP_CACHE_LINE_SIZE 128
>
> prefix missing
thanks for catching this. changed in v4
> > -#if defined __arm__ || defined __aarch64__
> > -
> > -#define ODP_CACHE_LINE_SIZE 64
Merged,
trivial change.
Maxim.
On 07/26/16 18:16, Maxim Uvarov wrote:
commit: ec0d570 linux-gen: pool: optimize thread local buffer cache
removed elements for prints but did not remove that
from format string.
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/odp_pool.c | 2 +-
1 fil
Thanks for reporting the bug, Mike.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Jul 26, 2016 at 5:44 PM, wrote:
> https://bugs.linaro.org/show_bug.cgi?id=2444
>
> --- Comment #1 from Mike Holmes ---
> (In reply to Mike Holmes from comment #0)
> > After some tests, I found the pr
https://bugs.linaro.org/show_bug.cgi?id=2444
--- Comment #2 from Oriol Arcas ---
Thanks for reporting the bug, Mike.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Jul 26, 2016 at 5:44 PM, wrote:
> https://bugs.linaro.org/show_bug.cgi?id=2444
>
> --- Comment #1 from Mike Holmes -
https://bugs.linaro.org/show_bug.cgi?id=2444
--- Comment #1 from Mike Holmes ---
(In reply to Mike Holmes from comment #0)
> After some tests, I found the problem in my setup. In my tests I have a
> pair of interfaces, and I want to have traffic management for each one.
> Since a odp_tm_t object
https://bugs.linaro.org/show_bug.cgi?id=2444
Bug ID: 2444
Summary: Can not use two PKTIOs with Traffic Manager
Product: OpenDataPlane - linux- generic reference
Version: v1.10.1
Hardware: Other
OS: Linux
Status: UNC
On 07/26/16 18:09, Maxim Uvarov wrote:
commit: 2d2e479 linux-gen: fix odp_pool_print
will send v2, wrong reference...
removed elements for prints but did not remove that
from format string.
Signed-off-by: Maxim Uvarov
---
somehow it passed odp-check run, I think we were lucky with stack a
commit: ec0d570 linux-gen: pool: optimize thread local buffer cache
removed elements for prints but did not remove that
from format string.
Signed-off-by: Maxim Uvarov
---
platform/linux-generic/odp_pool.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/platform/linux-generic
commit: 2d2e479 linux-gen: fix odp_pool_print
removed elements for prints but did not remove that
from format string.
Signed-off-by: Maxim Uvarov
---
somehow it passed odp-check run, I think we were lucky with stack allocation :)
platform/linux-generic/odp_pool.c | 2 +-
1 file changed, 1 inse
On 07/26/16 05:56, Brian Brooks wrote:
Define the ODP API for cache line size to the cache line size defined
in the internal architecture specific directories. Prefix internal
cache line size identifier with '_'.
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/default/odp/api/cpu_
merged to api-next branch with proper renaming.
Maxim.
On 07/22/16 16:19, Christophe Milard wrote:
Obviously a miss from commit c4aefb88d31452b3add8cf16f9eef152525c3e93
Signed-off-by: Christophe Milard
---
helper/test/.gitignore | 1 +
1 file changed, 1 insertion(+)
diff --git a/helper/te
Merged,
Maxim.
On 07/26/16 03:15, Bill Fischofer wrote:
For this series:
Reviewed-and-tested-by: Bill Fischofer
On Fri, Jul 22, 2016 at 5:04 AM, Matias Elo wrote:
Optimize local buffer cache performance which is critical to
many use cases - including packet IO.
Main parts of the optimizat
On 26 July 2016 at 14:50, Mike Holmes wrote:
>
>
> On 26 July 2016 at 03:32, Maxim Uvarov wrote:
>
>> On 07/26/16 09:35, Forrest Shi wrote:
>>
>>> Hi Mike,
>>>
>>> Why the cases under test/ are added termination path?
>>>
>>
> I am not up to speed, but all applications should close down graceful
On 26 July 2016 at 03:32, Maxim Uvarov wrote:
> On 07/26/16 09:35, Forrest Shi wrote:
>
>> Hi Mike,
>>
>> Why the cases under test/ are added termination path?
>>
>
I am not up to speed, but all applications should close down gracefully, is
that what you mean ?
>
>> I'm wondering if it is neede
On 26 July 2016 at 04:56, Brian Brooks wrote:
> Define the ODP API for cache line size to the cache line size defined
> in the internal architecture specific directories. Prefix internal
> cache line size identifier with '_'.
>
> Signed-off-by: Brian Brooks
The commit msg should say prefixed wit
Also got the following check-odp warnings:
Using patch: 0008-arm-odp_cpu_cycles-for-aarch64.patch
Trying to apply patch
Patch applied
WARNING: Missing a blank line after declarations
#42: FILE: platform/linux-generic/arch/arm/odp_cpu_arch.c:21:
+ uint64_t vct;
+ __asm__ volatile("i
I am confused about patch 3 which seems to remove arch specific for
arm and patch 6 which reintroduce it. Probably something I don't
understand... can you explain? Is this series working fine on ARM?
The API function:
uint64_t odp_cpu_hz_current(int id ODP_UNUSED)
Is still defined in the arch dire
Hi Maxim,
What's your idea bout below comment?
Increment one value from number of threads might be not good thing to do
due to 1) performance 2) atomic operations break.
It's better if each thread has it's own counters and control threads
summaries values.
odp_cpu_pause() is api defined here:
./include/odp/api/spec/cpu.h
Maxim.
On 07/26/16 05:56, Brian Brooks wrote:
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/default/odp/api/cpu_arch.h | 2 +-
platform/linux-generic/arch/mips64/odp/api/cpu_arch.h | 2 +-
platform/linux-gene
On 07/26/16 05:56, Brian Brooks wrote:
Signed-off-by: Brian Brooks
---
configure.ac | 12 ++--
platform/Makefile.inc| 6 +++---
platform/linux-generic/arch/{default => arm}/cpu_arch.h | 0
On 07/26/16 05:56, Brian Brooks wrote:
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/arm/odp_cpu_arch.c | 25 ++---
1 file changed, 6 insertions(+), 19 deletions(-)
diff --git a/platform/linux-generic/arch/arm/odp_cpu_arch.c
b/platform/linux-generic/arch/ar
On 07/26/16 05:56, Brian Brooks wrote:
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/powerpc/odp_cpu_arch.c | 26 --
1 file changed, 9 insertions(+), 17 deletions(-)
diff --git a/platform/linux-generic/arch/powerpc/odp_cpu_arch.c
b/platform/linux-generic/ar
On 07/26/16 05:56, Brian Brooks wrote:
--- a/platform/linux-generic/arch/mips64/cpu_arch.h
+++ b/platform/linux-generic/arch/mips64/cpu_arch.h
@@ -7,6 +7,8 @@
#ifndef ODP_PLAT_CPU_ARCH_H_
#define ODP_PLAT_CPU_ARCH_H_
+#include
+
#ifdef __cplusplus
extern "C" {
#endif
@@ -17,6 +19,1
On 07/26/16 05:56, Brian Brooks wrote:
Signed-off-by: Brian Brooks
---
platform/linux-generic/arch/x86/odp_cpu_arch.c | 19 ++-
1 file changed, 6 insertions(+), 13 deletions(-)
diff --git a/platform/linux-generic/arch/x86/odp_cpu_arch.c
b/platform/linux-generic/arch/x86/odp_
control thread mask and worker thread mask should give a hint to
implementation which cores might be used.
How to spread load inside that 2 masks it's the deal of implementation.
I think it's a little bit overload to create
one more mask for service threads
Maxim.
On 07/26/16 07:50, Bala
On 07/26/16 03:14, Bill Fischofer wrote:
> #define MSG_POOL_SIZE (4 * 1024 * 1024) /**< Message pool size */
>-#define MAX_ALLOCS35/**< Alloc burst size */
>+#define MAX_ALLOCS32/**< Alloc burst size */
>
Not a biggie, but it's not clear w
On 07/26/16 09:35, Forrest Shi wrote:
Hi Mike,
Why the cases under test/ are added termination path?
I'm wondering if it is needed in l3fwd.
Maxim is talking if l3fwd should be under test/ not example/, what's your
thought?
Before sending new version, I want to include this change.
thanks,
F
40 matches
Mail list logo
