It’s release_atomic needed when you have finished working on the current buffer
from an atomic queue, but you are not going to call another schedule call for a
while (like in this example when you step out from the main_odp_loop). Schedule
call releases atomic context implicitly, but now
Signed-off-by: Balasubramanian Manoharan bala.manoha...@linaro.org
---
V4: This patch incorporates previous review comments
and enhancements to ODP debug internal file description.
platform/linux-generic/include/api/odp_debug.h | 54 ++
On 15 October 2014 18:51, Ola Liljedahl ola.liljed...@linaro.org wrote:
What if a thread wants to stop consuming and processing events? We don't
(can't) leave some events prescheduled (and stashed in some per-core
portal) after the thread has stopped processing. So a thread must be able
to
On 10/15/2014 09:14 PM, Mike Holmes wrote:
Unit testing replicates a situation similar to an NFV case where the workload
is varied without rebooting the HW.
In this use case there have to be hooks for the application to inform the
implementation that HW needs to be released.
odp_term_global
On 15 October 2014 17:05, Bala Manoharan bala.manoha...@linaro.org wrote:
Hi,
On 15 October 2014 19:37, Anders Roxell anders.rox...@linaro.org wrote:
Question: should we add #include odp_classification.h into odp.h in
this patch as well or in a separate patch?
I would prefer to add this
Hi,
I think we don’t need to specify these in three different versions. It should
be enough that odp_atomic_xxx is defined as “relaxed”, as it's like that
already today, since we don’t promise (in API documentation) to include memory
barriers into those calls. I think relaxed is the common
I think the API need to be more explicit in what ordering guarantees are
made. As the original implementation is sequentially consistent, I think
there is a big risk of applications relying on this behavior without
knowing, SC is also what a naive programmer expects. Just modifying the
existing
On 16 October 2014 04:37, Taras Kondratiuk taras.kondrat...@linaro.org
wrote:
On 10/15/2014 09:14 PM, Mike Holmes wrote:
Unit testing replicates a situation similar to an NFV case where the
workload
is varied without rebooting the HW.
In this use case there have to be hooks for the
On 16 October 2014 02:39, Bala Manoharan bala.manoha...@linaro.org wrote:
On 15 October 2014 19:27, Mike Holmes mike.hol...@linaro.org wrote:
Sorry, found something
mike@fedora1:~/git/odp$ git am
~/incoming/lng-odp_ODP-PATCHv2_ODP_Macro_for_unimplemented_functions.mbox
Applying: ODP
One specific goal of this updated API is to access shared data correctly
(in the way defined by C11/C++11 standards, not using some ad-hoc model)
and to remove the need for explicit (HW and compiler) barriers. The API
should specify the actual intent of operations, the full memory barrier
does
Signed-off-by: Ciprian Barbu ciprian.ba...@linaro.org
---
doc/users-guide/guide.dox | 2 ++
1 file changed, 2 insertions(+)
diff --git a/doc/users-guide/guide.dox b/doc/users-guide/guide.dox
index 314d295..d5533a0 100644
--- a/doc/users-guide/guide.dox
+++ b/doc/users-guide/guide.dox
@@ -14,5
Yes I think redesigning the API to mimic C11 atomics and follow a
standardised memory model is the way to go. Over time, users will be more
familiar with the memory models of C11 and C++11 and doing something
different in ODP will not benefit from this. This patch is all it takes for
ODP to step
** Setup Details:
Kernel:
git://git.linaro.org/kernel/linux-linaro-tracking.git
(Branch:linux-linaro)
Toolchain:
gcc version 4.9.1 20140529 (prerelease)
(crosstool-NG linaro-1.13.1-4.9-2014.06-02 - Linaro GCC 4.9-2014.06)
NOTE: This codebase mainly for POC of Access perf counters using vDSO,
Signed-off-by: Yogesh Tillu yogesh.ti...@linaro.org
---
ARMv8_Module/Makefile |8
ARMv8_Module/README |1 +
ARMv8_Module/enable_arm_pmu.c | 96 +
3 files changed, 105 insertions(+)
create mode 100644 ARMv8_Module/Makefile
Signed-off-by: Yogesh Tillu yogesh.ti...@linaro.org
---
arch/arm64/kernel/vdso/Makefile |6 +++---
arch/arm64/kernel/vdso/vdso.lds.S |5 +
arch/arm64/kernel/vdso/vdso_perfc.c | 20
3 files changed, 28 insertions(+), 3 deletions(-)
create mode 100644
---
vdso_userspace_perf.c | 58 +
1 file changed, 58 insertions(+)
create mode 100644 vdso_userspace_perf.c
diff --git a/vdso_userspace_perf.c b/vdso_userspace_perf.c
new file mode 100644
index 000..42d2682
--- /dev/null
+++
I agree that it is better to make these changes now than later after more
code would potentially be impacted by them. Anything we can do to help
ensure correctness of parallel/concurrent code will pay many dividends
going forward. They seem well thought out and self-contained, and most of
these
On 2014-10-08 12:25, Mike Holmes wrote:
When the document is generated the prefix becomes repetitious.
Signed-off-by: Mike Holmes mike.hol...@linaro.org
---
classification_design.dox | 2 +-
crypto_design.dox | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git
https://bugs.linaro.org/show_bug.cgi?id=301
Mike Holmes mike.hol...@linaro.org changed:
What|Removed |Added
CC||mike.hol...@linaro.org
---
Signed-off-by: Balasubramanian Manoharan bala.manoha...@linaro.org
---
V4: moved odp_classification.h file in alphabetical order in Makefile.am
platform/linux-generic/Makefile.am | 2 +
.../linux-generic/include/api/odp_classification.h | 540 +
There are a few updates that have happened over the past week.
1. http://www.opendataplane.org/documentation/ - Restructured the
documentation page to separate release from draft versions of docs.; Added
API design guidelines and other draft docs.
2. Add ODP Roadmap page with dynamic
On Thu, Oct 16, 2014 at 10:26 AM, Balasubramanian Manoharan
bala.manoha...@linaro.org wrote:
Signed-off-by: Balasubramanian Manoharan bala.manoha...@linaro.org
Reviewed-by: Bill Fischofer bill.fischo...@linaro.org
---
V4: moved odp_classification.h file in alphabetical order in
On 16 October 2014 02:54, Balasubramanian Manoharan
bala.manoha...@linaro.org wrote:
Signed-off-by: Balasubramanian Manoharan bala.manoha...@linaro.org
Reviewed-by: Mike Holmes mike.hol...@linaro.org
---
V4: This patch incorporates previous review comments
and enhancements to ODP debug
The change to odp_pktio_pmr_match_set is a correction. Not sure why the
other changes to the typedefs. Typedefs in these documents are
illustrative, not normative. Each implementation creates them to suit its
needs.
On Thu, Oct 16, 2014 at 11:04 AM, Balasubramanian Manoharan
On 16 October 2014 09:00, Ciprian Barbu ciprian.ba...@linaro.org wrote:
Signed-off-by: Ciprian Barbu ciprian.ba...@linaro.org
---
doc/users-guide/guide.dox | 2 ++
1 file changed, 2 insertions(+)
diff --git a/doc/users-guide/guide.dox b/doc/users-guide/guide.dox
index 314d295..d5533a0
https://bugs.linaro.org/show_bug.cgi?id=779
Mike Holmes mike.hol...@linaro.org changed:
What|Removed |Added
Priority|--- |Normal
CC|
https://bugs.linaro.org/show_bug.cgi?id=779
Ola Liljedahl ola.liljed...@linaro.org changed:
What|Removed |Added
CC|
Unit testing replicates a situation similar to an NFV case where the workload
is varied without rebooting the HW.
In this use case there have to be hooks for the application to inform the
implementation that HW needs to be released.
Signed-off-by: Mike Holmes mike.hol...@linaro.org
---
On 16 October 2014 11:26, Balasubramanian Manoharan
bala.manoha...@linaro.org wrote:
Signed-off-by: Balasubramanian Manoharan bala.manoha...@linaro.org
---
V4: moved odp_classification.h file in alphabetical order in Makefile.am
platform/linux-generic/Makefile.am | 2 +
Reviewed but need the following changes:
1. odp_buffer_copy() - Add 2nd argument to reflect target buffer pool.
Syntax should be: odp_buffer_t odp_buffer_copy(odp_buffer_t buf,
odp_buffer_pool_t pool);
Thanks.
Bill
On Wed, Oct 8, 2014 at 2:50 AM, Balasubramanian Manoharan
The vDSO code is very small:
+__attribute__((no_instrument_function)) int __vdso__perf_read_counter(void)
+{
+int ret = 0;
+asm volatile(mrs %0, pmccntr_el0 : =r (ret));
+return ret;
+}
If this system call is called from an application instead of the
corresponding mrs instruction being inlined
Mike this patch needed update to the latest odp code.
Also it will be good to split it on 2: one for indent and other to
exactly problem. Just for better review.
When applying I think that your fix is not quite correct. And I suggest
to do it in that way:
---
Signed-off-by: Mike Holmes mike.hol...@linaro.org
---
configure.ac | 6 ++
1 file changed, 6 insertions(+)
diff --git a/configure.ac b/configure.ac
index aa94034..9d0b17d 100644
--- a/configure.ac
+++ b/configure.ac
@@ -197,3 +197,9 @@ AC_MSG_RESULT([
ldflags:
Based on discussions we had yesterday and today, I'd like to outline the
open issues regarding queues and synchronization/scheduling models. We'd
like to get consensus on this in time for next week's Tuesday call.
ODP identifies three different synchronization/scheduling models for
queues:
34 matches
Mail list logo
