Add test for odp_time_sum, odp_time_cmp, odp_time_to_u64 APIs.
Sophisticate a little tests for odp_time_diff, odp_time_local_from_ns,
odp_time_local_to_ns APIs. Check time on monotony.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/validation/time/time.c
This series is intended to modify time API to be used as monotonic wall
time in order to simplify time API usage and create prerequisites for
adding global time API and local wall time API.
Since v1:
- corrected name of the series to be api-next
- use thread local for holding start time
Ivan
Use signed long for ODP_TIME_*, that's enough and it allows to
compare it with signed types, used for instance in timespec type.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
include/odp/api/time.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff
time API.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 2 +-
include/odp/api/time.h | 30 --
.../linux-generic/include/odp/plat/time_types.h| 4 +-
platform/linux-generic/i
Sent v2 already.
On 12.11.15 19:05, Ivan Khoronzhuk wrote:
This series is intended to modify time API to be used as monotonic wall
time in order to simplify time API usage and create prerequisites for
adding global time API and local wall time API.
Ivan Khoronzhuk (7):
api: time: use signed
This series is intended to modify time API to be used as monotonic wall
time in order to simplify time API usage and create prerequisites for
adding global time API and local wall time API.
Ivan Khoronzhuk (7):
api: time: use signed long for ODP_TIME_*
validation: time: don't assign int
Under opaque type can be structure that cannot be used with direct
values, So, use conversion functions to get time_t to avoid build
issues.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/validation/time/time.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Add test for odp_time_sum, odp_time_cmp, odp_time_to_u64 APIs.
Sophisticate a little tests for odp_time_diff, odp_time_local_from_ns,
odp_time_local_to_ns APIs. Check time on monotony.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/validation/time/time.c
The linux-generic time API implementation shouldn't depend on cpu
cycle API wich is not stable enough to measure time period due to
dynamic frequency scaling.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/Makefile.am | 1 -
...
It's more convenient the local time to be a wall time. It's allows
to measure time beginning from start of the thread. This allows to use
local time in similar manner as it's supposed to be used with global
time and the 64-bit timer is enough to guarantee it.
Signed-off-by: Ivan Khoronzhuk
Use signed long for ODP_TIME_*, that's enough and it allows to
compare it with signed types, used for instance in timespec type.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
include/odp/api/time.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff
It's more convenient the local time to be a monotonic source.
The monotony allows to use local time in similar manner as it's
supposed to be used with global time and the 64-bit timer is enough
to guarantee it.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/gen
before entering a loop.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_schedule.c | 13 +++--
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/platform/linux-generic/odp_schedule.c
b/platform/linux-generic/odp_schedule.c
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org
<mailto:ivan.khoronz...@linaro.org>>
---
platform/linux-generic/odp_timer.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/platform/linux-generic/odp_timer.c
b/platform/linux-generic/odp_t
ation/scheduler/scheduler.h
index c869e41..bba79aa 100644
--- a/test/validation/scheduler/scheduler.h
+++ b/test/validation/scheduler/scheduler.h
@@ -14,6 +14,7 @@ void scheduler_test_wait_time(void);
void scheduler_test_num_prio(void);
void scheduler_test_queue_destroy(void);
void scheduler_
thread running on CPU0.
This patch series helps to close a bug:
https://bugs.linaro.org/show_bug.cgi?id=1449
Since v1:
- printed name of the timer pool while ticker overrun
- use "high" resolution instead of "low"
Ivan Khoronzhuk (3):
example: timer: warn if timeout less tha
that timeout is less then
resolution instead of silence.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/timer/odp_timer_test.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/example/timer/odp_timer_test.c b/example/timer/odp_timer_test.c
index 94619e4..aee01c8
On 10.11.15 16:47, Ola Liljedahl wrote:
On 10 November 2015 at 15:40, Ivan Khoronzhuk <ivan.khoronz...@linaro.org
<mailto:ivan.khoronz...@linaro.org>> wrote:
If tick is late then application should be warned about this.
It means that actual timer resolution is worse t
that ideally fits in 10ms.
But user can change it to be smaller, in case if CPU0 is isolated
and it handles only the timer ticks. This patch helps user to set
correct timer resolution.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_timer.c | 7 +++
thread is read.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_cpumask_task.c | 4
1 file changed, 4 insertions(+)
diff --git a/platform/linux-generic/odp_cpumask_task.c
b/platform/linux-generic/odp_cpumask_task.c
index 535891c..b4b4f23
or move that
defines from 4 to 3.
Please send updated version.
Maxim.
Maxim, I've split it intentionally as implementation side and app side.
There is no harm to merge it with commit message from the 3-d patch.
On 11/06/2015 16:08, Ivan Khoronzhuk wrote:
Current time API supposes
On 10.11.15 10:45, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of
EXT Ivan Khoronzhuk
Sent: Tuesday, November 10, 2015 10:34 AM
To: Maxim Uvarov; lng-odp@lists.linaro.org
Subject: Re: [lng-odp] [API
On 10.11.15 11:08, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Tuesday, November 10, 2015 10:59 AM
To: Savolainen, Petri (Nokia - FI/Espoo); Maxim Uvarov; lng-
o...@lists.linaro.org
Subject: Re
ignore it.
On 10.11.15 12:58, Ivan Khoronzhuk wrote:
If tick is lost then application should be warned about this.
It means that actual timer resolution is worse than expected.
The default timer resolution is set 10ms, that is equal to jiffy
on most systems. The default resolution is set
that ideally fits in 10ms.
But user can change it to be smaller, in case if CPU0 is isolated
and it handles only the timer ticks. This patch helps user to set
correct timer resolution.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
Since v1:
- changed ODP_ABORT on ODP_ERR
v1:
that ideally fits in 10ms.
But user can change it to be smaller, in case if CPU0 is isolated
and it handles only the timer ticks. This patch helps user to set
correct timer resolution.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
Since v1:
- changed ODP_ABORT on ODP_ERR
v1:
On 10.11.15 13:11, Maxim Uvarov wrote:
On 11/10/2015 14:01, Ivan Khoronzhuk wrote:
If tick is lost then application should be warned about this.
It means that actual timer resolution is worse than expected.
The default timer resolution is set 10ms, that is equal to jiffy
on most systems
/016772.html
Ivan Khoronzhuk (3):
example: timer: warn if timeout less than resolution
linux-generic: odp_timer: warn if tick is late
linux-generic: cpumask: warn that CPU0 is used by control and worker
thread
example/timer/odp_timer_test.c| 3 +++
platform/linux-generic
that timeout is less then
resolution instead of silence.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/timer/odp_timer_test.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/example/timer/odp_timer_test.c b/example/timer/odp_timer_test.c
index 94619e4..aee01c8
thread is read.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_cpumask_task.c | 4
1 file changed, 4 insertions(+)
diff --git a/platform/linux-generic/odp_cpumask_task.c
b/platform/linux-generic/odp_cpumask_task.c
index 535891c..b4b4f23
that ideally fits in 10ms.
But user can change it to be smaller, in case if CPU0 is isolated
and it handles only the timer ticks. This patch helps user to set
correct timer resolution.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_timer.c | 6 ++
On 10.11.15 13:37, Ivan Khoronzhuk wrote:
On 10.11.15 13:11, Maxim Uvarov wrote:
On 11/10/2015 14:01, Ivan Khoronzhuk wrote:
If tick is lost then application should be warned about this.
It means that actual timer resolution is worse than expected.
The default timer resolution is set 10ms
On 09.11.15 15:09, Maxim Uvarov wrote:
after that patch that scheduler fails:
FAIL: ../../../test/validation/scheduler/scheduler_main
for arm64.
Maxim.
Could you please point what exactly failed?
On 11/05/2015 18:33, Ivan Khoronzhuk wrote:
It's more accurate to compare ticks instead
On 09.11.15 15:29, Ivan Khoronzhuk wrote:
On 09.11.15 15:09, Maxim Uvarov wrote:
after that patch that scheduler fails:
FAIL: ../../../test/validation/scheduler/scheduler_main
for arm64.
Maxim.
Also, pat attention that for ARM64 the time API is based on cycle API
that is not accurate
es.
(don't forget v8 of 3 patch)
then
"[lng-odp] [API-NEXT PATCH] api: cpu: change order of arguments for diff
function"
On 09.11.15 15:29, Ivan Khoronzhuk wrote:
On 09.11.15 15:09, Maxim Uvarov wrote:
after that patch that scheduler fails:
FAIL: ../../../test/validation/scheduler/sch
patch?
Then I can check again.
But after this change it can reveal some other bug only.
Strange.
On 11/05/2015 18:33, Ivan Khoronzhuk wrote:
It's more accurate to compare ticks instead of ns in each
iteration, so calculate wait range before entering the loop.
Signed-off-by: Ivan
- FI/Espoo) wrote:
Reviewed-by: Petri Savolainen <petri.savolai...@nokia.com>
-Original Message-
From: lng-odp [mailto:lng-odp-boun...@lists.linaro.org] On Behalf Of
EXT Ivan Khoronzhuk
Sent: Thursday, November 05, 2015 6:01 PM
To: lng-odp@lists.linaro.org
Subject: [lng-odp] [API
, size of the handler and is only way to
draw attention on it.
But abort is not correct, I agree. I'll send new patch with warn instead of
abort.
On Fri, Nov 6, 2015 at 8:05 AM, Ivan Khoronzhuk <ivan.khoronz...@linaro.org
<mailto:ivan.khoronz...@linaro.org>> wrote:
On 31.10.15
cycle API wich is not stable enough to measure time period.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/Makefile.am| 1 -
platform/linux-generic/include/odp_cpu_internal.h | 29 --
platform/linux-generic/odp
req and then represented as max . So your
temp fix
is very close to this...approach..I'm not talking about freq deviations while
work.
On 06.11.15 12:16, Ivan Khoronzhuk wrote:
This will be looking like this patch. It allows to get rid of cpu_freq_max ... a
little. But temp freq actually breaks CPU cycle
t anyway.
So,
Reviewed-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_system_info.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/platform/linux-generic/odp_system_info.c
b/platform/linux-generic/odp_system_info.c
inde
On 06.11.15 13:00, Maxim Uvarov wrote:
On 11/06/2015 13:54, Ivan Khoronzhuk wrote:
On 06.11.15 12:44, hongbo.zh...@freescale.com wrote:
From: Hongbo Zhang <hongbo.zh...@linaro.org>
In the default dummy function systemcpu(), only cpu_hz[0] and model_str[0]
are set to dummy
On 31.10.15 00:56, Ivan Khoronzhuk wrote:
No need to continue test example if tick is lost. It means that
actual timer resolution is worse than expected. The default timer
resolution is set 10ms, that is equal to jiffy on most systems.
The default resolution is set bearing in mind that on a CPU
] On Behalf Of
EXT Ivan Khoronzhuk
Sent: Thursday, November 05, 2015 5:34 PM
To: lng-odp@lists.linaro.org
Subject: [lng-odp] [API-NEXT PATCH v7 0/6] api: time: unbind CPU cycles
from time API
This seres is intended to unbind time API names from CPU "cycles".
Also remove usage of word &qu
ed instead of odp_time_from_ns(0) for
comparison and initialization.
This patch changes only used time API, it doesn't change used var
names for simplicity.
This time API can be implemented with local timer counter, so
shouldn't be used between threads.
Signed-off-by: Ivan Khoronzhuk <
lign linux-generic implementation with
new time API.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/Makefile.am | 1 +
.../linux-generic/include/odp/plat/time_types.h| 36 +
platform/linux-generic/include/odp/time
It's more accurate to compare ticks instead of ns in each
iteration, so calculate wait range before entering the loop.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff
On 05.11.15 12:39, Ivan Khoronzhuk wrote:
On 05.11.15 12:29, Maxim Uvarov wrote:
Hello Hongbo,
are you looking to Ivan's solutions?
If will take time then please do:
1. Open bug for that issue describing problem and possible solutions.
2. Add ODP_DBG("") with description and bu
ed instead of odp_time_from_ns(0) for
comparison and initialization.
This patch changes only used time API, it doesn't change used var
names for simplicity.
This time API can be implemented with local timer counter, so
shouldn't be used between threads.
Signed-off-by: Ivan Khoronzhuk <
It's more convenient to pass parameters in order, like t2 - t1,
when t2 is supposed to be more.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 2 +-
include/odp/api/time.h| 4 ++--
platform/linux-generic/odp_sche
The word "cycle" is left from old API time names. The "cycle" is
ambiguous word, especially when it can be used for other purposes.
So better to use "tick" or "time" word or just "t" symbol.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz.
ktio_perf: fix potential overflow in wait loop
example: generator: compare ticks instead of ns in loop
Since v1, v2:
- changed series name a little bit
- added opaque type odp_time_t
- added missed spaces in printf
- removed legacy "count" words
Ivan Khoronzhuk (6):
performance: odp_pk
ons of periodic events in loops.
Allows to behave local and global functions in the same manner
All this hides at least two flags under time:
local/global and abs/interval.
Please, confirm if you are OK with this.
On 05.11.15 13:04, Ivan Khoronzhuk wrote:
This seres is intended to unbind time
Petri,
I've sent new version with cmp function correction.
Also sent separately patch which swaps args in diff for cpu_cycles
Please, review.
On 05.11.15 15:36, Ivan Khoronzhuk wrote:
On 05.11.15 15:15, Savolainen, Petri (Nokia - FI/Espoo) wrote:
+
+/**
+ * Compare two times
+ *
+ * @param
It's more convenient to pass parameters in order, like t2 - t1,
when t2 is supposed to be more. Also it's needed to keep in sync
with time API which uses same order.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
Based on series:
"[lng-odp] [API-NEXT PATCH v7 0/6
There cannot be used direct comparison of timestamps of counter
that can overflow, better to compare ranges.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/performance/odp_pktio_perf.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git
bed more clearly timestamps order in functions
- two new patches:
performance: odp_pktio_perf: fix potential overflow in wait loop
example: generator: compare ticks instead of ns in loop
Since v1, v2:
- changed series name a little bit
- added opaque type odp_time_t
- added missed spaces in p
It's more accurate to compare ticks instead of ns in each
iteration, so calculate wait range before entering the loop.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff
ed instead of odp_time_from_ns(0) for
comparison and initialization.
This patch changes only used time API, it doesn't change used var
names for simplicity.
This time API can be implemented with local timer counter, so
shouldn't be used between threads.
Signed-off-by: Ivan Khoronzhuk <
lign linux-generic implementation with
new time API.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/Makefile.am | 1 +
.../linux-generic/include/odp/plat/time_types.h| 36 +
platform/linux-generic/include/odp/time
It's more convenient to pass parameters in order, like t2 - t1,
when t2 is supposed to be more.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 2 +-
include/odp/api/time.h| 4 ++--
platform/linux-generic/odp_sche
The word "cycle" is left from old API time names. The "cycle" is
ambiguous word, especially when it can be used for other purposes.
So better to use "tick" or "time" word or just "t" symbol.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz.
a good step to the right direction.
-Petri
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
It's more convenient to pass parameters in order, like t2 - t1,
when t2 is supposed to be more.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
The same should be done for CPU cycle API
example/generator/odp_generator.c | 2 +-
include/odp/api/time.h
you can use another approach.
Again, for ARM, it should be retrieved from cpufreq first and only if it's off
-> dmi tables,
...then other approaches.
On 02.11.15 13:57, Ivan Khoronzhuk wrote:
Device tree has node for CPU that contains clock-frequency.
(ePAPR v1.1)
But it can be used
o/lng-odp
_______
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
. There is suggestion to make it dependent on core count.
Like if CPU count is more then 2, then exclude core 0.
On Fri, Oct 30, 2015 at 9:05 AM, Ivan Khoronzhuk <ivan.khoronz...@linaro.org
<mailto:ivan.khoronz...@linaro.org>> wrote:
ping
On 27.10.15 17:53, Ivan Khor
in 10ms. But user can change
it to be smaller, in case if CPU0 is isolated and handles the timer
ticks. This patch saves user from incorrect choice by aborting
the application.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_timer.c | 7 +++
The period cannot be less than timer resolution. Also the timer
accuracy is +- 1 tick, that means if period is [1;2) of tick the
timer can expire right after setting, but it doesn't break the
algorithm, so I didn't limit period to be more than 2*resolution.
Signed-off-by: Ivan Khoronzhuk
ping
On 27.10.15 17:53, Ivan Khoronzhuk wrote:
The worker cpu in most cases supposed to be used for one thread.
By default all control threads on CPU0, and odp_cpumask_default_control
returns it. It shouldn't overlap with worker cpumask.
Signed-off-by: Ivan Khoronzhuk <ivan.khor
alls.
-Petri
-Original Message-----
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Tuesday, October 27, 2015 2:22 PM
To: Savolainen, Petri (Nokia - FI/Espoo); lng-odp@lists.linaro.org
Subject: Re: [lng-odp] [API-NEXT PATCH v5 3/5] api: time: unbind CPU cycles
from t
ping.
On 23.10.15 11:35, Ivan Khoronzhuk wrote:
Hi, Petri
Thanks for the reply.
On 23.10.15 10:57, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Thursday, October 22, 2015 3:02 PM
To: lng-odp
On 27.10.15 18:14, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: EXT Nicolas Morey-Chaisemartin [mailto:nmo...@kalray.eu]
Sent: Tuesday, October 27, 2015 5:35 PM
To: Ivan Khoronzhuk; Savolainen, Petri (Nokia - FI/Espoo); lng-
o...@lists.linaro.org
Subject: Re
The worker cpu in most cases supposed to be used for one thread.
By default all control threads on CPU0, and odp_cpumask_default_control
returns it. It shouldn't overlap with worker cpumask.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/odp_cpumask_
*atom, uint32_t val)
{
__atomic_store_n(>v, val, __ATOMIC_RELAXED);
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
e need for converting between internal cycles and
sec+nsec (performance penalty).
sure. That's why conversion is made before loop.
We can discuss this on the calls.
-Petri
-Original Message-
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Tuesday, October 27,
ro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
On 23.10.15 14:18, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Friday, October 23, 2015 2:12 PM
To: Savolainen, Petri (Nokia - FI/Espoo); EXT Maxim Uvarov; lng-
o...@lists.linaro.org
Subject: Re
Hi, Petri
Thanks for the reply.
On 23.10.15 10:57, Savolainen, Petri (Nokia - FI/Espoo) wrote:
-Original Message-
From: EXT Ivan Khoronzhuk [mailto:ivan.khoronz...@linaro.org]
Sent: Thursday, October 22, 2015 3:02 PM
To: lng-odp@lists.linaro.org; Savolainen, Petri (Nokia - FI/Espoo
There cannot be used direct comparison of timestamps of counter
that can overflow, better to compare ranges.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/performance/odp_pktio_perf.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git
lign linux-generic implementation with
new time API.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
platform/linux-generic/Makefile.am | 1 +
.../linux-generic/include/odp/plat/time_types.h| 36 +
platform/linux-generic/include/odp/time
The word "cycle" is left from old API time names. The "cycle" is
ambiguous word, especially when it can be used for other purposes.
So better to use "tick" or "time" word or just "t" symbol.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz.
ed instead of odp_time_from_ns(0) for
comparison and initialization.
This patch changes only used time API, it doesn't change used var
names for simplicity.
This time API can be implemented with local timer counter, so
shouldn't be used between threads.
Signed-off-by: Ivan Khoronzhuk <
Since v1, v2:
- changed series name a little bit
- added opaque type odp_time_t
- added missed spaces in printf
- removed legacy "count" words
Ivan Khoronzhuk (5):
performance: odp_pktio_perf: fix potential overflow in wait loop
example: generator: compare ticks instead of ns in loop
It's more accurate to compare ticks instead of ns in each
iteration, so calculate wait range before entering the loop.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff
be checked in another way.
On 21 October 2015 at 18:15, Ivan Khoronzhuk <ivan.khoronz...@linaro.org> wrote:
On 21.10.15 15:14, Bala Manoharan wrote:
I skipped this coz this as I just realised that since this was an
error packet and it might not be required for platforms to parse an
error
On 21.10.15 08:14, Balasubramanian Manoharan wrote:
Common in-queue destroy function destroy_inq() is moved to
odp_classification_common.c file
Signed-off-by: Balasubramanian Manoharan <bala.manoha...@linaro.org>
Reviewed-by: Ivan Khoronzhuk <ivan.khoronz...@l
was received on error CoS
queue.
Regards,
Bala
It rather needed to identify packet,
and this packet shouldn't be corrupted also.
It think it was correct change.
On 21 October 2015 at 16:50, Ivan Khoronzhuk <ivan.khoronz...@linaro.org> wrote:
Seems in last patch the following was
On 21.10.15 08:14, Balasubramanian Manoharan wrote:
TCP data offset field is used to calculate the data offset for accessing
the sequence number field in the packet.
Signed-off-by: Balasubramanian Manoharan <bala.manoha...@linaro.org>
Reviewed-by: Ivan Khoronzhuk <ivan.khoronz...@l
his change.
For this concrete change:
Reviewed-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
On 21.10.15 08:14, Balasubramanian Manoharan wrote:
check for invalid sequence number is performed during packet create and
it is not required during packet receive
Signed-off-by: Balasubra
On 21.10.15 08:14, Balasubramanian Manoharan wrote:
pool lookup function is removed and pool_default variable is updated
directly during pool create
Signed-off-by: Balasubramanian Manoharan <bala.manoha...@linaro.org>
Reviewed-by: Ivan Khoronzhuk <ivan.khoronz...@l
On 20.10.15 13:36, Bala Manoharan wrote:
Hi Ivan,
Agreed. Splitting can be done.
Do you have any other comments apart from splitting?
Looks good.
Regards,
Bala
On 20 October 2015 at 15:52, Ivan Khoronzhuk <ivan.khoronz...@linaro.org> wrote:
Hi, Bala
Propose to split this on 4 p
_pmr_cos(void);
void configure_pktio_pmr_match_set_cos(void);
void test_pktio_pmr_match_set_cos(void);
-
+int destroy_inq(odp_pktio_t pktio);
#endif /* ODP_BUFFER_TESTSUITES_H_ */
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
+ * to enable applications to generate a printable value that represents
+ * an odp_time_t time.
*/
-uint64_t odp_time_ns_to_cycles(uint64_t ns);
+uint64_t odp_time_to_u64(odp_time_t time);
--
Regards,
Ivan Khoronzhuk
___
lng-odp mailing list
lng-od
Petri,
Could you please review this series.
On 15.10.15 16:02, Ivan Khoronzhuk wrote:
This seres is intended to unbind time API names from CPU "cycles".
Also remove usage of word "cycle" from appropriate places as it's
no more valid. This patch series only change API
Ok.
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
It's more accurate to compare ticks instead of ns in each
iteration, so calculate wait range before entering the loop.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff
cks instead of ns in loop
Since v1, v2:
- changed series name a little bit
- added opaque type odp_time_t
- added missed spaces in printf
- removed legacy "count" words
Ivan Khoronzhuk (5):
performance: odp_pktio_perf: fix potential overflow in wait loop
example: generator: compare tic
ed instead of odp_time_from_ns(0) for
comparison and initialization.
This patch changes only used time API, it doesn't change used var
names for simplicity.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
example/generator/odp_generator.c | 12 +++
includ
There cannot be used direct comparison of timestamps of counter
that can overflow, better to compare ranges.
Signed-off-by: Ivan Khoronzhuk <ivan.khoronz...@linaro.org>
---
test/performance/odp_pktio_perf.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git
301 - 400 of 612 matches
Mail list logo
