Thank you Bala for the clarifications.
However,
On Fri, Jan 12, 2018 at 3:08 PM, Bala Manoharan <bala.manoha...@linaro.org>
wrote:
>
> On 12 January 2018 at 17:43, Oriol Arcas <or...@starflownetworks.com>
> wrote:
>
>> Hi again,
>>
>> I two doubts
e
parameters (scheduling, type, etc.).
A second question is that for num_queue > 1 the queues are created
automatically by ODP, but in odp_cos_destroy() they are not destroyed. So I
assume either it should be done there, or the user is responsible for
retrieving and destroying those queues.
Re
Hi Bala,
I didn't find any hashing parameter in the CoS API, is it implemented or a
suggestion?
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jan 4, 2018 at 5:11 PM, Bala Manoharan <bala.manoha...@linaro.org>
wrote:
> Hi,
>
> In addition configuring the hashi
and sticking to a CPU.
Thank you all.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jan 4, 2018 at 11:01 AM, Bogdan Pricope <bogdan.pric...@linaro.org>
wrote:
> I guess, the issue is not how to hash traffic in different scheduled
> queues but how to lock a scheduled queue
queues?
I don't know if there is something similar in ODP already which I have
missed. I'll thank any feedback!
Best regards,
--
Oriol Arcas
Software Engineer
Starflow Networks
for the app to
query the state of the PKTIOs (keep calling odp_schedule() until
odp_lookup() says all the PKTIOs are closed).
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, May 2, 2017 at 9:32 AM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> Correct steps are:
> 1. od
h;
q = odp_tm_queue_create(tm, _params);
odp_tm_queue_connect(q, node)
odp_tm_sched_params_init(_params);
for (i = 0; i < ODP_TM_MAX_PRIORITIES; ++i) {
sch_params.sched_modes[i] = mode;
sch_params.sched_weights[i] = 255;
}
sch = odp_tm_sched_create("TM-SCH-3", _params);
odp_tm_node_sched_conf
Hi Bill, I'll try to be there.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, May 2, 2017 at 5:01 AM, Bill Fischofer <bill.fischo...@linaro.org>
wrote:
> Oriol: I've added this and the TM issue you raised to the agenda for
> Tuesday's ODP public call. Please join us if
it
still doesn't apply queuing (thus priorities and weights), but it is
clearly congested because it rejects packets.
Any thoughts on this?
Thank you in advance.
--
Oriol Arcas
Software Engineer
Starflow Networks
, is there a proper way of doing this?
--
Oriol Arcas
Software Engineer
Starflow Networks
ld be done, but other functions behave differently. See
https://www.securecoding.cert.org/confluence/pages/viewpage.action?pageId=6619179
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Apr 13, 2017 at 8:55 PM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> Why not just set errno
-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/_ishmphy.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/platform/linux-generic/_ishmphy.c
b/platform/linux-generic/_ishmphy.c
index d519af6..87c33ef 100644
--- a/platform/linux-generic/_ishmphy.c
+++ b/pl
Maxim, the fix in Monarch is quite different than in master. In Monarch, it
affects odp_shm_reserve() in odp_shared_memory.c. In master, it has been
moved to _odp_ishmphy_map() in _ishmphy.c.
Which patch should I submit? Or both?
--
Oriol Arcas
Software Engineer
Starflow Networks
On Wed, Apr 12
On Tue, Feb 21, 2017 at 6:50 PM, Bill Fischofer <bill.fischo...@linaro.org>
wrote:
>
>
> On Tue, Feb 21, 2017 at 11:35 AM, Oriol Arcas <or...@starflownetworks.com>
> wrote:
>
>> Hi,
>>
>> We have observed the problem with ODP and/or Netmap, with ker
is happening.
The performance can be lower than socket mmap, but they seem reliable. We
may add readv(2) and writev(2) to improve the performance.
Thank you for all your feedback. I'll add any news regarding this issue to
this thread.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Mon, Feb 20
transactions at kernel level, triggered by socket mmap or
netmap... I don't know if this matches your experience.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Mon, Feb 20, 2017 at 11:39 AM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> version from .travis file.
>
> On 20 Febr
Arcas
Software Engineer
Starflow Networks
On Fri, Feb 17, 2017 at 8:26 PM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> On 02/17/17 18:45, Oriol Arcas wrote:
> > I tried setting the MAC addresses. In my local test, the problem
> > disappeared, but I doubt that it's been
o try netmap.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Fri, Feb 17, 2017 at 1:23 PM, Elo, Matias (Nokia - FI/Espoo) <
matias@nokia-bell-labs.com> wrote:
>
> > On 17 Feb 2017, at 14:03, Oriol Arcas <or...@starflownetworks.com>
> wrote:
> >
> > Hi,
namespace.
I have tried it with ODP_PKTIO_DISABLE_SOCKET_MMAP, all the different
scheduling modes and -c 1.
Our next test would be using DPDK or netmap, if they can be used with veth
interfaces.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Fri, Feb 17, 2017 at 11:46 AM, Elo, Matias (Nokia - FI/E
PS: sorry for the longitude of the email, I tried to be as detailed as
possible given the weirdness of the problem...
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Feb 16, 2017 at 4:37 PM, Oriol Arcas <or...@starflownetworks.com>
wrote:
> Hi,
>
> We have been using
ith the TCP stack
- OpenVPN+ODP is somehow causing this
- ODP on a VM causes this behavior
I hope I explained the problem in an understandable way. Please do not
hesitate to ask for clarifications or new experiments.
We will appreciate any comments, questions and suggestions, especially
regarding the
Regarding the last question pointed out by Bill, max_queue_num =
tm_system->next_queue_num, which is always "current queues in the TM + 1".
So the loop, as it is now, visits all the elements in the array.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Wed, Aug 24, 2016 at
I'll try to be there. Thanks, Bill.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Fri, Jul 29, 2016 at 5:14 PM, Bill Fischofer <bill.fischo...@linaro.org>
wrote:
> That would be great and I like that approach, however this would be a
> larger restructuring we'd want to loo
.html
Anyway, thanks for the feedback, Bill.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jul 28, 2016 at 8:12 PM, Bill Fischofer <bill.fischo...@linaro.org>
wrote:
> That looks legit. Please do post a patch for it and let us know how that
> works as well as the other pa
iol Arcas
Software Engineer
Starflow Networks
On Fri, Jul 29, 2016 at 3:31 PM, Oriol Arcas <or...@starflownetworks.com>
wrote:
> I see. Then, if the 1-based _odp_int_pkt_queue_t is used as index to
> access pool->queue_num_tbl, and the 0 block is discarded, the malloc should
> al
I see. Then, if the 1-based _odp_int_pkt_queue_t is used as index to access
pool->queue_num_tbl, and the 0 block is discarded, the malloc should
allocate max + 1, right?
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jul 28, 2016 at 10:34 PM, Bill Fischofer <bill.fischo...@lina
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/odp_pkt_queue.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/platform/linux-generic/odp_pkt_queue.c
b/platform/linux-generic/odp_pkt_queue.c
index 7734ee9..be638a3
Hello Bill,
While checking your patch, I came up with this potential other bug:
https://bugs.linaro.org/show_bug.cgi?id=2449
If it is not a false positive, I can provide a fix and proceed with testing
your patch.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jul 28, 2016 at 12:22
Thanks for reporting the bug, Mike.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Jul 26, 2016 at 5:44 PM, <bugzilla-dae...@bugs.linaro.org> wrote:
> https://bugs.linaro.org/show_bug.cgi?id=2444
>
> --- Comment #1 from Mike Holmes <mike.hol...@linaro.org> --
would be an odp_schedule()-like interface that
would fairly schedule all the odp_tm_t.
There are some other details that could be discussed, like an optional
TM/PKTOUT-level priority. Anyway, I think that TM-system scheduling is
necessary.
Best,
--
Oriol Arcas
Software Engineer
Starflow Networks
On
it is relatively poorly
documented, and has no examples), like leaving some components
misconfigured or not configured at all, but I wanted to ask if this drop in
performance is considered normal in the current implementation state,
before going further with the performance analysis.
Thanks,
--
Oriol
Sorin, Maxim,
At Starflow we just had the same problem. I just sent the patch to increase
the number of pools to 64.
Oriol
--
Oriol Arcas
Software Engineer
Starflow Networks
On Mon, Jul 18, 2016 at 11:54 AM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> On 07/18/16 12:14, Sorin V
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/include/odp_config_internal.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/platform/linux-generic/include/odp_config_internal.h
b/platform/linux-generic/include/odp_config_internal.h
The API states that the return values of odp_tm_enq() and
odp_tm_enq_with_cnt() are different (the former returns 0 on success),
but the implementations of both were identical.
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/odp_traffic_mngr.c | 7
Understood, Maxim. Sorry, so many things to learn!
--
Oriol Arcas
Software Engineer
Starflow Networks
On Thu, Jun 9, 2016 at 10:04 PM, Maxim Uvarov <maxim.uva...@linaro.org>
wrote:
> Hi Oriol,
>
> I merged this path with description change.
>
> Usually we use or linux-gen
The GNU C++ compilers (4 to 6) don't accept _Static_assert (C11) when
compiling C++ code. A re-definition to static_assert (C++11) is
required. Clang seems not affected.
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/include/odp/api/debug.h | 9 ++
.
Since it is a deep modification, we have not tracked yet the exact problem,
but it may be related to the previous comments.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Jun 7, 2016 at 2:55 PM, Bala Manoharan <bala.manoha...@linaro.org>
wrote:
> Hi,
>
> Just realised that
Helper's chksum.h API had a void* to uint16_t* casting, which is not
legal in C++ (Clang and GCC).
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
Reviewed-by: Bill Fischofer <bill.fischo...@linaro.org>
---
helper/include/odp/helper/chksum.h | 2 +-
1 file changed, 1 ins
Helper's chksum.h API had a void* to uint16_t* casting, which is not
legal in C++ (Clang and GCC).
---
helper/include/odp/helper/chksum.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/helper/include/odp/helper/chksum.h
b/helper/include/odp/helper/chksum.h
index
commands from the schedule queue for
closed PKTIOs using a pktio_stop callback from the scheduler.
Best,
--
Oriol Arcas
Software Engineer
Starflow Networks
On Mon, May 30, 2016 at 12:57 PM, Oriol Arcas <or...@starflownetworks.com>
wrote:
> When a scheduled pktio entry is closed it wi
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
.../linux-generic/include/odp_packet_io_internal.h | 4 ++-
platform/linux-generic/odp_packet_io.c | 29 +-
2 files changed, 26 insertions(+), 7 deletions(-)
diff --git a/platform/linux
Best regards,
--
Oriol Arcas
Software Engineer
Starflow Networks
___
lng-odp mailing list
lng-odp@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/lng-odp
Thank you for the comments, Bill and Maxim.
I provided a PATCHv2. I addressed everything, but for odp_pool.c I left the
inline variable declarations to not break the readability.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Apr 5, 2016 at 9:54 PM, Bill Fischofer <bill.fis
The packet pool may run out of blocks before reaching the low buffer
watermark. A block low watermark is added to fix it.
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
v2: renamed variables and added stats
v1: fix for https://bugs.linaro.org/show_bug.cgi?id=2151
platform
The packet pool may run out of blocks before reaching the low buffer
watermark. A block low watermark is added to fix it.
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/include/odp_pool_internal.h | 21 +++--
platform/linux-generic/odp_
Sorry Bill, we were using 1.6. Let me send another patch for master.
--
Oriol Arcas
Software Engineer
Starflow Networks
On Tue, Apr 5, 2016 at 4:20 PM, Bill Fischofer <bill.fischo...@linaro.org>
wrote:
> This patch needs a rebase as it doesn't apply to the current odp.git
> mast
The packet pool may run out of blocks before reaching the low buffer
watermark. A block low watermark is added to fix it.
Signed-off-by: Oriol Arcas <or...@starflownetworks.com>
---
platform/linux-generic/include/odp_pool_internal.h | 21 +++--
platform/linux-generic/odp_
47 matches
Mail list logo