On Wed 23 Oct 2019 at 17:21, Jamal Hadi Salim wrote:
> On 2019-10-23 9:04 a.m., Vlad Buslov wrote:
>>
>> On Wed 23 Oct 2019 at 15:49, Jamal Hadi Salim wrote:
>>> Hi Vlad,
>>>
>
>>> I understand your use case being different since it is for h/w
>>> offload. If you have time can you test with bat
On 2019-10-23 9:04 a.m., Vlad Buslov wrote:
On Wed 23 Oct 2019 at 15:49, Jamal Hadi Salim wrote:
Hi Vlad,
I understand your use case being different since it is for h/w
offload. If you have time can you test with batching many actions
and seeing the before/after improvement?
Will do.
T
On Wed 23 Oct 2019 at 16:02, Jamal Hadi Salim wrote:
> I shouldve read the thread backward. My earlier email was asking
> similar question to Roman.
>
> On 2019-10-23 2:38 a.m., Vlad Buslov wrote:
>>
>> On Tue 22 Oct 2019 at 21:17, Roman Mashak wrote:
>
>> Hi Roman,
>>
>> I considered it, but d
I shouldve read the thread backward. My earlier email was asking
similar question to Roman.
On 2019-10-23 2:38 a.m., Vlad Buslov wrote:
On Tue 22 Oct 2019 at 21:17, Roman Mashak wrote:
Hi Roman,
I considered it, but didn't find good way to implement my change with
TCA_ROOT_FLAGS. I neede
On Wed 23 Oct 2019 at 15:49, Jamal Hadi Salim wrote:
> Hi Vlad,
>
> On 2019-10-22 10:17 a.m., Vlad Buslov wrote:
>> Currently, significant fraction of CPU time during TC filter allocation
>> is spent in percpu allocator. Moreover, percpu allocator is protected
>> with single global mutex which n
Hi Vlad,
On 2019-10-22 10:17 a.m., Vlad Buslov wrote:
Currently, significant fraction of CPU time during TC filter allocation
is spent in percpu allocator. Moreover, percpu allocator is protected
with single global mutex which negates any potential to improve its
performance by means of recent
On Tue 22 Oct 2019 at 20:09, Marcelo Ricardo Leitner
wrote:
> On Tue, Oct 22, 2019 at 03:52:31PM +, Vlad Buslov wrote:
>>
>> On Tue 22 Oct 2019 at 18:15, Marcelo Ricardo Leitner
>> wrote:
>> > On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
>> >> - Extend actions that are use
On Tue 22 Oct 2019 at 21:17, Roman Mashak wrote:
> Vlad Buslov writes:
>
>> On Tue 22 Oct 2019 at 17:35, Marcelo Ricardo Leitner
>> wrote:
>>> On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
Currently, significant fraction of CPU time during TC filter allocation
is spen
Vlad Buslov writes:
> On Tue 22 Oct 2019 at 17:35, Marcelo Ricardo Leitner
> wrote:
>> On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
>>> Currently, significant fraction of CPU time during TC filter allocation
>>> is spent in percpu allocator. Moreover, percpu allocator is protect
On Tue, Oct 22, 2019 at 03:52:31PM +, Vlad Buslov wrote:
>
> On Tue 22 Oct 2019 at 18:15, Marcelo Ricardo Leitner
> wrote:
> > On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
> >> - Extend actions that are used for hardware offloads with optional
> >> netlink 32bit flags field
On Tue 22 Oct 2019 at 18:15, Marcelo Ricardo Leitner
wrote:
> On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
>> - Extend actions that are used for hardware offloads with optional
>> netlink 32bit flags field. Add TCA_ACT_FLAGS_FAST_INIT action flag and
>> update affected actio
On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
> - Extend actions that are used for hardware offloads with optional
> netlink 32bit flags field. Add TCA_ACT_FLAGS_FAST_INIT action flag and
> update affected actions to not allocate percpu counters when the flag
> is set.
I just
On Tue 22 Oct 2019 at 17:35, Marcelo Ricardo Leitner
wrote:
> On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
>> Currently, significant fraction of CPU time during TC filter allocation
>> is spent in percpu allocator. Moreover, percpu allocator is protected
>> with single global mu
On Tue, Oct 22, 2019 at 05:17:51PM +0300, Vlad Buslov wrote:
> Currently, significant fraction of CPU time during TC filter allocation
> is spent in percpu allocator. Moreover, percpu allocator is protected
> with single global mutex which negates any potential to improve its
> performance by means
Currently, significant fraction of CPU time during TC filter allocation
is spent in percpu allocator. Moreover, percpu allocator is protected
with single global mutex which negates any potential to improve its
performance by means of recent developments in TC filter update API that
removed rtnl loc
15 matches
Mail list logo
