Richard Laing alliedtelesis.co.nz> writes:
> >>> Fair enough, I will look at making it a sysctl option. I guess the
> >>> default can be the current behaviour.
> Cheers
> Richard
>
Hi Richard,
Could you please share if this new patch for per-flow IPv4 ECMP has landed?
If so, which version s
On Tue, Aug 4, 2015 at 12:37 AM, Oleg A Arkhangelsky wrote:
>
>
> 04.08.2015, 04:29, "Richard Laing" :
>> Enable flow-based ECMP.
>
> Looks like your approach is only restricted to the case when sockets are
> involved. What about forwarded traffic (IP routing case)?
>
skb_get_hash_flowi4 can be ca
Hi Oleg,
I hadn't seen that patch, thanks, it looks like a pretty thorough solution.
Best Regards,
Richard
On 08/04/2015 07:37 PM, Oleg A Arkhangelsky wrote:
>
> 04.08.2015, 04:29, "Richard Laing" :
>> Enable flow-based ECMP.
> Looks like your approach is only restricted to the case when sockets
Hi Stephen,
Given that fib_nhs is currently an int I would rather keep live_nexthops
also as an int to match, probably fib_nhs could at least be set as
unsigned or changed to u16 or even u8 perhaps.
Best Regards,
Richard
On 08/04/2015 05:31 PM, Stephen Hemminger wrote:
> On Tue, 4 Aug 2015 13
04.08.2015, 04:29, "Richard Laing" :
> Enable flow-based ECMP.
Looks like your approach is only restricted to the case when sockets are
involved. What about forwarded traffic (IP routing case)?
Have you seen similar work done by Peter Nørlund?
http://comments.gmane.org/gmane.linux.kernel.api/1
On Tue, 4 Aug 2015 13:28:47 +1200
Richard Laing wrote:
> diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
> index 5fa643b..7db9f72 100644
> --- a/include/net/ip_fib.h
> +++ b/include/net/ip_fib.h
> @@ -117,6 +117,8 @@ struct fib_info {
> #ifdef CONFIG_IP_ROUTE_MULTIPATH
> int
Enable flow-based ECMP.
Currently if equal-cost multipath is enabled the kernel chooses between
equal cost paths for each matching packet, essentially packets are
round-robined between the routes. This means that packets from a single
flow can traverse different routes. If one of the routes experi
On 07/30/2015 04:12 AM, Tom Herbert wrote:
> On Tue, Jul 28, 2015 at 11:11 PM, Michal Kubecek wrote:
>> On Tue, Jul 28, 2015 at 09:20:02PM +, Richard Laing wrote:
> diff --git a/include/net/flow.h b/include/net/flow.h
> index 8109a15..d1d933d 100644
> --- a/include/net/flow.h
>>>
On Tue, Jul 28, 2015 at 11:11 PM, Michal Kubecek wrote:
> On Tue, Jul 28, 2015 at 09:20:02PM +, Richard Laing wrote:
>> >> diff --git a/include/net/flow.h b/include/net/flow.h
>> >> index 8109a15..d1d933d 100644
>> >> --- a/include/net/flow.h
>> >> +++ b/include/net/flow.h
>> >> @@ -79,6 +79,1
Hello,
On Tue, 28 Jul 2015, Richard Laing wrote:
> From: Richard Laing
>
> Enable flow-based ECMP.
>
> Currently if equal-cost multipath is enabled the kernel chooses between
> equal cost paths for each matching packet, essentially packets are
> round-robined between the routes. This
From: Michal Kubecek
Date: Wed, 29 Jul 2015 08:11:48 +0200
> Having both sysctl and per-route attribute would be perfect, of
> course.
+1
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger
On Tue, Jul 28, 2015 at 09:20:02PM +, Richard Laing wrote:
> >> diff --git a/include/net/flow.h b/include/net/flow.h
> >> index 8109a15..d1d933d 100644
> >> --- a/include/net/flow.h
> >> +++ b/include/net/flow.h
> >> @@ -79,6 +79,10 @@ struct flowi4 {
> >>#define fl4_ipsec_spiuli.sp
On 07/29/2015 09:15 AM, Tom Herbert wrote:
> On Tue, Jul 28, 2015 at 2:02 PM, Tom Herbert wrote:
>> On Mon, Jul 27, 2015 at 7:27 PM, Richard Laing
>> wrote:
>>> From: Richard Laing
>>>
>>> Enable flow-based ECMP.
>>>
>>> Currently if equal-cost multipath is enabled the kernel chooses between
>
On 07/29/2015 09:02 AM, Tom Herbert wrote:
> On Mon, Jul 27, 2015 at 7:27 PM, Richard Laing
> wrote:
>> From: Richard Laing
>>
>> Enable flow-based ECMP.
>>
>> Currently if equal-cost multipath is enabled the kernel chooses between
>> equal cost paths for each matching packet, essentially packe
On 07/29/2015 07:51 AM, Stephen Hemminger wrote:
> On Tue, 28 Jul 2015 02:27:57 +
> Richard Laing wrote:
>
>> From: Richard Laing
>>
>> Enable flow-based ECMP.
>>
>> Currently if equal-cost multipath is enabled the kernel chooses between
>> equal cost paths for each matching packet, essentia
On Tue, Jul 28, 2015 at 2:02 PM, Tom Herbert wrote:
> On Mon, Jul 27, 2015 at 7:27 PM, Richard Laing
> wrote:
>> From: Richard Laing
>>
>> Enable flow-based ECMP.
>>
>> Currently if equal-cost multipath is enabled the kernel chooses between
>> equal cost paths for each matching packet, essential
On Mon, Jul 27, 2015 at 7:27 PM, Richard Laing
wrote:
> From: Richard Laing
>
> Enable flow-based ECMP.
>
> Currently if equal-cost multipath is enabled the kernel chooses between
> equal cost paths for each matching packet, essentially packets are
> round-robined between the routes. This means t
On Tue, 28 Jul 2015 02:27:57 +
Richard Laing wrote:
> From: Richard Laing
>
> Enable flow-based ECMP.
>
> Currently if equal-cost multipath is enabled the kernel chooses between
> equal cost paths for each matching packet, essentially packets are
> round-robined between the routes. This me
On Tue, Jul 28, 2015 at 02:27:57AM +, Richard Laing wrote:
> From: Richard Laing
>
> Enable flow-based ECMP.
>
> Currently if equal-cost multipath is enabled the kernel chooses between
> equal cost paths for each matching packet, essentially packets are
> round-robined between the routes. Th
From: Richard Laing
Enable flow-based ECMP.
Currently if equal-cost multipath is enabled the kernel chooses between
equal cost paths for each matching packet, essentially packets are
round-robined between the routes. This means that packets from a single
flow can traverse different routes. If on
20 matches
Mail list logo
