[PATCH] multipath routing algorithm, better patch

[PATCH] multipath routing algorithm, better patch

[Patrick Jenkins]

Hi,

The last patch wont work, this should.

This patch assigns the multipath routing algorithm into the fib_info
struct's fib_mp_alg variable. Previously, the algorithm was always set to
IP_MP_ALG_NONE which was incorrect. This patch corrects the problem by
assigning the correct value when a fib_info is initialized.

This patch was tested against kernel 2.6.12.1 for all multipath routing
algorithms (none, round robin, interface round robin, random, weighted
random).

Please look this patch over and apply it to the kernel. I have been unable
to contact the creators of the multipath algorithm feature so this is why
I sent it to the lkml.

I am not a member of the list so please cc me in the reply.

Signed-off-by: Patrick Jenkins <patjenk@wam.umd.edu>

--- linux-2.6.12/net/ipv4/fib_semantics.c.orig  2005-06-30 
20:07:13.000000000 -0400
+++ linux-2.6.12/net/ipv4/fib_semantics.c       2005-06-30 
20:06:36.000000000 -0400
@@ -650,9 +650,22 @@ fib_create_info(const struct rtmsg *r, s
  #else
         const int nhs = 1;
  #endif
+
  #ifdef CONFIG_IP_ROUTE_MULTIPATH_CACHED
+
+#ifdef CONFIG_IP_ROUTE_MULTIPATH_RR
+       u32 mp_alg = IP_MP_ALG_RR;
+#elif CONFIG_IP_ROUTE_MULTIPATH_DRR
+       u32 mp_alg = IP_MP_ALG_DRR;
+#elif CONFIG_IP_ROUTE_MULTIPATH_RANDOM
+       u32 mp_alg = IP_MP_ALG_RANDOM;
+#elif CONFIG_IP_ROUTE_MULTIPATH_WRANDOM
+       u32 mp_alg = IP_MP_ALG_WRANDOM;
+#else
         u32 mp_alg = IP_MP_ALG_NONE;
-#endif
+#endif /* multipath algorithm determination */
+
+#endif /* CONFIG_IP_ROUTE_MULTIPATH_CACHED */

         /* Fast check to catch the most weird cases */
         if (fib_props[r->rtm_type].scope > r->rtm_scope)

Re: [PATCH] multipath routing algorithm, better patch

[Patrick McHardy]

Patrick Jenkins wrote:
> Hi,
> 
> The last patch wont work, this should.
> 
> This patch assigns the multipath routing algorithm into the fib_info
> struct's fib_mp_alg variable. Previously, the algorithm was always set to
> IP_MP_ALG_NONE which was incorrect. This patch corrects the problem by
> assigning the correct value when a fib_info is initialized.
> 
> This patch was tested against kernel 2.6.12.1 for all multipath routing
> algorithms (none, round robin, interface round robin, random, weighted
> random).

Multiple algorithms can be compiled in at once, so this patch is wrong.
mp_alg is supplied by userspace:

        if (rta->rta_mp_alg) {
                mp_alg = *rta->rta_mp_alg;

                if (mp_alg < IP_MP_ALG_NONE ||
                    mp_alg > IP_MP_ALG_MAX)
                        goto err_inval;
        }

If it isn't set correctly its an iproute problem. Did you actually
experience any problems?

Regards
Patrick

Re: [PATCH] multipath routing algorithm, better patch

[Thomas Graf]

* Patrick McHardy <42C4919A.5000009@trash.net> 2005-07-01 02:43
> Multiple algorithms can be compiled in at once, so this patch is wrong.
> mp_alg is supplied by userspace:
> 
>         if (rta->rta_mp_alg) {
>                 mp_alg = *rta->rta_mp_alg;
> 
>                 if (mp_alg < IP_MP_ALG_NONE ||
>                     mp_alg > IP_MP_ALG_MAX)
>                         goto err_inval;
>         }
> 
> If it isn't set correctly its an iproute problem. Did you actually
> experience any problems?

Well, my patch for iproute2 to enable multipath algorithm selection
is currently being merged to Stephen together with the ematch bits.
We had to work out a dependency on GNU flex first (the berkley
version uses the same executable names) so the inclusion was
delayed a bit.

Re: [PATCH] multipath routing algorithm, better patch

[Patrick McHardy]

Thomas Graf wrote:
> * Patrick McHardy <42C4919A.5000009@trash.net> 2005-07-01 02:43
> 
>>If it isn't set correctly its an iproute problem. Did you actually
>>experience any problems?
> 
> Well, my patch for iproute2 to enable multipath algorithm selection
> is currently being merged to Stephen together with the ematch bits.
> We had to work out a dependency on GNU flex first (the berkley
> version uses the same executable names) so the inclusion was
> delayed a bit.

So its no problem but simply missing support. BTW, do you know if
Stephen's new CVS repository is exported somewhere?

Regards
Patrick

Re: [PATCH] multipath routing algorithm, better patch

[Thomas Graf]

* Patrick McHardy <42C59ABA.1070305@trash.net> 2005-07-01 21:34
> So its no problem but simply missing support. BTW, do you know if
> Stephen's new CVS repository is exported somewhere?

cvs -d :pserver:cvsanon@developer.osdl.org/repos cvs login
cvs -d :pserver:cvsanon@developer.osdl.org/repos cvs co iproute2

Re: [PATCH] multipath routing algorithm, better patch

[Patrick Jenkins]

>Thomas Graf wrote:
>> * Patrick McHardy <42C4919A.5000009@trash.net> 2005-07-01 02:43
>>
>>>If it isn't set correctly its an iproute problem. Did you actually
>>>experience any problems?
>>
>> Well, my patch for iproute2 to enable multipath algorithm selection
>> is currently being merged to Stephen together with the ematch bits.
>> We had to work out a dependency on GNU flex first (the berkley
>> version uses the same executable names) so the inclusion was
>> delayed a bit.
>
>So its no problem but simply missing support. BTW, do you know if
>Stephen's new CVS repository is exported somewhere?

Yes, we did experience a problem. The routing doesnt work as advertised 
(i.e. it wont utilize the multiple routes). Patrick is correct in saying 
its missing support.

>From what you are saying it seems the problem will be corrected by a new 
feature in iproute2 that allows the user to select this ability. Is this 
correct? Also, does this mean my patch is not needed? It seems to me that 
it should be supported somehow in the kernel seeing as how everyone might 
not utilize iproute2.

Once again, please cc me in the reply because I am not a member of the 
list.

Thanks,
Patrick Jenkins