Re: add one more softnet taskq

2017-10-31 Thread Peter Hessler
reads fine, and works for me in light testing. OK phessler@ On 2017 Oct 30 (Mon) at 08:36:34 +0100 (+0100), Alexandr Nedvedicky wrote: :Hello, : :patch below adds additional softnet taskq. This will allow certain degree of :parallelism for packet processing in pf_test(). The current plan is to l

Re: add one more softnet taskq

2017-10-31 Thread Martin Pieuchot
On 31/10/17(Tue) 06:56, Alexandr Nedvedicky wrote: > Hello, > > just in case there is yet another brave soul running the diff below: > > Hrvoje has managed to trip the ASSERT() in ip_send() function at > line 1843. I assume the ip6_send() is prone to same error. > > I'll try to figure ou

Re: add one more softnet taskq

2017-10-31 Thread Alexandr Nedvedicky
Hello, On Tue, Oct 31, 2017 at 09:29:10AM +0100, Martin Pieuchot wrote: > On 31/10/17(Tue) 06:56, Alexandr Nedvedicky wrote: > > Hello, > > > > just in case there is yet another brave soul running the diff below: > > > > Hrvoje has managed to trip the ASSERT() in ip_send() function at > >

Re: add one more softnet taskq

2017-10-30 Thread Alexandr Nedvedicky
Hello, just in case there is yet another brave soul running the diff below: Hrvoje has managed to trip the ASSERT() in ip_send() function at line 1843. I assume the ip6_send() is prone to same error. I'll try to figure out what's going on and send updated diff shortly. Big thanks to Hrv

Re: add one more softnet taskq

2017-10-30 Thread Alexandr Nedvedicky
Hello, thank you for looking at my changes. updated patch is below. The list of changes is as follows: o SOFTNET_TASKS 1 is defined to 1 o call to task_add(): @@ -1839,5 +1839,5 @@ void ip_send(struct mbuf *m) { mq_enqueue(&ipsend_mq, m); - task_add(soft

Re: add one more softnet taskq

2017-10-30 Thread Martin Pieuchot
On 30/10/17(Mon) 10:32, Peter Hessler wrote: > :diff --git a/sys/netinet/ip_input.c b/sys/netinet/ip_input.c > :index 95c9194efcb..33cc3161bcb 100644 > :--- a/sys/netinet/ip_input.c > :+++ b/sys/netinet/ip_input.c > :@@ -1839,5 +1839,5 @@ void > : ip_send(struct mbuf *m) > : { > : mq_enqueue(&i

Re: add one more softnet taskq

2017-10-30 Thread Peter Hessler
On 2017 Oct 30 (Mon) at 08:36:34 +0100 (+0100), Alexandr Nedvedicky wrote: :Hello, : :patch below adds additional softnet taskq. This will allow certain degree of :parallelism for packet processing in pf_test(). The current plan is to let :packets received by even NICs (even ifindex) to be processe

Re: add one more softnet taskq

2017-10-30 Thread Martin Pieuchot
On 30/10/17(Mon) 08:36, Alexandr Nedvedicky wrote: > Hello, > > patch below adds additional softnet taskq. This will allow certain degree of > parallelism for packet processing in pf_test(). The current plan is to let > packets received by even NICs (even ifindex) to be processed by task0, packets

add one more softnet taskq

2017-10-30 Thread Alexandr Nedvedicky
Hello, patch below adds additional softnet taskq. This will allow certain degree of parallelism for packet processing in pf_test(). The current plan is to let packets received by even NICs (even ifindex) to be processed by task0, packets received by odd NICs (odd ifindex) by task1. big thanks sho

Re: One more rtfree(9)

2015-09-10 Thread Claudio Jeker
On Thu, Sep 10, 2015 at 10:25:59AM +0200, Martin Pieuchot wrote: > We're sure the parent is alive when a cloned route is removed from the > table because removing the parent kill all the children. > > Ok? OK > Index: net/route.c >

One more rtfree(9)

2015-09-10 Thread Martin Pieuchot
We're sure the parent is alive when a cloned route is removed from the table because removing the parent kill all the children. Ok? Index: net/route.c === RCS file: /cvs/src/sys/net/route.c,v retrieving revision 1.230 diff -u -p -r1.

one more fxp card

2015-06-26 Thread Kirill Bychkov
Hi. I've ordered two card from ebay and they are working fine on amd64. OK to commit? Index: fxp.4 === RCS file: /cvs/src/share/man/man4/fxp.4,v retrieving revision 1.45 diff -u -r1.45 fxp.4 --- fxp.4 15 Jan 2015 20:37:36 -

one more

2012-10-05 Thread Chris Cappuccio
proper ring size check when runt segment is added Index: if_vr.c === RCS file: /cvs/src/sys/dev/pci/if_vr.c,v retrieving revision 1.115 diff -u -r1.115 if_vr.c --- if_vr.c 18 Sep 2012 14:49:44 - 1.115 +++ if_vr.c 5 Oc