Re: Does FreeBSD have sendmmsg or recvmmsg system calls?

2016-01-13 Thread Konstantin Belousov
On Tue, Jan 12, 2016 at 04:53:25PM +0200, Boris Astardzhiev wrote: > Hello again, > > In my spare time I did the following simple libc-only implementation of the > syscalls. > I did some tests in a VM adapting these experiments: > https://blog.cloudflare.com/how-to-receive-a-million-packets/ > >

Re: Does FreeBSD have sendmmsg or recvmmsg system calls?

2016-01-13 Thread Sepherosa Ziehau
On Tue, Jan 12, 2016 at 10:53 PM, Boris Astardzhiev wrote: > Hello again, > > In my spare time I did the following simple libc-only implementation of the > syscalls. > I did some tests in a VM adapting these experiments: > https://blog.cloudflare.com/how-to-receive-a-million-packets/ On Dragonfl

Re: Kernel panic from lagg_ioctl and lagg_port_ioctl

2016-01-13 Thread Hans Petter Selasky
On 01/12/16 22:13, Lewis, Fred wrote: In earlier versions of lagg_ioctl() (e.g. stable/10/r171247) all of the callout vectors are set to NULL which I think will prevent the problem. Similar NULLing code is also in stable/7. I didn't check other releases. Don't forget to drain the callouts befo

Re: Kernel panic from lagg_ioctl and lagg_port_ioctl

2016-01-13 Thread Lakshmi Narasimhan Sundararajan
Hi Hans, Yes, we understand that piece of code. But Panasas wants this change to MFC to 10.3. This issue is not seen in 11. Would that happen? I have not seen acknowledgement of this issue or confirmation of MFC to 10.3. That information will help us plan. Regards, LN On Wed, Jan 13, 2016 at 6:0

Re: Kernel panic from lagg_ioctl and lagg_port_ioctl

2016-01-13 Thread Hans Petter Selasky
On 01/13/16 14:15, Lewis, Fred wrote: Hi Hans, Can you give an example of draining the callouts? Hi, man 9 callout_drain and man 9 callout_async_drain (11-current) callout_drain() might sleep and needs a context which allows for that. Else use callout_async_drain(). --HPS _

Re: Kernel panic from lagg_ioctl and lagg_port_ioctl

2016-01-13 Thread Lewis, Fred
Hi Hans, Can you give an example of draining the callouts? Thanks, -Fred On 1/13/16 7:33 AM, "Lakshmi Narasimhan Sundararajan" mailto:lakshm...@msystechnologies.com>> wrote: Hi Hans, Yes, we understand that piece of code. But Panasas wants this change to MFC to 10.3. This issue is not seen in

ipfw NAT, igb and hardware checksums

2016-01-13 Thread Karim Fodil-Lemelin
Hi, I've hit a very interesting problem with ipfw-nat and local TCP traffic that has enough TCP options to hit a special case in m_megapullup(). Here is the story: I am using the following NIC: igb0@pci0:4:0:0:class=0x02 card=0x8086 chip=0x150e8086 rev=0x01 hdr=0x00 And wh

[Bug 203524] TCP checksum failed on igb network adapter

2016-01-13 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=203524 fodillemlinka...@gmail.com changed: What|Removed |Added CC||fodillemlinka...@gmail.

Re: ipfw NAT, igb and hardware checksums

2016-01-13 Thread Adrian Chadd
This looks mostly sensible. hm! -a On 13 January 2016 at 11:55, Karim Fodil-Lemelin wrote: > Hi, > > I've hit a very interesting problem with ipfw-nat and local TCP traffic that > has enough TCP options to hit a special case in m_megapullup(). Here is the > story: > > I am using the following

Re: ipfw NAT, igb and hardware checksums

2016-01-13 Thread Alexander V . Chernikov
13.01.2016, 22:56, "Karim Fodil-Lemelin" : > Hi, > > I've hit a very interesting problem with ipfw-nat and local TCP traffic > that has enough TCP options to hit a special case in m_megapullup(). > Here is the story: > > I am using the following NIC: > > igb0@pci0:4:0:0: class=0x02 card=0x8

[Bug 206186] [patch][netgraph] New netgraph node for calculate IP IP6 TCP UDP checksums

2016-01-13 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=206186 Mark Linimon changed: What|Removed |Added Assignee|freebsd-b...@freebsd.org|freebsd-net@FreeBSD.org -- You are

[Bug 206185] [netgraph][patch] refactoring ng_patch

2016-01-13 Thread bugzilla-noreply
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=206185 Mark Linimon changed: What|Removed |Added Assignee|freebsd-b...@freebsd.org|freebsd-net@FreeBSD.org -- You are