Re: panic: 'Sleeping on "t4slptst" with the following non-sleepable locks held' when using cxgbe+lagg

2013-03-22 Thread Navdeep Parhar
Thanks for the bug report. I added this "t4 sleep test" sleep exactly for this purpose -- to catch cases where cxgbe(4) thinks it can sleep but it really can't. There is no need for a core, the stack has all the info I need. Regards, Navdeep On 03/22/13 12:43, Garrett Cooper wrote: > Unfort

panic: 'Sleeping on "t4slptst" with the following non-sleepable locks held' when using cxgbe+lagg

2013-03-22 Thread Garrett Cooper
Unfortunately I don't have swap setup on this machine (not sure why... I'll remedy that soon). Basically I was tcpdump'ing the interface that was doing intense NFS I/O at the same time, and I ran into this crash with cxgbe+lagg. Sources are a bit stale (~1.5 months old). Thanks, -Garrett PS Pl

RE: Re: kern/177139: [igb] igb drops ethernet ports 2 and 3

2013-03-22 Thread Pieper, Jeffrey E
One more question; Are these Dell i350 daughter cards or discrete NICs? Thanks, Jeff From: da...@dr.eclipse.co.uk [mailto:da...@dr.eclipse.co.uk] Sent: Friday, March 22, 2013 3:44 AM To: Pieper, Jeffrey E; freebsd-net@FreeBSD.org Subject: RE: Re: kern/177139: [igb] igb drops ethernet ports 2 and

Re: Netmap ixgbe driver problem on 10Gbps X540-AT2 adapters (Linux only?)

2013-03-22 Thread Walter de Donato
Thank you Michio, actually for me it was necessary to set -w 6 to avoid the adapter reset every second, but what I am observing is a big difference between what is being sent and what is being received on the other side, as reported in the following. This experiment relates to two separate Linux

Re: MPLS

2013-03-22 Thread Alexander V. Chernikov
On 18.03.2013 17:41, Andre Oppermann wrote: On 18.03.2013 13:20, Alexander V. Chernikov wrote: On 17.03.2013, at 23:54, Andre Oppermann wrote: On 17.03.2013 19:57, Alexander V. Chernikov wrote: On 17.03.2013 13:20, Sami Halabi wrote: ITOH OpenBSD has a complete implementation of MPLS out of

RE: Re: kern/177139: [igb] igb drops ethernet ports 2 and 3

2013-03-22 Thread david
  I'll try to arrange this but so far igb2/3 have been fine when going through the switch so tempted to take this as a solution. We loose some switch ports but that is no great loss.  Going to put some traffic through them over the weekend and see if they are still good on Monday. David - Or

Re: Netmap ixgbe driver problem on 10Gbps X540-AT2 adapters (Linux only?)

2013-03-22 Thread Michio Honda
You need to specify "-w 4" on running pkt-gen to wait for that the link resets. Cheers, - Michio On Mar 22, 2013, at 11:16 AM, Walter de Donato wrote: > Dear Tahir, > > I've already tried to disable Rx/Tx pause and autonegotiation but the > result in transmission is always the same. > > I did

Netmap ixgbe driver problem on 10Gbps X540-AT2 adapters (Linux only?)

2013-03-22 Thread Walter de Donato
Dear Tahir, I've already tried to disable Rx/Tx pause and autonegotiation but the result in transmission is always the same. I did the following: $ sudo ifconfig eth2 up $ sudo ethtool -A eth2 autoneg off $ sudo ethtool -A eth2 rx off $ sudo ethtool -A eth2 tx off These are the resulting dmesg o