Re: nve related LOR triggered by lots of small packets, and a hard hang

From: Danny Braniss <danny_at_cs.huji.ac.il>
Date: Wed, 07 Mar 2007 14:05:35 +0200
> On Wed, Feb 21, 2007 at 12:17:52PM +0200, Danny Braniss wrote:
>  > > Pyun YongHyeon wrote:
>  > > > Because your dmesg for nfe(4) shows 'obsoleted if_watchdog interface'
>  > > > message I think you've used stock nfe(4) on CURRENT.
>  > > > Try overhauld nfe(4) at the following URL.
>  > > > 
>  > > > http://people.freebsd.org/~yongari/nfe/if_nfe.c
>  > > > http://people.freebsd.org/~yongari/nfe/if_nfereg.h
>  > > > http://people.freebsd.org/~yongari/nfe/if_nfevar.h
>  > > > 
>  > > > The new nfe(4) has several protections for watchdog timeout errors
>  > > > reported on the driver. One user repored TSO related issues for
>  > > > the new driver so if you encounter strange errors in the new driver
>  > > > please turn off TSO capability(e.g. #ifconfig nfe0 -tso)
>  > > 
>  > hi,
>  > I'm having 2 - probably related - problems:
>  > 1- on a diskless host, it boots via PXE, attaches the NIC
>  >    but somehow the mount root nsf is bypassed.
>  > 2- booting of the disk, and with ifconfig_nfe0="DHCP", dhclient
>  >    fails.
>  > 
>  > if I run:
>  > 	dhclient nfe0
>  > all is fine.
>  > 
>  > adding 'net.inet.tcp.tso="0"' to sysctl.conf fixed 2, but
>  > adding the same to loader.conf did not help with the PXE boot.
>  > 
>  > danny
>  > 
>  > 
> 
> Sorry for late reply.
> Do you use MCP65 based hardware? It seems that MCP65 does not have
> checksum offload capability on the hardware so TSO should not be
> used on MCP65.

it's supposed to be a MCP51, but:
nfe0: <NVIDIA nForce 430 MCP12 Networking Adapter>

> (Due to lack of datasheet I've unconditionally enabled TSO for
> hardwares that have checksum offload capability.)
> dmesg will show you what hardware were attached.
> 
> If your hardware is not MCP65 based one please try latest tcp_output.c
> (rev, 1.125). Recently andre_at_ fixed TSO related bug.

I tried the latest, and it 'seems' to work ok if I boot of the disk, but still having
problems when booting via PXE:

...
nfe0: <NVIDIA nForce 430 MCP12 Networking Adapter> port 0xdc00-0xdc07 mem 0xfe02c000-0xfe02cfff irq 22 at device 20.0 on pci0
nfe0: Ethernet address: 57:6c:a9:f3:18:00
miibus1: <MII bus> on nfe0
e1000phy0: <Marvell 88E1111 Gigabit PHY> PHY 11 on miibus1
e1000phy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseTX-FDX, auto
nfe0: [ITHREAD]
...
nfs_diskless: no interface
...

	danny
PS: maybe it's time to change the Subject:?
Received on Wed Mar 07 2007 - 11:05:44 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:06 UTC