Ok, it's done it again and this time I have something on console : <3>Debug : Sleeping function called from invalid context at include/linux/rwsem.h:43 in_atomic:(0) irqs_disabled:(1) What's that all about ? On Wed, 2005-08-31 at 09:48 +0900, Naoki wrote: > The plot thickens, I now have two of these boxes running. One plugged into a Foundry switch > and the other into a cheap 10/100 switching hub. > > The box connected to the hub says : > > eth0: Tigon3 [partno(BCM95704A6) rev 2003 PHY(5704)] (PCIX:100MHz:64-bit) 10/100/1000BaseT Ethernet > tg3: eth0: Link is up at 100 Mbps, full duplex. > tg3: eth0: Flow control is on for TX and on for RX. > > The box hooked up to the foundry ( which is now nailed at 100-full ) says : > > eth0: Tigon3 [partno(BCM95704A6) rev 2003 PHY(5704)] (PCIX:100MHz:64-bit) 10/100/1000BaseT Ethernet > tg3: eth0: Link is up at 100 Mbps, full duplex. > tg3: eth0: Flow control is off for TX and off for RX. > > Now I just checked on of my production boxes also hooked up to a foundry but using a slightly different Broadcom chip : > > eth0: Tigon3 [partno(BCM95703A30) rev 1002 PHY(5703)] (PCIX:100MHz:64-bit) 10/100/1000BaseT Ethernet > tg3: eth0: Link is up at 100 Mbps, full duplex. > tg3: eth0: Flow control is off for TX and off for RX. > > So my only question now is this lack of flow control a tg3 driver problem, or a foundry switch problem? And how do I easily find out? > > Any suggestions from the group would be appreciated. > > > >From Aug 24th : > > You know what, I did, and it's been ok since. However I still have the > kernel message saying no flow control. I'm wondering if that's a > problem, it would sound like one to me. > > On Mon, 2005-08-22 at 13:47 +0200, Mogens Kjaer wrote: > > > > > Have you also disabled auto-negotiation on the switch? > > > > Mogens > > -- > > Mogens Kjaer, Carlsberg A/S, Computer Department > > Gamle Carlsberg Vej 10, DK-2500 Valby, Denmark > > Phone: +45 33 27 53 25, Fax: +45 33 27 47 08 > > Email: mk@xxxxxx Homepage: http://www.crc.dk >
