Hi Malik,
I have an idea about what is going on. It isn't a problem with the XW6000 per se, it is a problem with the switch not matching the speed/duplex setting of the XW6000.
If you check your switch, I think you'll find that either auto-negotiation is turned off on that port, or it is not set at 100FD.
You can see in your previous posting that XW6000 is not autonegotiating at boot, so either the switch must do it, or that port on the switch must be manually set to match the speed/duplex of the XW6000.
There was a time in the not too distant past, before N-Way auto-negotiation came out, when you would have the same result as you are seeing if your machine's NIC and the attached switch were BOTH trying to auto-negotiate. (but that probably has nothing to do with this problem, but might explain why your card is not set to auto-negotiate at boot.) N-way cleared up that problem.
-- Found in Sydney, Keith
Malik Jayawardena wrote:
Hi there,
I'm having this strange network problems with some new HP XW6000 workstations. They all have a Broadcom 5700 10/100/1000 onboard NIC.
All of these machines are running RedHat 7.3 with either kernel 2.4.18-18.7.xsmp or 2.4.18-27.7.xsmp and using the 'tg3' module
The original HP XW6000s seem to work fine, but we've got a few new ones recently (with OS installed) where, after boot up the network is REALLY slow..
We do alot of remote X-Serving and it is really apparent. The only way to fix this problem seems to be to stop networking unload the 'tg3' module, then reload it, restart networking and then it seems to be on par with the rest of the systems we have.
I've tried recompiling the new 'bcm57xx' module and running it, which works fine, but it has the same effect on boot as the 'tg3' driver. And yes, I have updated my modules.conf and it does install the new module on boot. But still is slow until the module is reloaded.
This is the output I've gotten from ethtool & mii-tool. The are somewhat different before and after reloading:
After boot(Slow):
# mii-tool eth0: 100 Mbit, full duplex, link ok
# ethtool eth0 Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: Not reported Advertised auto-negotiation: No Speed: 100Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: off Supports Wake-on: g Wake-on: d Link detected: yes
After Module Reload(Quick):
# mii-tool eth0: negotiated 100baseTx-FD, link ok
#ethtool eth0 Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Half 1000baseT/Full Advertised auto-negotiation: Yes Speed: 100Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on Supports Wake-on: g Wake-on: d Link detected: yes
I'm guessing it's a startup script kind of problem or something like that, but instead of tediously hunting through them, if someone is familiar with this problem or has some idea of what's going on any help would be greatly appreciated.
Cheers and thanks muchly,
Mal
-- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
