On Sat, 8 Jul 2006, Michael Vince wrote:
I thought I remember a developer working on the em driver saying just before 6.1 was released that this reset was needed and couldn't be avoided to ensure performance of the device to work at its best, I can't remember his explanation, but this topic has come up before, of course anything is possible to fix.
The thing is, and I may be mis-understanding the explanations so far, the 'reset' is to renegotiate the connection ... if that is the case, and both the switch and the interface are already locked at a speed (in my case, both are hard coded to "100baseTX full duplex", then what is there to re-negotiate?
And, why does it appear that *only* the em driver/interface requires this? I run bge and fxp interfaces on this same network, against the same switch, all locked at the same speed, and only the em driver exhibits this problem ... in fact, its only the *newer* em driver that does, as I have one server on the network, using an em interface, that is running an older FreeBSD 4.x kernel, that performs the same as the bge/fxp (ie. perfectly)
---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email . [EMAIL PROTECTED] MSN . [EMAIL PROTECTED] Yahoo . yscrappy Skype: hub.org ICQ . 7615664 _______________________________________________ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "[EMAIL PROTECTED]"
