At 10:59 PM 11/30/2001, you wrote: >Dennis wrote: > > > > At 02:50 PM 11/30/2001, Jeff Garzik wrote: > > >Dennis wrote: > > > > I can lock up the eepro driver in about 15 seconds firing 30k pps > at it in > > > > bridged mode. The Just FYI. The problem with most diagnostics is > that they > > > > test a normal case and not a worst case. Some solace is that the e100 > > > > driver is worse. > > > > > >FWIW Tim Hockin seems to have eepro100 docs -and- some fixes to eepro100 > > >which are more stable at heavy load than eepro100.c [2.4 or becker] and > > >e100.c. > > > > If he has the eepro100 docs....hes not allowed to release any code, unless > > intel has changed their NDA policy. > >It's a complicated legal situation that he can explain better. He >definitely has docs, and through the magic of the NDA his company signed >with Intel, apparently he can indeed release his work. Sun does have negotiating powers now known by many of us, I suppose. :-) Its too bad that intel doesnt have an interest in making their linux driver work better under load...its got alot of initialization and nway stuff in it that would be pretty useful...obviously Intel has the docs,and they cant make it work, so having the docs isnt a sufficient condition for making it work. The FreeBSD driver is virtually bulletproof, so it CAN be done, Dennis - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
