[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: New kernel's for Mac..



Hi,
 
> New kernel's from both 2.0.33pl1 and 2.1.120 have been posted to the
> site I've been watching...here's a short rundown on what I've found
> with both..
> 
> 2.0.33pl1 (build date 11/20/98)

That's mine (the 2.1.120 from 981014 definitely isn't mine, comments on 
that one are best directed to linux-mac68k@baltimore.wwaves.com).
Let's see... 

>   This kernel still does not function on the network.  Ping works, 
> but nothing else does (no telnet, no ftp, etc), BUT, iptraf does work
> in monitoring traffic on the ethernet when set to promiscuous mode.
> I've tried it on three different Mac's; Quadra 650, Quadra 700, and
> Quadra 840.  This is using the Asante NuBus etherent card.

Guessing (I can't imagine you running a full filesystem including telnet
etc. from a ramdisk): SCSI appears to be working fine on all these Quadras?
 
>   It still does not properly find the hardware address for the builtin
> controller (the sn device).  Sees it as Sonic, but reads the hardware
> address as an ILLEGAL MAC address, which is read as a different address
> on each boot.

Truely amazing. I've looked at the source this kernel was compiled from,
and there's no trace of any Sonic ethernet driver. Maybe it's not a 2.0.33
kernel after all? Could you please check the boot log messages for the
version - or for the Sonic output?
 
>   Telnet sets up the connection, but nothing else.  Ie, it says trying..
> Connected to...and the Escape char lines all appear, but thats it.  It
> hangs there forever.  The same behavior results when doing an ftp.
> 
>   The interesting thing about this is that iptraf DOES display network
> activity.

Can you use iptraf on one machine to log what happens when you initiate a
telnet connection on another? The point being that once you got the 
'connected' message, the initial SYN was successfully sent, and the 
ACK was received in return. Seems that sending and receiving work OK.
 
>   This behavior is essentially the same as the kernel build date 9/26.

Nothing changed in the network device code, no surprise there. I've not a
lot of leads to go on, and no crazy ethernet card to test. Someone with such 
a card needs to start hacking kernels.

	Michael


Reply to: