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

RE: hamm - network access problems



Hi,

>  I got my network connection working.  The new kernel ( vmlinux-2.1.105-980708)
>appears to work in the sense that I can "see the network" and ping machines. 
>However, whenever I try to telnet/ftp.  (The following was generated when I 
>tried to telnet to the machine from another machine)
>
>Data write fault at 0xfc0d0054 in Super Data (pc=0xf37ca)
>BAD KERNEL BUSERR
>Oops: 00000000
>PC: xxxx
>SRE xxxx SP:xxxxx a2:xxx
>d0: xxxx d1: xxxx d2: xxxx d3: xxxx
>d4: xxxx d5: xxxx a0: xxxx a1: xxxx
[...]
>Process swaooer (pid: 0, stackpage=xxxx)
>Aiee, killing interrupt handler
>Kernel panic: Attempted to kill the idle task!
>In swapper task - not syncing

:-( I apparently replied to the wrong ML on that one. Anyway: Alan Cox is the 
one to talk to about the network stuff on Mac, he wrote the Mac 8390 driver. 

You left out the interesting information: the xxxx stuff above. And the base
address of the 8390 chip, that's hopefully printed in the boot log, if not,
we'll have to make the kernel print it. Also required: the DrHw and DrSw
parameters printed there.

Please post this panic log to linux-mac68k@wave.lm.com, or to Alan directly
(alan@cymru.net is one of his addresses).

My analysis, based on the information you give: during interrupt processing,
the driver writes to a register location it thinks is legal for this type of
ethernet hardware, but the register isn't present on your card. What register
that is? I can't tell without the DrHW/DrSW stuff, as the Mac 8390 driver uses
different register mappings depending on the hardware type.

>I assume something similiar happens when trying to ftp/telnet from the machine
>(Mac IIci) however, it flashes the same message repeatedly, and I cannot read
>it as it keeps scrolling by..

It would stop scrolling (at least with my kernels from testing) if you manage
to generate an NMI. And it would resume scrolling on the next NMI. All Macs I
have tested Linux on do have a reset switch and a NMI (aka 'programmer')
switch. A number of Macs implement the NMI over the keyboard controller; this 
will only work if the ADB controller itself generates the NMI. I never tried
that...

	Michael


--  
To UNSUBSCRIBE, email to debian-68k-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org




Reply to: