Subject: sonic code samples + regular old ethernet
To: None <macbsd-development@NetBSD.ORG>
From: Brad Parker <brad@stemwinder.fcr.com>
List: macbsd-development
Date: 11/21/1994 20:24:20
[don't you hate it when people pop up out of the blue and dump a bunch
of wierd things to the list?  I sure do. ;-)]

1. I put the oft-promised-yet-never-delivered sonic ethernet sample code up
for ftp.

	ftp://ftp.fcr.com/pub/macbsd/sonic.tar

I did the tar under ms-dos.  Sorry, it's where the files where.
Probably means the files are all ending with ^Z's and have \r instead
of \n.  Hi ho.  Good excuse to remember the args to 'tr' ;-)

Also,

2. I've been building kernels from the sun-lamp sources (silly me).  I
got them to work on my 8mb IIcx.  I found I had to run Mode32 with the
new kernels I built.  Without mode32 the original 1.0-beta kernel ran
but then new ones I built would hang forking the initial shell...
someone probably knows what this means. (I suspect the kernel is
looking at the existing mmu setup.  sure. ok. whatever.)

I made some hacks to via.c and if_ae.c and now my Interlan ethernet
card works great.  It really works.  The first pass would wack adb,
however.  A tad more hacking and adb now works most of the time
(basically poll the level 1 via in the level 2 isr).  I think I have a
final wack (shut off level 2 while servicing the level 1 adb intr) and
in the next day or so I should be able to get adb to keep working
forever (oh sure).

anyway, I'll take home one of the these apple-clone cards I have and
make it work also so those with apple cards can also use this.  I seem
to recall someone had rewritten if_ae.c and had some luck.  Then no
word.  what's up with this?  (and, didn't you find the via2 nu-bus
interrupt code had some interlocking problems?  I did. also, the
if_ae.c I have was turning on interrupts before setting up the isr. bad mojo.)

3. I had to back out the patch to dev/grf.c.  When I put it in my dt
would start and then abort.  Something about the screenptr being zero.
What's up with this?  is my dt out of rev?

4. Who can explain to me why interrupts need to be off during
scrolling of the screen?  please do.  I like dt but have noticed that
the debugger does not use dt and never hang adb...

is it because interrupt routines want to printf to the console?  ok.
How much to they want to printf?  can't we use a semaphore and buffer
one line of output during the scroll?  Or maybe buffer into a ring?

I'll put my via.c and if_ae.c changes up for ftp in a day or so.
scream if that's not good enough.

-brad