On Thu, 5 Mar 1998, INOUE Katsumi wrote:

> Thanks SamMeEl ,Hamada-san,Bob,Yanagisawa-san for replying.
> But I still can't make it work.
> 
> My environment:
> PowerBook 550C, Japanese Keyboard, No internal modem, 12MB memory
> Mac OS 7.6 English, IBM 750MB SCSI HDD, almost dead internal backup battery.
> 
> Kernel: HWDIRECT-SBC#61
> 
> Boot still stops at 'zstty0' line most of the times even if I tick the 
> Booter option 'Serial Port->Open before...' on. But sometimes it just boots
> fine.
> 
> I tried to boot more than 50 times and I'm getting convinced that this
> is a Booter problem in that it fails to open serial port. Is it possible to 
> verify this?  I can step down to debugger when boot stops,

It's not the booter failing to open the serial port.

> zstty0 at zsc0 at channel 0
> Panic switch: PC is 0x117de0.
> Stopped at     _Debugger+....
> db> trace
> ..
> _lev7intr(?)
> _zs_iflush(...)
> _zs_loadchannelregs(...)
> ..

Thanks for the stack trace!!!!!

> 
> Boot always stops here when it stops.

But it sometimes makes it past this point??????

zs_iflush is a routine which looks to see if there are any bytes in the
input fifo, and deletes them if so. The problem is that it's a for loop w/
no bounds checking. If you really have an SCC, it will exit after 3 loops.
I bet, though, that you don't have an SCC at the "standard" modem port
address, so it's still looping.

There are two things you can do in addition to ditching the serial driver.

1) Change the #if 1 on line 135 of src/sys/dev/ic/z8530sc.c to "#if 0"

2) See if you can get the kernel to hang at zstty1. This might involve
re-compiling the booter (I don't know which serial driver it opens on
close, but re-compiling to open the other one might work).

We've never understood how Apple shuffled the serial port(s) around. MacOS
pretends there is both a printer and a modem port, but there's only one
SCC....

Do you have a volt meter? Can you check to see if there are voltages being
output from the serial port? That way we can (I hope) figure out if it's
there's no chip, or that MacOS still managed to turn off the power.

Take care,

Bill