Subject: Re: July first Kernels
To: Steve Revilak <root@parkdrive.home.net>
From: Colin Wood <cwood@ichips.intel.com>
List: port-mac68k
Date: 07/02/1999 22:28:45
Steve Revilak wrote:
> In light of the prior post (option-1,2,3), I tried out two of the
> kernels that were recently made available at 
> 
> 	ftp://ftp.netbsd.org/pub/NetBSD/arch/mac68k/binary/
> 
> (The date on the files is July 1).
> 
> netbsd.MRG-6
> 	Booted okay to the console in multi-user mode.  Interesting
> 	thing happened -- shortly after logging in, I edited a
> 	configuration file.  Wanting to HUP the daemon (syslogd), I
> 	used "ps -aux" to find it's process ID number.
> 
> 	The first line out ps' output:
> 
> 	ps: proc size mismatch: 8088  should be 666 

read the FAQ...
 
> 	At this time, the system had been up for less than 5 minutes.
> 	I went along and did a few things from the console (read last
> 	night's e-mail, etc), and checked back frequently.  The
> 	numbers changed slightly, but the first stayed in the range of
> 	8000 < x < 9000 and the second in the range of 600 < x <700.
> 
> 	Later, starting X lands me in the debugger....

this is annoying, but without a backtrace there is really no way of
knowing what went wrong.  type 'trace' at the debugger prompt so we can
see where things went wrong.

> 
> netbsd.GENERIC-6
> 
> 	Very interesting thing...somehow or other, stdin under this
> 	kernel eventually caught itself in a regressive loop.  For
> 	example, I'd type "ls -l" -- it would appear on the screen as 
> 
> 	lllllllllllllllllllsssssssssssssssssssssss           ------
> ---------------lllllllllllllllllllllll
> 
> 	followed by several dozen carriage returns.  Just as if every
> 	key on the keyboard had gotten stuck for a brief period after
> 	being pressed.
> 
> 	After this began happening, I killed the X server.  The
> 	phenomenon continued from the console.
>
> 	I logged out.... the problem persisted when the new login
> 	prompt came up.

this is actually fairly annoying.  there is now an open PR on it.  i'm not
quite sure where the problem is, tho (i.e. is it in the low-level ADB code
or the higher-level keyboard driver?)  if i can reproduce it, maybe i'll
see if i can fix it.  hopefully someone else will find it before me, tho.

later.

colin