[13055] in Athena Bugs
rsaix [7.7F]: AIX
daemon@ATHENA.MIT.EDU (yandros@MIT.EDU)
Thu Dec 29 05:11:31 1994
From: yandros@MIT.EDU
Date: Thu, 29 Dec 1994 05:11:24 +0500
To: bugs@MIT.EDU
System name: pickled-herring
Type and version: RS6k/220 7.7L
Display type: POWER_Gt1
What were you trying to do?
Paste a few lines of text into an xterm, on pickled-herring, an
RS/6000 model 220 in the SIPB office:
; machtype -a
rsaix
7.7
7.7G
7.7F
POWER
POWER_Gt1
fd0 unk 3.5inch4Mb
24576 K
What went wrong?
My xterm went away. I thought this a bit odd, since I hadn't heard
of that bug with the RS/6000. I then thought ``Since they haven't
fixed the bug in zephyr that I reported months ago about zwgc dieing
if it's parent dies, my zwgc should die any second now, too.''.
Then my zwgc died... ..along with my windowmanager, my emacs, my
xscreensaver, and my fclock (a personal invocation of dash). This
was quite odd, and left me with only a single xterm and xload. I
wondered if the file I had been editting was saved or not, so I ran
`vi .resources' in my lone remaining xterm. The xterm promptly went
away.
What should have happened?
None of that.
Anything else you'd like to add?
You better believe it. ;-)
Yeah, I think that the machine was trying to get me because I
figured out IBM's big secret: AIX is a sham. The POWER chip is a
sham. Remember those stories you've heard, about IBM creating
`WorkPlace OS', a merger of MS-Windows, OS/2, and AIX? That was all
a marketing gimmick, to make people think that they were starting
with UNIX. In actuallity, IBM is shipping 386sx25's in special
cases, running a very hacked version of OS/2 or MS-Windows. This is
the only way I can think of to explain the TRULY AWFUL
responsiveness of this hardware (which is an improvement over the
320! That bug report is next...)
How did I come to this conclusion, you might ask? Well, there were
two main clues:
o First of all, after trying to do a little work on the machine, I
quickly became aware of the fact that the OS was *NOT*
multitasking! I started a build in one window. I lowered it.
about 8 seconds later the window that had been under it finished
repainting. I typed `M-x man RETURN tgetent RETURN', all without
seeing what I was typing, because THE MACHINE COULDN'T KEEP UP!
Then I moved the mouse, intening to go into the window next to it.
The mouse never made it - it froze a few centimeters away, and for
the next three minutes or so the only response I got out of the
machine was the rising numbers on the front LEDs. This is NOT a
multitasking system.
o Second, and much simpler, is the failure mode of the machines.
This was the final hint that convinced me that the machines really
are running OS/2 or MS-Windows is the way they die: the screen
goes entirely black for at least several seconds. About thirty
seconds. About the same time as you'd expect for the timeout on a
dialog box. Yes, that's right; I see you've guessed it to.
During that thirty+ seconds of blackness, the machine is
displaying dialog box, in black on black with a black border and
cursor and a black background, because that's the only way they
could figoure out to make it not show up! So now you know; the
next time you see an unhappy ``Risc Systems 6000'' machine with a
``Performance Enhanced With Optimized Risc'' chip, you'll know
what it's *REALLY* saying. You'll know, that could you just
figure out the right settings to change those colors, that machine
would be happily trying to tell you:
+-------------------------------------------------------------------+
| AIX |
+-------------------------------------------------------------------+
| ____ |
| / \ This application has violated system integrity due to |
| | STOP | an invalid General Protection Fault and will be |
| \____/ terminated. |
| |
| Quit all applications, quit Windows, and restart your computer. |
| |
| +------+ |
| | OK | |
| +------+ |
| |
+-------------------------------------------------------------------+
