NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/sys/sun/admin/9806

Internet Message Format | 1992-12-30 | 5.3 KB

Xref: sparky comp.sys.sun.admin:9806 comp.sys.sun.hardware:6552
Newsgroups: comp.sys.sun.admin,comp.sys.sun.hardware
Path: sparky!uunet!stanford.edu!ames!elroy.jpl.nasa.gov!earle
From: [e-mail] (Greg Earle - Gainfully Unemployed)
Subject: POINTLESS SPECULATION re: Sunergy (SPARCclassic and LX) and 4.1.3
Message-ID: <[e-mail]>
Followup-To: comp.sys.sun.admin,comp.sys.sun.hardware
Summary: Enough!
Sender: isolar![e-mail]
Organization: Jet Propulsion Laboratory, Multi-Mission Image Processing Lab
References: <[e-mail]> <ellis.725324767@nova>
Date: Wed, 30 Dec 1992 09:01:29 GMT
Lines: 79

In article <ellis.725324767@nova> [e-mail] (Stew Ellis) writes:
>>------>   HAS ANYONE TRIED TO INSTALL 4.1.x ON A SPARCCLASSIC?   <------
>
>>I know 'Sun says' it won't work but Sun employees have been overheard
>>talking about what will hit the fan if it is even suggested that
>>4.1.x can be installed on a 'classic'.  I don't know about you
>>but this makes me curious.  It would sure be nice to get a classic _now_
>
>Should we be reading between the lines here that you are suggesting that Sun
>employees might know that you can boot a Classic with 4.1.2 or 4.1.3?
>
>If there is anyone out there who has a Classic, you could do everyone a big
>favor by trying to boot with 4.1.2 and with 4.1.3 and give the rest of us a
>report?

AUUUGGH!  Enough idle speculation, already!

Let's get this over with.  Maybe this'll save Guy Harris from having to make
even more canonical followup corrections to misguided speculative postings ...

(1) There was an internal Sun communication from either Scott McNealy or Joe
    Roebuck (depending on which rumor mill you heard it from) that supposedly
    emphasized the fact that the new machines were THE transition point from
    SunOS 4.1.x to Solaris 2.1, and therefore doing anything which aided and
    abetted 4.1.3 running on either constituted a fireable offense.  Capiche?

(2) It has been noted that attempting to install 4.1.3 as a sun4m architecture
    machine fails.  On the other hand, if you have installed 4.1.3 on another
    sun4m machine, such as a uniprocessor SPARCstation-10, and then transferred
    the disk drive to a SPARCClassic or SPARCstation LX, it may well boot and
    it may well even "run" (note quotes).  (To be honest, I have evidence that
    "it may well" is really more like "it does", but I'm being conservative.)

(3) Other people have aluded to this; most recently Guy Harris did.  The new
    machines have *new hardware in them*, and not just the MicroSPARC CPU.
    As Guy said, past SunOS releases have had code that was *specific to
    certain Sun platforms* in order to work around certain things that may
    have been peculiar to a particular SPARC processor or CPU board.  It is
    quite likely that the same will be true for the MicroSPARC.  But more
    importantly, Guy mentioned that there are new ASICs inside the new machines
    which are meant to be Vegematics - they slice, they dice, they do serial
    port I/O (-: as well as some other features.  I have it on highly reliable
    authority that the emulation of the devices they are intended to replace
    (like, e.g., the Zilog 85C30 serial chip) is not 100%!!  There is code in
    the Solaris 2.1 serial driver (and probably the other drivers for the other
    functions the ASICs perform; Guy aluded to the Lance Ethernet) that takes
    care of these anomalies; there is NO code in the 4.1.3 drivers that do so,
    since these chips didn't exist at the time!  Can you say, "Driver broken"?
    Sure, I knew you could ...

(4) Even if you get it running and you don't use or care about the serial ports
    or the Lance Ethernet (maybe you run Thinnet), what are you going to do if
    it starts crashing daily, with weird panics?  Call into Sun?  "Yeah, uh,
    I'm having crashing problems under 4.1.3, oh, er, uh, ah, I'm using a
    SPARCstation-10/30, yeah, that's the ticket!"  And when the Solution Center
    engineer tells you that they can't reproduce your problem on their own
    SPARCstation-10's, what then?  Beg from the net?  And hope the other 10
    people in the world also doing it can help you out?

The bottom line?  Well, I'll tell you, from where I sit, I wish that a 4.1.3
(or 4.1.4) existed to support these machines, especially the LX (I wish I could
do ISDN-over-PPP development work under 4.1.3, alas).  But an LX machine that
has serial ports which don't work because the serial drivers don't work is a
useless machine as far as I'm concerned.  Maybe you don't care about serial
ports; fine (my WorldBlazer and printer certainly do).  But maybe the Lance
Ethernet doesn't "quite" work, either (I dunno).  Do you really want to risk
it?!?  I don't think *I* want to ...

Get over it, already.  It's a dead issue.  If you really, really want 4.1.3 on
a new machine, buy source and port it yourself.  (Or wait for 4.4BSD and hope
that Chris Torek can reverse-engineer those machines at some later date; don't
hold your breath, though, since Sun hasn't given him any help with the port
he's done already for the older architectures.)

--
    - Greg Earle                | Disclaimer: These are my
      Itinerant Sun Consultant        | opinions ONLY, and do not
      isolar![e-mail]    | reflect the opinions of the
      [phone]            | organization I'm posting via!