Conference hydra::amiga_v1

    I've seen the MaxiPlan Plus problem you describe, but only on certain
    grids; specifically, the tax templates that I uploaded here.
    Grabbing the window size gadget and making the window small, then
    large, seems to clear the bad cells.
    
    I am using version 1.8e of MaxiPlan, but I hear that a new version
    is available with some bug fixes.  Call OXXI for upgrade details.
    
    I've also heard on PLINK that some A2000s don't handle a guru properly.
    Do you have the rev 4.3 motherboard?  
    
    
    

    re:.1
    
    The A2000 guru problem is bad code in the rom, there is a program,
    called SETALERT, you can run as part of the system startup that will
    fix that.  Next rev of the roms should have that problem corrected.
    My rev 4.2 motherboard has the problem rom, not too nasty once you know
    what is happening- the flashing power led means guru, then just
    C-A-A to reboot.
    
    -dave

Re: all

Yes, all Amiga 2000s have the guru problem.  The problem is that the
Guru code was written defensively!

The Guru code usually is called when something bad has happened to the
system; so it does certain consistency checks to determine if it will
be able to proceed.  One of the things that this code looks at is
the pointers in execbase and such.  Well, that "half-fast" ram in
the Amiga 2000 and (upgraded) 500 sits at an address that was illegal
in the original Amiga memory map.  The Guru code looks at execbase and
decides that nothing can live at address C00000 (the start of half
fast RAM), and so punts.

    Re: .3
    
    I've had a lot of fun with the Guru on my A500.  Are you
    implying that the bad ROM problem in the A2000 is part of the A500
    ROM?  ie. should I get SETALERT too?
    
    Tim (New-User-trying-to-escape-the-guru) Hicks 8^)
    

Re: .4

I believe that a 1 meg Amiga 500 will see the problem.  A 512k Amiga 500
will not.

Which do you have?

Note that the bug does not prevent the "Task held" requester from popping
up for errant tasks.

    To ALL of you thanks for your answers.  Now I can sleep easier at
    Nights.  I just did not  want to have to take it back to the dealer
    deal.  
    
    Thanks again
    
    ...gss...

Ok, So where's SETALERT?  :-)

I've been getting this *wierd* GURU who doesn't even see fit to pop up
the "task held" requester...

John 

As I understand it, SETALERT is an no longer needed if you use the PATCH_1
program that Bryce Nesbitt submitted to usenet.  (I received that Monday).
This program supposedly corrects a bug in Kickstart 1.2 even for A500/2000
owners.  I killed my copy, but it should be at YGDRSL::AMIGA_USENET if
noone else has it.

Question for those of you who know more about the Amiga operating system
than I:  Is this 'bug' real?  Does Bryce's patch seem legitimate??

						...richard

    Re: .5
    
    I have an A500 with 1 meg....8^) 

This note may be of some bearing on the A500/A2000 erroneous GURU handling.

Ed.



          <<< DUSTER::SYS$PUBLIC:[NOTES$LIBRARY]AMIGA_USENET.NOTE;1 >>>
                    -< comp.sys.amiga, .binaries, .sources >-
================================================================================
Note 1395.0  ** Read this message!! ** Bugs in Kickstart you should k  2 replies
DUSTER::AMIGO "eris.berkeley.edu!bryce"              75 lines  28-FEB-1988 02:32
--------------------------------------------------------------------------------

Newsgroups: comp.sys.amiga
Path: decwrl!labrea!agate!eris.berkeley.edu!bryce
Subject: ** Read this message!! ** Bugs in Kickstart you should know about!!
Posted: 27 Feb 88 09:32:55 GMT
Organization: University of California, Berkeley
 
 
fix_1.arc    Patch file #1 to fix bugs in V1.2 Kickstart
	     Written by Bryce Nesbitt.
 
Kickstart V1.2 has some bugs that can randomly Guru the machine.  Running
this file will protect you from these problems.  You must reboot to get rid
of the protection.
 
This program runs under Version 1.2 Kickstart ONLY.  If Version 1.3
Kickstart needs these fixes, then a separate patch file will be provided.
The bugs that are fixed:
 
1>  On machines with $C00000 memory (like the A500 and A2000), the Alert()
    function misbehaves.  This makes any alert fatal, even the
    "recoverable" ones.
2>  Every time a window is closed (or, in fact, any layer is deleted) the
    system frees a bit of memory then uses it again.  This can cause random
    crashes.
 
Please redistribute this patch file.  Use this naming convention:
 
fix_1.arc    Patch file #1 to fix bugs in V1.2 Kickstart
 
(Use "uudecode" from Fish disk #53 to decode this file)
 
begin 644 patch_1
M```#\P`````````!``````````````%&```#Z0```48L>0````23R4ZN_MHJT
M0'``(BT`K&8>0>T`7$ZN_H!![0!<3J[^C$ZN_WPB0DZN_H9P_TYU0_H!>$ZN"
M_F@CP````:HL0$ZN_\0J`&8$<!1.=2!Y````!`QH`"$`%&<20?H!>R(%)`AVY
M*TZN_]!P"DYU+'D````$(#P```!Z<@!.KO\Z*`!F!'`43G4B1$'Z`)(@/```!
M`'I.KOV0+'H!#D'Z`68B!20()CP```,03J[_T"QY````!$/Z`0).KOYH(\``;
M``&B0_H!`TZN_F@CP````:8B>0```:(@1"`(,'S_IDZN_EPCP````9HB3B!$3
MT?P````X(`@P?/^43J[^7"/````!GB)Z`)Q.KOYB(GH`F$ZN_F(B>@"43J[^2
M8GX`3G4O"B\+)$DO*@!$)FH`1$ZY`/X0@$J`6(]G%B\*3KD`_BR&+PM.N0#^F
M$*1P`5"/8`)P`"9?)%].=3/\0```W_":(#Q(14Q0L+D`````9R`CP`````!!2
M^0```0`@QR#5(#D````$(@`"@?\_^8EG!D[Y`/PP7D[Y`/PP!@``````````/
M````````````````9&]S+FQI8G)A<GD`;&%Y97)S+FQI8G)A<GD`:6YT=6ETY
M:6]N+FQI8G)A<GD`"E-O<G)Y+"!T:&ES($].3%D@=V]R:W,@9F]R(%8Q+C(@<
M2VEC:W-T87)T"F9I>%\Q+F%R8R`@("!0871C:"!F:6QE(",Q('1O(&9I>"!B%
M=6=S(&EN(%8Q+C(@2VEC:W-T87)T"E=R:71T96X@8GD@0G)Y8V4@3F5S8FETI
M="X@(%!L96%S92!R961I<W1R:6)U=&4@=&AI<R!F:6QE(0H*2VEC:W-T87)TO
M(%8Q+C(@:&%S('-O;64@8G5G<R!T:&%T(&-A;B!R86YD;VUL>2!'=7)U('1HP
M90IM86-H:6YE+B`@4G5N;FEN9R!T:&ES('!R;V=R86T@:&%S('!R;W1E8W1EE
M9"!Y;W4@9G)O;2!T:&5S90IP<F]B;&5M<RX@(%EO=2!M=7-T(')E8F]O="!TO
M;R!G970@<FED(&]F('1H92!P<F]T96-T:6]N+@H*5&AI<R!P<F]G<F%M(')UD
M;G,@=6YD97(@5F5R<VEO;B`Q+C(@2VEC:W-T87)T($].3%DN("!)9@I697)S/
M:6]N(#$N,R!+:6-K<W1A<G0@;F5E9',@=&AE<V4@9FEX97,L('1H96X@82!S^
M97!A<F%T90IP871C:"!F:6QE('=I;&P@8F4@<')O=FED960N("!4:&4@8G5GE
M<R!T:&%T(&%R92!F:7AE9#H*"C$^("!/;B!M86-H:6YE<R!W:71H("1#,#`PA
M,#`@;65M;W)Y("AL:6ME('1H92!!-3`P(&%N9`H@("`@03(P,#`I+"!T:&4@3
M06QE<G0H*2!F=6YC=&EO;B!M:7-B96AA=F5S+B`@5&AI<R!M86ME<R!A;GD*_
M("`@(&%L97)T(&9A=&%L+"!E=F5N('1H92`B<F5C;W9E<F%B;&4B(&]N97,N,
M"C(^("!%=F5R>2!T:6UE(&$@=VEN9&]W(&ES(&-L;W-E9"`H;W(L(&EN(&9A6
M8W0L(&%N>2!L87EE<B!I<PH@("`@9&5L971E9"D@=&AE('-Y<W1E;2!F<F5E6
M<R!A(&)I="!O9B!M96UO<GD@=&AE;B!U<V5S(&ET"B`@("!A9V%I;BX@(%1HA
M:7,@8V%N(&-A=7-E(')A;F1O;2!C<F%S:&5S+@H``````^P````&````````V
>`#X```#`````S@```-0```#F```!``````````/R\
``
end
size 1380
 
|\_/|  . ACK!, NAK!, EOT!, SOH!
{O_o} .     Bryce Nesbitt
 (")        BIX: mleeds (temporarily)
  U	    USENET: bryce@eris.berkeley.EDU -or- ucbvax!eris!bryce
 
========================================================================
Received: by decwrl.dec.com (5.54.4/4.7.34)
	id AA22335; Sat, 27 Feb 88 05:13:47 PST
================================================================================
Note 1395.1  ** Read this message!! ** Bugs in Kickstart you should know  1 of 2
DUSTER::AMIGO "bbn.com!cosell"                       36 lines  28-FEB-1988 02:57
--------------------------------------------------------------------------------

Newsgroups: comp.sys.amiga
Path: decwrl!decvax!ucbvax!pasteur!ames!hao!gatech!bloom-beacon!mit-eddie!bbn!bbn.com!cosell
Subject: Re: ** Read this message!! ** Bugs in Kickstart you should know about!!
Posted: 27 Feb 88 17:38:23 GMT
Organization: Bolt Beranek and Newman Inc., Cambridge MA
 
In article <7231@agate.BERKELEY.EDU> bryce@eris.berkeley.edu (Bryce Nesbitt) writes:
>
>fix_1.arc    Patch file #1 to fix bugs in V1.2 Kickstart
>	     Written by Bryce Nesbitt.
>
>Kickstart V1.2 has some bugs that can randomly Guru the machine.  Running
>this file will protect you from these problems.  You must reboot to get rid
>of the protection.
 
Huh? Did you mean this last sentence... why would you *want* to get rid of
the protection?
 
>This program runs under Version 1.2 Kickstart ONLY.  If Version 1.3
>Kickstart needs these fixes, then a separate patch file will be provided.
...
>1>  On machines with $C00000 memory (like the A500 and A2000), the Alert()
...
 
I notice that you mention the A2000 explictily.  I guess this means
that this isn't actually a patch for kickstart (since I think we'ed
need new ROMs to really do that, no?) .  When I get it onto my Amiga
what _will_ it be?  Perhaps I'm (as usual) missing the obvious, but I'm
not sure what I'm supposed to do with this guy -- put it in my
startup-sequence?  If so, does its placement have any bearing on other
things in s-s (e.g., Andy's previous alert() patch).?  Should it be
"run patch_1" out of s-s?  Sorry to be so dense...
   __
  /  )                              Bernie Cosell
 /--<  _  __  __   o _              BBN Labs, Cambridge, MA 02238
/___/_(<_/ (_/) )_(_(<_             cosell@bbn.com
================================================================================
Note 1395.2  ** Read this message!! ** Bugs in Kickstart you should know  2 of 2
DUSTER::AMIGO "eris!bryce"                           43 lines  28-FEB-1988 20:06
--------------------------------------------------------------------------------

Newsgroups: comp.sys.amiga
Path: decwrl!labrea!agate!eris!bryce
Subject: Re: ** Read this message!! ** Bugs in Kickstart you should know about!!
Posted: 27 Feb 88 21:59:00 GMT
Organization: University of California, Berkeley
 
In article <21290@bbn.COM> cosell@BBN.COM (Bernie Cosell) writes:
>In article <> bryce@eris.berkeley.edu (Bryce Nesbitt) writes:
|>
|>You must reboot to get rid of the protection.
|
|Huh? Did you mean this last sentence... why would you *want* to get rid of
|the protection?
 
You would not.  But let's say I implemented this like the SCA virus... then
you would need to do more than just reboot :-).  Wow!  What a neat way to
do automatic software updates :-).
 
 
|I notice that you mention the A2000 explictily.  I guess this means
|that this isn't actually a patch for kickstart (since I think we'ed
|need new ROMs to really do that, no?) .
 
Who needs to change ROMs to patch Kickstart?  The OS has this magic 
"SetFunction()" call all set up for the purpose.
 
 
>Perhaps I'm (as usual) missing the obvious, but I'm
>not sure what I'm supposed to do with this guy -- put it in my
>startup-sequence?
 
Yes.  Just put "patch_1" in the startup-sequence.  No need to run it.
 
 
|If so, does its placement have any bearing on other
|things in s-s (e.g., Andy's previous alert() patch).?
 
This fixes the Alert() bug, so Andy's "setalert" patch is not needed.
 
|\_/|  . ACK!, NAK!, EOT!, SOH!
{O_o} .     Bryce Nesbitt
 (")        BIX: mleeds (temporarily)
  U	    USENET: bryce@eris.berkeley.EDU -or- ucbvax!eris!bryce

    Re: .10
    
    OK...Now where can I get my hands on fix_1.arc?  I can't do downloads
    from the net, so if you could, please point me toward the person
    that might be kind enough to copy it on a diskette for me.
    
    Tim 8^)

Well, I'm not sure, but I think I just UUDECODEd AMIGA_PATCH.UUE.

I ran the VMS UUDECODE.EXE on it. It happened so fast I figured it didn't
work. But sure enough, it left a file called PATCH_1.


I'll try it out tonight and reply here. For anyone who's curious, it's
at WJG::AMIGA:PATCH_1


John

    Please excuse my ignorance but something is troubling me with this
    note!
    
    Where is Commodore??????   If this is a known bug (and it sure sounds
    like it) the patch should rightly come from them, no??  I am a new
    owner (3 months) of a 2000 and EXPECT known bugs to be fixed by
    the BUG OWNER.   
    
    It's great that the user community discovers and fixes things like
    this but as a novice to this machine I'm not totally comfortable
    with a patch that could have side effects.
    
    Is this support (LACK OF) typical of Commodore?  Is there an official
    supported patch?  Should I expect one?  

    Commodore generally does not release 'patches', and in fact seems
    to discourage this sort of thing.  They release major upgrades,
    such as from 1.1 to 1.2.
    
    Commodore engineers are very active on BIX and Usenet, as well as
    Plink, although their presence is 'official' only on BIX.
    
    Ed.

re .11 - Where is it?

Ed posted it in .10! :-)

Hey all, I was having some real serious problems last night with my 2000 hanging
on me.  It would just freeze - no mouse or keyboard input would do anything.
This happened four or five times.  I'm running BlitzFonts, MachClk v16, and
GOMF in the background and I had one CLI running shell.  Plenty of free RAM.
The only thing I had done differently is put PATCH_1 in my startup-sequence.

Well, I removed it and all seems well again.  Needless to say, this is not by
any means conclusive, but I'm curious.  I plan on testing this out some more.
The times that I would hang were always when I was manipulating an icon.

						Any ideas??

						...richard

    Re: .15
    
    Really!  Am I supposed to type that stuff into a file or something?
    Believe me, I'm not compu-literate enough to be sarcastic about your
    answer.  Are you serious?
    
    My A500 gets hung in one particular application (Disk Wizard). 
    Its usually the 1st thing I load after I boot.  I'll have an assortment
    of other things running, and I make a (D/W) request that goes to disk,
    and while its fetching that stuff I might casually move the pointer
    and click on a gadget, and >>>ZZZAAAPPPP<<<, frozen solid.  No pointer,
    no keyboard, nothin'.  Ctl-C,-D,-E,-F have no effect.  I just don't
    do that sort of thing after the third time it happened (takes me
    a while to catch-on).  Maybe its my fault.  But shouldn't I get
    a guru or something?
    
    ...Tim |^( 

Re: .13, .14

Actually, Commodore DID release a patch for this bug last October.  I'd
be willing to bet that half of patch_1 came from its author disassembling
the official Commodore patch.

Commodore seems pretty good at fixing problems.  Within a week or so of
the Guru bug being reported, a Commodore patch was released.  Commodore
was also very prompt in releasing the virus checking programs.

The second bug that patch_1 claims to fix I have never heard anyone
mention before, and in my year+ of running 1.2 (9 months on a 1000,
4 months on a 2000), I have never seen.  Personally, I see no reason
to run patch_1 instead of the Commodore patch.  (And I don't bother
running the Commodore patch since in my opinion getting a delayed
guru is a pretty small problem.)

I does seem that Commodore prefers fixing bugs by periodically updating
the operating system.  As long as bugs are fairly minor, this seems
reasonable to me.

    the Commodore fix is called Set_Alert and was written by Andy Finkel
    if I remember correctly.  I've been running it for a while on my
    A2000 with no problems.  Hopefully, Patch_1 is the Commodore fix
    plus a couple more, I haven't tried it yet.  I don't normally run
    a shell, gomf, or even popcli.
    
    -dave