Number: 1790

Date: 10-Aug-84 14':42':25

Submitter: Sannella.PA

Source: Charnley.pa

Subject: New Dlion ucode in Harmony Release

Assigned To: 

Attn: 

Status: Fixed

In/By: 

Problem Type: 

Impact: 

Difficulty: 

Frequency: 

Priority: 

System: Language Support

Subsystem: Microcode

Machine: 

Disk: 

Lisp Version: 

Source Files: 

Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 

Disposition: '
["Sannella.PA" "21-Aug-84 10':07':11" Description':]'
["Sannella.PA" "28-Aug-84 09':37':32" Description':]

Description: '
Date': 11 Jul 84 09':12':03 PDT (Wednesday)'
From': Charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': , Charnley'
'
  A new set of microcode for the dandelion has been spliced into {Phylum}<Lispcore>next>Lisp.sysout and {Phylum}<Lispcore>next>Full.sysout.'
 It has not been placed on {eris}<lispcore> yet because it''s full.'
 '
  These files contains both 4K and 12K microcode.'
  This fixes the Dandelion freeze problem.'
  '
  Problems to me.'
  '
  don'
'
-----'
'
Date': 11 Jul 84 11':45':38 PDT (Wednesday)'
From': Charnley.pa'
Subject': Re': New DLion uCode'
To': jellinek.pa'
cc': lispcore^, Charnley'
'
'
Of course, I meant Eris, not Phylum.  I have now stored the microcode to {ERIS}<lispcore>next>dLispDomino.db.'
'
The new 12K microcode has the three speech byte codes in it.'
'
     don'
'
-----'
'
Date': 18 Jul 84 15':52':41 PDT (Wednesday)'
From': Charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': Charnley'
'
  A new set of microcode for the dandelion has been spliced into {ERIS}<Lispcore>next>Lisp.sysout   and {ERIS}<Lispcore>next>Full.sysout   and'
 It has been placed on {ERIS}<lispcore>next>dLispDomino.db'
 '
  These files contains both 4K and 12K microcode.'
   This hopefully fixes the 4K Dandelion non-booting problem.'
   I introduced the bug via some code which was "supposed to be" commented out.  My apologies to those of you who were inconvenienced.'
 '
  Problems to me.'
  '
  don'
'
-----'
'
Date': 23 Jul 84 18':34':16 PDT (Monday)'
From': Charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': , Charnley'
'
  A new set of microcode for the dandelion has been spliced into {ERIS}<Lispcore>next>Lisp.sysout   and {ERIS}<Lispcore>next>Full.sysout   and'
 It has been placed on {ERIS}<lispcore>next>dLispDomino.db;5'
 '
  These files contains both 4K and 12K microcode.'
  The non-caching bug has been squashed.  Should only impact performance.'
  The 4K sysout should recover from ^C now.'
 '
  Problems to me.'
  '
  don'
'
-----'
'
Date': 26 Jul 84 15':13':13 PDT (Thursday)'
From': Charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': Charnley'
'
  A new set of microcode for the dandelion has been spliced into {ERIS}<Lispcore>next>Lisp.sysout   and {ERIS}<Lispcore>next>Full.sysout   and'
 It has been placed on {ERIS}<lispcore>next>dLispDomino.db;6'
 '
  These files contains both 4K and 12K microcode.'
  Minor change in the Page-Cross logic in the 12K Sysout.  May improve things.  Let me know of any repeatable errors.'
 '
  Problems to me.'
  '
  don'
'
-----'
'
Date':  2 Aug 84 14':47':08 PDT (Thursday)'
From': charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': Charnley'
'
  A new set of microcode for the dandelion has been spliced into {ERIS}<Lispcore>next>Lisp.sysout   and {ERIS}<Lispcore>next>Full.sysout   and'
 It has been placed on {ERIS}<lispcore>next>dLispDomino.db;6'
 '
  These files contains both 4K and 12K microcode.'
  Final version {?} of the Array op-codes.  The 12K includes the Trident disk microcode.'
 '
  Problems to me.'
  '
  don'
'
-----'
'
Date':  3 Aug 84 13':06':10 PDT (Friday)'
From': charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': charnley'
'
  A new set of microcode for the dandelion has been spliced into {ERIS}<Lispcore>next>Lisp.sysout   and {ERIS}<Lispcore>next>Full.sysout   and'
{ERIS}<Lispcore>next>Demo.sysout   and'
 It has been placed on {ERIS}<lispcore>next>dLispDomino.db;8'
 '
  These files contains both 4K and 12K microcode.'
  Both the 4K and the 12K have a bug fix to HardReturn.  This caused the 9005 MP code during the sketch demo.'
 '
  Problems to me.'
  '
  don'
'
-----'
'
Date': 17 Aug 84 21':17 PDT'
From': Masinter.pa'
Subject': microcode bug in JUMPXX which crosses page boundary'
To': Charnley'
cc': LispSupport, Shrager'
Format': TEdit'
'
The following function fragment jumps to byte 2574 rather than 1574':'
'
 542':       147  45 215      ACONST    ASSOCRECORD'
 545':       360              EQ'
 546':       242              TJUMP-> 552'
 547':       261  |*|  2  25      JUMPXX->1574'
'
'
where the |*| denotes where the page boundary is. This only is a problem the first time the code is run, presumably because there is a page fault.'
'
Thanks to Jeff Shrager for helping track this one down.'
'
-----'
'
Date': 20 Aug 84 14':40':30 PDT (Monday)'
From': Charnley.pa'
Subject': New DLion uCode'
To': Lispcore^'
cc': , Charnley'
'
  A new set of microcode has been placed on {ERIS}<Lispcore>next>dLispDomino.db!9'
  and spliced into {ERIS}<Lispcore>next>Lisp.Sysout'
  and spliced into {ERIS}<Lispcore>next>Full.Sysout'
'
    This fixes a long standing bug in the JUMPXX microcode which caused the jumped to address to be off by one page when the target page page-faulted.'
'
    This was the bug reported by Larry in his message of 17 Aug 84 21':17 PDT.'
'
       Problems to me.'
'
          don'
'


Workaround: 

Test Case: 

Edit-By: Sannella.PA

Edit-Date: 28-Aug-84 09':37':32