draco@chinet.UUCP (Kent D. Meyers) (08/04/86)
In Article <132@wlbreng1.UUCP> pete@wlbreng1.UUCP (Pete Lyall) writes: > By the way, have you (you = You || Dave) had any trouble using the > NEWDISK under version 2.0?? A few of the users on CompuServe have > installed it under 2.0, and have had problems.. Initially, after > cold start, they fall out of the startup shell script because no > paths are set up (!!). If after this, the user MANUALLY does a > chd and chx, the problem is resolved and all functions normally. > Another interim fix for this is that the coco 2.0 SYSGO module > contains a hook that will attempt to CHD to '/h0' if the user has > one in the system. We have found that if these are patched to be > '/D0' that all is fine .. Strange, eh...?? The reason that there are no "paths" is because the I$chgdir call in OS9p2 fails and /D0 is not installed in the device table. I was able to determine that the device initialization code was being executed with no error, but the whole thing had me stumped for a while. Then it suddenly hit me that the F$Virq call was located in the Clock module and any call to F$Virq before the Clock was started would cause an error. This was what was happen- ing with Newdisk. Sysgo has already started the Clock by the time that it attempts the I$chgdir call, so the patches to Sysgo fixed the problem. Removing one line from the source will permanently cure the problem. This line is located in subroutine DSKSTART. OS9 F$VIRQ PULS Y BCS RETURN3 <---- Remove this line!!! There is also a problem with Newdisk if you have your drives jumpered for head load on device select. Newdisk deselects the drive upon completion of a command. This leads to an inordinate amount of head banging. Make the following changes to correct the offending code. CCDEXIT LDA #8 STA >DPORT PULS A,X,PC Change this to: CCDEXIT PULS A,X,PC Kent D. Meyers PO Box 266 Le Roy, MN 55951 (507)-324-5836 (After 8:30PM CDT) Net> ihnp4!chinet!draco