AMODE(24) in the below should read RMODE(24), right? Also, some routines won't run AMODE(31), even if they are RMODE(24). Programs with tables with words of the format X'flags',AL3(data) for example, if they load the word into a register and then use the contents of the register as an address.
However, you can solve that by front-ending the routine with a stub that goes AMODE(31) to (24) before the call and (24) to (31) on the way back out - assuming of course that the passed parameters are below the line. What a pain! Was it Gene Amdahl who said the biggest mistake of the 360 architecture was the 24-bit addresses? Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of john gilmore Sent: Thursday, June 22, 2006 9:47 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Old Hashing Routine If you can use a 'new fangled' program object instead of a load module as your executable, then RMODE(SPLIT) provides a nice resolution of such problems as you describe: AMODE(24) for a few intractable, difficult to convert routines and RMODE(31) for all the rest. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html