On Jun 30, 2012, at 09:59, Edward Jaffe wrote: > > 24-bit I/O. Don't do it! > Some years ago, attempting to write to a RECFM=U data set from a Rexx program, using ISPF LM services (LMPUT, IIRC), I encountered an ABEND. PMR. APAR. The Problem Description stated that in this case LMPUT couldn't deal with the Rexx buffer above the line. The resolution was that LMPUT would copy the data below the line to perform the I/O.
Is there something wrong with this picture? IBM seems to be doing the right thing with the BPX1*/BPX4* system calls. For every(?) BPX1* syscall there's a corresponding BPX4* that accepts a 64-bit plist. z/OS 1.13 is reported to support execution of code above the bar. Interrupts will be handled properly, but (most?) SVCs won't be supported when issued from above the bar. I wonder whether BPX4* syscalls can be issued from above the bar? Does the z hardware support 64-bit I/O? -- gil
