When you open a DCB for a JES file, OPEN creates an ACB in the same address space. The JES code in support of the open ACB runs partially in the user's address and partially in the JES address space. The details depend, among other things, on which JES you are using.
-- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 ________________________________________ From: IBM Mainframe Assembler List <[email protected]> on behalf of Paul Gilmartin <[email protected]> Sent: Wednesday, January 31, 2018 12:17 AM To: [email protected] Subject: Re: Device Independence On 2018-01-30, at 21:30:33, Jon Perryman wrote: >> Steve Thompson wrote:> It causes me to wonder, why DSORG causes a problem >> with a path or > >> a file. After all, unless you know something special about the >> file, aren't all files DSORG=PS -- Physical Sequential in EXT, >> I believe that Using Data Sets states that PO.DATA.SET(MEMBER) opens as sequential. But I discover that OPEN still fills the DCB with PO, not PS. Experts say this is normal. >> EXT2, btrfs, hfs, zfs, etc.? Ok, in hfs and zfs they are >> emulating an FBA device, but still... > > Programs do not access Unix files. Specifying path causes a Unix address > space to handle the file. It communicates with the program thru the subsystem > file > Isn't the same true for JES files? > (not disk). It is up to the subsystem to follow the protocol that the DCB > requires. Obviously, they decided not to implement every feature which is why > it fails. > Pretty much my understanding. And that subsystem does a creditable job of following that protocol. HLASM handles UNIX (subsystem) files well for SYSIN, SYSLIN, SYSPRINT, and SYSLIB. (It took a couple of APARs before nested COPY members worked for me -- some wild POINTs, some ABENDs.) My SR on Rexx SYSEXEC went about like this: gil> Rexx fails with a UNIX directory as SYSEXEC. Using Data Sets says any combination of PDS, PDSE, and UNIX directories is a valid partitioned concatenation. IBM> Rexx checks and (in this case) requires DSORG=PO. WAD. gil> That check is superfluous. It should be removed. IBM> Then OPEN might ABEND if DSORG were wrong. WAD. gil> Let it ABEND. A failure is a failure. IBM> ABEND is a more severe failure and must be avoided. WAD. Spitefully, I tried a concatenation of a PDS followed by a VSAM data set as SYSEXEC. Rexx OPEN ABENDed. I opened a new SR on the ABEND gil> ABEND occurs when SYSEXEC is a concatenation of PO and VSAM. IBM> Predictable. User error. gil> (Citing my earlier SR) IBM has told me that Rexx must not abend. IBM> User error. SR closed. I suppose I forgot Emerson on Consistency. I didn't go to RFE. I suppose my chances were slim. -- gil
