On 2 December 2015 at 06:47, Peter Hunkeler <[email protected]> wrote: > When I look at system trace entries for translation exceptions (PBG 10, > 11, etc) I see for example > > PGM....011 00000000_25A00E32 00060011 00000000.................. 07852001 > 80000000 .................. 40000400 > > > (replaced blanks with dots, to somehow keep the data aligned) > > > From the "MVS Tools & Service Aids" manual I understand that the last > words on the two lines are the address which cause the tramslation > exception (TEA, High word on upper line, low word on lower line). In the > above example my program was accessing storage at x'00000000_40000000' > (first reference). The trace entry shows this as x'00000000_40000400'. > > The TEA seems to always point to the start of the page where within the > page the access was (which makes sense), but what is the x'400' in the low > 12 bit telling me? The manual only talks about the high order bit > indicating primary or secondary address space. I guess it some flags, but > where is this documented? >
In the Principles of Operation Chapter 3, under Assigned Storage Locations 168-175 (A8-AF hex). This part also refers to the descriptions of [Enhanced] Suppression on Protection earlier in the same chapter. Perhaps we're back to testing the facility bits before you can be sure what goes on... Bits 0-51 are the address (to the page level), and there are several flag bits defined and other bits to the right that may be "unpredicatable" under various circumstances. This all assumes that the data at A8-AF is copied exactly to the trace record, and that the trace record as formatted by IPCS shows exactly those same bits. I guess that these are both the case, but I don't know. Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
