Presumptuously, I have verified that what John Ehrman says about the use of SECTALIGN(16) is entirely correct. (I use this value routinely for my AMODE(64) code, and I had not been aware of this dependency.) I am nevertheless left unhappy. A reader of the discussion of CNOP in the LR could be forgiven for coming away from it supposing that CNOP supports quadword alignment unconditionally, and there are even more egregious offenses elsewhere in the LR. For example, the text | LQ | quadword | [DC LQ'0.1'] | quadword | in which both implicit alignment and alignment are specified as quadword in the table on page 129 of the current HLASM 1.6 LR is, I think, misleading. [Note that the bracketed material actually appears, improperly, in the line above this one.]
John Gilmore Ashland, MA 01721-1817 USA