On Thu, 27 Jun 2013 09:25:40 -0700, Ed Jaffe wrote: > >The area between 2G and 4G is guaranteed as a hole (unless you use an >undocumented parameter to acquire store there). I suggest that 64-bit >code load 3G into its "unused" registers to guarantee 0C4 even when long >displacements are used. > There might be some argument for initializing obtained storage to 7FFFF000 (or perhaps 3G) rather than to 0.
-- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
