This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: question regarding regrename and failure of 950704-1.c on main


In message <200205092222.g49MMkrC016094@hiauly1.hia.nrc.ca>, "John David Anglin
" writes:
 > > True, but if you look you'll find that they're packed in the wrong manner
 > > as register arguments (ie the order for passing them in registers is not
 > > the order in which we want to use them elsewhere).  They get copied out an
 > d
 > > repacked into proper order immediately after the prologue.
 > 
 > The Runtime Architecture document indicates that 64-bit integers are
 > passed with the high order word in an odd arg word (ie., arg1 == %r25 or
 > arg3 == %r23).  Thus, it would appear that special packing is not necessary.
 > The return value has the high order word in ret0 == %r28.
Maybe I had it backwards.  Arguments OK, reurn values reversed?

jeff
jeff


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]