https://bugs.kde.org/show_bug.cgi?id=426014
--- Comment #4 from Mark Wielaard <m...@klomp.org> --- Created attachment 132887 --> https://bugs.kde.org/attachment.cgi?id=132887&action=edit Fix ARM64Instr_VTriS/D argument order Turned out we had one typo assigning arguments (skipping arg2 and using arg3 twice) and the order of the addend, multiplier product. Which obviously caused the result to be totally incorrect. With this fixed the results look sane. There is a question whether the original argument order of the Iop_MAdd/SubF32/64 were correct in the first place. We better make sure we use the same as other arches (s390 and mips) so any VEX IR optimization PASS gets/keeps things in the correct order. -- You are receiving this mail because: You are watching all bug changes.
