Joern Rennecke <joern.renne...@embecosm.com> writes: > On 2 April 2014 13:08, Anthony Green <gr...@moxielogic.com> wrote: > >> I though the answer was to simply add something like this... >> >> (define_insn "zero_extendqisi" >> [(set (match_operand:SI 0 "register_operand" "=r") >> (zero_extend:SI (match_operand:QI 1 "register_operand" "r")))] >> "" >> "; ZERO EXTEND (comment for debugging)") > > That pattern is obviously not outputting valid code.
It's actually just a valid assembler comment so I can see if the pattern is used. > You should make this a define_insn_and_split, with an r/r alternative > that is split (after reload) as necesary into shifts, and an m/r alternative > that outputs a load. sprinkle with rtx_cost adjustments as necessary. Thanks for the tip. I have it working now! AG
