https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93738
--- Comment #17 from Segher Boessenkool <segher at gcc dot gnu.org> --- (In reply to Kishan Parmar from comment #16) > Apart from this, i notice that other arch like ia64, aarch64, etc.. lowers > zero_extract to respective bit-field extract insns.. should we do the same > for rs6000? No. Our machine insns are way more general and way more useful than this very partial not-so-useful abstraction, which always can be expressed with simpler RTL anyway (just like the way more general mask insert things we also support). The only thing that would happen if our backen would allow zero_extract is that we need twice as many patterns for things touching on mask insert, one in simple terms (as we have now), and one using zero_extract.
