https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124019

--- Comment #6 from Hans-Peter Nilsson <hp at gcc dot gnu.org> ---
(In reply to Jeffrey A. Law from comment #5)
> Then that's a target issue HP.
> 
> From a canonicalization standpoint Andrew's approach is sound and checking
> RTX costs is inappropriate here as we're in gimple and very much want to
> avoid introducing those kinds of target dependencies into gimple.

Thus the argument that "X is always more expensive than Y for all targets" is
also inappropriate (as well as incorrect here, for X=shift and Y=and), leaving
the canonicalization argument standing.  I'll buy that.

Reply via email to