On Mon, Apr 23, 2012 at 7:04 PM, Ulrich Weigand <uweig...@de.ibm.com> wrote: > Uros Bizjak wrote: > >> I don't have any problems with proposed order, but in my failed >> attempt to convert x86 to post-reload compare elimination, >> constrain_operands failed to recognize generated combined add+compare >> pattern, when commutative matched operands were swapped (e.g. the >> perfectly valid x86 add pattern above). constrain_operands was called >> from the call to validate_change (please see compare-elim.c) when >> flags clobber was substituted with compare. > > I would consider the pattern you mention above: > >> >> (set (reg: ax) (plus (reg: bx)(reg: ax))) > > "perfectly valid" only *before* reload, but invalid after reload.
You are right. I didn't notice that addsi_1 has additional r/r/0 alternative, and this alternative is in fact missing from addsi_2. I automatically assumed that something is wrong with pattern recognizers... Thanks for your explanation, I will fix this oversight in i386.md ASAP. Uros.
