On Apr 5, 2011, Bernd Schmidt <ber...@codesourcery.com> wrote: > On 04/05/2011 11:05 PM, Alexandre Oliva wrote: >> While debugging the -fcompare-debug regression that turned out to be >> independently reported as 48403, I got slightly annoyed that >> rank_for_schedule would walk a sequence of debug insns over and over, >> once per compare. I figured it might be useful to waste a few bits of >> static memory to save a pointer to the latest nondebug scheduled insn, >> so we wouldn't have to skip over debug insns at that point.
> Yeah, the thought did occur to me afterwards... >> Regstrapped on x86_64- and i686-linux-gnu. Ok to install? > + if (flag_sched_last_insn_heuristic > + && NONDEBUG_INSN_P (last_nondebug_scheduled_insn)) > Isn't that always true now Not if it's the initial NOTE. > (except we should probably initialize it to NULL_RTX and check for > that here)? Good idea. Will try that, retest, and apply if it passes. -- Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/ You must be the change you wish to see in the world. -- Gandhi Be Free! -- http://FSFLA.org/ FSF Latin America board member Free Software Evangelist Red Hat Brazil Compiler Engineer