Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!metro!news.cs.su.oz.au!harbinger.cc.monash.edu.au!msuinfo!agate!howland.reston.ans.net!vixen.cso.uiuc.edu!sdd.hp.com!apollo.hp.com!bloom-beacon.mit.edu!ai-lab!life.ai.mit.edu!mycroft From: mycroft@duality.gnu.ai.mit.edu (Charles Hannum) Newsgroups: comp.os.386bsd.misc Subject: Re: Why uses /usr/src -O instead of -O2? Date: 17 Feb 1994 23:29:04 GMT Organization: MIT Artificial Intelligence Lab Lines: 25 Message-ID: <MYCROFT.94Feb17182904@duality.gnu.ai.mit.edu> References: <1994Feb16.103448.6398@paramount.nikhefk.nikhef.nl> NNTP-Posting-Host: duality.gnu.ai.mit.edu In-reply-to: eloy@paramount.nikhefk.nikhef.nl's message of Wed, 16 Feb 1994 10:34:48 GMT In article <1994Feb16.103448.6398@paramount.nikhefk.nikhef.nl> eloy@paramount.nikhefk.nikhef.nl (Eloy Domingos) writes: When rebuilding the 'world' I noticed that the makefiles use -O instead of the better optimizing -O2. Is there a reason not to use the better optimization? Sometimes the `higher' levels cause bugs to appear that no one has noticed before; in particular, there was a bug in GCC 2.4.5 that caused the kernel to sometimes get stuck in an infinite loop in the signal processing code which only seemed to appear with -O2 or so, and only on some 486s. You don't see this now because I got tired of my Vectra crashing and fixed it. Other than compiler bugs, there is no particular reason. The first thing I ever did with 386BSD was shove it down GCC 2's throat, screaming all the way, and I've never used GCC 1 since (except for occasional testing). All my compiles are done with `-O6'. -- - Charles Hannum NetBSD group Working ports: i386, hp300, amiga, sparc, mac68k, pc532. In progress: pmax, sun3.