Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 30 Jul 2003 20:06:18 -0700
From:      David Schultz <das@FreeBSD.ORG>
To:        Scott Long <scottl@FreeBSD.ORG>
Cc:        cvs-all@FreeBSD.ORG
Subject:   Re: cvs commit: src/release Makefile
Message-ID:  <20030731030618.GA7795@HAL9000.homeunix.com>
In-Reply-To: <3F27EFCF.4070605@freebsd.org>
References:  <200307301446.h6UEklCE044311@repoman.freebsd.org> <3F27EFCF.4070605@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Jul 30, 2003, Scott Long wrote:
> Ruslan Ermilov wrote:
> >ru          2003/07/30 07:46:46 PDT
> >
> >  FreeBSD src repository
> >
> >  Modified files:
> >    release              Makefile 
> >  Log:
> >  The -Os kernel build with GCC 3.3.1 also requires -fno-strict-aliasing
> >  to survive.
> >  
> >  Discussed with: bde
> >  Tested on:      i386 and pc98
> >  
> >  Revision  Changes    Path
> >  1.806     +2 -1      src/release/Makefile
> >
> 
> Thanks for doing this.  However, I think that Peter was looking at doing
> this globally in /sys/conf, not just the release.  The reasoning is that
> it might not be safe to use -fstrict-aliasing in any context right now
> until we understand the actual impact of those warnings.  I'd also
> prefer if it was done there and not special-cased for the release.

It has never been safe to compile the kernel with
-fstrict-aliasing (or consequently -O2) in gcc3.
Peter fixed a place where the warning was bogus,
not a place where the kernel violated C's aliasing
rules in a way that would cause gcc to make incorrect
optimizations.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030731030618.GA7795>