Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 28 Jul 2002 21:10:29 -0700 (PDT)
From:      Archie Cobbs <archie@dellroad.org>
To:        Clifton Royston <cliftonr@lava.net>
Cc:        freebsd-stable@FreeBSD.ORG
Subject:   Re: Is use of -O2 still deprecated for buildworld in -stable?
Message-ID:  <200207290410.g6T4AU910168@arch20m.dellroad.org>
In-Reply-To: <20020727191945.A10231@lava.net> "from Clifton Royston at Jul 27, 2002 07:19:45 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
Clifton Royston writes:
>   My understanding is that whether or not -O2 *should* produce exactly
> equivalent code to -O, there are recurrent reports of people having
> trouble doing a buildworld with -O2 in -stable or in 4.x release
> trains, where their problems go away when they revert back to -O.  In
> this case, we simply shouldn't be using it on production servers.  I'm
> using it on my home machines (-O2 for world, -O for kernel) but they
> has altogether different needs for reliability.

I've wondered about this a lot myself and never seen a clear answer.

My best guess as to what's going on is that the FreeBSD kernel contains
some (broken?) asm() statements that cause incorrect code to be generated
when -O2 is used, but not when -O is used.

This would explain how -O2 could cause FreeBSD to not work without
there actually being a bug in gcc causing it.

-Archie

__________________________________________________________________________
Archie Cobbs     *     Packet Design     *     http://www.packetdesign.com

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200207290410.g6T4AU910168>