Date: 02 Nov 2000 11:37:09 -0500 From: Randell Jesup <rjesup@wgate.com> To: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> Cc: <freebsd-stable@FreeBSD.ORG>, Cy Schubert - ITSD Open Systems Group <Cy.Schubert@uumail.gov.bc.ca>, "David O'Brien" <obrien@FreeBSD.ORG> Subject: Re: "Malloc type lacks magic" show-stopper solved Message-ID: <ybubsvyfhze.fsf@jesup.eng.tvol.net.jesup.eng.tvol.net> In-Reply-To: Gerald Pfeifer's message of "Thu, 2 Nov 2000 12:54:23 %2B0100 (CET)" References: <Pine.BSF.4.30.0011021247480.81655-100000@taygeta.dbai.tuwien.ac.at>
next in thread | previous in thread | raw e-mail | index | archive | help
Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> writes:
>David O'Brien wrote:
>> Forgot to mention. I'm considering removing the ability to do -O2 and
>> -O3 from the system C compiler again.... people are just proving over
>> and over how much they want to shoot their foot off.
>
>Please don't. We are already seeing enough problems due to differences
>between FSF GCC and FreeBSD's patched version. Please let's reduce those
>differences, not introduce anything further.
Removing -O2 and -O3 from userland would be a Very Bad Thing, and
would cause us considerable consternation. Proper use of -O2, -O3, and
various -f's and -m's can get us 10's of % on some heavily compute-bound
code. This translates to considerable differences in $$$ in hardware.
>Besides, -O2 and -O3 are perfectly legitimate options for user code, and
>I really don't see why one shouldn't use the system compiler for that as
>well but install another one.
Absolutely.
>Finally, I do agree about -O3, but -O2 *should* definitely work for the
>kernel. If it does not, this is a bug, either in GCC or FreeBSD, and ought
>to be tracked down and, in the former case, at least be reported to the
>GCC folks.
-O2 should work. I've never seen a bug with -O2 code generation
from GCC (not that they can't happen, of course). Most bugs I've seen in
my career with optimization levels were actually timing holes in the
source code, where someone was accessing a shared resource and counting
on ordering/etc, or bugs with accessing hardware registers where structures
weren't properly marked volatile.
If there's a problem, find it and report it. IMHO.
--
Randell Jesup, Worldgate Communications, ex-Scala, ex-Amiga OS team ('88-94)
rjesup@wgate.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?ybubsvyfhze.fsf>