Gentoo Archives: gentoo-dev

From: Bart Verwilst <verwilst@g.o>
To: gentoo-dev@g.o
Subject: Re: [gentoo-dev] Re: GCC3.1 bootstrap fails
Date: Sun, 02 Jun 2002 16:38:27
Message-Id: 200206022332.09275.verwilst@gentoo.org
In Reply to: [gentoo-dev] Re: GCC3.1 bootstrap fails by Prashanth Aditya Susarla
1 Would be cool if you could try and bootstrap again with -fomit-frame-pointer,
2 and now use the 1.3-gcc3.1 tarball from ibiblio
3 (http://www.ibiblio.org/gentoo/releases/build/) i just made. If it fails, i
4 guess we should filter out most of the optimisations... I think the new
5 gcc3.1-compiled tarball will make all the gcc3.1-users much happier :o)
6 I also updated the packages file from the gcc 3.1 profile, and the
7 make.defaults. Please keep us posted on any progress!
8
9 Thanks!
10
11
12 On Sunday 02 June 2002 19:42, Prashanth Aditya Susarla wrote:
13 || >do -not- use -fomit-frame-pointer during bootstrap! (can anyone who
14 || >actually got this to -work- using omit-frame-pointer speak up??)
15 || >
16 || >This causes breakage in glibc.
17 ||
18 || That's interesting. When I used default-1.0, I used all the flags which
19 || one particular FAQ on gentoo.org mentioned (-fforce-addr
20 || -fomit-frame-pointer -funroll-loops blah blah blah along with -march=i686
21 || -O3 of course) and everything worked fine (atleast compiled fine though I
22 || was not truly happy with the performance (lots of memory leaks and
23 || subsequent crashes of kswapd primarily). So maybe this is an issue only
24 || with gcc-3.x and/or binutils-2.12.x. But I toned down the optimizations
25 || and the bootstrap worked fine.
26 ||
27 || >this is due to a broken glibc (omit frame pointer and -O3. I think glibc
28 || >strips out -O3 but not the -f flags. )
29 ||
30 || gcc does this. glibc retains whatever you specify. Anyway, that's beside
31 || the point - only for the sake of information ;-).
32 ||
33 || >> problem is that most binaries segfault soon after glibc is built.
34 || >
35 || >Try again with -O2 -pipe and you will get a working system :)
36 ||
37 || Infact, I finally used the default options for Athlon systems (-march=i686
38 || -pipe -O3.... not -O2 but -O3) and glibc didn't break. As I said earlier,
39 || this is strange because all those crazy optimizations worked in the case
40 || of the default-1.0 profile.
41 ||
42 || >emerge system
43 || >/(watch it break down and cry in xfree because you overoptimized)
44 ||
45 || So you're saying that if I use -march=athlon-tbird -fomit-frame-pointer
46 || -O3 [-pipe] -mmmx -m3dnow (the latter only for the sake of multimedia apps
47 || like xmms, mplayer etc.) the xfree ebuild b0rks out? That's too bad,
48 || really. And the thing with xfree is that experimenting stops being an
49 || option if the very first emerge itself doesn't work :-(. Not all of us
50 || have all the time in the world...
51 ||
52 || Regards,
53 || Prashanth Aditya Susarla
54
55 --
56 Bart Verwilst
57 Gentoo Linux Developer, Desktop Team
58 Gent, Belgium

Replies

Subject Author
Re: [gentoo-dev] Re: GCC3.1 bootstrap fails Prashanth Aditya Susarla <aditya@×××××××.in>
Re: [gentoo-dev] Re: GCC3.1 bootstrap fails Prashanth Aditya Susarla <aditya@×××××××.in>
Re: [gentoo-dev] Re: GCC3.1 bootstrap fails Brian Webb <webbb@××××××××××××.net>