Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
GCC 4.1 (Part 2)
View unanswered posts
View posts from last 24 hours

Goto page 1, 2, 3  Next  
Reply to topic    Gentoo Forums Forum Index Portage & Programming
View previous topic :: View next topic  
Author Message
Maedhros
Bodhisattva
Bodhisattva


Joined: 14 Apr 2004
Posts: 5511
Location: Durham, UK

PostPosted: Sat Mar 18, 2006 12:21 pm    Post subject: GCC 4.1 (Part 2) Reply with quote

Continued from GCC 4.1[beta].
_________________
No-one's more important than the earthworm.
Back to top
View user's profile Send private message
irondog
l33t
l33t


Joined: 07 Jul 2003
Posts: 715
Location: Voor mijn TV. Achter mijn pc.

PostPosted: Sun Mar 19, 2006 12:19 am    Post subject: Reply with quote

I've got some experiences to share with you guys interested using GCC 4.1:
* GCC 4.1.0 is still hard masked at this time. You'll need to unmask it to get this compiler.
* The 4.1.0 patchset seems to have been updated at least one time since the ebuild appeared in portage, but nothing changed in de ebuild's -r versioning. Make sure you are using the latest patchset before complaining here.
* You don't have to expect many C/C++ errors when compiling up-to-date packages. GCC 4.0 and 4.1 mostly agree what C/C++ should look like. You might remember the time switching from 3.3 to 3.4 or from 3.x to 4.x? Switching from 4.0 to 4.1 hardly introduces new C/C++ errors. The world (everybody using Mandriva, Fedora, Suse, Debian) has been using GCC 4.0 for some time now, so many maintained projects (including Gentoo) have already been working on code fixes to support GCC4.
* You should be able to build a complete system with this compiler.
* Compiling will be slower compared to almost any other version of GCC. If you are only interested in fast compiling to have a smaller code/compile/test cycle, think twice about using this compiler.
* The code generated by this compiler seems to be faster than previous versions.
* As far as I know there are no compiler profiles when using GCC 4.1.0. So you won't get a hardened GCC profile when using portage ebuilds.
_________________
Alle dingen moeten onzin zijn.
Back to top
View user's profile Send private message
Matteo Azzali
Retired Dev
Retired Dev


Joined: 23 Sep 2004
Posts: 1133

PostPosted: Sun Mar 19, 2006 2:03 pm    Post subject: Reply with quote

Hum, since I'm maybe the one here that found most un-compiling packages in the last days (I'm supposing this reading previous thread, I may be wrong),
here's a little summary of my experience unmasking GCC-4.1.0 and using glibc-2.4-r1:

Out of 856 packages, about 20 weren't compiling anymore at my first try (all were not-system and not-vital packages).
Of these 20 packages:
about 6 are now fixed directly in portage (firefox, thunderbird and enigmail for glibc-2.4, baghira, xawtv, mupen64-riceplugin),
while 11 have working patch in gentoo bugzilla (jpeg-mmx got from metalshark overlay),
and 3 packages are still not compiling (xmms-ir , erlang, trickle).

I admit that 3 in 856 isn't zero, but given these aren't dependancies for anything (so don't affect other packages) ,
situation seems relatively good here.....
_________________
Every day a new distro comes to birth. Every day a distro "eats" another.
If you're born distro, no matter what, start to run.
---- http://www.linuxprinting.org/ ---- http://tuxmobil.org/
Back to top
View user's profile Send private message
Gergan Penkov
Veteran
Veteran


Joined: 17 Jul 2004
Posts: 1464
Location: das kleinste Kuhdorf Deutschlands :)

PostPosted: Sun Mar 19, 2006 2:32 pm    Post subject: Reply with quote

jpeg-mmx was fixed without a bump:
Quote:
15 Mar 2006; Mark Loeser <halcy0n@gentoo.org>
+files/jpeg-mmx-0.1.6-gcc41.patch, jpeg-mmx-0.1.6-r1.ebuild:
Add patch to fix compilation with gcc-4.1 by Ryan Hill <dirtyepic.sk AT
gmail DOT com>; bug #119851

I have similiar numbers for the failing packages (from around 1300), but I'm still sorting the failed ones :)
_________________
"I knew when an angel whispered into my ear,
You gotta get him away, yeah
Hey little bitch!
Be glad you finally walked away or you may have not lived another day."
Godsmack
Back to top
View user's profile Send private message
Halcy0n
Developer
Developer


Joined: 17 Sep 2003
Posts: 1682
Location: Freehold, NJ

PostPosted: Sun Mar 19, 2006 6:40 pm    Post subject: Reply with quote

Gergan Penkov wrote:
jpeg-mmx was fixed without a bump:
Quote:
15 Mar 2006; Mark Loeser <halcy0n@gentoo.org>
+files/jpeg-mmx-0.1.6-gcc41.patch, jpeg-mmx-0.1.6-r1.ebuild:
Add patch to fix compilation with gcc-4.1 by Ryan Hill <dirtyepic.sk AT
gmail DOT com>; bug #119851

I have similiar numbers for the failing packages (from around 1300), but I'm still sorting the failed ones :)

Compile fixes will be fixed without a bump, since the patch is only needed by people that aren't able to compile the application :)
_________________
Mark Loeser
http://www.halcy0n.com
Back to top
View user's profile Send private message
Vlad.Sharp
Guru
Guru


Joined: 08 Dec 2004
Posts: 337
Location: Cambridgeshire, UK

PostPosted: Sun Mar 19, 2006 7:27 pm    Post subject: Reply with quote

How much of a performance boost does -finlines-hidden give, at least on Gcc 4.1?
Back to top
View user's profile Send private message
Exner
Tux's lil' helper
Tux's lil' helper


Joined: 08 Apr 2003
Posts: 128
Location: Melbourne, Australia

PostPosted: Sun Mar 19, 2006 11:07 pm    Post subject: Reply with quote

On page
Quote:
http://gcc.gnu.org/gcc-4.1/changes.html

it says when using
Quote:
The x86-64 medium model (that allows building applications whose data segment exceeds 4GB)...

you must use
Quote:
Binutils 2.16.91 or newer...

_________________
- Exner (Antony Suter)
Back to top
View user's profile Send private message
metalshark
Apprentice
Apprentice


Joined: 07 Jan 2006
Posts: 197
Location: Bournemouth, England

PostPosted: Tue Mar 21, 2006 6:25 am    Post subject: Reply with quote

Vlad.Sharp wrote:
How much of a performance boost does -finlines-hidden give, at least on Gcc 4.1?


Just finished an
Code:
emerge -e world
and it doesn't seem to affect in program speed at all but the binaries are certainly smaller.

You might find adding
Code:
if [ "$CATEGORY" == "sys-libs" ] ; then
   export CFLAGS=${CFLAGS/-fvisibility=hidden}
fi
to your /etc/portage/bashrc will stop your system borking if you are enabling it in your /etc/make.conf :D.

For instance a simple:
Code:
source /etc/make.conf
CFLAGS="$CFLAGS -fvisibility=hidden" emerge file zlib util-linux

will fail on fsck.cramfs.c as /usr/lib/libmagic.so.1.0.0 will not add zlib's inflate command into the ld list and fsck for cramfs relies on inflate. This is the common scenario for using it in sys-libs, there is rare breakage outside that category but be careful with x11 stuff (KDE only tested here).

PS: -fprofile-generate does work for a fair few packages - but you have to disable ccache and mod the ebuilds to install the .gcdb file (in my amateur experience - cannot find much documentation on this feature). Also -fwhole-program does work on the occasional package - building in support requires copying the source code from the package(s) you use the libraries of and modifying the configures/headers/makefiles - however - the performance is immensely improved (2-3% is stated everywhere but from package to package I get 0-15%).
_________________
Gentoo's Portage: Any ideas for reform?
Back to top
View user's profile Send private message
nxsty
Veteran
Veteran


Joined: 23 Jun 2004
Posts: 1556
Location: .se

PostPosted: Tue Mar 21, 2006 8:29 am    Post subject: Reply with quote

Don't put -fvisibility=hidden in your CFLAGS. That's a certain way of borking things. The -fvisibility-inlines-hidden flag is farily safe though.
Back to top
View user's profile Send private message
metalshark
Apprentice
Apprentice


Joined: 07 Jan 2006
Posts: 197
Location: Bournemouth, England

PostPosted: Tue Mar 21, 2006 8:57 am    Post subject: Reply with quote

nxsty wrote:
Don't put -fvisibility=hidden in your CFLAGS. That's a certain way of borking things. The -fvisibility-inlines-hidden flag is farily safe though.


Not doubting you but could you please explain the reasoning and/or packages for your comment?

I am just trying to get a better grasp of what these new features do and do badly.
_________________
Gentoo's Portage: Any ideas for reform?
Back to top
View user's profile Send private message
walrus_55
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jul 2003
Posts: 122
Location: Melbourne, Australia

PostPosted: Tue Mar 21, 2006 10:02 am    Post subject: Reply with quote

I get this error while trying to emerge gcc-4.1.0

Code:

Automaton `pentium'
       48 NDFA states,            138 NDFA arcs
       48 DFA states,             138 DFA arcs
       20 minimal DFA states,      82 minimal DFA arcs
      273 all insns         17 insn equivalence classes
   88 transition comb vector els,   340 trans table els: use comb vect
   88 state alts comb vector els,   340 state alts table els: use comb vect
  340 min delay table els, compression factor 2

Automaton `pentium_fpu'
       80 NDFA states,            172 NDFA arcs
       80 DFA states,             172 DFA arcs
       75 minimal DFA states,     162 minimal DFA arcs
      273 all insns          8 insn equivalence classes
  164 transition comb vector els,   600 trans table els: use comb vect
  164 state alts comb vector els,   600 state alts table els: use comb vect
  600 min delay table els, compression factor 1

Automaton `ppro_decoder'
        4 NDFA states,             12 NDFA arcs
        4 DFA states,              12 DFA arcs
        4 minimal DFA states,      12 minimal DFA arcs
      273 all insns          4 insn equivalence classes
   13 transition comb vector els,    16 trans table els: use simple vect
   13 state alts comb vector els,    16 state alts table els: use simple vect
   16 min delay table els, compression factor 8

Automaton `ppro_core'
      105 NDFA states,            376 NDFA arcs
      105 DFA states,             376 DFA arcs
      105 minimal DFA states,     376 minimal DFA arcs
      273 all insns         13 insn equivalence classes
  481 transition comb vector els,  1365 trans table els: use comb vect
  481 state alts comb vector els,  1365 state alts table els: use comb vect
 1365 min delay table els, compression factor 1

Automaton `ppro_idiv'
       38 NDFA states,             79 NDFA arcs
       38 DFA states,              79 DFA arcs
       38 minimal DFA states,      79 minimal DFA arcs
      273 all insns          5 insn equivalence classes
   82 transition comb vector els,   190 trans table els: use simple vect
   82 state alts comb vector els,   190 state alts table els: use simple vect
  190 min delay table els, compression factor 1

Automaton `ppro_fdiv'
       38 NDFA states,             79 NDFA arcs
       38 DFA states,              79 DFA arcs
       38 minimal DFA states,      79 minimal DFA arcs
      273 all insns          5 insn equivalence classes
   82 transition comb vector els,   190 trans table els: use simple vect
   82 state alts comb vector els,   190 state alts table els: use simple vect
  190 min delay table els, compression factor 1

Automaton `ppro_load'
        3 NDFA states,              8 NDFA arcs
        3 DFA states,               8 DFA arcs
        3 minimal DFA states,       8 minimal DFA arcs
      273 all insns          4 insn equivalence classes
    9 transition comb vector els,    12 trans table els: use simple vect
    9 state alts comb vector els,    12 state alts table els: use simple vect
   12 min delay table els, compression factor 4

Automaton `ppro_store'
       16 NDFA states,             56 NDFA arcs
       16 DFA states,              56 DFA arcs
       11 minimal DFA states,      44 minimal DFA arcs
      273 all insns          7 insn equivalence classes
   51 transition comb vector els,    77 trans table els: use simple vect
   51 state alts comb vector els,    77 state alts table els: use simple vect
   77 min delay table els, compression factor 4

Automaton `k6_decoder'
        4 NDFA states,             11 NDFA arcs
        4 DFA states,              11 DFA arcs
        3 minimal DFA states,       9 minimal DFA arcs
      273 all insns          4 insn equivalence classes
   10 transition comb vector els,    12 trans table els: use simple vect
   10 state alts comb vector els,    12 state alts table els: use simple vect
   12 min delay table els, compression factor 8

Automaton `k6_load_unit'
       11 NDFA states,             24 NDFA arcs
       11 DFA states,              24 DFA arcs
       11 minimal DFA states,      24 minimal DFA arcs
      273 all insns          4 insn equivalence classes
   26 transition comb vector els,    44 trans table els: use simple vect
   26 state alts comb vector els,    44 state alts table els: use simple vect
   44 min delay table els, compression factor 2

Automaton `k6_store_unit'
       68 NDFA states,            233 NDFA arcs
       68 DFA states,             233 DFA arcs
       37 minimal DFA states,     126 minimal DFA arcs
      273 all insns          6 insn equivalence classes
  140 transition comb vector els,   222 trans table els: use simple vect
  140 state alts comb vector els,   222 state alts table els: use simple vect
  222 min delay table els, compression factor 1

Automaton `k6_integer_units'
      114 NDFA states,            396 NDFA arcs
      114 DFA states,             396 DFA arcs
      114 minimal DFA states,     396 minimal DFA arcs
      273 all insns         11 insn equivalence classes
  471 transition comb vector els,  1254 trans table els: use comb vect
  471 state alts comb vector els,  1254 state alts table els: use comb vect
 1254 min delay table els, compression factor 1

Automaton `k6_fpu_unit'
       58 NDFA states,            120 NDFA arcs
       58 DFA states,             120 DFA arcs
       57 minimal DFA states,     118 minimal DFA arcs
      273 all insns          5 insn equivalence classes
  120 transition comb vector els,   285 trans table els: use simple vect
  120 state alts comb vector els,   285 state alts table els: use simple vect
  285 min delay table els, compression factor 1

Automaton `k6_branch_unit'
        2 NDFA states,              5 NDFA arcs
        2 DFA states,               5 DFA arcs
        2 minimal DFA states,       5 minimal DFA arcs
      273 all insns          3 insn equivalence classes
    6 transition comb vector els,     6 trans table els: use simple vect
    6 state alts comb vector els,     6 state alts table els: use simple vect
    6 min delay table els, compression factor 8

Automaton `athlon'
      518 NDFA states,           1668 NDFA arcs
      518 DFA states,            1668 DFA arcs
       76 minimal DFA states,     328 minimal DFA arcs
      273 all insns         10 insn equivalence classes
  359 transition comb vector els,   760 trans table els: use simple vect
  359 state alts comb vector els,   760 state alts table els: use simple vect
  760 min delay table els, compression factor 2

Automaton `athlon_load'
      162 NDFA states,            855 NDFA arcs
      162 DFA states,             855 DFA arcs
      162 minimal DFA states,     855 minimal DFA arcs
      273 all insns         10 insn equivalence classes
 1047 transition comb vector els,  1620 trans table els: use simple vect
 1047 state alts comb vector els,  1620 state alts table els: use simple vect
 1620 min delay table els, compression factor 2

Automaton `athlon_mult'
       16 NDFA states,             48 NDFA arcs
       16 DFA states,              48 DFA arcs
       16 minimal DFA states,      48 minimal DFA arcs
      273 all insns          4 insn equivalence classes
   50 transition comb vector els,    64 trans table els: use simple vect
   50 state alts comb vector els,    64 state alts table els: use simple vect
   64 min delay table els, compression factor 2

Automaton `athlon_fp'
    15522 NDFA states,          99908 NDFA arcs
    15522 DFA states,           99908 DFA arcs
      483 minimal DFA states,    3275 minimal DFA arcs
      273 all insns         21 insn equivalence classes
 3474 transition comb vector els, 10143 trans table els: use comb vect
 3474 state alts comb vector els, 10143 state alts table els: use comb vect
10143 min delay table els, compression factor 1

17545 all allocated states,     102661 all allocated arcs
32702 all allocated alternative states
 6673 all transition comb vector els, 17200 all trans table els
 6673 all state alts comb vector els, 17200 all state alts table els
17200 all min delay table els
    0 locked states num

  transformation: 0.008001, building DFA: 9.192573
  DFA minimization: 0.460030, making insn equivalence: 0.000000
 all automaton generation: 9.784611, output: 0.104007
/bin/sh /var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../move-if-change tmp-attrtab.c insn-attrtab.c
echo timestamp > s-attrtab
gcc   -O -DENABLE_CHECKING -DENABLE_ASSERT_CHECKING -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedantic -Wno-long-long -Wno-variadic-macros -Wold-style-definition -Wmissing-format-attribute    -DHAVE_CONFIG_H -I. -I. -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/. -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../include -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../libcpp/include     -c insn-attrtab.c \
  -o insn-attrtab.o
/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/config/i386/i386.md: In function 'internal_dfa_insn_code':
/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/config/i386/i386.md:21422: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://bugs.gentoo.org/> for instructions.
The bug is not reproducible, so it is likely a hardware or OS problem.
make[2]: *** [insn-attrtab.o] Error 1
make[2]: Leaving directory `/var/tmp/portage/gcc-4.1.0/work/build/gcc'
make[1]: *** [stage1_build] Error 2
make[1]: Leaving directory `/var/tmp/portage/gcc-4.1.0/work/build/gcc'
make: *** [profiledbootstrap] Error 2

!!! ERROR: sys-devel/gcc-4.1.0 failed.
Call stack:
  ebuild.sh, line 1557:   Called dyn_compile
  ebuild.sh, line 966:   Called src_compile
  ebuild.sh, line 1272:   Called toolchain_src_compile
  toolchain.eclass, line 24:   Called gcc_src_compile
  toolchain.eclass, line 1487:   Called gcc_do_make

!!! emake failed with profiledbootstrap
!!! If you need support, post the topmost build error, and the call stack if relevant.


and this is my emerge --info

Code:
gentoo cam # emerge --info
Portage 2.1_pre6-r3 (default-linux/x86/2006.0, gcc-4.0.3, glibc-2.3.6-r3, 2.6.15-gentoo-r1 i686)
=================================================================
System uname: 2.6.15-gentoo-r1 i686 AMD Athlon(tm) XP 3200+
Gentoo Base System version 1.6.14
ccache version 2.3 [enabled]
dev-lang/python:     2.3.4-r1, 2.4.2
sys-apps/sandbox:    1.2.12
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.16.1-r2
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -mtune=athlon-xp -pipe -O3 -fweb -frename-registers -fforce-addr -momit-leaf-frame-pointer -fomit-frame-pointer -ftracer"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/lib/X11/xkb /usr/lib/mozilla/defaults/pref /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/splash /etc/terminfo /etc/env.d"
CXXFLAGS="-march=athlon-xp -mtune=athlon-xp -pipe -O3 -fweb -frename-registers -fforce-addr -momit-leaf-frame-pointer -fomit-frame-pointer -ftracer -fvisibility-inlines-hidden"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache distlocks metadata-transfer sandbox sfperms strict"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LDFLAGS="-Wl,-O1 -Wl,--enable-new-dtags -Wl,--sort-common -s"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 3dfx 3dnow 3dnowext X alsa apache2 apm arts asf avi berkdb bitmap-fonts cdb cdr cli crypt ctype cups dba divx4linux dri dts dvb dvd dvdr dvdread eds emboss encode esd expat fastbuild firefox foomaticdb force-cgi-redirect fortran ftp gd gdbm gif gpm gstreamer gtk gtk2 hal imlib ipv6 java joystick jpeg kde libg++ libwww lirc mad memlimit mikmod mmx motif mozilla mp3 mpeg mysql mythtv ncurses nls nptl nptlonly nsplugin nvidia ogg opengl oss pam pcre pdflib perl pic png posix ppds python qt quicktime readline real samba sdl session simplexml soap sockets spell spl sse sse2 ssl symlink tcpd tokenizer truetype truetype-fonts type1-fonts udev usb vorbis win32codecs xine xml xmms xsl xv xvid zlib elibc_glibc kernel_linux userland_GNU"
Unset:  ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LANG, LC_ALL, LINGUAS


I first tried to emerge 4.0.3 but that didnt work, so then i have been trying 4.1.0 to no avail, so i downgraded glibc to 2.3 and this is where i am. I have read that glibc 2.4 is not very stable so am i right to try to upgrade to gcc 4 first and then upgrade to glibc 2.4 later on?
Back to top
View user's profile Send private message
metalshark
Apprentice
Apprentice


Joined: 07 Jan 2006
Posts: 197
Location: Bournemouth, England

PostPosted: Tue Mar 21, 2006 10:10 am    Post subject: Reply with quote

Shot in the dark - but it is stopping on gcc/config - have you run
Code:
emerge -u gcc-config
?
_________________
Gentoo's Portage: Any ideas for reform?
Back to top
View user's profile Send private message
Matteo Azzali
Retired Dev
Retired Dev


Joined: 23 Sep 2004
Posts: 1133

PostPosted: Tue Mar 21, 2006 10:17 am    Post subject: Reply with quote

#xmms-ir (fixed with patch at https://bugs.gentoo.org/show_bug.cgi?id=126591).

out of 856 packages only erlang and trickle are still not ported to gcc-4.1, trickle is likely to become a masked package
(unmantained by years), so the last un-emerging package is erlag.
_________________
Every day a new distro comes to birth. Every day a distro "eats" another.
If you're born distro, no matter what, start to run.
---- http://www.linuxprinting.org/ ---- http://tuxmobil.org/
Back to top
View user's profile Send private message
Bob P
Advocate
Advocate


Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs

PostPosted: Tue Mar 21, 2006 10:32 am    Post subject: Reply with quote

dirtyepic wrote:
Bob P wrote:
i have to say that in my experience the combination of GCC 4.1.0, Glibc 2.3.6 and nptlonly works great, but that the combination of GCC 4.1.0, Gentoo's Glibc 2.4 and nptlonly is horribly broken right now.

well, you're set firmly in the minority.

well, time has proven me right about that.

glibc-4.2-r1 is out and the truth has been revealed that Gentoo's glibc-2.4-r0 did indeed have a major league bug in it. there was an entire line of code that was inadvertently left out of Gentoo's version of glibc-2.4-r0 becuase a defective patch set that was copied from SuSE inappropriately removed a critical line of code.

from the bug report:
Code:
Gentoo apply a patch, 5021_all_2.3.6-fnmatch.patch, as part of
glibc-2.3.6-patches-1.8.tar.bz2.  Amongst other things, this removes the
following line from the fnmatch() function in posix/fnmatch.c:

    wstring = (wchar_t *) alloca ((n + 1) * sizeof (wchar_t));

This leaves the variable wstring uninitialised if strlen(string) >= 1024, which
appears to cause of segfaults in GNU ld when heavily templated C++ results in
very long ELF section names.


so even if i was firmly in the minority about the problems with 2.4-r0, at least i was right about them. :lol:

IME after rebuilding about a dozen boxen, Glibc 2.4-r1 works great with GCC 4.1 and all of the problems i had been having with Glibc 2.4-r0 have been resolved. IMHO, its now safe to go back into the water. :wink:
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks
Back to top
View user's profile Send private message
Vlad.Sharp
Guru
Guru


Joined: 08 Dec 2004
Posts: 337
Location: Cambridgeshire, UK

PostPosted: Tue Mar 21, 2006 9:08 pm    Post subject: Reply with quote

Ok, thanks for the info on -fvisibility-inlines-hidden - I've used it before, but now I have had a new system install and I don't really think that a few extra MBs of -fomit-frame-pointer will make any difference in reality. Oh well, time to have a good wait for -Bdirect support in non-broken versions of binutils (I really don't want to go too unstable for a while ;) )
Back to top
View user's profile Send private message
walrus_55
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jul 2003
Posts: 122
Location: Melbourne, Australia

PostPosted: Wed Mar 22, 2006 6:22 am    Post subject: Reply with quote

i have run gcc-config, and set it back to 3.4.5 and re-emerged all the dependancys of gcc and now i get a simmilar but different error

Code:
gcc -c   -O -DENABLE_CHECKING -DENABLE_ASSERT_CHECKING -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -Wmissing-format-attribute    -DHAVE_CONFIG_H -DGENERATOR_FILE -I. -Ibuild -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/build -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../include -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../libcpp/include     -o build/genautomata.o /var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/genautomata.c
gcc -c   -O -DENABLE_CHECKING -DENABLE_ASSERT_CHECKING -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -Wmissing-format-attribute    -DHAVE_CONFIG_H -DGENERATOR_FILE -I. -Ibuild -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/build -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../include -I/var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/../libcpp/include     -o build/varray.o /var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/varray.c
gcc   -O -DENABLE_CHECKING -DENABLE_ASSERT_CHECKING -DIN_GCC   -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -Wmissing-format-attribute    -DHAVE_CONFIG_H -DGENERATOR_FILE -Wl,-O1 -Wl,--sort-common -s -o build/genattrtab \
 build/genattrtab.o build/genautomata.o \
 build/rtl.o build/read-rtl.o build/ggc-none.o build/min-insn-modes.o build/gensupport.o build/insn-conditions.o build/print-rtl.o build/errors.o \
 build/varray.o ../build-i686-pc-linux-gnu/libiberty/libiberty.a -lm
build/genattrtab /var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/config/i386/i386.md > tmp-attrtab.c
/bin/sh: line 1:   661 Segmentation fault      build/genattrtab /var/tmp/portage/gcc-4.1.0/work/gcc-4.1.0/gcc/config/i386/i386.md >tmp-attrtab.c
make[2]: *** [s-attrtab] Error 139
make[2]: Leaving directory `/var/tmp/portage/gcc-4.1.0/work/build/gcc'
make[1]: *** [stage1_build] Error 2
make[1]: Leaving directory `/var/tmp/portage/gcc-4.1.0/work/build/gcc'
make: *** [profiledbootstrap] Error 2

!!! ERROR: sys-devel/gcc-4.1.0 failed.
Call stack:
  ebuild.sh, line 1565:   Called dyn_compile
  ebuild.sh, line 974:   Called src_compile
  ebuild.sh, line 1280:   Called toolchain_src_compile
  toolchain.eclass, line 24:   Called gcc_src_compile
  toolchain.eclass, line 1487:   Called gcc_do_make

!!! emake failed with profiledbootstrap
!!! If you need support, post the topmost build error, and the call stack if relevant.




Can anyone help?
Back to top
View user's profile Send private message
Vlad.Sharp
Guru
Guru


Joined: 08 Dec 2004
Posts: 337
Location: Cambridgeshire, UK

PostPosted: Wed Mar 22, 2006 7:10 am    Post subject: Reply with quote

Hmmmn, what is your actual architecture and what are the contents of emerge --info and /etc/make.conf?
Back to top
View user's profile Send private message
walrus_55
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jul 2003
Posts: 122
Location: Melbourne, Australia

PostPosted: Wed Mar 22, 2006 10:02 am    Post subject: Reply with quote

i have an athlon xp 3200+,

emerge --info
Code:
emerge --info
Portage 2.1_pre6-r5 (default-linux/x86/2006.0, gcc-3.4.5, glibc-2.3.6-r3, 2.6.15-gentoo-r1 i686)
=================================================================
System uname: 2.6.15-gentoo-r1 i686 AMD Athlon(tm) XP 3200+
Gentoo Base System version 1.6.14
ccache version 2.3 [enabled]
dev-lang/python:     2.3.4-r1, 2.4.2
sys-apps/sandbox:    1.2.12
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.16.1-r2
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -mtune=athlon-xp -pipe -O2 -frename-registers -fomit-frame-pointer"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/lib/X11/xkb /usr/lib/mozilla/defaults/pref /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/splash /etc/terminfo /etc/env.d"
CXXFLAGS="-march=athlon-xp -mtune=athlon-xp -pipe -O2 -frename-registers -fomit-frame-pointer -fvisibility-inlines-hidden"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache distlocks metadata-transfer sandbox sfperms"
GENTOO_MIRRORS="http://distfiles.gentoo.org http://distro.ibiblio.org/pub/linux/distributions/gentoo"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -s"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 3dfx 3dnow 3dnowext X alsa apache2 apm arts asf avi berkdb bitmap-fonts cdb cdr cli crypt ctype cups dba divx4linux dri dts dvb dvd dvdr dvdread eds emboss encode esd expat fastbuild firefox foomaticdb force-cgi-redirect fortran ftp gd gdbm gif gpm gstreamer gtk gtk2 hal imlib ipv6 java joystick jpeg kde libg++ libwww lirc mad memlimit mikmod mmx motif mozilla mp3 mpeg mysql mythtv ncurses nls nptl nptlonly nsplugin nvidia ogg opengl oss pam pcre pdflib perl pic png posix ppds python qt quicktime readline real samba sdl session simplexml soap sockets spell spl sse sse2 ssl symlink tcpd tokenizer truetype truetype-fonts type1-fonts udev usb vorbis win32codecs xine xml xmms xsl xv xvid zlib elibc_glibc kernel_linux userland_GNU"
Unset:  ASFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, LANG, LC_ALL, LINGUAS


And my /etc/make.conf

Code:
cat /etc/make.conf
# Copyright 2005 Jackass! Production Team
# Distributed under the terms of the GNU General Public License v2
# File: /etc/make.conf
# Task: Basic configuration for the Jackass-optimized Gentoo toolkit
#
#  /\          /\  Jackass 2005.0 for Athlon-XP
# ( \\        // )
#  \ \\      // /  This system configuration should be used for all testing
#   \_\\||||//_/  of the Jackass! Project.  Please do not  make any changes
#    \/ _  _ \/  to the following settings (especially CFLAGS!), other than
#   \/|(O)(O)|  to edit the location of your Gentoo mirror.
#  \/ |      |
#_\/  \      /  If you should have questions, please refer to: man gcc, man
#      |____|  make.conf, and /etc/make.conf.example for more information.
#     /      \
#     \ 0  0 /  Have fun with your new Jackass! Toolkit for Gentoo 2005.0!
#    / \____/
#   /
#_/
#
#(donkey by hard '96)
#
CHOST="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -mtune=athlon-xp"
#
CFLAGS="${CFLAGS} -pipe"
CFLAGS="${CFLAGS} -O2"
#CFLAGS="${CFLAGS} -fweb"
CFLAGS="${CFLAGS} -frename-registers"
#CFLAGS="${CFLAGS} -fforce-addr"
#CFLAGS="${CFLAGS} -momit-leaf-frame-pointer"
CFLAGS="${CFLAGS} -fomit-frame-pointer"
#CFLAGS="${CFLAGS} -ftracer"
#
CXXFLAGS="${CFLAGS} -fvisibility-inlines-hidden"
#
ACCEPT_KEYWORDS="x86"
PORTAGE_TMPDIR=/var/tmp
PORTDIR=/usr/portage
DISTDIR=${PORTDIR}/distfiles
PKGDIR=${PORTDIR}/packages
PORT_LOGDIR=/var/log/portage
PORTDIR_OVERLAY=/usr/local/portage
MAKEOPTS="-j2"
USE="3dfx 3dnow 3dnowext alsa arts asf avi cdb cdr cups divx4linux dts dvb dvd dvdr dvdread encode firefox foomaticdb -gnome gtk gtk2 hal java joystick jpeg kde lirc nptl mysql mp3 mpeg mmx mozilla mythtv nptlonly nsplugin nvidia opengl oss pam pic ppds real samba sse sse2 symlink truetype usb win32codecs xine xmms xvid"
#GENTOO_MIRRORS="<your-mirror-goes here> http://gentoo.osuosl.org http://www.ibiblio.org/pub/Linux/distributions/gentoo"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
RSYNC_RETRIES="3"
RSYNC_TIMEOUT=180
PORTAGE_NICENESS=3
AUTOCLEAN="yes"
FEATURES="-strict distlocks sandbox ccache"
CCACHE_SIZE="512M"
LDFLAGS="-Wl,-O1 -Wl,--sort-common -s"
LIRC_OPTS="--with-driver=livedrive_midi"
Back to top
View user's profile Send private message
Bob P
Advocate
Advocate


Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs

PostPosted: Wed Mar 22, 2006 10:52 am    Post subject: Reply with quote

wow. a Jackass! user. 8) Well, sort of. Its interesting to note that although your make.conf says Jackass! 2005.0, your emerge --info shows that you are NOT using the Jackass! 2005.0 toolkit. :? In fact you've changed toolkit components at least a couple of times, which complicates the issue.

I'm just wondering, prior to performing the GCC migration, did you have a chance to look at my Unauthorized Migration Guide for GCC?

When emerging hard masked toolkit components, for the sake of completeness its always helpful to post the contents of your /etc/portage/package.* folders, as they contain important configuration information that is not contained in make.conf.

looking over your previous posts, i see something that i didn't see at first -- that you had migrated from GCC 3.4.x to GCC 4.0.x, and in attempting to migrate to GCC 4.1 you started off with GCC 4.0.3 and you've fallen back to GCC 3.4.5. Wow. That really complicates things.

Quote:
and this is my emerge --info

Code:
gentoo cam # emerge --info
Portage 2.1_pre6-r3 (default-linux/x86/2006.0, gcc-4.0.3, glibc-2.3.6-r3, 2.6.15-gentoo-r1 i686)
=================================================================
System uname: 2.6.15-gentoo-r1 i686 AMD Athlon(tm) XP 3200+
Gentoo Base System version 1.6.14
ccache version 2.3 [enabled]
dev-lang/python:     2.3.4-r1, 2.4.2
sys-apps/sandbox:    1.2.12
sys-devel/autoconf:  2.13, 2.59-r7
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.16.1-r2
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.11-r2


your situation is complicated by the fact that you've migrated the toolkit back and forth between several versions of GCC. what method did you use to do the updates after you performed your initial installation? its absolutely essential that you've done everything right, or your toolchain may not be in order, and your problems could well be related to a corrupted toolkit.

at this point, i think your first priority has to be to assure the integrity of your compiler and your system toolkit. before you try to upgrade to anything, i'd recommend verifying your system config files, environment and libraries.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks
Back to top
View user's profile Send private message
walrus_55
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jul 2003
Posts: 122
Location: Melbourne, Australia

PostPosted: Wed Mar 22, 2006 11:10 am    Post subject: Reply with quote

Yes it is indeed a mess now, i originally tried to upgrade to gcc 4.0.3 and it compilled fine, but then failed when i tried to re-compile it with new gcc-4.0.3, and also glibc would fail.
So i then went back to 3.4.5 in an effort to try to emerge gcc-4.1.0. but this will not compile at all

package.keywords
Code:
cat /etc/portage/package.keywords
# for the toolkit
~sys-apps/coreutils-5.94 ~x86
~sys-apps/groff-1.19.2 ~x86
#
sys-devel/binutils ~x86
sys-devel/gcc-config ~x86
~sys-devel/gcc-4.1.0
sys-libs/glibc ~x86
#
~sys-libs/glibc-2.3.6 ~x86
~sys-apps/busybox-1.1.0 ~x86
sys-libs/libstdc++-v3 ~x86
sys-libs/timezone-data ~x86
dev-libs/mpfr ~x86
dev-libs/gmp ~x86
#
# for portage:
sys-apps/portage ~x86
>=dev-python/pycrypto-2.0.1 ~x86


package.use
Code:
cat /etc/portage/package.use
# Copyright 2005 Jackass! Production Team
# Distributed under the terms of the GNU General Public License v2
# File: /etc/portage/package.use
# Task: package-specific USE flags for building the Jackass! Toolkit
#
sys-libs/glibc userlocales nomalloccheck nptl nptlonly pic
sys-devel/gcc multislot
dev-lang/perl ithreads
sys-devel/libperl ithreads
dev-libs/gmime mono
app-misc/beagle pdf spreadsheet webservices wv
dev-java/swt -firefox


package.unmask
Code:
cat /etc/portage/package.unmask
=x11-base/xorg-x11-6.8.99.15-r4
=sys-devel/gcc-4.1.0
Back to top
View user's profile Send private message
Bob P
Advocate
Advocate


Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs

PostPosted: Wed Mar 22, 2006 11:37 am    Post subject: Reply with quote

well, i recognize some of those other files. :wink: although i'm glad to see them, i'm a bit concerned about some of the mixing and matching that we're dealing with. are you using any ebuilds are overlays that are not a part of the official gentoo portage tree? even though i like the toolkit components and config files you're trying to use now, i can't say anything about the underlying toolkit that you have on your system. as far as fixing it goes, are you more interested in an approach that minimizes CPU time or seat time? if your toolkit is functional and you're willing to sacrifice CPU cycles, I'd recommend trying a redundant rebuild of the system with your GCC 3.4.5 compiler if it will rebuild itself. then you'll have to enter the GCC migration path all over again. this approach would take a LONG time. OTOH, if you're willing to reinstall there's always the option of just untarring a toolkit that is known to work, and using it for the foundation of your system.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks
Back to top
View user's profile Send private message
geki
Advocate
Advocate


Joined: 13 May 2004
Posts: 2387
Location: Germania

PostPosted: Wed Mar 22, 2006 3:49 pm    Post subject: Reply with quote

To compile STLport-5.0.2 with gcc4:
https://bugs.gentoo.org/127188
_________________
hear hear
Back to top
View user's profile Send private message
mirek
Guru
Guru


Joined: 20 Sep 2004
Posts: 489
Location: Oslo Norway

PostPosted: Thu Mar 23, 2006 6:33 am    Post subject: Reply with quote

I have got over to GCC 4.1.0 and now I can not emerge no-sources from an overlay. I have got segmentation fault
Code:

>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/s2io.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/sctp.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/shaper.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/sis900.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/sk98lin.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/skfp.txt
>>> /usr/src/linux-2.6.16-rc5-no3/Documentation/networking/slicecom.hun
Segmentation fault

I have tried archck-sources with the same error
Back to top
View user's profile Send private message
travlr
Apprentice
Apprentice


Joined: 20 Apr 2005
Posts: 158
Location: USA

PostPosted: Thu Mar 23, 2006 10:44 am    Post subject: Reply with quote

Anyone else having problems with sys-libs/timezone-data-2006(a or b). I'm using these
Code:
LDFLAGS="-Wl, -O1 -Wl, --as-needed -Wl, --sort-common -s"
CFLAGS="-march=pentium4 -O2 -fomit-frame-pointer -pipe"


No LDFLAGS, no problem.

travlr
_________________
ProDataLab.com -- A Focus On Efficiency
Back to top
View user's profile Send private message
Gergan Penkov
Veteran
Veteran


Joined: 17 Jul 2004
Posts: 1464
Location: das kleinste Kuhdorf Deutschlands :)

PostPosted: Thu Mar 23, 2006 11:39 am    Post subject: Reply with quote

travlr wrote:
Anyone else having problems with sys-libs/timezone-data-2006(a or b). I'm using these
Code:
LDFLAGS="-Wl, -O1 -Wl, --as-needed -Wl, --sort-common -s"
CFLAGS="-march=pentium4 -O2 -fomit-frame-pointer -pipe"


No LDFLAGS, no problem.

travlr

It works here without the
Code:
-Wl, --sort-common -s

I use only
Code:
-Wl,-O1 -Wl,--as-needed

_________________
"I knew when an angel whispered into my ear,
You gotta get him away, yeah
Hey little bitch!
Be glad you finally walked away or you may have not lived another day."
Godsmack
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Portage & Programming All times are GMT
Goto page 1, 2, 3  Next
Page 1 of 3

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum