View previous topic :: View next topic |
Author |
Message |
vipernicus Veteran
Joined: 17 Jan 2005 Posts: 1462 Location: Your College IT Dept.
|
Posted: Mon Nov 06, 2006 12:37 am Post subject: Re: Am I crazy???? |
|
|
ryszardzonk wrote: | I just tried to play with the ebuilds for the little wild and this is what I come up with.
1) I tried to get new binutils to merge with bidirect flag, however unsuccesfully. then I tried to use it with just a bidirect patch and not the hash+dynsort patch. It failed co merge as well
2) What I tried to merge binutils with bidirect+hashvals patch only (not dynsort) and it merged properly !!! Using just a -Wl,-hashvals produced gcc error in configure that in can not create executables, but what I found out is that placing -Wl,-Bdirect and -Wl,--hash-style=both WORKED! Either I am insane and it is just my sight playing tricks on me or I have it acctually working (Bdirect and new hashvals patch)
3) I am not sure if this is important, but what I did is emerged the toolchain first without old hashlvals and then tried to emerge it with the new ne and it worked. I am almost at the end of emerge world and yet to see a problem
4) I am not sure if plain ebuild from nxsty would work as I added extra OpenSuse patches. If the would be someone that can take further look at it and say why those patches are not in the Gentoo patchset it would be great.
Ebuilds I used
http://bigvo.dyndns.org/binutils.tar.bz2
http://bigvo.dyndns.org/glibc.tar.bz2 |
Hate to break it to you, but you're missing the glibc patch for -bdirect. _________________ Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog |
|
Back to top |
|
|
nirax Guru
Joined: 06 Jul 2004 Posts: 319 Location: Germany, old Europe
|
Posted: Mon Nov 06, 2006 10:44 pm Post subject: |
|
|
tomee wrote: | R-Type wrote: | Code: | gcc -DHAVE_CONFIG_H -I. -I/var/tmp/portage/binutils-2.17.50.0.5/work/binutils-2.17.50.0.5/bfd -I.
make[4]: *** [bfdsort.lo] Error 1
make[4]: Leaving directory `/var/tmp/portage/binutils-2.17.50.0.5/work/build/bfd'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/var/tmp/portage/binutils-2.17.50.0.5/work/build/bfd'
make[2]: *** [all] Error 2 |
I get this when building the new binutils-2.17.50.0.5 build from your overlay. the bdirect useflag is enabled. This is on amd64. |
Bump, same here. Binutils-2.17.50.0.5 from official tree emerged OK. |
same error trying to compile with bdirect flag binutils 2.17.50.0.5. without bdirect flag it compiles though, but having the whole system already bdirected i woudl like to stay with it. strange that this error is still there. (Read about it nearly one month before _________________ quot licet iovi non licet bovi |
|
Back to top |
|
|
ryszardzonk Apprentice
Joined: 18 Dec 2003 Posts: 225 Location: Rzeszów, POLAND
|
Posted: Tue Nov 07, 2006 6:06 pm Post subject: Re: Am I crazy???? |
|
|
vipernicus wrote: | Hate to break it to you, but you're missing the glibc patch for -bdirect. |
I'm glad that You found the time to take a closer look at the patchset. I must say that from the beginning I was thinking it is to good to be true But if this is true how come neither opensuse patchset does not have it nor the toolchain overlay based on them. Maybe it is not needed? (I know -bdirect is not in the official glibc) When I first switched to -Bdirect stuff I noticed quite a jump in app loading time and I must say it sure didn't decrease now at least visually while I am using these two. _________________ Sky is not the limit... |
|
Back to top |
|
|
nowinter Tux's lil' helper
Joined: 30 Apr 2006 Posts: 90
|
Posted: Thu Nov 09, 2006 9:37 pm Post subject: |
|
|
ok i'm done with Bdirect, removed it from flags and re-emerged the world with ---.6 portage binutils. this thread seems to freeze... _________________ All In All Is All We All Are |
|
Back to top |
|
|
Spruit_elf n00b
Joined: 30 Jul 2005 Posts: 33
|
Posted: Sat Nov 11, 2006 10:56 am Post subject: |
|
|
nxsty wrote: | abhay wrote: | nxsty wrote: | Everbody using binutils 2.17.50.0.5 from portage should use this ebuild instead as the regression is pretty serious (binutils bug #3314) Get it from toolchain_overlay. |
Damn!!! I just completed moving to new hashstyle and compiled whole world using this binutils version. Does this mean I have to recompile everything AGAIN?
How bad is it? |
If you´re on 64 bit you should or all your binaries will be unnecessary large. It doesn´t seem to be a problem on 32 bit. |
is the official 2.17.50.0.6 also broken or is it safe? |
|
Back to top |
|
|
na641 Apprentice
Joined: 27 Jun 2002 Posts: 169 Location: Eugene, OR
|
Posted: Wed Nov 15, 2006 6:52 pm Post subject: Anyone else having fetch problems |
|
|
I'm having difficulty pulling in the glibc patches/files. I figured i should wait a few days before posting to see if the issue would resolve itself, but alas it has not. Anyway, i browsed through to the online directory that hosts the files and the site is functioning fine, the files just aren't there. Could anyone perhaps offer up these missing files? Particularly 'glibc-2.5-patches-1.2.tar.bz2 |
|
Back to top |
|
|
S. Traaken Tux's lil' helper
Joined: 14 Nov 2003 Posts: 135
|
Posted: Wed Nov 15, 2006 9:38 pm Post subject: |
|
|
Try here for the time being. |
|
Back to top |
|
|
na641 Apprentice
Joined: 27 Jun 2002 Posts: 169 Location: Eugene, OR
|
Posted: Wed Nov 15, 2006 10:28 pm Post subject: |
|
|
S. Traaken wrote: | Try here for the time being. |
Thanks for the link, but i fixed the problem shortly after posting. I pulled the files for the portage glibc and saw that the patches were at 1.3, so i changed the overlay ebuild from 1.2 to 1.3 and it worked |
|
Back to top |
|
|
nesl247 Veteran
Joined: 15 Jun 2004 Posts: 1614 Location: Florida
|
Posted: Thu Nov 16, 2006 12:48 am Post subject: |
|
|
na641 wrote: | S. Traaken wrote: | Try here for the time being. |
Thanks for the link, but i fixed the problem shortly after posting. I pulled the files for the portage glibc and saw that the patches were at 1.3, so i changed the overlay ebuild from 1.2 to 1.3 and it worked |
Thanks for confirming it works, I've bumped the ebuild to use 1.3. |
|
Back to top |
|
|
S. Traaken Tux's lil' helper
Joined: 14 Nov 2003 Posts: 135
|
Posted: Thu Nov 16, 2006 1:04 am Post subject: |
|
|
Still waiting for a fix for binutils on amd64... |
|
Back to top |
|
|
nirax Guru
Joined: 06 Jul 2004 Posts: 319 Location: Germany, old Europe
|
Posted: Sun Nov 19, 2006 11:37 am Post subject: |
|
|
i am recompiling the gentoo standard glibc and binutils as the key people embedding changes to this builds seem to have lost interest and we have open issues with amd64 for long time now.
thanks for the great compilation nxsty though and the time spent to put all together ! _________________ quot licet iovi non licet bovi |
|
Back to top |
|
|
JoKo Tux's lil' helper
Joined: 16 May 2004 Posts: 141 Location: Xanthi, Greece
|
Posted: Sun Nov 19, 2006 5:08 pm Post subject: |
|
|
Code: | !!! A file listed in the Manifest could not be found: '/usr/local/overlays/toolchain_overlay/sys-libs/glibc/glibc-2.5.ebuild' |
This is what I get if I try to emerge glibc-2.5.90.20061022-r1. I've fixed it by digesting the ebuild again... |
|
Back to top |
|
|
tscolin n00b
Joined: 07 Dec 2005 Posts: 11
|
Posted: Wed Dec 06, 2006 4:25 pm Post subject: |
|
|
is this still being maintained? |
|
Back to top |
|
|
FireBurn Apprentice
Joined: 19 Sep 2004 Posts: 170 Location: Edinburgh, UK
|
Posted: Wed Dec 06, 2006 10:49 pm Post subject: |
|
|
It had been v quite on here
I don't feel like I'm on the bleeding edge with a glibc that's more than a month old!
He he |
|
Back to top |
|
|
nesl247 Veteran
Joined: 15 Jun 2004 Posts: 1614 Location: Florida
|
Posted: Wed Dec 06, 2006 10:52 pm Post subject: |
|
|
nxsty doesn't seem to be online anymore. I'll have to do some poking at him via private message. |
|
Back to top |
|
|
pal_gene n00b
Joined: 15 Nov 2005 Posts: 54
|
Posted: Sun Dec 10, 2006 8:41 am Post subject: |
|
|
FireBurn wrote: | I don't feel like I'm on the bleeding edge with a glibc that's more than a month old! |
yes, i felt also. so, i create with cvs-head by composite nxsty ebuild.
it is completely same patches from last 1022-r1.ebuild exclude branch update.
but minor sync with portage on ebuild detail strings.
i test on only x86-system and follow near safety env.
Code: | Portage 2.1.2_rc3 (default-linux/x86/2006.1/desktop, gcc-4.1.1, glibc-2.5.90.20061210-r0, 2.6.19-gentoo-r1 i686)
System uname: 2.6.19-gentoo-r1 i686 AMD Athlon(tm)
sys-devel/binutils: 2.17.50.0.8
virtual/os-headers: 2.6.19
CFLAGS=CXXFLAGS="-O2 -march=athlon-xp -fomit-frame-pointer -pipe"
LDFLAGS="-Wl,-O1 -Wl,--as-needed -Wl,--hash-style=gnu"
|
http://palgene-overlay.googlecode.com/svn/trunk/overlay/sys-libs/glibc/
http://palgene-overlay.googlecode.com/svn/trunk/distfiles/
- glibc-2.5-branch-update-20061210.patch.bz2 branch update for 20061210
- glibc-2.5-patches-1.4.tar.bz2 only minor fix against 20061210 patches based 1.3.
- glibc-2.5-extra_patches-1.4.tar.bz2 no-any difference nxsty extra patches
|
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
tscolin n00b
Joined: 07 Dec 2005 Posts: 11
|
Posted: Wed Dec 13, 2006 4:26 pm Post subject: |
|
|
im just trying to keep this on page 1 of forum list so it isn't forgotten. I hope nxsty comes back soon. |
|
Back to top |
|
|
boris64 Veteran
Joined: 04 Oct 2003 Posts: 1770 Location: Vechelde/Peine
|
Posted: Sat Dec 16, 2006 4:46 pm Post subject: |
|
|
kernelOfTruth wrote: | testing it right now, thanks |
Did the new glibc kill your computer or are you just a
bit too busy to give a short report?
Thx in advance _________________ boris64.net 200x / visit my desktop / try these tiny kernel patches |
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
Posted: Sat Dec 16, 2006 6:49 pm Post subject: |
|
|
sorry for the long delay, yeah, I'm a little busy:
here my experiences so far:
works fine with USE="glibc-omitfp hardened nls nptl nptlonly " (tested on x86)
works fine with gcc 4.1, 4.2, 4.3 , it's feels pretty stable & responsive, haven't run a benchmark though, so I can't tell you if there were any improvements in performance ...
pal_gene has done a good job
kudos to nxsty for his patches & good work, hope he continues this ... _________________ https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004
Last edited by kernelOfTruth on Sun Dec 17, 2006 10:57 am; edited 2 times in total |
|
Back to top |
|
|
nbvcxz Guru
Joined: 02 Sep 2005 Posts: 379 Location: Kraków / PL
|
Posted: Sat Dec 16, 2006 6:57 pm Post subject: |
|
|
here also works fine for few days (compiled with 4.2.0), but I can not compile it (and any other glibc) with 4.3.0
thx pal_gene _________________ nBVCXz
zen-kernel (bfq compcache) | /tmp -> tmpfs | ext4 | zsh | xfce | schedtool |
|
Back to top |
|
|
mbar Veteran
Joined: 19 Jan 2005 Posts: 1990 Location: Poland
|
Posted: Sun Dec 17, 2006 7:51 am Post subject: |
|
|
pal_gene wrote: |
i test on only x86-system and follow near safety env.
|
So I'm going to test it on AMD64 |
|
Back to top |
|
|
pal_gene n00b
Joined: 15 Nov 2005 Posts: 54
|
Posted: Sun Dec 17, 2006 4:32 pm Post subject: |
|
|
i modified to pass glibc testsuite and somehing wrong against gentoo glibc-compat.
FEATURES="test" will pass now.
elf/check-localplt test needs some modify for glibc-ssp-compat if only don't have USE="glibc-compat20".
so, scripts/data/localplt-{i386,x86_64}-linux-gnu.data was changed in this ebuild.
But i can't test on amd64.
*additional export func.
0010_all_glibc-ssp-compat.patch:__stack_chk_fail
6901_all_2.4-amd64-strings-20061210:__bzero
FEATURES="test" USE="alltest" will pass with required toolchain.
testing_note
Code: |
you can test glibc by glibc testsuite set enviroment FEATURES="test" when emerging glibc.
skip test nptl/tst-cancel1 by default. you can test nptl/tst-cancel1 with USE="alltest".
WARNING! During test, system become high load on some test.
-nptl/tst-mutex5,tst-cond*:
failing sometimes randomly for me.
it is timing problem, i think it depend on kernel timer or preemptible or scheduler.
application must engage timing can use rt(realtime) code, thus it should be harmless.
see http://www.osdl.org/developer_bugzilla/show_bug.cgi?id=34
- nptl/tst-cancel1
stop with "Didn't expect signal from child: got `Aborted'"
this seem to fail with not supported toolchain (gcc,binutils and other debugging with gdb).
DW_CFA_val_{offset{,_sf},expresion} is supported since gcc-4.2, binutils-2.17.
see http://sourceware.org/ml/libc-alpha/2006-09/msg00039.html
|
glibc-2.5.90.20061214
- Synced with portage on minor string diff
- New snapshot (20061214)
- Correct distfiles URL
- Fix some test error and maintain default ignore tests
|
|
Back to top |
|
|
kernelOfTruth Watchman
Joined: 20 Dec 2005 Posts: 6111 Location: Vienna, Austria; Germany; hello world :)
|
|
Back to top |
|
|
tscolin n00b
Joined: 07 Dec 2005 Posts: 11
|
Posted: Wed Dec 20, 2006 6:42 pm Post subject: |
|
|
pal_gene, thank you for maintaining glibc!!! your a godsend in the absence of nxsty. Could you make your own thread tho? cant say its very manageable w/ your ebuilds being on page 30 |
|
Back to top |
|
|
|