Gentoo Forums :: View topic - CONRAD Installation Guide 2.5.1 "Hajimemashite"

CONRAD Installation Guide 2.5.1 "Hajimemashite"

View unanswered posts
View posts from last 24 hours

Goto page Previous 1, 2, 3, 4, 5, 6, 7, 8, 9

Gentoo Forums Forum Index

Unsupported Software

View previous topic :: View next topic

Author

Message

vipernicus
Veteran
Veteran

Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.

Posted: Thu Apr 06, 2006 9:31 pm Post subject:

cheater1034 wrote:

1) kde-meta won't build all the way through, ever. (so many packages fail)
2) System actually seems slower?
3) Over-Optimizing is never a good solution
4) It does break things, breaking 1 or 2 things is alright, but running so unstable does indeed break things
5) Tons of strange errors on emerges for more than 3 people? (increases the compile times)

1. I've emerged kde-meta several times, no fails
2. Improvements here, how could that slow it down?
3. CFLAGS Over-Optimize, not LDFLAGS
4. As long as you are using flags like -frename-registers -fweb -tracer, etc, some things will break slightly, or at least make debugging impossible.
5. I have had strange errors while using no-sources that I didn't get while using a ck-based kernel.
_________________
Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog

Last edited by vipernicus on Fri Apr 07, 2006 4:23 pm; edited 1 time in total

cheater1034
Veteran
Veteran

Joined: 09 Sep 2004
Posts: 1558

Posted: Fri Apr 07, 2006 12:02 am Post subject:

vipernicus wrote:

cheater1034 wrote:

vipernicus wrote:

What kind of problems are you having with those ldflags?

Quite a bit actually
1) kde-meta won't build all the way through, ever. (so many packages fail)
2) System actually seems slower?
3) Over-Optimizing is never a good solution
4) It does break things, breaking 1 or 2 things is alright, but running so unstable does indeed break things
5) Tons of strange errors on emerges for more than 3 people? (increases the compile times)

Not worth it basically, it isn't a bad idea, it is a good idea to replace pre-linking, but using simple ldflags is way more stable

Things do break, people don't realize them though because they are so used to it :-p, and they have a bashrc to remove the flag, etc.

I want my guide to be the best, and that is what it will be (How many packages do you need to go LDFLAGS="" emerge pkgname for?)

1. I've emerged kde-meta several times, no fails
2. Improvements here, how could that slow it down?
3. CFLAGS Over-Optimize, not LDFLAGS
4. No matter what as long as you are using flags like -frename-registers -fweb -tracer, etc, some things will break.
5. I have had strange errors while using no-sources that I didn't get while using a ck-based kernel.

1. I haven't, maybe gentoo fixed their kde ebuilds to exclude the -Bdirect flag (aka, it doesn't work)
2. Of course there are improvements, just not noticable
3. They link the binaries, etc. things like -Bdirect are too experimental
4. Nothing I've experienced has broken due to my CFLAGS
5. What? I've had some strange experiences with ck and staircase v15/plugsched 6.3.2 (nicksched blew it up in interbench?) - Compile errors won't come from the kernel :-p
6. How many times have you had to do LDFLAGS="" to fix a package, or add a bashrc entry?
7. Most packages don't break anymore because they exclude the flag, so there isn't much point to using it anyway.

----
Sidenotes:

I ran interbench, fresh-booted command line, equal load (nothing), with staircase and nicksched (2.6.16-no1, with staircase v15 and smpnice - as seen in plugsched 6.3.2)

I was pleased with what I saw when I benchmarked the two, nicksched won in every thing except, compilation on Audio test, and X test (lost by a small margin), and it lost everything in the video test by a slight margin. Nicksched had better latencies at no load, video load, X load, in all categories, except video. Nicksched also won the read/write categories (in Audio/X/Gaming, not Video) by a significant margin. Nicksched won everything in the gaming category, some by bigger margins than others.

My conclusion is, on my system (athlon64, 754 2800+, 1gb pc3200 ram, sata150 hdd, gigabye k8ns-pro mobo) Nicksched is definitely the better option for me.
Link: http://pastebin.com/644922
----

The new guide is going to have an option, a gcc 4.1.1 svn option, and a 4.1.0 from portage option.

For the gcc 4.1.1 svn option (Default) I made a gcc-svn-update script, basically nxsty's small script, with questions (about installation svn, creating an ebuild, is the date correct, and some re-inforcing words) It works very well, and will be used for updating gcc 4.1 svn, and the ebuild. And it will be used for creating the snapshot and the ebuild. For people who do not want to do this, they can just ignore the gcc overlay and emerge gcc.

Guide coming soon!

_________________
IRC!: #zen-sources on irc.rizon.net
zen-kernel.org
--
Lost in android development land.

nesl247
Veteran
Veteran

Joined: 15 Jun 2004
Posts: 1614
Location: Florida

Posted: Fri Apr 07, 2006 3:40 am Post subject:

cheater1034 wrote:

1. You are probably using a broken or oudated version of glibc, and your binutils probably isn't patched correctly. (You need to use .6, not .7, not sure what you use)
2. I notice it, maybe you forgot to include LD_BIND_DIRECT="1" in /etc/env.d/99local
3. Only problem that I have has was libXfont which is fixed in portage now
4. Your right, its probably because of your toolchain
5. I have had performance issues with anything mm based, ck has worked fine.
6. My bashrc is empty now.
7. Not necessarily true. To my knowledge, -Bdirect didn't break much.

yngwin
Retired Dev

Joined: 19 Dec 2002
Posts: 4572
Location: Suzhou, China

Posted: Fri Apr 07, 2006 12:54 pm Post subject:

suur13 wrote:

yngwin wrote:

Quote:

But... during booting, somewhere after the volumes are mounted, I get like 20 errors, all saying something "unable to load xxx before sysinit is completed"... where xxx stands for diffrent system components - eth0, gpm, numlock...

I've got the same problem. There is more about it here:
https://forums.gentoo.org/viewtopic-t-435535.html
https://forums.gentoo.org/viewtopic-p-3086639.html#3086639

So far no solution for me...

Got over it without doing anything (after 4th reboot). Evrything works now.

I logged in as root, started all the services by hand (/etc/init.d/xxx start) then rebooted. Works here as well now.

suur13 wrote:

Only udev is complaining something. Errors on lines 8xx and 8xx (numbers are changing).
udevd-event[8xx]: find_free_number: %e
That this is deprecated, will not work and will be deleted...

Yes, I noticed that as well. It's a problem in /etc/udev/rules.d/50-udev.rules, but since I know zip about writing udev-rules, I don't know how to fix it. The is issue is known (of course), as you can see in https://bugs.gentoo.org/show_bug.cgi?id=127815 and https://bugs.gentoo.org/show_bug.cgi?id=128601. But I don't get why #127815 is 'RESOLVED INVALID' - because the issue is neither resolved nor invalid!
_________________
"Those who deny freedom to others deserve it not for themselves." - Abraham Lincoln
Free Culture | Defective by Design | EFF

sirformatalot
n00b
n00b

Joined: 03 Feb 2006
Posts: 38

Posted: Fri Apr 07, 2006 4:49 pm Post subject: restart

So Cheater I have this toolkit built with gcc 4.1 & binutils 2.16.91.0.6 just waiting for some fun smart ideas( but with all LDFLAGS commented out ). Should I just format yet again & look for a saner guide or try to resurrect what's left of your last nutty overlay?

bigbob73
Guru
Guru

Joined: 31 Dec 2004
Posts: 332
Location: Under the Lone Star

Posted: Fri Apr 07, 2006 5:05 pm Post subject: Re: restart

sirformatalot wrote:

So Cheater I have this toolkit built with gcc 4.1 & binutils 2.16.91.0.6 just waiting for some fun smart ideas( but with all LDFLAGS commented out ). Should I just format yet again & look for a saner guide or try to resurrect what's left of your last nutty overlay?

seems like it would be easier to wait for the new guide.

_________________
A computers attention span is only as long as it's electrical cord (Murphy)

seren
Guru
Guru

Joined: 27 Aug 2005
Posts: 448
Location: Wisconsin

Posted: Fri Apr 07, 2006 5:40 pm Post subject:

i just finished upgrading to a new gcc 4.1.1 snapshot, and recompiled everything with NO LDFLAGS, and sane CFLAGS and my system feels way more responsive than with -Bdirect -hashvals and all that fun stuff. :roll:

Code:

ortage 2.1_pre7-r3 (default-linux/x86/2006.0, gcc-4.1.1-pre20060405, glibc-2.4.20060329-r0, 2.6.16-no1 i686)
=================================================================
System uname: 2.6.16-no1 i686 Intel(R) Pentium4(R) CPU 2.40GHz
Gentoo Base System version 1.12.0_pre16
ccache version 2.4 [enabled]
dev-lang/python: 2.3.5-r2, 2.4.2-r1
sys-apps/sandbox: 1.2.17
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-r2
sys-devel/binutils: 2.16.91.0.6
sys-devel/libtool: 1.5.22
virtual/os-headers: 2.6.16
ACCEPT_KEYWORDS="x86 ~x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-O2 -march=pentium4 -pipe -fomit-frame-pointer -frename-registers -fweb"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/kde/3/share/config /usr/lib/mozilla/defaults/pref /usr/share/X11/xkb /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/eselect/compiler /etc/gconf /etc/revdep-rebuild /etc/splash /etc/terminfo /etc/env.d"
CXXFLAGS="-O2 -march=pentium4 -pipe -fomit-frame-pointer -frename-registers -fweb -fvisibility-inlines-hidden"
DISTDIR="/usr/portage/distfiles"
EMERGE_DEFAULT_OPTS="--alphabetical"
FEATURES="autoconfig candy ccache distlocks metadata-transfer multilib-strict parralel-fetch sandbox sfperms strict"
GENTOO_MIRRORS="ftp://gentoo.mirrors.pair.com/"
LANG="en_US"
LINGUAS="en"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
PORTDIR_OVERLAY="/usr/local/portage /usr/local/beyond-dev-sources-1"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 X aac akode alsa apache2 apm aspell avi bash-completion berkdb bitmap-fonts bzip2 cairo cli crypt ctype dba dri eds emboss encode esd expat fastbuild firefox flac flash foomaticdb force-cgi-redirect fortran ftp gd gdbm gif glibc-omitfp gpm graphviz gstreamer gtk gtk2 imlib insecure-savers insecure_drivers insecure_savers ipv6 isdnlog ispell java javascript jpeg kde kdeenablefinal kdehiddenvisibility kdexdeltas libg++ libwww mad memlimit mikmod motif mozilla mp3 mpeg mplayer ncurses nls nptl nptlonly nsplugin ogg opengl oss pam pcre pdflib perl png posix postgres pppd python qt quicktime readline sdl session simplexml soap sockets spell spl sse sse2 ssl symlink tcpd tiff tokenizer truetype truetype-fonts type1-fonts udev vorbis win32codecs xcomposite xine xml xmms xscreensaver xsl xv zlib elibc_glibc input_devices_keyboard input_devices_mouse kernel_linux linguas_en userland_GNU video_cards_ati video_cards_vga video_cards_vesa"
Unset: ASFLAGS, CTARGET, INSTALL_MASK, LC_ALL, LDFLAGS

bigbob73
Guru
Guru

Joined: 31 Dec 2004
Posts: 332
Location: Under the Lone Star

Posted: Fri Apr 07, 2006 6:51 pm Post subject:

seren wrote:

Code:

I have suspected this would be the case. can't wait to do the 2.6. guide

_________________
A computers attention span is only as long as it's electrical cord (Murphy)

cheater1034
Veteran
Veteran

Joined: 09 Sep 2004
Posts: 1558

Posted: Fri Apr 07, 2006 7:57 pm Post subject:

nesl247 wrote:

cheater1034 wrote:

1. I experienced problems with nxsty's latest gcc 2.4 snapshot
2. I was using binutils 2.16.91.0.6 - I know patches and misc. things, I can do things right
3. most packages in portage filter out -Bdirect, therefore it isn't used anyway
4. I do indeed put LD_BIND_DIRECT in 99local
5. Well, the only -mm based kernel out now on fgo is no-sources, and benchmarks say nicksched is more interactive than staircase 15 (my system, and I've had other reports (v14 though) ). And it includes the performance patches from -ck.
So, I don't know what you're talking about with this statement, but w/e.
a) -CK uses an entirely different Scheduler
b) -MM Uses ingosched with some scheduler patches(I like ingo because it is some of nick piggins work)
c) -NO Uses plugsched, so you get staircase if you want it, though benchmarks said nicksched performs better (haven't tried ingosched yet)
6. Portage did that work for you, and it shouldn't be because kdesktop is broken
7. It is filtered out by most packages and rendered ineffective
-----
I know what I am doing (Atleast I hope so :-p)
Sidenotes:
today, 2.6.17-rc1-no1, and, guide 2.6.0. Next week, LiveCD 1.0.0 (Almost done)
_________________
IRC!: #zen-sources on irc.rizon.net
zen-kernel.org
--
Lost in android development land.

Display posts from previous:

	Gentoo Forums Forum Index Unsupported Software	All times are GMT Goto page Previous 1, 2, 3, 4, 5, 6, 7, 8, 9
Page 9 of 9

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

Copyright 2001-2024 Gentoo Foundation, Inc. Designed by Kyle Manna © 2003; Style derived from original subSilver theme. | Hosting by Gossamer Threads Inc. © | Powered by phpBB 2.0.23-gentoo-p11 © 2001, 2002 phpBB Group
Privacy Policy