Stage 1/3 Installation Support - Gentoo 2004.3 & GCC 343

Message

powderific · Post by **powderific** » Mon Mar 21, 2005 7:21 pm

Pentium-M support appears to have been added between stable gcc and gcc3.4, thanks for pointing me in the right direction

kamisama · Post by **kamisama** » Mon Mar 21, 2005 8:19 pm

I finished this install method with complete success a little over a week ago as far as I remember, just had to come back and share the success I "definately" could not have with you.

Kimchi_sg

You will be shocked to hear that not only did I complete this installation method using genkernel, but my system runs great. And I did run into that perl bug stopping emerge -e system from completing. And somehow, even though I don't compile my own kernel, I manage to fix that without any trouble, manually editing the file which had references to previous kernel versions, commenting out 3 lines did the trick. After what you said about genkernel users being inept fixing that perl bug made me feel like a super genius.

Bob P · Post by **Bob P** » Mon Mar 21, 2005 10:27 pm

ruach wrote:'m leaving for pycon tomorrow, but if you thought we might be able to get this working, and you thought that there would be a good performance advantage to using 3.4 and ntpl I would be willing to give it another shot. If you want I'd be willing to give you a logon so you (bob P) can see what is happening.

Let me know what you think

hmmm. there have been lots of reports of problems with the glibc-2.3.4.20050125-r1 ebuild. did you try compiling an older version of glibc as you had thought about doing? to do that, you could just make the glibc line in package.keywords look like this:

Code: Select all

=sys-libs/glibc-2.3.4.20050125

that would force your system to use the version of glibc that is currently the penultimate version of glibc in the portage tree. its the equivalent of adding the current -r1 version to the masking file.

just a thought...

btw, if i were working on that box right now, i'd set the makeopts value to j3.

Timdog · Post by **Timdog** » Tue Mar 22, 2005 1:31 am

I have another problem. This time I did read through the thread and did a search to see if anything came up. I'm sorry for my previous post. :/

I'm at the part where you have to cd /usr/src/linux. When I do this it says that there is no such file or directory. However, if I do ls /usr/src/linux, it says it is there, and if I do mkdir /usr/src/linux, it says that the file already exists. Gentoo seems to be contradicting itself here, and I'm not sure what I should do.

kimchi_sg · Post by **kimchi_sg** » Tue Mar 22, 2005 1:38 am

Timdog wrote:I have another problem. This time I did read through the thread and did a search to see if anything came up. I'm sorry for my previous post. :/

I'm at the part where you have to cd /usr/src/linux. When I do this it says that there is no such file or directory. However, if I do ls /usr/src/linux, it says it is there, and if I do mkdir /usr/src/linux, it says that the file already exists. Gentoo seems to be contradicting itself here, and I'm not sure what I should do.

What is the output of

Code: Select all

ls -l /usr/src/linux

kimchi_sg · Post by **kimchi_sg** » Tue Mar 22, 2005 1:41 am

kamisama wrote:After what you said about genkernel users being inept fixing that perl bug made me feel like a super genius.

Apparently, not enough of a "genius" to know your own hardware. Or simply to write down the output of lspci and lsmod and use these to make a kernel.

So the score is still kimchi_sg 1, kamisama 0

Timdog · Post by **Timdog** » Tue Mar 22, 2005 1:44 am

kimchi_sg wrote:
Timdog wrote:I have another problem. This time I did read through the thread and did a search to see if anything came up. I'm sorry for my previous post. :/

I'm at the part where you have to cd /usr/src/linux. When I do this it says that there is no such file or directory. However, if I do ls /usr/src/linux, it says it is there, and if I do mkdir /usr/src/linux, it says that the file already exists. Gentoo seems to be contradicting itself here, and I'm not sure what I should do.
What is the output of
Code: Select all
ls -l /usr/src/linux

It's lrwxrwxrwx 1 root root 22 Mar 20 22:53 /usr/src/linux -> linux-2.6.10-gentoo-r2

kimchi_sg · Post by **kimchi_sg** » Tue Mar 22, 2005 1:53 am

Timdog wrote:
kimchi_sg wrote:What is the output of
Code: Select all
ls -l /usr/src/linux
It's
Code: Select all
lrwxrwxrwx 1 root root 22 Mar 20 22:53 /usr/src/linux -> linux-2.6.10-gentoo-r2

You don't have an actual /usr/src/linux-2.6.10-gentoo-r2 directory.
Do an

Code: Select all

ls -l /usr/src

and see what is the actual kernel source directory.

One of the lines in the output should look similar to the following (XX and Y are numbers).

Code: Select all

drwxr-xr-x  19 root root 1248 Mar 21 09:00 linux-2.6.XX-gentoo-rY

Then based on the about output, remove the broken /usr/src/linux symlink and create a new one to the actual kernel directory that you have.

Code: Select all

rm /usr/src/linux
ln -s /usr/src/linux-2.6.XX-gentoo-rY /usr/src/linux

Remember, XX and Y are to be replaced with the actual numbers based on the "ls" output.

Timdog · Post by **Timdog** » Tue Mar 22, 2005 1:56 am

Thanks for the help, got it working.

Bob P · Post by **Bob P** » Tue Mar 22, 2005 2:01 am

kimchi_sg wrote:Apparently, not enough of a "genius" to know your own hardware. Or simply to write down the output of lspci and lsmod and use these to make a kernel.

ooh. use genkernel, get spanked.

hielvc · Post by **hielvc** » Tue Mar 22, 2005 2:53 am

As curtis119 would say "Harder Harder":lol:

IcEr3K · Post by **IcEr3K** » Tue Mar 22, 2005 2:54 am

I happen to encounter an error when rebuilding the system toolkit (step 7.2.4).

Code: Select all

Swap cache: add 0, delete 0, find 0/0, race 0+0
Out of Memory: Killed process 2469 (cc1).
make[2]: *** Deleting file `/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/iconvdata/iso-2022-cn-ext.os'
{standard input}: Assembler messages:
{standard input}:84968: Warning: end of file not at end of a line; newline inserted
{standard input}:86615: Error: suffix or operands invalud for `mov'
make[2]: *** [/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/iconvdata/gb18030.os] Terminated
make[2]: *** [/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/iconvdata/iso-2022-cn-ext.os] Terminated
make[1]: *** [iconvdata/others] Terminated
make[1]: Leaving directory `/var/tmp/portage/glibc-2.3.4.20050125-r1/work/glibc-2.3.4'
make: *** [all] Error 2

This is the second time i've tried to compile glibc.
Please let me know if you guys need any additional information.

hielvc · Post by **hielvc** » Tue Mar 22, 2005 3:05 am

Post the output of

Code: Select all

 df -h            <and>
free

Timdog · Post by **Timdog** » Tue Mar 22, 2005 3:51 am

What exactly do the dnsdomainname and nis.domainname settings do? Also, what does the hostname do? I semi understand that the hostname is what the comp will appear as on the network. I have a router with DHCP, and I can ping google.com, I'm just wondering why I have to set the dnsdomainname when I already have set my DNS servers in the resolv.conf file? Should I even worry about setting these?

kimchi_sg · Post by **kimchi_sg** » Tue Mar 22, 2005 5:58 am

Bob P wrote:
kimchi_sg wrote:Apparently, not enough of a "genius" to know your own hardware. Or simply to write down the output of lspci and lsmod and use these to make a kernel.
ooh. use genkernel, get spanked.

Come on... here we just need to write the output of a command on a piece of paper then read it back when doing the kernel config.

kimchi_sg · Post by **kimchi_sg** » Tue Mar 22, 2005 6:04 am

Timdog wrote:What exactly do the dnsdomainname and nis.domainname settings do? Also, what does the hostname do? I semi understand that the hostname is what the comp will appear as on the network. I have a router with DHCP, and I can ping google.com, I'm just wondering why I have to set the dnsdomainname when I already have set my DNS servers in the resolv.conf file? Should I even worry about setting these?

Actually, you can skip the dns and nis domain name setup. Not the hostname though.

But if your ISP does not assign you a domain name via DNS, it will be set to something ugly like "localdomain", which is the reason you should set one. NIS domain name is purely unnecessary if you do not know what NIS is.

curtis119 · Post by **curtis119** » Tue Mar 22, 2005 4:35 pm

hielvc wrote:As curtis119 would say "Harder Harder":lol:

@Bob P, Now Bob we both know that genkernel isn't that bad.

Ok, well it is bad but still....

kamisama · Post by **kamisama** » Tue Mar 22, 2005 7:22 pm

So the score is still kimchi_sg 1, kamisama 0

I didn't know we were keeping score.

And if we are I'm curious as to how I'm the one with the working gentoo system you said I would "Definately" never have.

"Go back to the standard installation handbook"

If thats what gentoo users have to say to people to score points than I'd rather not score any.

Bob P · Post by **Bob P** » Tue Mar 22, 2005 8:34 pm

/* yawn */

torchZ06 · Post by **torchZ06** » Wed Mar 23, 2005 1:08 am

ok, i was going for my 3rd successful 1-on-3 when something went bad... the first emerge of gcc 3.4 (step 7.1) failed. here is the error:

Code: Select all

/var/tmp/portage/gcc-3.4.3.20050110-r1/work/gcc-3.4.3/gcc/libgcc2.c: In function `__ashrdi3':
/var/tmp/portage/gcc-3.4.3.20050110-r1/work/gcc-3.4.3/gcc/libgcc2.c:485: internal compiler error: Illegal instruction
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://bugs.gentoo.org/> for instructions.
Preprocessed source stored into /var/tmp/portage/gcc-3.4.3.20050110-r1/temp/ccflgCEi.out file, please attach this to your
bugreport.
make[3]: *** [libgcc/./_ashrdi3.o] Error 1
make[3]: Leaving directory `/var/tmp/portage/gcc-3.4.3.20050110-r1/work/build/gcc'
make[2]: *** [libgcc.a] Error 2
make[2]: Leaving directory `/var/tmp/portage/gcc-3.4.3.20050110-r1/work/build/gcc'
make[1]: *** [stageprofile_build] Error 2
make[1]: Leaving directory `/var/tmp/portage/gcc-3.4.3.20050110-r1/work/build/gcc'
make: *** [profiledbootstrap] Error 2

!!! ERROR: sys-devel/gcc-3.4.3.20050110-r1 failed.
!!! Function gcc_do_make, Line 1165, Exitcode 2
!!! make failed with profiledbootstrap
!!! If you need support, post the topmost build error, NOT this status message.

the only thing different this time is i'm building this on a P4 but targeting a via 586 class chip (with mmx and 3dnow) and i guess that could be the cause since a P4 does not support 3dnow but rather sse. only reason i didn't immediately suspect that was it compiled linux26-headers, gcc-config, glibc and binutils just fine before it blew up doing gcc itself.

here are more details:

Code: Select all

livecd / # cat /etc/make.conf                                                                                             
# ppk 3/21/05                                                                                                             
# machine: advantech 9575                                                                                                 
# from stage1-on-3 with robmoss's CFLAGS                                                                                  
                                                                                                                          
CHOST="i586-pc-linux-gnu"                                                                                                 
# gcc 3.4.x cflags                                                                                                        
#CFLAGS="-Os -march=c3 -ftracer -pipe"                                                                                    
#CXXFLAGS="${CFLAGS} -fvisibility-inlines-hidden"                                                                         
# gcc 3.3.x                                                                                                               
CFLAGS="-Os -march=c3 -pipe"                                                                                              
CXXFLAGS=${CFLAGS}                                                                                                        
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                                                                                        
GENTOO_MIRRORS="ftp://ftp.ussg.iu.edu/linux/gentoo http://gentoo.osuosl.org http://www.ibiblio.org/pub/Linux/distributions/gentoo"                                                                                                                  
SYNC="rsync://desiderata/gentoo-portage"
RSYNC_RETRIES="3"
RSYNC_TIMEOUT=180
MAKEOPTS="-j2"
PORTAGE_NICENESS=3
AUTOCLEAN="yes"
FEATURES="sandbox ccache distlocks"
CCACHE_SIZE="512M"
RSYNC_EXCLUDEFROM=/etc/portage/rsync_excludes
USE="-gnome -kde 3dnow alsa acpi ithreads mmx nptl pam pthreads python readline ssl tcpd"

Code: Select all

livecd / # emerge info                                                                                                    
Portage 2.0.51-r2 (default-linux/x86/2004.3, gcc-3.3.4, glibc-2.3.4.20050125-r1, 2.6.9-gentoo-r1 i686)                    
=================================================================                                                         
System uname: 2.6.9-gentoo-r1 i686 Intel(R) Pentium(R) 4 CPU 1.80GHz                                                      
Gentoo Base System version 1.4.16                                                                                         
Autoconf: sys-devel/autoconf-2.59-r5                                                                                      
Automake: sys-devel/automake-1.8.5-r1                                                                                     
Binutils: sys-devel/binutils-2.15.92.0.2-r1                                                                               
Headers:  sys-kernel/linux26-headers-2.6.8.1-r2                                                                           
Libtools: sys-devel/libtool-1.5.2-r5                                                                                      
ACCEPT_KEYWORDS="x86"                                                                                                     
AUTOCLEAN="yes"                                                                                                           
CFLAGS="-Os -march=c3 -pipe"                                                                                              
CHOST="i586-pc-linux-gnu"                                                                                                 
COMPILER=""                                                                                                               
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control"                
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"                                                                 
CXXFLAGS="-Os -march=c3 -pipe"                                                                                            
DISTDIR="/usr/portage/distfiles"                                                                                          
FEATURES="autoaddcvs autoconfig ccache distlocks sandbox sfperms"                                                         
GENTOO_MIRRORS="ftp://ftp.ussg.iu.edu/linux/gentoo http://gentoo.osuosl.org http://www.ibiblio.org/pub/Linux/distributions
/gentoo"                                                                                                                  
MAKEOPTS="-j2"                                                                                                            
PKGDIR="/usr/portage/packages"                                                                                            
PORTAGE_TMPDIR="/var/tmp"                                                                                                 
PORTDIR="/usr/portage"                                                                                                    
PORTDIR_OVERLAY="/usr/local/portage"                                                                                      
SYNC="rsync://desiderata/gentoo-portage"                                                                                  
USE="x86 3dnow X acpi alsa apm arts avi berkdb bitmap-fonts crypt cups emboss encode font-server foomaticdb fortran gdbm g
if gpm gtk gtk2 imlib ipv6 ithreads jpeg libg++ libwww mad mikmod mmx motif mp3 mpeg ncurses nls nptl oggvorbis opengl oss
 pam pdflib perl png pthreads python qt quicktime readline sdl spell ssl svga tcpd truetype truetype-fonts type1-fonts xml
2 xmms xv zlib"

i'm going to try doing the install on the target machine itself to see if it makes any difference, but it might take awhile as its a 800MHz via ezra.

btw my other 1-on-3s are working great and fast as hell!

[SOLVED]
OK, i tried doing the install on the actual machine instead of a compile-surrogate P4 and it worked fine.... started it last night and now its successfully completed step 7.1! took a lot longer, but it worked.

Bob P · Post by **Bob P** » Wed Mar 23, 2005 2:24 am

torchZ06 wrote:ok, i was going for my 3rd successful 1-on-3 when something went bad... the first emerge of gcc 3.4 (step 7.1) failed. here is the error:
Code: Select all
internal compiler error:
the only thing different this time is i'm building this on a P4 but targeting a via 586 class chip (with mmx and 3dnow) and i guess that could be the cause since a P4 does not support 3dnow but rather sse.

building on one machine for a totally different architecture isn't covered in the guide, so i'm going to have to give you an answer of "unsupported." actually, this flag is so wildly deviant from the Guide that i should give you a penalty flag on this one.

your flags are wrong if that's what you're trying to accomplish. this question really deserves its own thread.

hielvc · Post by **hielvc** » Wed Mar 23, 2005 2:31 am

Im on a 899 via c3 samuel 2 right and if you figure out make this as snapy as my athlon-xp 2000 Id be real happy

And as to tit for tat " sticks and stones may brake my bones but words will never hurt me cuz I cant read

or ninnnnner ninnner

Bob P · Post by **Bob P** » Wed Mar 23, 2005 2:36 am

hielvc wrote:Im on a 899 via c3 samuel 2 right and if you figure out make this as snapy as my athlon-xp 2000 Id be real happy

And as to tit for tat " sticks and stones may brake my bones but words will never hurt me cuz I cant read or ninnnnner ninnner

"HIEL, BACK AWAY SLOWLY FROM THE KEYBOARD!"

IcEr3K · Post by **IcEr3K** » Wed Mar 23, 2005 3:19 am

I don't know if the entry to the one below my previous one is in reply to mine or not, but df -h gives me:

Code: Select all

df: cannot read table of mounted filesystems: No such file or directory

as for free, it tells me that my total memory is 126016, with 51032 used, and 74984 free. 0 is shared, 17744 is buffers, and 19248 is cached.
the -/+ buffers/cache has 14040 used and 111976 free.
0 for all on swap.

so i got through the initial emerging of glibc by removing some cflags in my make.conf.

but, it happens again, and i get the exact same error message about being out of memory or something.
note that this is a pentium 3 laptop with 128MB of ram. i know, its not much, but i didn't think i would need that much.
this will be the third time that i will emerge -e system, and this time i will get the exact error message.[/code]

kimchi_sg · Post by **kimchi_sg** » Wed Mar 23, 2005 4:04 am

IcEr3K wrote:df -h gives me:
Code: Select all
df: cannot read table of mounted filesystems: No such file or directory

@Bob P:

Please, please include the cp /proc/mounts /etc/mtab step from the standard Installation Handbook into your tutorial. It is the cause of the silly error message quoted above.

Code: Select all

# mv /etc/mtab /etc/mtab.bak
# df
df: cannot read table of mounted filesystems: No such file or directory
# mv /etc/mtab.bak /etc/mtab
# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/hda2             39149200   3819160  35330040  10% /
udev                    452696       532    452164   1% /dev
none                    452696         0    452696   0% /dev/shm

In summary, the arguments FOR and AGAINST including this step are:

Arguments FOR including "cp /proc/mtab /etc/fstab":

It is in the original Gentoo Installation Handbook.
It prevents possible boot errors on some SATA, SCSI and maybe even IDE drives.
It prevents annoying "Can't create lock file /etc/mtab~623" error messages from being shown when booting on ALL systems.
It prevents df from bombing out with the error message quoted above.

Arguments AGAINST including said command:

We want to be as different from the original Gentoo Installation Handbook as possible. Hopefully Sven Vermuellen will notice us and incorporate this into the Alternative Installation HOWTO.
We want all those silly people running SATA drives not to be able to boot when following this tutorial. The tutorial is written based on a Pentium classic with IDE drives after all. Who told these chaps to burn themselves trying the latest and greatest unsupported hardware?
We love scaring n00bs with scary-looking "Can't create lock file" messages. This is an advanced installation tutorial after all, so those n00bs better keep their hands off.
We want people to stop checking the free space on their mounted filesystems using df. In order for linux to look like a typical powerful OS straight from the '70s, all users must use a calculator to sum up the total size of each and every file, then deduct this from the size of the filesystem to get the free space on each one.

Have I driven my point across?