View previous topic :: View next topic |
Author |
Message |
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Mon Aug 26, 2002 4:26 pm Post subject: new sparc64 stage1 |
|
|
Has anyone tried this? Are all the binaries in 64bit since gcc3.2 fully supports 64 bit on sparc?
Any more info
TIA
xming |
|
Back to top |
|
|
PimpNasty n00b
Joined: 05 Jun 2002 Posts: 36
|
Posted: Tue Aug 27, 2002 3:00 pm Post subject: |
|
|
Whats the link? I could through it on my machine. |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Wed Aug 28, 2002 7:21 pm Post subject: sparc |
|
|
finally got bootstraped, emerge system (CFLAGS=-O2 -pipe", compiled kernel and rebooted.
Kernel is compiled with egcs (64bit) userland with gcc3.2 still 32bit, I will try later to compile everything in 64bit, don't know it everything is 64bit clean.
Mysql is masked very strange, unmaking in the mask did't help, so I just emerge the .ebuild file.
Tomorrow I will have more update
xming |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Mon Sep 02, 2002 11:15 am Post subject: update |
|
|
I have got gentoo 1.4 running on sparc T1 as mail/web/ftp, it runs great. There are some work arounds needed to install Gentoo on a Sparc and certanly on a headless one.
There are a lot of pgk masked, I thought the problem was solved in the portage.
Who is the maintainer of Sparc iso? I would like to share my experience with him so we can have a better ISO
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Mon Sep 02, 2002 6:59 pm Post subject: |
|
|
So this sparc64 stage1 does include GCC 3.2?
I have a Sunblade 100 that's currently unused. My hope is to put something on it, but so far, CD Installs of FreeBSD, NetBSD, Solaris 9 and Gentoo 1.1a-sparc64 failed, something about the clock being inaccurate followed by a system hang. Hopefully, gentoo 1.4-sparc64 will boot/work on it.
// Edit
Also, what is a good way to boot a Sunblade 100 for an install anyways? I'm guessing a netboot, but I am unsure of what kernel I need on another linux machine for the sunblade to boot, as well as any extra bits. I cracked tftp/rarp/bootp once before, in trying to netboot FreeBSD, but that had similar problems as CD installs.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Mon Sep 02, 2002 9:05 pm Post subject: sparc |
|
|
yep 3.2
I find net boot is too much work, I like cds. I just booted from gentoo 1.1 iso, ftp stage-1 untared.
If sol9 is saying something about the clock, maybe you should check the HW
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Mon Sep 02, 2002 10:31 pm Post subject: Re: sparc |
|
|
xming wrote: | yep 3.2
I find net boot is too much work, I like cds. I just booted from gentoo 1.1 iso, ftp stage-1 untared.
If sol9 is saying something about the clock, maybe you should check the HW
xming |
Actually, Sol9 installer just couldn't find X, for some odd reason, and then it dropped me into a /bin/sh shell. The Blade is previously owned, so there is an existing install of Sol8 on the drive., but I'd rather start fresh.
FreeBSD 5.0-DP1 boots the kernel, then when I try to mount the root drive (the CD), it states the clock is off by a certain number of days, and spits a screenful of debug garbage out and hangs.
NetBSD also stated the clock was off, but hung when asking me to choose a terminal.
Last time I tried it w/ the gentoo 1.1a CD, I don't think I ever got to booting the kernel. Something about memory or somesuch. I'll have to try again later and post the exact output (by typing blindly; only one Monitor).
As for the clock, booting into the existing sol8 and getting to the root prompt, a "date" command shows me today's date, only off by a couple of seconds with my other machines, so I've no idea what the BSD Kernels are complaining about.
// Edit
Ahh, when I try to boot Gentoo 1.1a on the Sunblade 100, I get to the SILO prompt, and type "ramdisk"
It then states "Uncompressing...", and then a second or two later, I get a "Fast Data Access MMU Miss", and get dropped back off at the OpenPROM boot prompt.. The only thing I can then do from there is "reset-all" and try again w/ a different OS.
// Edit #2
I have managed to upgrade to OpenBOOT to 4.5.9, using the existing Sol8 on the drive. IF I can get Gentoo on here, I am gonna have to find a way to save a tiny UFS partition for upgrading OpenBOOT. And the rror still occurs (despite a Debian mailing list suggestion stating upgrading openboot would probably fix it.)
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Tue Sep 03, 2002 8:25 am Post subject: |
|
|
Well, I managed to get a Debian 3.0 netboot running (damn netboot is annoying to setup), and currently, it's bootstrapping (Sun Blade 100).
I've used the following make.conf directives:
Code: | CHOST="sparc64-unknown-linux-gnu"
CFLAGS="-mcpu=ultrasparc -mtune=ultrasparc -O3 -pipe -fomit-frame-pointer"
CXXFLAGS="-mcpu=ultrasparc -mtune=ultrasparc -O3 -pipe -fomit-frame-pointer"
MAKEOPTS="-j2" |
Hopefully it'll compile (Knowing my luck......), but I'm curious, how does one get GCC 3.2 to build in 32bit or 64bit? I'll be sure to build the first kernel in egcs64, but I am definately going to be interested in it building under 3.2-64bit, or getting most, if not all the machine in 64bit mode. More advice on this would be nice.
Secondly, while I've noticed the kernel includes code for other archs, I've never seen them in the "make menuconfig" options, so I'm guessing the menuconfig code auto-determines what CPUs to show based upon the known arch?
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Tue Sep 03, 2002 10:59 am Post subject: sparc |
|
|
Code: | Hopefully it'll compile (Knowing my luck......), but I'm curious, how does one get GCC 3.2 to build in 32bit or 64bit? I'll be sure to build the first kernel in egcs64, but I am definately going to be interested in it building under 3.2-64bit, or getting most, if not all the machine in 64bit mode. More advice on this would be nice.
|
default: kernel uses ecgs in 64, userland uses 32 bit gcc3.2. I will try everything in 64bit with gcc3.2 with my new sparc.
Code: | Secondly, while I've noticed the kernel includes code for other archs, I've never seen them in the "make menuconfig" options, so I'm guessing the menuconfig code auto-determines what CPUs to show based upon the known arch?
|
make *config is smart
PS: Let me know if you run into troubles
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Tue Sep 03, 2002 2:59 pm Post subject: |
|
|
Well, it died early on in the "scripts/bootstrap" stage:
Code: | gcc -O3 -pipe -fomit-frame-pointer -DHAVE_UNISTD_H -DUSE_MAP -o example example.o -L. -lz
collect2: ld terminated with signal 11 [Segmentation fault], core dumped
/usr/bin/ld: skipping imcompatible ./libz.a when searching for -lz
/usr/bin/ld: warning: sparc:v9 architecture of input file `example.o' is incompatible with sparc output
make: *** [example] Error 1
make: *** Waiting for unfinished jobs...
!!! ERROR: The ebuild did not complete successfully.
!!! Function src_compile, Line 3, Exitcode 2
!!! (no error message) |
I tried again by unpacking the stage1-sparc64-1.4 tarball again, and removing "-mcpu=ultrasparc -mtune=ultrasparc" from the make.conf file, leaving only "-O3 -pipe -fomit-frame-pointer", but it still failed at that same error.
From what I can tell, it emerges binutils before doing libz, and once it emerges that binutils, it switches over to using the 'ld' from that binutils, and that seems to be choking. My guess it that binutils is expecting zlib to match or something, but I'm not quite sure.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Tue Sep 03, 2002 4:23 pm Post subject: sparc |
|
|
try without -fomit-frame-pointer.
and while we are on binutils, try to see if you can compile anything with the newly compiled binutils (small hello.c), there were problem with it.
Generally speaking a sig 11 means hardware failure, but if it always happens on the same spot, it is probably something else
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Wed Sep 04, 2002 4:01 am Post subject: |
|
|
Well, so far I've tried the following settings:
Code: | Original (2nd, 3rd, && 4th only show changed values):
CHOST="sparc64-unknown-linux-gnu"
CFLAGS="-mcpu=ultrasparc -mtune=ultrasparc -O3 -pipe -fomit-frame-pointer"
CXXFLAGS="-mcpu=ultrasparc -mtune=ultrasparc -O3 -pipe -fomit-frame-pointer"
MAKEOPTS="-j2"
PORTAGE_TMPDIR="/usr/obj"
Second Try:
CFLAGS="-O3 -pipe -fomit-frame-pointer"
CXXFLAGS="-O3 -pipe -fomit-frame-pointer"
Third Try:
CFLAGS="-O3 -pipe"
CXXFLAGS="-O3 -pipe"
Fourth Try:
CFLAGS="-O2 -pipe"
CXXFLAGS="-O2 -pipe" |
None Worked. All dies at the same exact spot, so my guess is it's something in that particular version of Binutils.
I also tried a simple "Hello World" file consisting of the following:
Code: | #include <stdio.h>
int main(void) {
printf("Hello World!\n");
return 0;
} |
Which also failed to link, when running "gcc hello.c -o hello". The same error, seg fault, etc..
I'm not quite sure what's causing it. I doubt hardware, so it's most likely something in the binutils package. I did boot the machine using a debian woody/3.0 netboot kernel, located via a link in the Gentoo Sparc install guide, so I don't know if that has any profound impact on the gentoo tools itself or not.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Wed Sep 04, 2002 7:39 am Post subject: binutils |
|
|
It is the binutils IMHO, try not to bootstrap binutils (edit bootstrap.sh) after everythings works try to make a backup of the ar, ld, as, ... then emerge binutils, hope this helps
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Wed Sep 04, 2002 8:29 am Post subject: Re: binutils |
|
|
xming wrote: | It is the binutils IMHO, try not to bootstrap binutils (edit bootstrap.sh) after everythings works try to make a backup of the ar, ld, as, ... then emerge binutils, hope this helps
xming |
Ran an "emerge rsync" in the vain hope binutils may have been recently updated, and then edited scripts/bootstrap.h to comment/remove references to binutils, and re-ran the boot strap.
It appears to have gotten somewheres within building GCC 3.2 when it died with the following error (I didn't blind-type the actual compiliation block):
Code: | /usr/sparc64-unknown-linux-gnu/bin/ld: cannot open /usr/lib64/crti.o: No such file or directory
collect2: ld returned 1 exit status
make[3]: *** [libgcc_s.so] Error 1
make[3]: *** Waiting for unfinished jobs.... |
And after that failure, make/emerge gracefully tell me of a new error.
I'm guessing the binutils installs this /usr/lib64/crti.o file, and GCC apparently needs it, yet I have no /usr/lib64 directory even created. I dunno how you got it to build, I think the Netra series has alot in common with the Blade series in terms of hardware (and price), but I may be thinking of the Netra X1, not T1. I'm guessing my only choice now is to maybe try a binary install of binutils, or just skip stage 1 and use a stage 2 tarball.
If that even fails.....well, There is always Stage 3, or if truly neccessary, Debian.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Wed Sep 04, 2002 9:25 am Post subject: sparc install guide |
|
|
here is how I did it, please be carefull, I haven't verified it yet
Code: |
gentoo on sparc64 (netra T1)
iso 1.1
ramdisk init=/bin/sh
ifconfig eth0 x.x.x.x netmask y.y.y.y broadcast z.z.z.z
route add default gw a.a.a.a
(net1=eth0, net0=eth1)
/sbin/devfsd /dev -np
/etc/init.d/rcS
fdisk
mkfs
swapon /dev/sda5
mkdir /mnt/cdrom
mount /dev/cdrom /mnt/cdrom
cd /mnt
mkdir gentoo
mount /dev/sda2 /mnt/gentoo
cd /mnt/gentoo
mkdir boot usr var
mount /dev/sda1 boot
mount /dev/sda4 usr
mount /dev/sda6 var
wget http://152.2.210.81/pub/Linux/distributions/gentoo/releases/build/.1.4/sparc64/stage1-sparc64-1.4.tbz2
bzip2 -cd stage1-sparc64-1.4.tbz2 | tar xvf -
mount -o bind /proc /mnt/gentoo/proc
chroot /mnt/gentoo /bin/bash
date -s
tar jxfp stage1-sparc64-1.4.tbz2
source /etc/profile
nano /etc/resolv.conf
emerge sync
emerge rsync
CHOST=sparc-unknown-linux-gnu
#myPORTAGE=`cat ${MYPROFILEDIR}/packages | grep -v '^#' | grep sys-apps/portage
FEATURES="ccache"
(not updating portage, 'cause sandbox isn't working, so commenting the line in bootstrap.sh and set FEATURES)
cd /usr/portage
scripts/bootstrap.sh
export CONFIG_PROTECT=""
emerge system
emerge vanilla-source
cd /usr/src/linux
make menuconfig
make dep;amke clean;make vmlinux;make modules
emerge syslog-ng
rc-update add syslog-ng default
emerge vcron
rc-update add vcron default
emerge reiserfsprogs
cd /etc
ln -sf ../usr/share/zoneinfo/Europe/Brussels localtime
nano -w /etc/hostname
nano -w /etc/hosts
nano -w /etc/conf.d/net
rc-update add net.eth0 default
nano -w /etc/rc.conf
emerge silo
cp /usr/src/linux vmlinux /boot
cd boot
gzip -9 vmlinux
nano -w /etc/silo.conf
mkdir /boot/etc
cp /etc/silo.conf /boot/etc
silo -C /boot/etc/silo.conf
|
|
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Wed Sep 04, 2002 9:28 am Post subject: My param |
|
|
Code: |
USE="+sparc64 +mysql +innodb -opengl -alsa -esd -X -java -gpm -berkdb -qt -gmp"
ARCH="sparc64"
PLATFORM=sparc64-unknown-linux-gnu
CHOST="sparc-unknown-linux-gnu"
CFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
CXXFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
|
|
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Wed Sep 04, 2002 8:02 pm Post subject: |
|
|
Hmm, I'll give the above a try. I never realized that passing /bin/sh as an init shell would've probably worked, but if it fails, I do know that netbooting will indeed work to get me back to a shell.
I'll alter my USE flags and give it one more shot under the debian installer, and then if that fails, see about the 1.1a bootdisk.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Thu Sep 05, 2002 2:57 am Post subject: |
|
|
xming wrote: | Code: |
USE="+sparc64 +mysql +innodb -opengl -alsa -esd -X -java -gpm -berkdb -qt -gmp"
ARCH="sparc64"
PLATFORM=sparc64-unknown-linux-gnu
CHOST="sparc-unknown-linux-gnu"
CFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
CXXFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
|
|
Well, it appears the first error I've posted in here was my own fault. I used the USE flags you posted, and I'm guessing by specifying my ${CHOST} var as "sparc64-unknown-linux-gnu" was cauing it to build in 64 bit mode, and this would probably explain the whining from the binutils. So it appears to have completed most of the boot strapping, including building GCC 3.2 over again, and got somewheres into the glibc compile when it choked and died on a fairly new error, seen below:
Code: | /var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.S: Assembler messages:
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.S:25: Error: Architecture mismatch on "be,pn %icc,.LL3".
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.S:25: (Requires v9|v9a|v9b; requested architecture is sparclite.)
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.S:31: Error: Architecture mismatch on "return".
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.S:31: (Requires v9|v9a|v9b; requested architecture is sparclite.)
make[2]: *** [/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crti.o] Error 1
make[2]: *** Waiting for unfinished jobs
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.S: Assembler Messages:
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.S:9: Error: Architecture mismatch on "return".
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.S:9: (Requires v9|v9a|v9b; requested architecture is sparclite.)
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.S:17: Error: Architecture mismatch on "return".
/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.S:17: (Requires v9|v9a|v9b; requested architecture is sparclite.)
[/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/buildhere/csu/crtn.o] Error 1
make[2]: Leaving directory `/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5/csu'
make[1]: *** [csu/subdir_lib] Error 2
make[1]: Leaving Directory `/var/tmp/portage/glibc-2.2.5-r6/work/glibc-2.2.5'
make: *** [all] Error 2 |
I got this far by changing my ${CHOST} to "sparc-unknown-linux-gnu", and Judging by the error messages, glibc is somehow thinking I want sparclite architecture now, when the ${CFLAGS} clearly state "ultrasparc", so I'm not quite sure how it got confused. I'll try a google search on the errors to see what I can turn up, but any ideas are welcomed.
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Thu Sep 05, 2002 8:40 am Post subject: |
|
|
After trying several other values for CFLAGS/CXXFLAGS, including -mcpu=v9 -mtune=ultrasparc, and adding -Wa,-Av9,-xarch=v9, it still didn't want to work, so begrudgingly, I removed the -mcpu=ultrasparc -mtune=ultrasparc, and so far (as of this posting), glibc is compiling fine.
I wonder what kind of performance may be lost by not optimizing directly for ultrasparc. Probably minimal, but well, whatever compiles. I'll try emerging the system using -mcpu=ultrasparc -mtune=ultrasparc and see how that fares (provided this bootstrap actually finishes).
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Thu Sep 05, 2002 8:50 am Post subject: ok my fault |
|
|
when I bootstraped I used:
Code: | CFLAGS="-O2 -pipe"
CXXFLAGS="-O2 -pipe"
|
after I used
Code: | CFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
CXXFLAGS="-O3 -pipe -mcpu=ultrasparc -mtune=ultrasparc"
|
I didn't think it would matter, but apparently it does,
try this first to get your system up
Code: | CFLAGS="-O2 -pipe"
CXXFLAGS="-O2 -pipe"
|
|
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Thu Sep 05, 2002 8:01 pm Post subject: |
|
|
The bootstrap finished using -O3 -pipe -fomit-frame-pointer. It's emerging the system with the -mcpu=ultrapsarc -mtune=ultrasparc, so I'll see how that goes.
Now once I get this all up and running, what's the feasability of a 64-bit Userland like?, how would one pull that off anyways?
// Edit:
Forgot to ask, is XFS (the Filesystem) supported yet? I know the current sparc install guide says it's not, but that relates to the gentoo 1.1a release, have there been any developments since then?
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Thu Sep 05, 2002 8:41 pm Post subject: userlan 64 bit |
|
|
I would like to have that too, but it seems you've tried and failed I don't know if every thing is 64 bit clean? I am trying to install a second T1, and maybe I will try to figure out the 64bit thinggy.
I don't trust XFS (see my other posts) had a major fs corruption, I am using reiser and ext2 for /boot. Anyway reiser is much faster and personaly I never had any problems with it
cheers
xming |
|
Back to top |
|
|
Kumba Developer
Joined: 16 Jul 2002 Posts: 393 Location: Sigma 957
|
Posted: Thu Sep 05, 2002 8:44 pm Post subject: |
|
|
Hmm, I'm using XFS on an x86 box running Gentoo, and so far, it seems to do well, although the box spends most of it's time just idling, and the Gentoo documents stated ReiserFS wasn't really trustworthy (again, on an x86 Machine). Currently, I'm using ext3 on all partitions on the Blade 100, and aside from the disk being a bit loud (and slow it seems), ext3 works okay. Reading the Unofficial Sun Blade 100 FAQ has pointed out that the disk is itself the problem, replacing it with a "real" 7200rpm drive makes it much faster. (It currently has a 15.3GB Seagate ATA II Baracuda drive).
As for 64-bit, has the kernel been compiled in 64-bit mode using GCC 3.2, or is egcs64 the only viable alternative currently?
--Kumba _________________ "The past tempts us, the present confuses us, the future frightens us. And our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic |
|
Back to top |
|
|
xming Guru
Joined: 02 Jul 2002 Posts: 441
|
Posted: Thu Sep 05, 2002 9:12 pm Post subject: |
|
|
well, reiser is more stable the XFS IMHO, apparently reiser has some design fault which can lead to fs corruption and yes it happens (too little change) but XFS is totally not compatible with some kernel options (especially preempt) and it does currupt fs (bigger change). Ext3 is a safe bet.
I haven't tried kernel with gcc3.2, tomorrow I will, I let you know
xming |
|
Back to top |
|
|
|
|
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
|
|