You can subscribe to this list here.
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(17) |
Dec
(7) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
(5) |
Feb
(2) |
Mar
|
Apr
|
May
(4) |
Jun
(23) |
Jul
(3) |
Aug
(5) |
Sep
(1) |
Oct
|
Nov
(1) |
Dec
|
| 2004 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(19) |
Jul
|
Aug
(14) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2005 |
Jan
|
Feb
(5) |
Mar
(3) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2007 |
Jan
(11) |
Feb
(4) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From:
<sg....@ay...> - 2007-02-12 07:34:45
|
Hi! > "console=3Dtty0 mem=3D64M root=3D/dev/ram0 rw init=3D/linuxrc = ip=3Ddhcp". > he tosses out my "root=3D" and my "init=3D" and puts in > root=3D/dev/nfs nfsroot=3D/pfad/zum/rootfs > ltsp one. If I use his initrd with my kernel the /linuxrc does not = seem > to start (even though its in the ramdisk) when I pass > root=3D/dev/ram0 rw init=3D/linuxrc. > Question@all: Are there known problems/limitation with the ramdisks on > the Netstation? Is this a possible reason why Garcia uses cramfs > ramdisks? Has someone made similar experience with cmd-line = parameters? Well, lots of cmdlines and different results and I'm going to add yet a = couple of them more... I used to have this cmdline when I used to have the system boot out of a = root NFS filesystem, without using a ramdisk: console=3DttyS0,9600 console=3Dtty0 root=3D/dev/nfs nfsroot=3D/linuxppc = ip=3Ddhcp But now that I have gone to a ramdisk based boot with a smbfs filesystem = all I have is this: console=3DttyS0,9600 console=3Dtty0 ip=3Ddhcp As you can see I'm not specifying any root or init, the kernel finds it = all by itself, I suppose that may be caused because the ramdisk is being = included on the kernel itself when we are compiling it using = zImage.initrd as the target for the kernel build. As for the kernel messages I'm seeing they should be something like this = (I'm trying not to change anything as I type them): RAMDISK: cramfs filesystem found at block 0 RAMDISK: Loading 2464 blocks [1 disk] into ram disk... done. Freeing initrd memory: 2464k freed VFS: Mounted root (cramfs filesystem). Freeing unused kernel memory 248k init 36k pmac 8k chrp 4k openfirmware striofb_op_set_var con=3D6 1024x768-1024x768-15bpp mode=3D12690 176 16 = 28 1 96 3 -- act=3D2 ... As far as I know there should be no problems with ext2 filesystem, I = just chose cramfs because it was easier for me to build that ramdisk = than others, as you don't need to mount it for creation and a normal = user can do it, no need for root, and at that time it was the standard = ramdisk on Debian, which is what I'm using anyway. Regards... -- Santiago Garc=EDa Manti=F1=E1n Servicio de Inform=E1tica (Redes) AYUNTAMIENTO DE LA CORU=D1A Tlf: 981184200/22 Ext 10997 |
|
From: Stefan U. H. <st...@he...> - 2007-02-11 18:45:40
|
Hi Folks,
Michael (a German buddy of mine) and myself are busy updating lots of
code for the ltsp stuff ...
However, we have come accross one problem when it comes to the initrd
and the cmd-line.
My cmd-line (on a ethernet 8362-a53) for the initrd is
"console=3Dtty0 mem=3D64M root=3D/dev/ram0 rw init=3D/linuxrc ip=3Ddhcp".
So I can specify the location of the nfs-root at boot-time. That works
fine (but only for me an only with the "older" ltsp-4.1 initrd).
Michael (who has a tokenring version of the 8362) needs to specify
nfs-root in the kernel, otherwise the Netstation won't boot properly, so
he tosses out my "root=3D" and my "init=3D" and puts in
root=3D/dev/nfs nfsroot=3D/pfad/zum/rootfs
otherwise his system hangs.
In the process of our efforts, he has rebuild an initrd with recent
tools and libraries. And that works for him the same way, as the old
ltsp one. If I use his initrd with my kernel the /linuxrc does not seem
to start (even though its in the ramdisk) when I pass
root=3D/dev/ram0 rw init=3D/linuxrc.
All I get then at boot-time is the following:
[...]
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 864k freed
Mounted devfs on /dev
Freeing unused kernel memory:
172k init 36k pmak 4k chrp 4k openfirmware=20
=20
... and then my system hangs.
Despite that, I've had frequently problems with modified initrds, even
if they looked fine (with e2fsck) and all was in, it wouldn't work.
Then I did go back to an older version, did the same stuff again (often
just change of one line in a script) and then it worked.
Question@all: Are there known problems/limitation with the ramdisks on
the Netstation? Is this a possible reason why Garcia uses cramfs
ramdisks? Has someone made similar experience with cmd-line parameters?
Maybe someone has here some insight in the far back of his mind that
might save us a couple dozen hours of work.
Thanks a lot!
Stefan.
--=20
/ Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \
| st...@he..., my Cyberhome http://www.hegner-web.de|
\ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 /
|
|
From: Stefan U. H. <st...@he...> - 2007-02-11 08:18:57
|
Hi Folks, have been quiet (and busy) for a while. In the meantime I got my 8362 running on LTSP 4.1 with remote sound - and that was (looking back) not so difficult at all. Here's a step-by-step howto, in case someone wants to try it out: http://wiki.ltsp.org/twiki/bin/view/Ltsp/NetworkStation1000 Shuouldn't take more than an hour or two ;-) Any feedback, corrections, additions, ... warmly welcome! Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Stefan U. H. <st...@he...> - 2007-02-03 10:36:38
|
Hi Garcia, thank you for you assistance! Am Montag, den 29.01.2007, 12:24 +0100 schrieb Garc=EDa Manti=F1=E1n, Santiago: > place your ramdisk with the name "ramdisk.image.gz" at > arch/ppc/boot/images/ and then you must compile the kernel with "make > zImage.initrd". That way the final kernel.1000 will carry your image > inside without needing to do any other thing. What do you put then on your kernel-cmd-line? I've got something like this: CONFIG_CMDLINE=3D"console=3Dtty0 mem=3D64M root=3D/dev/ram0 rw init=3D/linu= xrc ip=3Ddhcp" Did I miss out something? Thanks Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From:
<ma...@ay...> - 2007-01-29 11:24:45
|
>> In case it might help, our kernel is 3321323 bytes in length and it >> carries inside an initrd of 2523136 bytes. > My current problem is packing kernel and initrd that it works on ns1000= . > mkelf-linux doesn't help here (see also other post). Well, I don't think that will ever work, I don't know if you have to use that mkelf-linux for any reason, if not... then you can do as we do here. We are using a cramfs ramdisk, but you can use whatever fits you better, what you do is place your ramdisk with the name "ramdisk.image.gz" at arch/ppc/boot/images/ and then you must compile the kernel with "make zImage.initrd". That way the final kernel.1000 will carry your image inside without needing to do any other thing. Even though the file name i= s .gz it doesn't have to be compressed with gzip if you don't want, for example, the cramfs we use is already a compressed file so compressing it with gzip wouldn't help, so we don't compress it. Hope that helps you with your final problem. Regards... --=20 Santiago Garc=EDa Manti=F1=E1n Servicio de Inform=E1tica (Redes) AYUNTAMIENTO DE LA CORU=D1A Tlf: 981184200/22 Ext 10997 |
|
From: <th...@bo...> - 2007-01-28 14:42:25
|
Hi Stefan, unfortunately I'm not familiar with LTSP. So my answer is not more than a guess .. > Now I'm still stuck with the way ltsp normally boots using a /linuxrc > script that resides (with optional modules I don't need for a > monolithic > kernel) on an initrd. - See my posts yesterday. > By the way, what is an initrd good for, if I don't need it to load > modules right after boot? The use of an initrd is more to cover a range of hardware like networkadapters on a specific platform with one standard kernel and only the modules needed for that particular machine. > Can't I make things way easier, if I take the LTSP /linuxrc and > put it into my nfsroot? - Looking at the hardware provided in the > NS1000, a monolithic kernel should suffice always, even if all > possible > hardware support is compiled in. That's what I did, I do not use an initrd at all, everything resides in a NFS root and the kernel is monolithic. You need to find out what scripts are run in the initrd and include these scripts in your NFS based root. But keep in mind, that all executable files need to be compiled for the ppc platform. This might be the case in the LTSP project but I don't know. > Is it enough, to copy the /linuxrc to the nfsroot an pass the > "nfsroot=/some/place init=/linuxrc" parameters to the cmdline? - That > would be easy. > > -> Some hints if this is feasible and how to do it would be greatly > appreciated. > Please note, that you can not easily pass command line options to the kernel, as the BIOS of the Networkstation boots the kernel directly off NFS or TFTP. Therefore you can add commandline options to the kernel, when compiling: General Setup --> Initial kernel command string: ( I have: console=ttyS0,9600 console=tty0 root=/dev/nfs nfsroot=/ share/projects/8362/rootfs ip=dhcp ) You need to have: Networking Options --> TCP/IP neworking --> IP: Kernel level autoconfiguration --> IP: DHCP support File Systems --> Network Filesystems --> NFS Filesystem support --> Root filesystem on NFS That should boot the kernel, get an IP address from the DHCP server and use the "next server" option in DHCP to choose the correct server for the NFS-root, mount it and run whatever is in there. Regards, Thomas |
|
From: Stefan U. H. <st...@he...> - 2007-01-27 22:03:50
|
Hi Thomas, thanks for your really detailled howto. - That's almost exactly what I did, and if I had had this two weeks earlier, it had saved me ~10 hours of figuring things out myself ;-( - Nevertheless, I believe I learned quite a bit, doing it the "hard" way. ;-) Am Samstag, den 27.01.2007, 18:15 +0100 schrieb Thomas Boeber: > The Kernel image is now under /share/projects/8362/kernel/kernel.1000 > and you should be able to boot it via NFS or TFTP. Now I'm still stuck with the way ltsp normally boots using a /linuxrc script that resides (with optional modules I don't need for a monolithic kernel) on an initrd. - See my posts yesterday. By the way, what is an initrd good for, if I don't need it to load modules right after boot? Can't I make things way easier, if I take the LTSP /linuxrc and put it into my nfsroot? - Looking at the hardware provided in the NS1000, a monolithic kernel should suffice always, even if all possible hardware support is compiled in. Is it enough, to copy the /linuxrc to the nfsroot an pass the "nfsroot=3D/some/place init=3D/linuxrc" parameters to the cmdline? - That would be easy. -> Some hints if this is feasible and how to do it would be greatly appreciated. Thanks Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Thomas B. <th...@bo...> - 2007-01-27 17:15:23
|
Hello, as it seems there are still a few working with the Networkstation! I was looking for some hardware to play with a different architecture and came across the Networkstation. With it's ppc architecture you need to create a cross compiler toolchain in order to get software compiled on a standard i386 desktop. Due to the lack of time I had to stop but maybe the following is of use to the one or other trying to do the same.. 1. Building the toolchain(s) The development environment has been set up on the development machine in /share/projects/8362. The following directory layout was used: directory layout /share/projects/8362 |-docu (documentation for this project) |-crosstool-0.43 (unpacked version of crosstool-0.43) |-kernel (kernel sources and binaries) |-powerpc-603-linux-gnu (the resulting toolchain) |-rootfs (NFS Root with all binaries) |-source (all source packages) |-sysapps (place to compile all apps for target) We need to create a working toolchain for the ppc target in order to compile software on a fast i686 host for the target. Crosstool is a collection of scripts to create such a toolchain. The toolchain will be created using crosstool-0.43 from http://kegel.com/crosstool. Actually we need 2 toolchains. As the NetworkStation 1000 relies on the 2.4.xx Kernel we need a gcc 2.95.3 to successfully compile the kernel for it. Tests with 3.3.5 and 3.4.2 failed to produce a bootable kernel image. All other binaries for the NFS root are compiled using a recent toolchain with gcc 3.4.2 and glibc 2.3.3. Switching the compiler to be used is as simple as changing the search path... The two toolchains have been compiled using a script networkstation.sh which is based on demo-powerpc603.sh in the crosstool archive. You should be able to use the scripts as-is beside from changing the TARBALLS_DIR and RESULT_DIR in demo-ppc603.sh according to your environment. cp demo-powerpc-603.sh networkstation.sh change the path for TABALLS_DIR and RESULT_TOP and uncomment the needed toolchains in networkstation.sh ---------------------------- #!/bin/sh # This script has one line for each known working toolchain # for this architecture. Uncomment the one you want. # Generated by generate-demo.pl from buildlogs/all.dats.txt set -ex TARBALLS_DIR=/share/projects/8362/source RESULT_TOP=/share/projects/8362 export TARBALLS_DIR RESULT_TOP GCC_LANGUAGES="c,c++" export GCC_LANGUAGES # Really, you should do the mkdir before running this, # and chown /opt/crosstool to yourself so you don't need to run as root. mkdir -p $RESULT_TOP #eval `cat powerpc-603.dat gcc-2.95.3-glibc-2.1.3.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-2.95.3-glibc-2.2.2.dat` sh all.sh --notest eval `cat powerpc-603.dat gcc-2.95.3-glibc-2.2.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.2.3-glibc-2.2.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.2.3-glibc-2.3.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.2.3-glibc-2.3.2-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.2.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.2.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.2-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.5-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.6.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.3.6-glibc-2.3.6-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.2.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.2.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.2-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.5-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.6.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-3.4.5-glibc-2.3.6-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.2.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.3.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.3.2-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.3.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.3.6.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.0.2-glibc-2.3.6-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.2.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.2.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.2-tls.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.5.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.5-tls.dat` sh all.sh --notest eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.6.dat` sh all.sh --notest #eval `cat powerpc-603.dat gcc-4.1.0-glibc-2.3.6-tls.dat` sh all.sh --notest echo Done. ---------------------------- The source packages are needed in /share/projects/8362/source. If the source packages are not present, crosstool tries to download them. building the 2 toolchains: sh networkstation.sh 2>&1 | tee networkstation.log Depending on the host system this can take a few hours... 2 Compiling the Kernel To compile the Kernel for the NetworkStation 1000 we first have to patch the Kernel source. Download the patch for the 2.4.29 Kernel: http://prdownloads.sourceforge.net/networkstation/patch-2.4.29-ns1000-20050209.gz?download The "export PATH=/share/projects/8362/powerpc-603-linux-gnu/gcc-2.95.3-glibc-2.2.5/bin:$PATH" sets the toolchain with the gcc 2.95.3 first in the search path. The ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- switches force the use of the ppc Architecture and the use of the cross compile toolchain. Kernel-2.4.29 cd /share/projects/8362/kernel tar xvjf ../source/linux-2.4.29.tar.bz2 export PATH=/share/projects/8362/powerpc-603-linux-gnu/gcc-2.95.3-glibc-2.2.5/bin:$PATH patch Kernel in order to compile a kernel for NS1000 (8362) patch -p1 < ../../source/patch-2.4.29-ns1000-20050209 make ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- mrproper cp ../../source/kernel-config-2.4.29 .config make ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- oldconfig make ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- menuconfig make ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- dep make ARCH=ppc CROSS_COMPILE=powerpc-603-linux-gnu- zImage cp .config ../../source/kernel-config-2.4.29 cp arch/ppc/boot/images/kernel.1000 .. The Kernel image is now under /share/projects/8362/kernel/kernel.1000 and you should be able to boot it via NFS or TFTP. This way you can at least get the kernel compiled, in order to compile a small Linux system you need to cross compile a few packages and are likely to end up in some trouble. Regards, Thomas |
|
From: Stefan U. H. <st...@he...> - 2007-01-26 15:38:39
|
Hi Garc=EDa, Am Dienstag, den 23.01.2007, 13:35 +0100 schrieb Garc=EDa Manti=F1=E1n, Santiago: > Well, all I can say is that we still use the old patch-2.4.29-ns1000 > released on 20050209 and we use it with 2.4.34 right now without any > problem (other than the smbfs unix extensions are broken on 2.4.34). Great hint. - Yes I got 2.4.34 to work (see my recent post on the list). > We are > compiling with gcc 2.95 as the code we have on the NetworkStation patch > doesn't work with more recent compilers. Used that hint for cross-compiling with crosstool form kegel.com and with 2.95 it works! > In case it might help, our kernel is 3321323 bytes in length and it > carries inside an initrd of 2523136 bytes. My current problem is packing kernel and initrd that it works on ns1000. mkelf-linux doesn't help here (see also other post). Thanks a lot Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Stefan U. H. <st...@he...> - 2007-01-26 15:34:58
|
Hi Jochen, Am Mittwoch, den 24.01.2007, 13:33 -0800 schrieb Jochen Roth: > Still out here, just really really busy. me too ;-) > >> maybe it's the wrong compiler version? > > no but a broken cross-compile-chain. I got 2.4.34 compiled now using gcc-2.95.3 and glibc-2.2.5. kernel.1000 also boots. Great! > The ppc platform alternate mnemonics at some point changed in binutils,=20 > and the kernel folks adopted these changes when they started=20 > recommending gcc 3 for compilation. This is from memory, and I could be=20 > wrong on the gcc version. What does this mean to me? Is there a "practial" difference in the gcc version used for building the kernel? Does it limit me in any way using the compiled kernel? > Here are some kernels that all worked for me in the past: [...] > -rw-rw-r-- 1 jochen users 3613204 Dec 10 2002 k_initrd That sounds to me like a kernel with an initrd. I got now a kernel that should be ltsp-compatible together with a ltsp-initrd. However, mkelf-linux does not accept the kernel.1000 for creating the tagged kernel image and complains about the format: > mkelf-linux kernel.1000 initrd-ppc-monolithic-ltsp-1.gz >vmlinuz-2.4.34-l= tsp-1.1000 kernel.1000: not a Linux kernel image (I did try this on i686 though - is this a problem?) What do I do now, to build the boot-image with initrd? > It may be that the ELF header generated by the crosstool-based > linker is not compatible with the ns1000 firmware loader. The fw-loader seems to be picky. That's probably the reason for the special "kernel.1000" format and also the cause, why so few people managed to get a Netstation 1000 up under linux so far (before I came accross this project I found none. - On ltsp.org it is documented as not working). > Also, did the ncdcs tool complain about anything? What is this? Never heard of and didn't use it. Did a Google search and found some posts by you that make me believe, that's the tool I need for building the image with the initrd. Can you point me to a ncdcs download-location? Thanks Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Jochen R. <jo...@pa...> - 2007-01-24 21:30:44
|
Hi Stefan, Still out here, just really really busy. The 2.4.21 kernel with patch never compiled properly with the older gcc tool chain from the tarball on sourceforge. Stefan Ulrich Hegner wrote: > powerpc-linux-gcc -D__ASSEMBLY__ -D__KERNEL__ > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/include > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/arch/ppc -c -o > cpu_setup_6xx.o cpu_setup_6xx.S > cpu_setup_6xx.S: Assembler messages: > cpu_setup_6xx.S:405: Error: Unrecognized opcode: `mftbl' > cpu_setup_6xx.S:406: Error: Unrecognized opcode: `mftbl' > make[1]: *** [cpu_setup_6xx.o] Fehler 1 In cpu_setup_6xx.S, change 'mftbl' to 'mftb' and at least this file should compile. Stefan Ulrich Hegner wrote: > Am Sonntag, den 21.01.2007, 13:18 +0100 schrieb Armin Schindler: >> maybe it's the wrong compiler version? > no but a broken cross-compile-chain. I think that's a wild guess at best, right? I know that these things can be extremely frustrating but I find it important to always look at things objectively. Like Armin said, wrong compiler version. That does not make it a broken tool. The ppc platform alternate mnemonics at some point changed in binutils, and the kernel folks adopted these changes when they started recommending gcc 3 for compilation. This is from memory, and I could be wrong on the gcc version. Stefan Ulrich Hegner wrote: > The Problem is now, that the resulting kernel (taking the 2.4.29 > vanilla sources + patch from the project and the config from > coruna as "oldconfig") is too big. > > "Address out fo range. File is bad or too large for installed memory." > > The new kernel has 847.521 bytes whereas the old one had only > 786.920. Here are some kernels that all worked for me in the past: [jochen@jt ns1000]$ l kernels/ibm/kernel.1000 -rw-rw-r-- 1 jochen users 4275584 Jul 20 2000 kernels/ibm/kernel.1000 [jochen@jt ns1000]$ l kernels/fb total 7032 -rw-rw-r-- 1 jochen users 1180180 Dec 11 2002 k_hda2 -rw-rw-r-- 1 jochen users 1180180 Dec 18 2002 k_hda3 -rw-rw-r-- 1 jochen users 1180180 Dec 18 2002 k_hda4 -rw-rw-r-- 1 jochen users 3613204 Dec 10 2002 k_initrd It may be that the ELF header generated by the crosstool-based linker is not compatible with the ns1000 firmware loader. Also, did the ncdcs tool complain about anything? Regards, Jochen |
|
From:
<ma...@ay...> - 2007-01-23 12:36:31
|
Sorry for the late reply: > The Problem is now, that the resulting kernel (taking the 2.4.29 vanill= a > sources + patch from the project and the config from coruna as > "oldconfig") is too big. Well, all I can say is that we still use the old patch-2.4.29-ns1000 released on 20050209 and we use it with 2.4.34 right now without any problem (other than the smbfs unix extensions are broken on 2.4.34). We are compiling it on a Debian Sarge PowerPC based machine, not cross compiling, so I don't know if cross compiling should work or not. We are compiling with gcc 2.95 as the code we have on the NetworkStation patch doesn't work with more recent compilers. > The new kernel has 847.521 bytes whereas the old one had only 786.920. = I > now tossed out "advanced partition types" and hope this will help. - > Otherwise I need to go to use modules. In case it might help, our kernel is 3321323 bytes in length and it carries inside an initrd of 2523136 bytes. If you are not able to build a kernel and all you want me to build for yo= u a standard 2.4.34 kernel send me the config file and I'll compile it for you. Regards... --=20 Santiago Garc=EDa Manti=F1=E1n Servicio de Inform=E1tica (Redes) AYUNTAMIENTO DE LA CORU=D1A Tlf: 981184200/22 Ext 10997 |
|
From: Stefan U. H. <st...@he...> - 2007-01-22 21:06:12
|
Hi, Am Sonntag, den 21.01.2007, 13:18 +0100 schrieb Armin Schindler: > maybe it's the wrong compiler version? no but a broken cross-compile-chain. > I didn't try it yet (I still have my Networkstation unused on my desk=20 > because of lack of time), but have you tried other cross tools? I now tried with crosstool. The Problem is now, that the resulting kernel (taking the 2.4.29 vanilla sources + patch from the project and the config from coruna as "oldconfig") is too big. "Address out fo range. File is bad or too large for installed memory." The new kernel has 847.521 bytes whereas the old one had only 786.920. I now tossed out "advanced partition types" and hope this will help. - Otherwise I need to go to use modules. Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Armin S. <ar...@me...> - 2007-01-21 12:56:21
|
Hi, maybe it's the wrong compiler version? I didn't try it yet (I still have my Networkstation unused on my desk because of lack of time), but have you tried other cross tools? I would propose ELinOS. If you don't want to buy the whole package, there is a free version available for download at www.sysgo.com which includes full cross-tools and some pre-compiled target binaries. Armin On Sat, 20 Jan 2007, Stefan Ulrich Hegner wrote: > Hi, > > is there still anybody out there? > > I'm trying to cross-compile a kernel for my Networkstation using the ppc > compiler tarball from the project. > > The idea is to build a kernel suitable for ltsp (www.ltsp.org). > > However, no matter if I try the patched 2.4.21 or the patched 2.4.29 > kernel, I get more or less the same error condition: > > 2.4.21: > powerpc-linux-gcc -D__ASSEMBLY__ -D__KERNEL__ > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/include > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/arch/ppc -c -o > cpu_setup_6xx.o cpu_setup_6xx.S > cpu_setup_6xx.S: Assembler messages: > cpu_setup_6xx.S:405: Error: Unrecognized opcode: `mftbl' > cpu_setup_6xx.S:406: Error: Unrecognized opcode: `mftbl' > make[1]: *** [cpu_setup_6xx.o] Fehler 1 > make[1]: Leaving directory > `/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/arch/ppc/kernel' > make: *** [_dir_arch/ppc/kernel] Error 2 > > 2.4.29: > powerpc-linux-gcc -D__ASSEMBLY__ -D__KERNEL__ > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/include > -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/arch/ppc -c -o > cpu_setup_6xx.o cpu_setup_6xx.S > cpu_setup_6xx.S: Assembler messages: > cpu_setup_6xx.S:444: Error: Unrecognized opcode: `mftbl' > cpu_setup_6xx.S:445: Error: Unrecognized opcode: `mftbl' > make[1]: *** [cpu_setup_6xx.o] Fehler 1 > make[1]: Leaving directory > `/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/arch/ppc/kernel' > make: *** [_dir_arch/ppc/kernel] Error 2 > > Can anyone help me out on this one? > > Your assistance is appreciated! > > Thanks > > Stefan. > > > > -- > / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ > | st...@he..., my Cyberhome http://www.hegner-web.de| > \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / > > > |
|
From: Stefan U. H. <st...@he...> - 2007-01-20 20:07:23
|
Hi, is there still anybody out there? I'm trying to cross-compile a kernel for my Networkstation using the ppc compiler tarball from the project. The idea is to build a kernel suitable for ltsp (www.ltsp.org). However, no matter if I try the patched 2.4.21 or the patched 2.4.29 kernel, I get more or less the same error condition: 2.4.21: powerpc-linux-gcc -D__ASSEMBLY__ -D__KERNEL__ -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/include -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/arch/ppc -c -o cpu_setup_6xx.o cpu_setup_6xx.S cpu_setup_6xx.S: Assembler messages: cpu_setup_6xx.S:405: Error: Unrecognized opcode: `mftbl' cpu_setup_6xx.S:406: Error: Unrecognized opcode: `mftbl' make[1]: *** [cpu_setup_6xx.o] Fehler 1 make[1]: Leaving directory `/home/hegi/netstation/ns1000/usr/src/linux-2.4.21/arch/ppc/kernel' make: *** [_dir_arch/ppc/kernel] Error 2 2.4.29: powerpc-linux-gcc -D__ASSEMBLY__ -D__KERNEL__ -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/include -I/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/arch/ppc -c -o cpu_setup_6xx.o cpu_setup_6xx.S cpu_setup_6xx.S: Assembler messages: cpu_setup_6xx.S:444: Error: Unrecognized opcode: `mftbl' cpu_setup_6xx.S:445: Error: Unrecognized opcode: `mftbl' make[1]: *** [cpu_setup_6xx.o] Fehler 1 make[1]: Leaving directory `/home/hegi/netstation/ns1000/usr/src/linux-2.4.29/arch/ppc/kernel' make: *** [_dir_arch/ppc/kernel] Error 2 Can anyone help me out on this one? Your assistance is appreciated! Thanks Stefan. --=20 / Stefan Ulrich Hegner aka "Hegi" - Loehne/ Westfalen - Germany \ | st...@he..., my Cyberhome http://www.hegner-web.de| \ GPG-Fprint D9DB 51BD 2DA6 9B3A 41CB 0287 05A1 8D11 38BA CE91 / |
|
From: Brian T. <tea...@ya...> - 2005-03-20 18:02:22
|
I want to ask the help of anyone who has worked, or is
working, on the
low-level kernel stuff for the NS1000, or powerpc
stuff in general.
For the last month or so, I've been trying to port the
NS1000 changes
into the 2.6 kernel. I ported everything back to 4.15
so I could switch
over to the 2.5 branch, and from there I applied all
the vanilla kernel
patches up to 2.6.4, fixing compile errors and other
stuff along the
way. The NS will at least recognize the images as
valid, but then the
kernel itself fails in various ways.
The last version I got to work properly was 2.5.10.
Somewhere between
there and 2.5.30, the boot sequence stopped working.
I've put in some
serial debugging messages similar to the what the
networkstation patch
does to the boot code, and so I've got it narrowed
down to where the
problem occurs.
In 2.5.30, the bootloader runs, accepts the command
line, unzips the
kernel image to 0x00000000, and passes control to
_start in
arch/ppc/kernel/head.S. It runs through mmu_off,
flush_tlbs, etc. . .,
up to turn_on_mmu. At this point, it sets the bits to
enable memory
mapping, and jumps to start_here via an rfi
instruction. That is where
it fails.
I have verified that the kernel is running at physical
location
0x00000000, not 0xc0000000. I have verfied that the
BAT registers are
0x00000002 (lower) and 0xc0001ffe (upper), which means
the mmu should
map calls in the 0xc0000000 range down to 0x00000000.
I checked the
value of the msr register just before the rfi, and it
does in fact have
MSR_DR and MSR_IR turned on. I've also tried this:
mfmsr r0
ori r0,r0,MSR_DR|MSR_IR
mtmsr r0
bl start_here
but that seems to die immediately after the mtmsr.
If I don't change the msr, and just do a plain old bl
to start_here, it
does get there, although it dies shortly after that
since by then it
needs the mmu.
I've tried all kinds of other stuff, and I just don't
get it. Comparing
the disassembled kernels using objdump -d, I can see
that these sections
of code are byte-for-byte identical (except for a
slightly different
address of start_here) to 2.4.x and 2.5.10, both of
which work. I don't
see what's going on here; it seems like there must be
something going on
in the processor that my documentation doesn't show.
I don't have a
JTAG debugger and I haven't had any luck getting the
kgdb kernel
debugger to work.
I'm just doing this for fun, so there's no urgent need
here. I'm going
to keep plugging away at this, and if I ever actually
get a 2.6 kernel
going, I'll put a patch up here. Until then, if
anyone has any advise,
I'd appreciate it.
__________________________________
Do you Yahoo!?
Yahoo! Small Business - Try our new resources site!
http://smallbusiness.yahoo.com/resources/
|
|
From: Thomas B. <th...@bo...> - 2005-03-09 19:13:13
|
Hello, the new patch for Linux 2.4.29 is working just fine. Did some tests and it worked out very well.. Good work! Is there a usable patch for swapping over NFS for this version of the kernel? There is one for 2.4.21 but I could not find one for 2.4.29. With only 64 MB of RAM some compiler runs are just not working. Is there an easy to use documentation regarding cross compiling the kernel?? Regards Thomas |
|
From: Thomas B. <th...@bo...> - 2005-03-04 19:28:36
|
Hello, is there an option to use the IDE Port in the 8362 NS1000? It seems that the place, where the connector for the IDE port would have gone is empty. Maybe one could put the missing parts into the Networkstation to use the IDE interface, but not only the connector needs to be added, at least 2 Chips are missing as well. Maybe someone else already found a way to get it working ... It would be cool to use a 2.5 inch HDD or a CompactFlash card! Regards Thomas |
|
From: Jochen R. <jo...@pa...> - 2005-02-22 18:35:54
|
Hi Veli-Pekka, I think I asked for the dmesg output, way back, if memory serves. Would it be possible for you to email me the corresponding kernel.300 image as well? Regards, Jochen Veli-Pekka Peltola wrote: > Hello! > > This is my first post to this list. > > While (one and half year) ago someone asked NetBSD dmesg-log from NC300. > So here it is: > > Copyright (c) 1998, 1999 > Network Computer, Inc. All rights reserved. > Copyright (c) 1996, 1997, 1998, 1999 > The NetBSD Foundation, Inc. All rights reserved. > Copyright (c) 1982, 1986, 1989, 1991, 1993 > The Regents of the University of California. All rights reserved. > > NetBSD 1.4-NCOS (CLIENT.NS3C) #0: Fri Nov 3 15:24:52 MST 2000 > > buildmaster@buildns1k:/buildmaster/modules/NCOS/build-history/build-3.0.beta-200011011341/NCOS/os-src/sys/arch/ns3c/compile/CLIENT.NS3C > > cpu0: PPC403GCX 0.1 > real mem = 50331648 > avail mem = 44756992 > using 614 buffers containing 2514944 bytes of memory > mainbus0 (root) > atnvram0 at mainbus0 > nvram0 at atnvram0 > low_ram->boot_source = C0000002 > not a PCMCIA_CISTPL_DEVICE > , mem 0x0x74600000, io 0x0x74400000 not configured > ne0 at mainbus0, io 0x0x74400c00, irq 1 > ne0: NE2000 Ethernet > ne0: Ethernet address 00:00:00:00:00:00 > fb0 at mainbus0, mem 0x0x72c00000, io 0x0x72200000: 640 x 480 x 8bpp > wsdisplay0 at fb0: console (std, vt100 emulation) > isa0 at mainbus0, io 0x0x74000000 > com0 at isa0 port 0x2f8-0x2ff irq 3: ns16550a, working fifo > lpt0 at isa0 port 0x378-0x37b irq 4 > pckbc0 at isa0 port 0x60-0x64 > pckbd0 at pckbc0 (kbd slot) > pckbc0: using irq 4 for kbd slot > wskbd0 at pckbd0: console keyboard > pms0 at pckbc0 (aux slot) > pckbc0: using irq 4 for aux slot > wsmouse0 at pms0 > wslpn0 at pms0 > glue0 at mainbus0 > audio0 at glue0: half duplex > pcic0 at mainbus0, mem 0x0x73000000, io 0x0x74000000 > pcmcia0 at pcic0 > pcic0: pcic_detect_card > sc=0x8002fc00 > rfs driver level: 19991105 > unknown device major 0xfff00 > Trying device ppp1 > unknown device major 0xfff00 > Trying device ppp0 > Trying device ne0 > RFS root not mounted. errno=53 > (192.168.1.1:/NetworkStationV2/prodbase/ppc) > root on 192.168.1.1:/NetworkStationV2/prodbase/ppc > root file system type: nfs > > I edited Ethernet MAC-address to zeros. > > I hope this will help you to compile the kernel for NC300. > |
|
From: Veli-Pekka P. <vel...@ik...> - 2005-02-18 09:20:26
|
On Thu, 10 Feb 2005, Santiago Garcia Mantinan wrote: > It is my plesure to announce that we have a new kernel patch, and that > it is > for 2.4.29, it seems to work ok here, if you have any troubles with it, > just > say it. I got it working after moving isa_init_dma function (starting from line 935) before line 909. This hack made it to compile fine. This was my first modification in the kernel source ever so I can't say is this a fix or just a piece of luck. I'm very beginner what comes to coding in C... Now Networkstation says: "Address out of range. File is bad or too large for installed memory." I guess it may fail to boot because I used a wrong compiler. Sorry my not so fluent English. -- Veli-Pekka Peltola |
|
From: Veli-Pekka P. <vel...@ik...> - 2005-02-14 17:55:05
|
Hello!
This is my first post to this list.
While (one and half year) ago someone asked NetBSD dmesg-log from NC300.
So here it is:
Copyright (c) 1998, 1999
Network Computer, Inc. All rights reserved.
Copyright (c) 1996, 1997, 1998, 1999
The NetBSD Foundation, Inc. All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
The Regents of the University of California. All rights reserved.
NetBSD 1.4-NCOS (CLIENT.NS3C) #0: Fri Nov 3 15:24:52 MST 2000
buildmaster@buildns1k:/buildmaster/modules/NCOS/build-history/build-3.0.beta-200011011341/NCOS/os-src/sys/arch/ns3c/compile/CLIENT.NS3C
cpu0: PPC403GCX 0.1
real mem = 50331648
avail mem = 44756992
using 614 buffers containing 2514944 bytes of memory
mainbus0 (root)
atnvram0 at mainbus0
nvram0 at atnvram0
low_ram->boot_source = C0000002
not a PCMCIA_CISTPL_DEVICE
, mem 0x0x74600000, io 0x0x74400000 not configured
ne0 at mainbus0, io 0x0x74400c00, irq 1
ne0: NE2000 Ethernet
ne0: Ethernet address 00:00:00:00:00:00
fb0 at mainbus0, mem 0x0x72c00000, io 0x0x72200000: 640 x 480 x 8bpp
wsdisplay0 at fb0: console (std, vt100 emulation)
isa0 at mainbus0, io 0x0x74000000
com0 at isa0 port 0x2f8-0x2ff irq 3: ns16550a, working fifo
lpt0 at isa0 port 0x378-0x37b irq 4
pckbc0 at isa0 port 0x60-0x64
pckbd0 at pckbc0 (kbd slot)
pckbc0: using irq 4 for kbd slot
wskbd0 at pckbd0: console keyboard
pms0 at pckbc0 (aux slot)
pckbc0: using irq 4 for aux slot
wsmouse0 at pms0
wslpn0 at pms0
glue0 at mainbus0
audio0 at glue0: half duplex
pcic0 at mainbus0, mem 0x0x73000000, io 0x0x74000000
pcmcia0 at pcic0
pcic0: pcic_detect_card
sc=0x8002fc00
rfs driver level: 19991105
unknown device major 0xfff00
Trying device ppp1
unknown device major 0xfff00
Trying device ppp0
Trying device ne0
RFS root not mounted. errno=53 (192.168.1.1:/NetworkStationV2/prodbase/ppc)
root on 192.168.1.1:/NetworkStationV2/prodbase/ppc
root file system type: nfs
I edited Ethernet MAC-address to zeros.
I hope this will help you to compile the kernel for NC300.
--
Veli-Pekka Peltola
|
|
From: Armin S. <ar...@me...> - 2005-02-10 10:39:12
|
On Thu, 10 Feb 2005, Santiago Garcia Mantinan wrote: > Hi! > > It is a long time since the last time somebody posted anything here, and > even longer since my last post, but it was about time I did this. > > It is my plesure to announce that we have a new kernel patch, and that it is > for 2.4.29, it seems to work ok here, if you have any troubles with it, just > say it. > > This allows us to avoid all the nasty bugs (specially those security ones) > that had been found in previous versions of the 2.4 kernels. > > I'm also kooking some new stuff for the userland, but I prefer not to say > what it is till I finish it. > > Hope you enjoy this ;-) Any work on 2.6 yet ? I'm new with networkstation, but when I have some time, I would try it. Armin |
|
From: Santiago G. M. <ma...@ay...> - 2005-02-10 09:14:20
|
Hi! It is a long time since the last time somebody posted anything here, and even longer since my last post, but it was about time I did this. It is my plesure to announce that we have a new kernel patch, and that it is for 2.4.29, it seems to work ok here, if you have any troubles with it, just say it. This allows us to avoid all the nasty bugs (specially those security ones) that had been found in previous versions of the 2.4 kernels. I'm also kooking some new stuff for the userland, but I prefer not to say what it is till I finish it. Hope you enjoy this ;-) Regards! |
|
From: Enrique <eo...@ig...> - 2004-08-30 13:03:50
|
El Lun 30 Ago 2004 14:49, escribi=F3: > I have changed all my settings as per your instructions . However, whe= n > booting the NS cannot find a file called nvram.mig. Any ideas where th= is > is? I am trying to get these working without using DHCP - if this fail= s > I'll start over again on a Windows 2003 machine instead of XP (can't wo= rk > out how to configure DHCP turbo!) Don't worry about nvram.mig. This file is generated by the original Netwo= rk=20 Station Manager (NSM) supplied by IBM. It's purpose is to change BIOS=20 settings such as the language or the BIOS password. If you don't use NSM,= =20 there is no way of generating this file, so, don't mind. I remember DHCP turbo was very simple. Much more than Win2003 DHCP=20 administration. Try to define a subnet and create some IP addresses there= =2E=20 For each one, specify the same extra configuration options as I wrote at = the=20 website for Windows 2003. If you have any particular question about DHCP Turbo, tell me. If I've ti= me=20 enough, I will install it and write a tutorial. > Last question for the moment. When I finally get these to boot will > everything be in Spanish or is there a language selection screen? (My > Spanish is not up to much I'm afraid!) It will be in spanish. But don't worry. Its very simple to change te lang= uage=20 in Linux. We've made a particular configuration for a school that ask us=20 about, that let the users choose between english and spanish at startup (= just=20 to learn english). > I'll try using Knoppix on a different machine (I tried it on my own but= it > wouldn't let me mount my hard drive and uncompress the tar (it wanted t= o do > it in RAMdisk but as I don't have 1gb of RAM it kept locking up.... may= be, > as you suggest, it would work over an NFS share better). I think copying or untarring by NFS is the only way to get the file=20 permissions properly set up. See you. --=20 Enrique Oca=F1a Gonz=E1lez Ingeniero en Inform=E1tica=20 Gesti=F3n de la Internet Municipal Educativa del Ayuntamiento de A Coru=F1= a mailto:eo...@ig... Igalia S.L. http://www.igalia.com - Yo tambi=E9n quiero una Europa libre de Patentes de Software - -------------------------------------------------------------- EuropeSwPatentFree - http://EuropeSwPatentFree.hispalinux.es ------------------------------------------------------------- |
|
From: Matthew M. <m_...@bl...> - 2004-08-30 12:50:28
|
Hi Enrique,
Please feel free to send these messages onto the forum. I would be very
pleased if what we have discussed helps someone else.
I have changed all my settings as per your instructions . However, when
booting the NS cannot find a file called nvram.mig. Any ideas where this
is? I am trying to get these working without using DHCP - if this fails
I'll start over again on a Windows 2003 machine instead of XP (can't work
out how to configure DHCP turbo!)
Last question for the moment. When I finally get these to boot will
everything be in Spanish or is there a language selection screen? (My
Spanish is not up to much I'm afraid!)
Thanks again,
Matt
P.S
I'll try using Knoppix on a different machine (I tried it on my own but i=
t
wouldn't let me mount my hard drive and uncompress the tar (it wanted to =
do
it in RAMdisk but as I don't have 1gb of RAM it kept locking up.... maybe=
,
as you suggest, it would work over an NFS share better).
----- Original Message -----
From: "Enrique Oca=F1a Gonz=E1lez" <eo...@ig...>
To: "Matthew Miles" <m_...@bl...>
Sent: Monday, August 30, 2004 8:11 AM
Subject: Re: [nwst-dev] (no subject)
El S=E1b 28 Ago 2004 12:35, escribi=F3:
> I've just found out that I can get into the NS setup utility by booting
the
> terminal without connecting the network cable. When it tries to find a
> DHCP server press escape (it states booting cancelled by user) and it w=
ill
> boot into the setup utility. The default password is 'password' withou=
t
> the speechmarks. From here you can set up booting from NVRAM etc.
Yes, I know. In fact, you can change the password and some other settings
(even reflash the BIOS) using the original NetworkStation server software
written by IBM.
> Problem is that I have no idea what to enter into these settings!
This is our configuration...
Network Parameters:
- IP: Network
- DHCP: 1
- BOOTP: Disabled
- Ethernet Standard: Version 2
Set boot parameters:
- Boot file: kernel.1000
- TFTP boot directory:
- NFS boot directory: /networkstationv2/prodbase/ppc/
- Boot host protocol:
TFTP order: Disabled
NFS order: 1
Local order: Disabled
Set configuration parameters:
- Configuration file:
- Configuration directory:
First:
Second:
- Configuration host protocol:
First: NFS
Second: Default
Set monitor parameters:
- Set monitor resolution: 1024x768@85Hz
- Set monitor power management: Enabled
Set language parameters:
- Select keyboard language: Spanish (in our particular case)
- Select startup language: Spanish (in our particular case)
Set verbose diagnostic messages: Enabled
> Is there any chance that you could send me the files from the main tar
> after extraction (i.e sutiable for windows on a couple of CD's at my
> expense as I am having lots of problems extracting them myself!)
If you want, I can publish or send you a ZIP file containing all files.
Unfortunately, there is a BIG problem about doing this. A ZIP file, a TAR
file or whatever that is going to be uncompressed from Windows, will lose
the
ACL (Access Control List =3D permissions) and symbolic link information.
The ACL that each file has is very very important, because it determines =
the
file permissions when it is viewed from Linux through NFS. If the files h=
ave
a wrong owner or a wrong ACL, Linux will probably be unable to access the=
m.
This was the MAIN problem we had when trying to move from a
Linux-server-based configuration to a Windows-server-based one.
Moreover, standard Windows applications (zip and tar included) don't know
how
to deal with symbolic links. Symbolic links are something "emulated" in
Windows by Services For Unix.
For this reason we don't know any other method to uncompress and install
Corunix other than mounting the NFS export into a Linux machine and runni=
ng
there the TAR utility. As far as we know, only doing this we ensure that
permissions and symbolic links are restored properly in the Linux side, a=
nd
ACL, owner and "symbolic link emulation files" are restored properly in t=
he
Windows side.
If you know about some utility that can store (and restore) ACLs in a
reliable way, we'll happy to know about it and use it to made an archive =
for
you.
Try to get a laptop or another computer from a friend, boot it using Knop=
pix
or another live distribution, and do the installation steps that require
Linux. We don't know other method. :-(
See you.
P.S.: Would you mind if I send all the thread to the nwst-dev list? I thi=
nk
all these emails would be very interesting for the people subscribed to t=
he
list.
--
Enrique Oca=F1a Gonz=E1lez
Ingeniero en Inform=E1tica
Gesti=F3n de la Internet Municipal Educativa del Ayuntamiento de A Coru=F1=
a
mailto:eo...@ig...
Igalia S.L. http://www.igalia.com
- Yo tambi=E9n quiero una Europa libre de Patentes de Software -
--------------------------------------------------------------
EuropeSwPatentFree - http://EuropeSwPatentFree.hispalinux.es
-------------------------------------------------------------
|