RS-6000 43P
PLEASE HELP!!! I have an RS-6000 43P Model 240 that I'm trying to install Debian onto, and I cannot get a single kernel to boot off of this blasted thing. Every time I try to boot a kernel image it pukes up an error, which varies depending on what kernel you're using. First, if anybody knows what architecture this is (CHRP or PReP), please let me know, because I'm still clueless. If you could point out how to tell which it is that would be even better. Second, if anybody knows of any kernels that will work with this machine (and the OF commands to load it) I would be greatly appreciative if you would let me in on the secret. I have visited all of the Debian documentation sites, and they all unfortunately say something to the effect of "Not ready yet Please add documentation" (which isn't too helpful). If anybody could shine some light on my situation I'd greatly appreciate. If you want specific errors and such let me know and I'll give them to you. Thanks in advance! Eric
Building PPC Kernel
OK I just got done reading through IBM's sketchy information pages, and it appears as though I need to build a 2.4 kernel in order to be able to boot off of my RS-6000 43P 240 machine. Can I build a kernel for a PPC architecture on an i386 machine, or do I have to find (fat chance) a powerpc machine to build the kernel on? Better yet, does anybody have a pre-compiled PPC kernel version 2.4?
Re: Building PPC Kernel
At 10:38 PM 10/11/00 , Ethan Benson wrote: On Wed, Oct 11, 2000 at 10:27:35PM -0400, Eric Reischer wrote: > OK I just got done reading through IBM's sketchy information pages, and it > appears as though I need to build a 2.4 kernel in order to be able to boot > off of my RS-6000 43P 240 machine. Can I build a kernel for a PPC > architecture on an i386 machine, or do I have to find (fat chance) a > powerpc machine to build the kernel on? Better yet, does anybody have a > pre-compiled PPC kernel version 2.4? you could build a cross compiler for powerpc, i once looked into doing this when i was stuck with linuxppc and its broken compiler, i decided waiting until i could get debian installed was a better idea ;-) (iow its a royal pain) i assume you can build an RS-6000 kernel on any powerpc machine including macs right? id build one for you (or let you build it yourself) but i have a slow connection and no 2.4 source trees around.. Well, I'm assuming that it would work, just as long as it uses the PowerPC arch for the build. I tried using IBM's config file they provide at http://oss.software.ibm.com/developerworks/opensource/linux/projects/ppc/con fig.power3, but when I did the make it blew up when it went to compile in 'offb.c' with about 2 dozen errors. I did notice that it was running gcc with the args '-march=i686', which can't be doing anything good here. If you want to try to build a kernel then cheers to you.you'll probably be the first other than IBM to be able to do it, and I'd be very grateful. Kernel sources are at ftp.kernel.org under the v2.4 folder. The most recent one is 2.4.0-test9. My last resort would be to pester the hell out of IBM until they give me some type of response, because we're renting these machines from them and as of now they are just chewing up our money.. Eric
Tweaking Kernel Makefile
If I "tweak" the top-level makefile for a 2.3 kernel to force the ARCH variable to be PPC instead of running uname -m, would that produce a successful compile for a PowerPC kernel, providing I get the makefiles correct? It seems too easy, and I'm probably wrong, but this seems to be the only variable affecting how the kernel is built. Eric
Kernel image
If anybody is running an RS-6000 43P that has a 2.4 kernel built, could you please sent it to me (off list)? I don't have a PowerPC machine to build a kernel on, and none of the ones offered with any of the distros are working with my machine. Thanks. Eric
Re: Kernel image
I have tried setting up cross-compiling, but never found all the necessary source files to make the compiler. Eric At 07:49 AM 10/24/00 , Rob Andrews wrote: In newsgate.debian.powerpc, Eric Reischer wrote: > If anybody is running an RS-6000 43P that has a 2.4 kernel built, could you > please sent it to me (off list)? I don't have a PowerPC machine to build a > kernel on, and none of the ones offered with any of the distros are working > with my machine. Thanks. Have you tried cross-compiling? -- Rob [EMAIL PROTECTED] note: send off-list/newsgroup replies to address listed in this footer! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Innd problems
Try checking the permissions and ownership on it. I've had that problem before, but it wasn't with lnnd. Eric At 03:53 PM 10/26/00 , Paul Talacko wrote: I can't get Innd to start and keep getting this error in my syslog: SERVER cant bdzinit /var/lib/news/history No such file or directory Now, the file IS there. I've also run makehistory as user news to make it, but I keep getting this error. Is there something else I should be doing? Thanks Paul - Paul Talacko http://www.seditiousdiaries.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernel Image
/usr/src/binutils/target $ ../configure --prefix=/usr/src/cross --target=ppc-unknown-linux-gnu Does it matter that when I type in the above command it says "Configuring for a i586-pc-linux-gnu host?" (And it makes the ppc-unknown-linux-gnu folder in the /usr/src/binutils/target folder (which is where I ran the ../configure from) My ultimate aim, obviously from the above, is to make a ppc cross-compiler on an intel machine. Eric
Network Boot Problems
I'm having a small problem that is turning out to be a bigger pain than I expected. I have an IBM RS/6000 machine that I'm installing Linux on, but when I try to get it to boot off the network, it contacts my bootp server (which happens to be an intel Debian box) and starts to download the file. The strange thing is, it only fully completes the download about 1/4 of the times I try to get it to download. Usually, when I type in 'boot net' to get it to boot from the network, it loads anywhere between 1500 and 2000 packets, and then stops. About 20 seconds later, it gives me a packet error (which on IBM machines is just a bunch of numbers) and the network boot fails. Other times, if I'm lucky, it'll go through fine. This is a pain because it makes booting kernel images >1Meg really hard because the bigger it is the more of a chance of it failing part-way through. Anybody seen this before? -- Eric Reischer "You can't depend on your eyes [EMAIL PROTECTED] if your imagination is out of focus." [EMAIL PROTECTED]-- Mark Twain --
Kernel build crapped out
I tried to build a kernel on my newly built cross-compiler this evening, and after about an hour of compiling the zImage, it gave the the following error: powerpc-unknown-linux-gnu-ld -T arch/ppc/mm/mm.o command for about 11 lines) drivers/input/inputdrv.o: In function 'keybdev_event': drivers/input/inputdrv.o(.text+0x16bc): undefined reference to 'emulate_raw' drivers/input/inputdrv.o(.text+0x16bc): relocation truncated to fit: R_PPC_REL24 emulate_raw make: *** [vmlinux] Error 1 Anybody know what I'm doing wrong; or if it's a bug, is there a patch? -- Eric Reischer "You can't depend on your eyes [EMAIL PROTECTED] if your imagination is out of focus." [EMAIL PROTECTED]-- Mark Twain --
Re: Kernel build crapped out
Sorry about that.I'm compiling a 2.4test9 kernel. And when I did the config, I issued the command: make CROSS_COMPILE=powerpc-unknown-linux-gnu- ARCH=ppc config which automatically put the CONFIG_PPC entry in the autoconf.h file. Eric At 02:33 AM 10/31/00 , Martin Costabel wrote: Eric Reischer wrote: > > I tried to build a kernel on my newly built cross-compiler this evening, > and after about an hour of compiling the zImage, it gave the the following > error: > > powerpc-unknown-linux-gnu-ld -T arch/ppc/mm/mm.o command for about 11 lines) > drivers/input/inputdrv.o: In function 'keybdev_event': > drivers/input/inputdrv.o(.text+0x16bc): undefined reference to 'emulate_raw' > drivers/input/inputdrv.o(.text+0x16bc): relocation truncated to fit: > R_PPC_REL24 emulate_raw > make: *** [vmlinux] Error 1 > > Anybody know what I'm doing wrong; or if it's a bug, is there a patch? You don't even say what kernel version you are trying to compile, and I know nothing about cross-compiling, but I'll try a guess anyway: Do you have "#define CONFIG_PPC 1" in your include/linux/autoconf.h? If not, why not? -- Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernel build crapped out
Is this something that should be submitted to the Linux Kernel mailing list to be fixed by the next kernel prerelease? Eric At 06:49 PM 10/31/00 , Martin Costabel wrote: Eric Reischer wrote: > > Sorry about that.I'm compiling a 2.4test9 kernel. And when I did the > config, I issued the command: > make CROSS_COMPILE=powerpc-unknown-linux-gnu- ARCH=ppc config > which automatically put the CONFIG_PPC entry in the autoconf.h file. [] > > > drivers/input/inputdrv.o: In function 'keybdev_event': > > > drivers/input/inputdrv.o(.text+0x16bc): undefined reference to > > 'emulate_raw' OK, since what you are describing is impossible with the version of drivers/input/keybdev.c I have (from the bitkeeper linuxppc tree), I had a look at the version supplied with the "official" 2.4.0test9. That's a really weird one, and the word "crap" you used seems justified: The function emulate_raw is used without any ifs, but its definition some lines earlier is enclosed in either #if defined(CONFIG_X86) || defined(CONFIG_IA64) || defined(__alpha__) || defined(__mips__) or #elif defined(CONFIG_ADB_KEYBOARD) So in your case you would need to put CONFIG_ADB_KEYBOARD=y into your .config file. Or change these weird #ifs. The bitkeeper version of the file is somewhat better in that it uses only one set of conditionals, #if defined(CONFIG_X86) || defined(CONFIG_IA64) || defined(__alpha__) || defined (__mips__) || defined(CONFIG_PPC) but the function is still used without any condition. -- Martin
kernel compile
Perhaps somebody can help me with this problem. I'm building antoher 2.4test9 kernel, and for the life of me cannot figure out why I can't enable the SYM53C8XX low-level SCSI driver in the kernel. No matter what I select and/or deselect, I just can't get that field to un-grey-out in the xconfig window. Is there a dependency that I don't know about here? Eric Reischer "You can't depend on your eyes [EMAIL PROTECTED] if your imagination is out of focus." [EMAIL PROTECTED]-- Mark Twain
INIT frezzes on boot
Has anybody heard of INIT freezing just as it is starting up on a PowerPC machine? I'm using my own custom kernel, but I'm using Debian's root.bin image to mount as root. After it mounts the root partition, the computer frees the unused kernel memory, and then hangs (I'm guessing it's trying to load init or something. Ctrl-Alt-Del still works, and it does a clean reboot, but I don't get the usual init messages that go along with it (e.g. INIT: Switching to runlevel 6), so I'm assuming that the machine is not "hung", but is instead waiting for something??? -- Eric Reischer "You can't depend on your eyes [EMAIL PROTECTED]if your imagination is out of focus." [EMAIL PROTECTED]-- Mark Twain --
INIT under 2.4
Does anybody have an INIT binary compiled under the 2.4 kernel? The one provided with Debian's CD's is not working (it hangs before it loads, just after freeing unused kernel memory). I'm not sure what the problem is, but I think that is it, since I can get one from YellowDog Linux to work (but it's custom YDL and won't work with debian). Thanks. ------ -- Eric Reischer "You can't depend on your eyes [EMAIL PROTECTED] if your imagination is out of focus." [EMAIL PROTECTED]-- Mark Twain --
Re: CHRP boot disks
> Where is the CHRP boot disk image?
>
> I looked here but could not find one:
>
http://ftp.debian.org/debian/dists/potato/main/disks-powerpc/current/chrp/
images-1.44/
>
> It can not be rescue.bin because that image will not fit on a
dos-format'ed floppy.
>
You need to download that file, along with root.bin, and use the following
commands:
dd if=rescue.bin of=/dev/fd0 bs=1024
dd if=root.bin of=/dev/fd0 bs=1024
{obviously, switching disks in between} The disk images are intended to be
written block by block onto the disk, which is why they can't be just
copied over. I believe that the destination disk type is either ext2 or
minix, depending on the date of the distribution, and the mood of the
developer I suppose.
--
Eric Reischer
CCIL Network Administrator
[EMAIL PROTECTED]
--
Re: 3COM 3c905c-TX-M.... works, or not?
I might be able to give you a little bit of help here. First of all, the 3C905 series does work on PowerPC machines, thanks to Donald and all of his hard work on getting drivers made for the entire 3Com series. As far as your problems, it appears as though it may be a problem with how your motherboard assigns IRQs. First of all, I can tell you straight off that it is very strange to see that your card is being given an IRQ of 1.it should be a lot higher than thatonboard resources usually take that one, so that might be your problem. Try putting the NIC in a Intel machine, and grab ahold of the DOS diagnostics from either the disk you got with the card or the 3Com website. Run the diagnostics from a straight DOS prompt (NOT from a DOS prompt from within WindowsF8 at boot time will get you there). From there you can manually set the IRQ settings on your card, and they will be written to NVRAM, so the changes will still be there when you move the card back to your PPC machine. If this doesn't work, try looking at the documentation of your motherboard on assigning IRQs. * Eric Reischer "You can't depend on your eyes CCIL Network Administrator if your imagination is out of focus." [EMAIL PROTECTED] -- Mark Twain *
Re: Need help installing Potatoe on G4
Ok Al Gore. Try manually entering in information in the /etc/passwd file and see if it saves ok. It appears as though the install program is attempting to read configuration information from the disk but isn't finding a file, so it defaults back to ask you for the information again. I am trusting you chose to use shadow passwords, yes? Unless you have a reason not to use them, I recommend using MD5 passwords, although that won't cause the problem you're experiencing. Make sure that your partitions are all squared away, and if that still doesn't work, then completely erase the partition table of the disk (dd if=/dev/zero of=/dev/sda count=10), replacing sda with whatever drive your root resides on, repartition and reformat, and then try to install again. Eric At 04:42 PM 12/3/00 , you wrote: Hello everybody, I've got a really weird problem : While installing the potatoe, everything goes rigth until the moment I have to choose the root password. When I enter a pass and its confirmation, the install program kinda loop and reasks me the same question. I solved the problem by loging in as root and setting the pass. Then I've got the same problem with the Adduser stuff... So I dont add any user, choose to remove the PCMCIA, not install PPP, not install anything for Apt, scan the 3 cds, and then when I choose install (simple or advanced), I go back to the choose MD5 pass ? Anyone can help please ? I have a G4/400 AGP no Giga Ethernet, and I use the r0 isos to install. pejvan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: 3COM 3c905c again
As for the PCI slots, I'm afraid I can't be much of any help. The error you are describing in your syslog, that basically means that your DHCP server does not have any more IP addresses to give out to clients. If you administer this server, then you might want to consider adding more addresses to the scope. When the machine comes up, I'll bet if you do an ifconfig, and take a look at your eth1 adapter settings (if it's there), it won't have an IP address, which explains why packets aren't bouncing back to your ping source. Eric At 07:49 PM 12/3/00 , Jonathan Lane wrote: Okay, I swapped cards around, and it turns out that *anything* in PCI slots 3-6 are not correctly configured (this is on a UMAX S900). I moved the USB card and NIC to slots 1 and 2, and my video card to slot 4, which appears to be working (although I've not tried using X yet, as XF86 4.0.1 is currently broken...). However, now I am having another problem: When I ping the machine, I see the RX light on the NIC flash, but the source machine doesn't get any responses. A few things to note: Every once in a while, I see the following message: eth1: Setting promiscuous mode. Also, in my syslog, dhcpd complains with: Aug 27 16:42:20 colour dhcpd-2.2.x: DHCPDISCOVER from 08:00:46:05:9c:c8 via eth1 Aug 27 16:42:20 colour dhcpd-2.2.x: no free leases on subnet 192.168.1.0 (Ignore the date, my clock had yet to be set). But this indicates that packets *are* getting thru, as the MAC address shown above is correct. If anyone can shed some light on this matter, I'd greatly appreciate it. I'd also like to know why the last four PCI slots don't work properly, although it appears it's not incredibly important now, unless X doesn't work... TIA, Jonathan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sorry, I can't build 4.0.1-9pre10v1 for powerpc
Appears to be a memory/swap related problem. Open up another tty and run 'top' while the make is going on, and watch your total free memory and your total free swap..they might get too small for cc to run successfully. (VM is the virtual memory driver) Eric At 12:17 AM 12/4/00 , Branden Robinson wrote: My build box has a slight problem: make[7]: Entering directory `/home/branden/xfree86/xfree86-4.0.1/build-tree/xc/lib/GL/mesa/src/drv/gamma' rm -f gamma_gl.o gcc -c -O2 -g -fsigned-char -I../../../../../../lib/X11 -I../../../../../../include/extensions -I. -I.. -I../../include-I../../../../dri/drm -I../../../../dri -I../../../../glx -I../../../dri -I../../../../../../include -I../../../../../../include/GL -I../../../../../../programs/Xserver/hw/xfree86/os-support -I../../../../../../programs/Xserver/GL/dri -I../../../../../../programs/Xserver/hw/xfree86/drivers/glint -I../../../include -I../.. -I../../X -I../common -I../../../../../.. -I../../../../../../exports/include -Dlinux -D__powerpc__ -D_POSIX_C_SOURCE=199309L -D_POSIX_SOURCE -D_XOPEN_SOURCE -D_BSD_SOURCE -D_SVID_SOURCE -D_GNU_SOURCE-DFUNCPROTO=15 -DNARROWPROTO -DXTHREADS -D_REENTRANT -DXUSE_MTSAFE_API-DMALLOC_0_RETURNS_NULL -DGLXEXT -DXF86DRI -DGLX_DIRECT_RENDERING -DGLX_USE_DLOPEN -DGLX_USE_MESA -DDRIVERTS -fPIC gamma_gl.c gcc: Internal compiler error: program cc1 got fatal signal 9 make[7]: *** [gamma_gl.o] Error 1 Dec 3 16:17:14 eris kernel: VM: do_try_to_free_pages failed for kswapd... Dec 3 16:17:16 eris kernel: VM: do_try_to_free_pages failed for cc1... Dec 3 16:17:16 eris kernel: VM: killing process cc1 This is reproducible. It hits this wall on gamma_gl.c every time. Can someone else build this pre-version of -10 for powerpc, and email me any changes that need to be made to the debian/*.powerpc files? -- G. Branden Robinson | When I die I want to go peacefully in Debian GNU/Linux| my sleep like my ol' Grand Dad...not [EMAIL PROTECTED] | screaming in terror like his passengers. http://www.debian.org/~branden/ |
Re: floppy low-level format
If you mean you want to erase every last cluster on the disk, then you can use the following command to do that, no matter what type of drive it is: dd if=/dev/zero of=/dev/fd0 count=clear whole disk> This should essentially create a floppy disk with absolutely no information on it; not even a boot sector. From there you can format as usual to create a usable filesystem on the disk. Eric At 11:13 AM 12/5/00 , Andre Berger wrote: I haven't figure out yet how to do a low-level format on /dev/fd0 on PPCs, can anybody help me? My machine is a PowerBook 3400/potato/2.2.17, the floppy drive comes as a module (hardware module of course :)). Thank you, Andre -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
