Re: ANN: initial sparc potato boot floppies ready for testing
Ben Collins wrote: 3. booting from network+nfsroot (linux-a.out from sun4cdm/ and root.tar.gz) --- Report: can boot the kernel but was unable to mount the nfsroot fs because this kernel is not configured for root on NFS ;(( (need CONFIG_IP_PNP=y, CONFIG_IP_PNP_BOOTP=y, CONFIG_IP_PNP_RARP=y and CONFIG_ROOT_NFS=y). This was on purpose since a lot of people complained about the long wait looking for a BOOTP server that wasn't there. I think we can just get around this with simple RARP and NFSROOT right? Can't we just pass the nfs server/root and IP address on the boot command (not as pretty, but it allows us to avoid the long hang)? Could we build a specific kernel for it then? Btw, I guess nfsroot is only needed for really old, low memory, systems. Maybe adding it to the sun4c kernel image would be sufficient. I'm still not sure about the nfs mounting. I remember having this problem a long time ago on my first slink install for an IPC. I can't remember if I got around it, or just did it another way. I built my own boot-floppies based on 2.2.13 today, and I can mount nfs partitions with no problem at all. Could you compare the options you have selected with those from 2.2.13 ? I will bring my 1+ home next week so I can start dealing with some of these issues. Eric, you noticed the sun4c rescue image right? I was able to squeeze that thing down all the way to be small enough for a single floppy and still be uncompressed. My 1+ had issues with the old floppy boot, so I will be able to test this image too. I will revive my old sparc2 to do some trial tomorrow. Regards. -- Eric Delaunay | S'il n'y a pas de solution, c'est qu'il n'y [EMAIL PROTECTED] | a pas de problème. Devise Shadok.
question about floppy drive support
I've been having bad troubles with the FDD on my Sparc IPX. Well, I booted the 2.2.14 a.out kernel from the new testing run of install disks, and after a fresh install with the old 2.2.1-based Slink disks I can access the FDD. I slipped a floppy in and could cat it without any trouble. Then I installed `fdutils' from potato. Accessing the same disk yields |floppy0: sector not found: track 0, head 0, sector 1, size 2 |floppy0: sector not found: track 0, head 0, sector 1, size 2 |end_request: I/O error, dev 02:00 (floppy), sector 0 |cat: /dev/fd0: Input/output error and of course superformat isn't working. Still, it's working better than it was. is there something besides superformat I can use on the Sparc, or do I just have a funky machine?
Troubles booting sparcstation over a network
Hello everybody, After successfully installing debian on my PC, I thought I'd give my old Sparcstation a try. Since I don't have a monitor or floppy drive for it, I tried to boot it over my home network. After getting the tftpboot.img file and installing tftp on my PC, I got the following error message when booting the sparcstation: ok boot net Booting from: le(0,0,0) 2e3200 Illegal Instruction ok The hardware is a sparcststion 1+, 16 Mb of memory, ROM revision 1.3. It's currently running Solaris 2.5 without any problems. Dows anybody have any idea what might be the problem? Maarten Vink [EMAIL PROTECTED]
Re: Troubles booting sparcstation over a network
On Sun, 9 Jan 2000, Maarten Vink wrote: After successfully installing debian on my PC, I thought I'd give my old Sparcstation a try. Since I don't have a monitor or floppy drive for it, I tried to boot it over my home network. After getting the tftpboot.img file and installing tftp on my PC, I got the following error message when booting the sparcstation: It's kinda tricky to boot via net. Every machines looks for a specific bootkernel with a specific name. In the TFTP boot docfile is discribed how you can calculate the name - another way would be to start a tcp sniffer and look what kinda requests the sparc is sending out. There you can see the filename. If you have another Sun in the subnet try snoop host. Another helpful source would be the INSTALL Doc from debian.org. I hope I helped a bit R. Jaeschke
Re: Troubles booting sparcstation over a network
On Sun, 9 Jan 2000, Maarten Vink wrote: After successfully installing debian on my PC, I thought I'd give my old Sparcstation a try. Since I don't have a monitor or floppy drive for it, I tried to boot it over my home network. After getting the tftpboot.img file and installing tftp on my PC, I got the following error message when booting the sparcstation: It's kinda tricky to boot via net. Every machines looks for a specific bootkernel with a specific name. In the TFTP boot docfile is discribed how you can calculate the name - another way would be to start a tcp sniffer and look what kinda requests the sparc is sending out. There you can see the filename. If you have another Sun in the subnet try snoop host. Another helpful source would be the INSTALL Doc from debian.org. I hope I helped a bit R. Jaeschke It does actually get the boot image; I see an incrementing counter, which stops at a certain moment. After about a second, it displays the error message. I tried the same thing with a sparc32 tftp-image from RedHat, which also resulted in this error. I'm beginning to think it's some kind of hardware problem; however, it works fine using Solaris, and ftp-ing a file to and from the machine doesn't lead to any file corruption so I think the network is OK. Maarten Vink
SPARC 1+ installation report (potato)
Hi. I tried the disks from http://xia01.kachinatech.com/~buildd/sparc-2.2.14-potato-boot/ on my SPARC 1+ (SUN 4c and SUN 4cdm rescue disks seem to produce identical results.) First, a warning is displayed when the rescue disk is automatically ejected: WARNING floppy change called early Second, after inserting the root disk, the VFS /root filesystem is reported to be installed correctly, but then the floppy drive stops and my SPARC hangs ... Ciao Sven
Re: ANN: initial sparc potato boot floppies ready for testing
[EMAIL PROTECTED] wrote: Machine is a Sparc IPX with generic 4x16MB 36-bit SIMMS and OEM hard drive on serial console. Booted by TFTP Console is `minicom' running on Slink i386. Noticed that dbootstrap is not drawing with the IBM `box' characters like 2.1 dbootstrap did. Pass TERM=vt100 or TERM=vt102 at boot time. Current init process does not enforce TERM settings anymore ;(( I don't know wether I will enable it again or just let the user pass TERM=xxx at boot time. Maybe a serial target for silo could also be helpful for floppy boot. On this topic, do you know about a way of identifying the type of console connected to the serial line ? Is there an escape sequence that returns vt100 or the like ? -- Eric Delaunay | S'il n'y a pas de solution, c'est qu'il n'y [EMAIL PROTECTED] | a pas de problème. Devise Shadok.
Re: ANN: initial sparc potato boot floppies ready for testing
Eric Delaunay wrote: Well, my first try is not successful at all :(( Hardware: SparcClassic (sun4m), 24MB RAM, 207MB harddisk. 1. booting from network (tftpboot.img) -- Report: image downloaded, then messages are displayed very slowly (could be PROM console ?). However the kernel reports cgthree detected and switching to fb0 (85x39). ... Switching to colour frame buffer device 85x39 fb0: cgthree at 0.6000 ... I further get some kernel oops but the welcome box from dbootstrap is displayed. ... Starting pid 6, console /dev/tty1: '/etc/init.d/rcS ' Unable to handle kernel paging request at virtual address 0xfdde tsk-mm-context = 0002 tsk-mm-pgd = f016f800 ... Running /etc/init.d/rcS ... more oops ... /etc/init.d/rcS done. starting dbootstrap Well, I investigated a bit more, and it seems the oopses are coming from all accesses to /dev/ttyN in /etc/init.d/rcS. Since the video appears to be PROM console, I guess VT is not initialized properly, so the oopses. I tried a serial console install both on my sparc2 and my SparcClassic and these oopses are not shown. But I cannot go further because I still cannot mount an nfs partition. Btw, I would want to build my own bootdisks based on your 2.2.14 kernel, so where can I download these kernel-image packages? Another potential pb: # ll sparc-2.2.14-potato-boot/sun4c total 1306 drwxr-xr-x2 eric perso1024 Jan 7 22:44 disks-1.44 -rw-r--r--1 eric perso 1328759 Jan 7 00:39 linux-a.out ^^^ should be a multiple of 4 if you want to netboot from it (eventually after you will add rarp nfsroot to the kernel). Regards. -- Eric Delaunay | S'il n'y a pas de solution, c'est qu'il n'y [EMAIL PROTECTED] | a pas de problème. Devise Shadok.
Re: ANN: initial sparc potato boot floppies ready for testing
On Sun, Jan 09, 2000 at 04:29:40PM +0100, Eric Delaunay wrote: Pass TERM=vt100 or TERM=vt102 at boot time. Current init process does not enforce TERM settings anymore ;(( I don't know wether I will enable it again or just let the user pass TERM=xxx at boot time. Maybe a serial target for silo could also be helpful for floppy boot. On this topic, do you know about a way of identifying the type of console connected to the serial line ? Is there an escape sequence that returns vt100 or the like ? IMHO, it's safe enough to assume vt100 as long as there's a way to override that assumption if necessary. How many people actually use terminals that don't speak vt100? -- Mike Stone pgpTTscGRaw5y.pgp Description: PGP signature
Re: ANN: initial sparc potato boot floppies ready for testing
On Sun, 9 Jan 2000, Eric Delaunay wrote: [EMAIL PROTECTED] wrote: Machine is a Sparc IPX with generic 4x16MB 36-bit SIMMS and OEM hard drive on serial console. Booted by TFTP Console is `minicom' running on Slink i386. Noticed that dbootstrap is not drawing with the IBM `box' characters like 2.1 dbootstrap did. Pass TERM=vt100 or TERM=vt102 at boot time. Current init process does not enforce TERM settings anymore ;(( I don't know wether I will enable it again or just let the user pass TERM=xxx at boot time. Maybe a serial target for silo could also be helpful for floppy boot. On this topic, do you know about a way of identifying the type of console connected to the serial line ? Is there an escape sequence that returns vt100 or the like ? There is such a beastie as far as I've heard, but I don't know what it might be. You might check the `ctlseqs.ms' file in the xterm source to see what xterm uses, then check it against minicom. The file is 14k compressed, in case anyone needs a copy. I've read the thing, but I'm not even a novice at reading escape codes. ;
Re: ANN: initial sparc potato boot floppies ready for testing
On Sun, 9 Jan 2000, Michael Stone wrote: On Sun, Jan 09, 2000 at 04:29:40PM +0100, Eric Delaunay wrote: Pass TERM=vt100 or TERM=vt102 at boot time. Current init process does not enforce TERM settings anymore ;(( I don't know wether I will enable it again or just let the user pass TERM=xxx at boot time. Maybe a serial target for silo could also be helpful for floppy boot. On this topic, do you know about a way of identifying the type of console connected to the serial line ? Is there an escape sequence that returns vt100 or the like ? IMHO, it's safe enough to assume vt100 as long as there's a way to override that assumption if necessary. How many people actually use terminals that don't speak vt100? I know of quite a few Wyse (50|60) terminals still alive and well being used as serial terminals. Don't know of any being used on Sparcs right now, but a colleague is using one on his i386. ;
Re: ANN: initial sparc potato boot floppies ready for testing
On Sun, 9 Jan 2000, Eric Delaunay wrote: On this topic, do you know about a way of identifying the type of console connected to the serial line ? Is there an escape sequence that returns vt100 or the like ? Unfortunately, there is no universal mechanism for detecting what kind of terminal you're attached to. Even worse, you can't try to autodetect it, since some terminals (now thankfully rare) will do Bad Things(tm) like lock up when they're prodded in the wrong way. Just look through /etc/termcap sometime, it's fascinating reading... ;) There is, however, an escape sequence pair you can use to determine what model of VTxxx you're talking to. It's documented rather nicely in ctlseqs.ms, which is in /usr/share/xterm/doc/, so I won't waste bandwidth by detailing it here. I'd tend to say that assuming VT100 until told otherwise by the user isn't a bad idea. In fact, you can probably get away with assuming VT102. [I had the opportunity several years to get very well-acquainted with not only all the VT1xx escape sequences but all the bugs you have to emulate, too, while writing a VT100 emulator in Turbo Pascal - even Telix was too slow on my XT...] -- Jeffrey Sean Connell __ | VP of Software Development, Convergence Equipment [EMAIL PROTECTED] \/ | PGP key: http://www.canuck.gen.nz/~ankh/pgpkey.html +
potato success on various SPARCstations and two questions.
Hi all, I just installed the new disks on 4 SS and everything looks good. I installed on SS2, SS5/110 SS10/40 single CPU and SS20/50 single CPU. After installation I started installing single pkgs using dpkg to buil a customized system. I just have two questions: 1) After installing the disks I also installed all the pkgs mentioned in the doc Software you need for develpment. Making this I had to upgrade some installed pkg from the dist disks. I just had a problem while installing g++ and libstdc++. Each one of the two pkgs asks for the other one to be installed and configured first. How can I do to install them? 2) I've always installed xfee from distributions and I'm not really sure about all the pkgs are needed for the X system to run properly. Server, fonts, wm etc. All my workstations have a TGX video board. Can some of you show me the list i strictly need, a part from every wm like gnome etc. thank you in advance for your help. Andre Martano Italy Infinitamente gratis, Infinitamente meglio http://www.infinito.it