@Hypex

I'll try with no DMA first and see what happens.
The drive and its partitions are seen and formattable just fine in OS4, so it isn't connected wrong. Yes, I am using one of those thin UDMA cables.
I don't recall seeing a LNX\00 option in media toolbox filesystems.. I'll take a look again.
Is SWP the same as for setting up an OS4 swap partition?
As for the hard drive being okay...hmmm...that brings to mind something....it may be that this hard drive failed a few years ago and a guy in IT at work ended up reformatting it to exclude bad sectors that were failing.....it suddenly comes back to me.
Won't debian allow me to do a low level format that handles any bad blocks by marking them as bad? I did try to format the swap drive but the hard drive started making clinging noises and locked up - I had to shut down the AmigaOne with the switch.
Will OS4 allow me to lock out bad sectors if I format the drives in Os4?


Thanks!

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

Okay...looking at this now on my AmigaOne.

It looks like the two linux drive partitions on the 8 gig drive are set up properly.
I see dha as a SWP\00 and dhb as a LNX\00 in media toolbox.

Both automount. dhb is bootable.
Are sector sizes of 512 etc all okay? There's a lot of settings that could be messed with here.

Anyway...should I do a lowlevel format of this entire drive from Amiga OS and then go back to debian and try to redo them, or should I set everything up in AmigaOS after I do a low level format? I think the low level format will check for bad sectors and disable them...won't it?

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

Well, it seems that my 8G Western Digital Caviar 38400 drive has some bad sectors starting around 301969 - several of them in fact. I was able to grab a program from Aminet named checkdisk that verifies this.
When I do a full format, it just stops after a little while and shows bad cylinder 6291 then quits. Isn't there some sort of format command that lets you format AND mark as used all bad sectors so the OS won't use them? Of course, I guess I need the Debian version of this tool anyway to be useful.
So, if I make a dummy unused partition that covers the bad sectors, then put my hda SWAP right after that, then dhb root after that, will Debian be okay with that?
What ever happened to the third partition that Woody had - the boot partition? Is that now combined with root?

Last edited by scabit on 31-Oct-2012 at 02:09 PM.
Last edited by scabit on 31-Oct-2012 at 02:08 PM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Just a quick reply for now. Fully formatting a drive is supossed to get a drive to detect bad sectors and remap them. Or so I was told.

You could try low level format I guess but that low level format was a SCSI thing and you're mot mean to to do it on IDE. IIRC MTB had all these format options but can't see any now.

You could also try smartctl. OS4 has it in the CLI.

Last edited by Hypex on 31-Oct-2012 at 04:21 PM.
Last edited by Hypex on 31-Oct-2012 at 04:19 PM.

@Hypex

Quote:

Just a quick reply for now. Fully formatting a drive is supossed to get a drive to detect bad sectors and remap them. Or so I was told.

Yes, as I mentioned above, I did a complete format on my dha and dhb partitions in OS4 and they work just fine (took something like 2 hours to do a full format of the larger partition!). I left the lower part of the drive alone like the first 8192 sectors - to ensure I wouldn't hit any bad sectors.
Tried it under the debian install CD and as usual, if I try to format in EXT3 (is that what its called?) it locks up at 33%. If I use EXT2 it finishes, but then locks up during the install later.
I tried setting ide to nodma...didn't matter.

I'll have to give up for now....I've spent almost a week trying to get this working so far to no avail. Let me know if you have any other ideas. I don't really think it is the hard drive itself, because it works fine in Os4 (at least the part I have set up to use for dha and dhb).

Thanks,

Scott

Last edited by scabit on 01-Nov-2012 at 12:01 AM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@Hypex

Okay..finally figured it out. I was about to give up and I remembered that when I tried to install the "woody" release of Debian, it would lock up if I tried to use a swap partition...so for fun I ran through everything again just as I had done before...but did not mount a swap partition. Presto! It works....well, as far as it can before I need more help...
It goes through and installs the whole base system stuff ...takes a LONG time....then I tell it no kernel, continue. It tries to locate a mirror to use to download files, but fails no matter what mirror I try to use, saying "does not have a valid release file on it. Try a different mirror". So I have to tell it to skip using go back.
I then select install package manager, and this time tell it NOT to use a mirror. It then fails saying "the security updates on security.debian.org couldn't be accessed." I select continue.
I am asked to take a survey...I select no.
I select Desktop Environment and Standard System as my software selection. It loads and unpacks 487 files. I go take a nap...
The install is finished. I am told I need to set boot as /dev/hdb2 and root=/dev/hdb3 as a kernel argument.
I finish and then select go back to try to perform step #4 in posting #1 of this thread - that is copy the kernel. Now, where do I copy it to? root or boot? boot is the bootable smaller partition I set up - about 400 megabytes I think. root is the 7 gigs where all the software got installed and is now 17% full.
I go to the menu to change CDs, insert and mount the install CD you provided so I can copy off the kernel...now what?
When I do a df I see:
/dev/ide/host0/bus0/target1/lun0/part3 6735708 etc etc 17% /target
/dev/ide/host0/bus0/target1/lun0/part2 563901 etc etc 0% /target/boot

I try to "ls" to those and it just shows nothing..no errors, nothing.
I don't see my cdrom mounted...how do I do that? I've tried doing dmesg and all I see is errors...no mounted partitions or devices.
Here is the problem.....if I reboot the machine, I go back to uboot which does NOT boot me into linux, nor can it since there is no kernel. So I have to go back and insert the intall CD you provided and use diskboot to boot from it, at which point no matter how hard I try, I always end up HAVING to reformat my hard drive partitions, even when I tell it not to, because it sees errors on them if I do not for some reason. So I sit and wait for everything to install for another hour and try it all over again....

Could you please expound on step #4 above? Assume I am at the point where the install is complete but I have not rebooted the amigaone yet. I have selected go back and am sitting at the main install menu.
How can I copy the kernel file from the install CD to my boot (or root?) partition? I have mounted the CD using mount /dev/cdrw -t iso9660 -r /cdrom and now I can see and copy files from the CD.
Where exactly is the boot drive? That is where I want to copy the kernel to correct? Or do I need to use dd to transfer the kernel image to the boot partition?

Thanks!

Scott

Last edited by scabit on 02-Nov-2012 at 11:42 AM.
Last edited by scabit on 01-Nov-2012 at 09:24 PM.
Last edited by scabit on 01-Nov-2012 at 07:57 PM.
Last edited by scabit on 01-Nov-2012 at 07:56 PM.
Last edited by scabit on 01-Nov-2012 at 07:55 PM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

That's great to hear you got it working! This deserves a:

Sorry you 've been typing a long story of your experience without any response. And no one took my place. Shocking!

You shouldn't need to reinstall it all again. But if there are errors reported a warning sign is there something is wrong.

I recommend running smartctl from OS4 or Linux. If you boot to a Linux shell you can it download it if not in the installer files by: (AFAIK)
apt-get install smartctl

Or it could be named smartmontools.

Now as to forgetting about swap. Good idea and well done. Why didn't I think of that!? Leaving you to solve it yourself.

Okay, sorry about that #4, yes you are right, I am a bit vague!

Now, this method it uses requires an SLB that can boot Linux. Mine is called L:slb 1.16. You don't need a boot partition or to directly write to any raw blocks. It's all file based.

What those instructions were suppossed to tell you is that in your freshly installed Linux volume you will need to put the kernel files into a directory called /boot. There should already be a /boot in the root level of the Linux volume.

So after you install, or since you already have; you need to do this from the installer. So boot my installer disc again. Just skip any installing and select to mount the root file system. Locate the volume and mount it. Then select to drop to a shell. Mount my CD. BTW it should be mountable as /cdrom or /cdrom1 depending where it is but you got it working so that's fine.

You can CD into the kernel dir on the CD. And now copy all the files to the /boot dir. Or just the HD kernel for faster booting. The root should be mounted on /target so the following should copy it assuming you CD'd to kernel:

cp *.* /target/boot/

Or even:

cp /cdrom/kernel/kernelhd-2.6.18.img /target/boot/

Now you need a line to boot it from SLB. Also in /boot. I'll use your drive as an example. This should get it working. Type this to get an editor for the boot menu:

nano /target//boot/a1boot.conf

Enter a line like this:
Etch /boot/kernelhd-2.6.18.img root=/dev/hdb3

That should get you started! Add other options as you usually have. And make sure to press return after the line so the file has a newline after the menu entry otherwise SLB will ignore it and you won't be able to boot! Happened to me and spent a while looking for the error

You can also give the volume a label and use that as a mount reference. Such as if you called it Etch then you could use a root spec like "root=/dev/disk/by-label/Etch" but I don't know if it works in kernel 2.6.18.

Hope you now have a happy Linux!:

Last edited by Hypex on 03-Nov-2012 at 12:58 AM.
Last edited by Hypex on 02-Nov-2012 at 04:11 PM.

@Hypex

Note - there are 4 entries in my fstab - none of them mount any devices.
What is happening is that I am never ever actually booting into the hard drive partition set up by the installer. I am simply in the RAM based partition in which there are 3 tmpfs devices of 257460 blocks each. Eventually if I go through enough of the install (using the etch CD-1" and go to a shell, I can see that the two hard drive partitions exist. But I have never yet been able to mount them such that I can make the root hard drive partition my root.

When I boot from your install CD, there is no option to mount the hard drive system from my 8 Gig drive. In fact, my hard drive isn't even seen until I switch to the linux CD-1 cd and go through a large part of the install process. When it is seen, it has the path of:
/dev/ide/host0/bus0/target1/lun0/part3 and I cannot LS it, CD to it, or access it. It needs to be mounted somehow, but there is no way to do that from any of the install menu functions, and I have tried for many hours to do so from the shell with no success.
As for putting a kernel on the partition...I will never know until I can somehow actually access it. THe boot procedure does not in any way - even when completed successfully to the end, allow you access to the hard drive partition you just created...it expects the reboot to boot you into it apparently., which won;t happen unless a kernel is installed...but you cannot install a kernel without booting into the root partition...a catch 22 - never possible.

So, I think this is not actually possible to do. Sadly, when I tried to look at my hard drive using the smartctl command in OS4, I got a message saying my hard drive does not support smart commands, and then it proceeded to run the test command I asked it to do anyway, which made huge churning and crashing noises on my hard drive, which now makes continual noise whenever I boot up. Looks like just using the smartctl command with a drive that doesn't recognize it destorys the hard drive?
I'm glad it was a spare I pulled out of an old PC!
But that ends my attempt to get Debian running.

A few lessons learned about trying to install Debian:
1) You MUST have a network connected and active or the install will halt with no indications of anything - no errors, nothing - it just locks up.
2) You cannot use ide=nodma because etch tries to install USB2 drivers and if you have ANY USB devices plugged in the install locks up with no indications of any problems. If you plug anything into a USB port, it locks up if nodma is enabled.
3) You CANNOT use a swap partition. Don't know why, but it will not work and locks up when the base system files start to get installed.

Perhaps I will try the yellow Dog linux installation available from Intuitionbase. Does anyone know if that works or is still valid? I see a link to an install ISO and links to 3 different CD isos that you need to download for it - yikes! Do I really need 3 cds to get this to work?

Well, this experience has certainly increased my appreciation for Amiga OS4! I used to use linux regularly before it branched into dozens (hundreds?) of different unique non compatible versions. Of course, I was never successful in setting it up on any computer before..only tried a couple of times. Poor Linus must be grieved over what has become of his concepts....this is certainly not the way he intended things to be.

Thanks for trying to help.

Scott

Last edited by scabit on 03-Nov-2012 at 02:55 AM.
Last edited by scabit on 03-Nov-2012 at 01:12 AM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Did you change your post? I read it earlier but it doesn't look the same. The info is missing on the dha/dhb.dhc.

I noticed something missed in the original guide. Quite possibly because you needed an existing working setup. And lots of people updated their kernels. So most were the new method.

That said, the partition you installed Linux on, IOW the root volume, must be set to bootable in the RDB. The new method uses an RDB formatted drive to boot Linux off. So the partition must be marked as bootable Media Toolbox. You only need one volume to install it on (minus swap). It doesn't need a boot partition. So only the root volume must be marked as bootable, Automount need not be set.

Once this is set up with Linux installed to a volume with Ext2/3 and an a1boot menu file the SLB will pick it up and you will see an entry for your Linux volume underneath your OS4 volumes. It's automatic from then on. Just select it and press return.

You could also try rescue mode which won't try to install the base system and will try to mount a volume for you and drop to a shell which is what you need. Add "rescue/enable=true" to your bootargs for this.

When booting my CD, once it has loaded it all in, and UBoot is booting it, you can remove it and replace it with the Debian CD. The installer only detects Debian in the first drive I found. It will then detect the CD and load packages from it. From there on you should see the menu. Now to mount a volume you must go to the Partitioner screen. You select your volume and mount as "/" which is root. Then you exit and. It should then be mounted. Making sure all it will do is mount it and not format anything!
Otherwise cancel and go back.

After that you can drop to a shell and copy the kernel. And also modules if you know the format to unpack a tar.gz off hand.,

Sorry about the HD. Looks like it was tool old to be SMART. 8GB was small even ten years ago so I hope it is no great loss and you had no personal data on it? Looks like smartctl found a weakness in the drive. Better finding out now than later. But no, it's not known to destroy drives. It can be used to grab the drive error log if it supports the SMART standard.

Lessons:

1: Yes, Linux relies on internet. As per my warning about the network in install guide.

2: Didn't know that about USB2. But the installer was made before we all put USB2 cards or at least years before OS4 supported it. And Micro is known not to work with USB2 in OS4 at all so I'm not surprised. But I am surprised it came up with no DMA! Also I wasn't aware kernel 2.6.18 had USB2 support. I thought it was only in 2.3.31.

3: I haven't heard of this. I wonder if it is a Micro only problem? However your HD was not in a perfect condition so the swap formatter might have caught a hard error. Once Linux catches a hard error forget it. Ii will delay for minutes. It happened to me. BTW if you Ctrl-Alt-F1 do you see lots of error messages?

I've heard of Yellow Dog working but haven't tried myself. Don't know the install process. But, what will you install it on?

At this point I suggest two things. One is to locate a working HD with no hard errors! Second is to try again with that. Or wait for my next installer which will be way easier. Well easier than now.

If you have space you could also try installing to your OS4 volume All you need is to create a spare partition. 5GB is scraping past. 10GB is fine size. And use manual partitioning.

@Hypex

Yes, after examining what was going on I changed my post a bit to reflect what I discovered. A lot that I originally wrote didn't actually apply.
No worries about the hard drive..it was an old spare ...nothing useful on it. I am purchasing a newer Seagate 160Gig - should have it within a week. Seems I managed to fill up my 80Gig Amiga drive in only 7 years! Amazing!
As for Debian.....I think I was really close to getting it to work but was just missing a few key details. I understood from reading your first post that you are expecting this to be a "reinstall" from established linux partitions. But I figured it wouldn't be too hard to use partitioner to create new linux partitions....and it wasn't. The real problem was accessing them once they were created. After I used partitioner to mount the hard drive partitions, a boot and a root, I saw the following when I did a df from the shell:

/dev/ide/host0/bus0/target1/lun0/part3 6735708 etc etc 17% /target
/dev/ide/host0/bus0/target1/lun0/part2 563901 etc etc 0% /target/boot

But, I was still set to the root partition of the ram based disk when I was doing this. Doing an ls or trying to cd to these partitions did not work. They were there, but could not be accessed. I think they required some sort of additional mount command to make them become accessible. Partitioner sure did not provide that capability.
The reason I made two linux partitions is because, to my knowledge, there is no way to make both a /root partition and a /boot partition. You can make a /root partition and then set the bootable bit, is that all that is needed? That obviously changed since woody, which required separate /boot and /root partitions to boot, and also, it seems, required use of the dd command to make the boot partition match the boot image file exactly in order to work.
Anyway, having these target drives accessible as shown in the paths above....just so I know for next time....what do I need to do to be able to access them or mount them? Partitioner makes them visible - but they are not accessible. This has nothing to do with my faulty hard drive...remember, I had blocked off the bad sectors - in fact, it had been set up that way for years on my PC and worked fine just using sectors from 8192 and above. The real problem came when I used smartctl and it went to test the drive.
Making the debian system boot from uboot was a different problem. I never set up any sort of multi system boot in uboot - I left it to just boot to AmigaOS4.1 by default, and if I stopped in uboot by hitting a key, to allow me to manually type in boot commands to access the linux drives. I was never able to access anything other than the bootable install CD you provided, since I always got a "bad magic number" error whenever I tried.
One other point of interest .... I was never able to get the journaling file systems to work with debian etch. I could only use EXT2. I tried many times to set up and format with both EXT3 and the JXFS, always failing. That should be noted for anyone trying to set up new partitons. Perhaps that was due to my bad hard drive - sectors 1 to 8192 being blocked off as a separate BAD partition. Not sure about that.

So I really would like to know just this one thing for any future attempts at debian.... when I go through and use partitioner to set everything up, make a /root partition with a bootable bit set, install the drivers and base system (but do not yet have a kernel insalled), and have the following partitions (along with the temp fs partitions in ram) displayed when I do a df command from a shell...

/dev/ide/host0/bus0/target1/lun0/part3 6735708 etc etc 17% /target
/dev/ide/host0/bus0/target1/lun0/part2 563901 etc etc 0% /target/boot

.. what do I need to do to access these partitions? CD to them does not work. I cannot LS them...they show nothing. What is the trick to mounting them so I can actually access them?Partitioner DOES NOT mount them...if mounting means that they are made accessible. If you can tell me how to access them, it would help a lot. This is the major problem that prevented me from getting this all to work.

Thanks!

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

I think they required some sort of additional mount command to make them become accessible. Partitioner sure did not provide that capability.

If you got to the Partitioner then the installer was able to load it off the Debian disc and extra modules. At this point, were you trying to mount existing partitions?

If so then you need to make sure you have told it to "Use as" Ex2 file system in your case and the mount point. If you mount it as root it will mount as /target for installation. Making sure it doesn't want to format it and blank the volume

If you have done all this and still cannot CD to /target in the shell or list the directory then there is something wrong. That is all you need to do. The Partitioner mounts it for you and that is all that is normally needed.

One thing I noticed is that your device node is unusual. This should be more direct and point to the IDE block driver. For example like: /dev/hdb3

I wonder if that is affecting it. But, you were able to install earlier and so you should be able to simply mount the volume! Otherwise it doesn't make any sense.

Quote:

You can make a /root partition and then set the bootable bit, is that all that is needed?

Yes, all that is needed is one root partition, you don't need a special boot partition for the kernel. This changed since the OS4 SLB supported booting Linux. So you use the SLB to pick Linux alongside your other OS4 partitions.

Internally SLB will still use UBoot to load and boot the kernel, but it makes the process transparent. And all needed files can remain on the main Linux volume.

Quote:

Anyway, having these target drives accessible as shown in the paths above....just so I know for next time....what do I need to do to be able to access them or mount them?

I'm afraid I am at a loss here as to why they won't mount but appear as already mounted volumes. Perhaps section 6.3.3.3. Manual Partitioning on here cam give a hint:
http://www.debian.org/releases/stable/powerpc/ch06s03.html.en#di-partition

You could try to manually mount the volumes without using the Partitioner by going into a shell and using my example block device:

mount /dev/hdb3 /target

or

mount -t ext2 /dev/hdb3 /target

You could also use /mnt or create one yourself if it doesn't exist from the ramdisk "/" mount-point. Like with: mkdir /test

Quote:

I never set up any sort of multi system boot in uboot - I left it to just boot to AmigaOS4.1 by default

Making sure bootable is set for both the Amiga RDB volume and the Linux volume from the Partitioner is all you should need alongside setting the UBoot Boot option to be Amiga Multiboot. Which should already be done to boot OS4 anyway.

Then the OS4 bootloader can pick up the Linux volume.

Quote:

I tried many times to set up and format with both EXT3 and the JXFS, always failing.

Really?! I thought JXFS was only an Amiga file system! Is JXFS available for Linux?

Quote:

I could only use EXT2

I'm not sure why this would be. Ex3 should work. The drivers should be in Etch.

Quote:

.. what do I need to do to access these partitions?

I can only offer the advice above. Also, did you try the rescue mode? Although more typing to boot the installer again it is geared more towards an existing system. Since somehow the usual methods of mounting from the installer are not working.

Just one last note about your /boot and /root volumes for clarification. As discussed /boot isn't needed when using the OS4 SLB. But note that "/root" isn't the root volume, /root" is a reference to a directory called "root" sitting off the "/' root volume.

The root is always "/" and will point to the installer ramdisk. When the Partitioner mounts a volume you specify as root it will mount it as "/target" as it s the root volume for the installation target. It still needs to access it's own files off the ramdisk

Last edited by Hypex on 06-Nov-2012 at 08:38 AM.
Last edited by Hypex on 06-Nov-2012 at 08:25 AM.

@Hypex

Okay...using my new hard drive, nothing has changed at all.
I went through and tried reinstalling debian etch 4 times, got the same thing each time. When the install is done I always end up in a situation where I am set to a root which is a temporary ram device used to install, and cannot change the root to /dev/hda2 (my hard drive root partition I set up with partitioner), and even worse, I am unable to mount the CD anymore using the command that worked for me before.
So I decided to try your recommendation of booting into a rescue mode.
This helped. I booted from your install CD and went into a shell after telling the installer to set /dev/hda2 as my root partition.
Sure enough...there I am on the actual hard drive! Good!
I found that mount /media/cdrom mounted your install cd. Good!
I copied the kernelhd-2.6.18.img from your install cd kernel directory to /boot on my hard drive partiton. Good!
I went to start nano to make an a1.conf file in /boot. Ooops...error opening terminal b!
Hmmmmm..... lets trick it... echo "Etch /boot/kernelhd-2.6.18.img etc etc" > a1boot.conf
Aha! Got it. Cat /boot/a1boot.conf shows the correct text!
So the kernelhd-2.6.18.img is in /boot and the a1.conf file is in /boot on my hard drive.

Just a note - in order to try to make echo insert a linefeed at the end of the line, I put a \n at the end, so it looked like this...

echo "Etch /boot/kernelhd-2.6.18.img /root=/dev/hda2 ide=reverse hde=noprobe hdf=noprobe l2cr=0x80000000\n" >/boot/a1boot.conf

Reboot. Of course, it goes back to AmigaOS4....that's what I have set up as in uboot by default.

Now, I thought if I booted from that drive/partition using diskboot, it would boot into debian. I am running from a hard drive connected to an sii0680ide controller, so my card is #1. When I do ide reset 1 it resets the hard drive partition and the dvd drive, which are device 0 and 1 respectively.
Here is what I do and what I see:

] diskboot 1000000 0:2 1

(I expect this to work. 2 is the partition number of my linux (LNX) partition

I see...

Loading from ide device 0, partition 2: name: LNX type uboot

Bad Magic Number

Thats all I get. This hard drive is bootable from OS4.

What do I need to do next?

Thanks,

Scott

Last edited by scabit on 10-Nov-2012 at 09:25 PM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@Hypex

Okay, I discovered another issue here.....

In order for SLB2 to see a linux partition and add it to the boot menu, you MUST go into mediatoolbox in OS4 and set the linux partition filesystem to 0x4C4E5800 LNX\00
Mine was set as I had partitioned it in linux, 0x45585403 EXT\3 !!! This will not work!

Now when I boot the AmigaOne, I see Etch as a boot option, and when I select it, I boot into Etch....

First time it locked up....

Now it works...I am typing this from Debian Iceweasel!


Thanks,

Scott

Last edited by scabit on 11-Nov-2012 at 03:49 AM.

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

] diskboot 1000000 0:2 1 (I expect this to work. 2 is the partition number of my linux (LNX) partition

That won't work because the SLB needs to load it and UBoot will just load the sectors from the partition which contains the initial blocks of an Ext2 file-system and reject it because you didn't have a UBoot image stored there!

Once you don't have a UBoot partition containing only a raw UBoot image there is no point trying to diskboot it! The SLB must pick it up.

Quote:

n order for SLB2 to see a linux partition and add it to the boot menu, you MUST go into mediatoolbox in OS4 and set the linux partition filesystem to 0x4C4E5800 LNX\00

Yes I had mentioned this setting this in post #99:
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=26623&forum=34&start=80&viewmode=flat&order=0#687089

Quote:

Mine was set as I had partitioned it in linux, 0x45585403 EXT\3 !!! This will not work!

I see. I don't know if EXT\2 would work as DOS type. But it only understands Ex2 in any case. It can read Ext3 as the format is the same but with extra journals. It's backwards compatible.

Quote:

Now it works...I am typing this from Debian Iceweasel!

Hooray! It works!

I'll be sure to use your experience in preparation for my next installer.

The newline thing and "LNX\00" are both important, good advice. But I can't (won't) check those right now, perhaps tomorrow (see other thread).

Last edited by Thematic on 11-Nov-2012 at 07:18 PM.

_________________
: AmigaOneXE (unmod.) 750FX/512 MB +stuff & AmigaOS 4.(0|1)
: A1200/68060&96MB/SCSI/EM1200-Voodoo3 & OS 3.5
: A500/1MB
: Pegasos (ff) 512 MB & MorphOS
Praise seitan.

@Thematic

I remember that. So you got it working?

@Hypex

No. There is a newline at the end of a1boot.conf and the partition type is LNX\00 but it's being ignored.

Last edited by Thematic on 13-Nov-2012 at 07:09 PM.

_________________
: AmigaOneXE (unmod.) 750FX/512 MB +stuff & AmigaOS 4.(0|1)
: A1200/68060&96MB/SCSI/EM1200-Voodoo3 & OS 3.5
: A500/1MB
: Pegasos (ff) 512 MB & MorphOS
Praise seitan.

It's probably too late to add something to the discussion here, but nonetheless:

Quote:

scabit wrote: echo "Etch /boot/kernelhd-2.6.18.img /root=/dev/hda2 ide=reverse hde=noprobe hdf=noprobe l2cr=0x80000000\n" >/boot/a1boot.conf

If you have a G4 CPU, then it may be better to disable the L2 cache. That will slow down the system, but may make it a little bit more stable.

Regarding USB2:
Yes, this is already supported by the A1 v2.6.18 Linux kernel and it should not lock up the system. Was there any kernel panic due to the USB2 device or was it a straight lockup?

BTW: Etch was always very unstable on my A1. Lenny worked much better with the same kernel version. Thus I usually recommend to do an upgrade to Lenny. The necessary Debian package repositories are still available on archive.debian.org.

Last edited by Geri on 13-Nov-2012 at 10:05 PM.
Last edited by Geri on 13-Nov-2012 at 10:05 PM.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

Quote:

It's probably too late to add something to the discussion here...

Nah...never too late! Thanks for joining in the fun!

Quote:

If you have a G4 CPU

I wish...only a lowly G3 I'm afraid.

Quote:

Regarding USB2: Yes, this is already supported by the A1 v2.6.18 Linux kernel and it should not lock up the system. Was there any kernel panic due to the USB2 device or was it a straight lockup?

Straight HARD lockups. Usually required a physical reset button push to restart.

No kernel panic...no info shown anywhere.

Quote:

BTW: Etch was always very unstable on my A1. Lenny worked much better with the same kernel version. Thus I usually recommend to do an upgrade to Lenny. The necessary Debian package repositories are still available on archive.debian.org.

Yes, I am finding Etch to be unstable for me as well. Particularly when I plug in a USB device...even just using my built in micro A1 USB ports. I was going to try turning off the internal USB ports and seeing if it becomes more stable.

As for Lenny....say I go get the CD-1 for Lenny install. How do I install this over my Etch install? It won't (I'm guessing) work with the Etch Amiga Installer linked to in posting #1 of this thread. Is there an Amiga Lenny Install CD I could use?
I'm also thinking I can't just use aptitude to do such a major upgrade could I?

Thanks!

Scott

_________________
AmigaOne uA1-c 512M RAM - Only Amiga Makes It Possible!
Check my blog AmigaOne Computing

@scabit

Quote:

scabit wrote: @Geri Quote: It's probably too late to add something to the discussion here... Nah...never too late! Thanks for joining in the fun!

It's good to see that users still give Linux a try on the A1.

Quote:

Quote: Regarding USB2: Yes, this is already supported by the A1 v2.6.18 Linux kernel and it should not lock up the system. Was there any kernel panic due to the USB2 device or was it a straight lockup? Straight HARD lockups. Usually required a physical reset button push to restart. No kernel panic...no info shown anywhere.

Do you have a USB2 card installed in your A1? I puzzle over how the ide=nodma kernel option could affect the USB2/ehci drivers, since this is an option for the IDE subsystem. The only approach I know to straight lockup Linux on the A1 is to switch rear and front USB ports in U-Boot.

Quote:

Quote: BTW: Etch was always very unstable on my A1. Lenny worked much better with the same kernel version. Thus I usually recommend to do an upgrade to Lenny. The necessary Debian package repositories are still available on archive.debian.org. Yes, I am finding Etch to be unstable for me as well. Particularly when I plug in a USB device...even just using my built in micro A1 USB ports. I was going to try turning off the internal USB ports and seeing if it becomes more stable.

I wonder, if this is a kernel or userspace problem - maybe you could only login to a virtual console, stop the hal daemon (which AFAIK is reponsible for automouting hotplug devices under Etch) and try to mount the device manually with:

Quote:

mount -t vfat /dev/sda[0-9] /mnt/

If you try it out, let me know, if it works.

Quote:

As for Lenny....say I go get the CD-1 for Lenny install. How do I install this over my Etch install? It won't (I'm guessing) work with the Etch Amiga Installer linked to in posting #1 of this thread. Is there an Amiga Lenny Install CD I could use?

You don't need a Lenny CD. You just need to change the lines in /etc/apt/sources.lists to the Debian repositories on archive.debian.org:

Quote:

deb http://archive.debian.org/debian/ lenny main contrib non-free deb-src http://archive.debian.org/debian/ lenny main contrib non-free deb http://archive.debian.org/debian-security/ lenny/updates main contrib non-free deb-src http://archive.debian.org/debian-security/ lenny/updates main contrib non-free #deb http://archive.debian.org/debian-volatile/ lenny/volatile main contrib non-free #deb-src http://archive.debian.org/debian-volatile/ lenny/volatile main contrib non-free #deb http://archive.debian.org/debian-backports/ lenny-backports main contrib non-free #deb-src http://archive.debian.org/debian-backports/ lenny-backports main contrib non-free

Note that it may be better to disable (#) the volatile and backports repositories during the upgrade.

Quote:

I'm also thinking I can't just use aptitude to do such a major upgrade could I?

Sure! apt-get or aptitude are used for this job. I think aptitude should be prefered over apt-get for the upgrade from etch to lenny. It is usually a good idea to update apt-get and aptitude first and then do a full system upgrade with apt-get dist-upgrade or aptitude full-upgrade. Please take a look at the Debian Lenny release notes, which should contain further information on how to upgrade a system from etch to lenny.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -