@Geri

Here is my lspci -nn output:

00:00.0 Host bridge [0600]: Mai Logic Incorporated Articia S Host Bridge [10cc:0660] (rev 01)
00:01.0 PCI bridge [0604]: Mai Logic Incorporated Articia S PCI Bridge [10cc:0661]
00:05.0 Multimedia audio controller [0401]: C-Media Electronics Inc CM8738 [13f6:0111] (rev 10)
00:06.0 Ethernet controller [0200]: 3Com Corporation 3c905C-TX/TX-M [Tornado] [10b7:9200] (rev 78)
00:07.0 ISA bridge [0601]: VIA Technologies, Inc. VT82C686 [Apollo Super South] [1106:0686] (rev 40)
00:07.1 IDE interface [0101]: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE [1106:0571] (rev 06)
00:07.2 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 1a)
00:07.3 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 1a)
00:07.4 Non-VGA unclassified device [0000]: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] [1106:3057] (rev 40)
00:07.5 Multimedia audio controller [0401]: VIA Technologies, Inc. VT82C686 AC97 Audio Controller [1106:3058] (rev 50)
00:07.6 Communication controller [0780]: VIA Technologies, Inc. AC'97 Modem Controller [1106:3068] (rev 30)
00:08.0 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 61)
00:08.1 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 61)
00:08.2 USB Controller [0c03]: VIA Technologies, Inc. USB 2.0 [1106:3104] (rev 63)
00:08.3 RAID bus controller [0104]: VIA Technologies, Inc. VT6421 IDE RAID Controller [1106:3249] (rev 50)
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE] [1002:5159]

Yes, hda is using the onboard IDE controller on the micro a1.
Too bad we can't get the USB2 stuff working by remapping the I/O area.
I have formatted and mounted a partition on my SATA drive, and copied files to/from it just fine.
hdparm shows the harddrive i/o speed of the SATA as 28.22 MB/s. My hda shows 29MB/s. Since I am using a SATA to IDE converter, this is about what I would expect.

Let me know what you think about getting the RAID/IDE portion of the VT6421 working on this board.

Thanks,

Scott

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

@Geri

Quote:

Okay. You mean /proc/cmdline of the U-Boot bootargs variable?

Yes, that was the one I was looking for,

Quote:

True, the cache setting (l2cr=) would make sense.

Something else to consider later.

Quote:

Right. But it is questionable, if the hardware even supports cable detection.

Good point. Even if the chip supports it doesn't mean it is wired in.

Quote:

The "boot arguments"!

Aha, a merged club mix.

Quote:

Which means we have to do more changes to the initial boot process of the debian-installer.

Being able to use a standard module args like ide_core would help to reduce changes needed to just boot args.

Quote:

On a normal installed system we could work around it by only putting the ehci_hcd module in the initramfs, which would defer the loading of the ohci/uhci_hcd modules until the root filesystem becomes available.

Possible yes. I also tested later kernels and saw same messages.

Last edited by Hypex on 31-Dec-2012 at 03:39 PM.

@Geri

Quote:

You can use "dd" to make an exact 1:1 copy of a partition, if the target partition has the same size or is bigger.

I tried that, but it doesn't seem to work. My bootable linux drive is 9.7 gig, and I am using a SATA partition that is 11.7 gig. I tried this: dd if=/dev/hda2 of=/dev/sda10 ds=4096
It seemed to work, I saw the light on the SATA card periodically flash, etc. After about 7 or 8 minutes linux locked up. Powered off, formatted the sda10 drive to ext3 (same as hda2) and tried it like this: dd if=/dev/hda2 of=dev/sda10
Same thing. coped files for 7 or 8 minutes then linux locked up. No mouse pointer, no flashing cursor...nothing. When I look at the drive sda10 many of the files from hda2 are there, but many are missing. gparted shows the drive as having 6.7 G used, where-as on my hda2 partition it shows 5.9 G used.
Why would the target drive have more data on it than the source drive?
I looked at the device info for each partition, they show the same sector size and track sizes. Anyway, for whatever reason dd doesn't do it.
I cannot find a PPC version of clonezilla anywhere, although some MAC forums reference it as if it existed at some point in time.
Interestingly, I can copy files to the sda10 drive and read them just fine. Its just the dd command that doesn't work for it.

Scott

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

@scabit

Quote:

scabit wrote: @Geri ... 00:08.0 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 61) 00:08.1 USB Controller [0c03]: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller [1106:3038] (rev 61) 00:08.2 USB Controller [0c03]: VIA Technologies, Inc. USB 2.0 [1106:3104] (rev 63) 00:08.3 RAID bus controller [0104]: VIA Technologies, Inc. VT6421 IDE RAID Controller [1106:3249]

I checked the Linux kernel source. Unfortunately the PATA (IDE) portion of the VIA6421 is only supported since kernel v2.6.25. The sata_via module also handles the PATA portion, as both functionalities (SATA and PATA) are integrated into the same PCI device.

Quote:

Too bad we can't get the USB2 stuff working by remapping the I/O area.

The strange thing is that I have EHCI working here in PCI MMIO address space (at least it doesn't lockup the machine). So I wonder a little bit, what makes the difference.

Quote:

I have formatted and mounted a partition on my SATA drive, and copied files to/from it just fine. hdparm shows the harddrive i/o speed of the SATA as 28.22 MB/s. My hda shows 29MB/s. Since I am using a SATA to IDE converter, this is about what I would expect.

Yes, that sounds reasonable for an AmigaOne.

Quote:

scabit wrote: I tried that, but it doesn't seem to work. My bootable linux drive is 9.7 gig, and I am using a SATA partition that is 11.7 gig. I tried this: dd if=/dev/hda2 of=/dev/sda10 ds=4096

You mean "bs=4096", for blocksize? I would keep its default value of 512 bytes.

Quote:

It seemed to work, I saw the light on the SATA card periodically flash, etc. After about 7 or 8 minutes linux locked up. Powered off, formatted the sda10 drive to ext3 (same as hda2) and tried it like this: dd if=/dev/hda2 of=dev/sda10

You don't need to format the sda10 partition. "dd" will overwrite any existing file system regardless of the consequences.

Quote:

Same thing. coped files for 7 or 8 minutes then linux locked up. No mouse pointer, no flashing cursor...nothing. When I look at the drive sda10 many of the files from hda2 are there, but many are missing. gparted shows the drive as having 6.7 G used, where-as on my hda2 partition it shows 5.9 G used.

Can you retry it with PIO mode for the onboard IDE controller? (with ide=nodma for kernel 2.6.18).

Quote:

Why would the target drive have more data on it than the source drive?

I wouldn't trust the file system information at all, if the copy is incomplete.

Quote:

I looked at the device info for each partition, they show the same sector size and track sizes. Anyway, for whatever reason dd doesn't do it. I cannot find a PPC version of clonezilla anywhere, although some MAC forums reference it as if it existed at some point in time.

You would need a special AmigaOne version of Clonezilla anyway, as usually all PowerPC related boot images only support PPC Macs.

Quote:

Interestingly, I can copy files to the sda10 drive and read them just fine. Its just the dd command that doesn't work for it.

It could be the heavy traffic on the PCI bus which locks up the machine. That's why I propose to try it in PIO mode.

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

- A1 Linux support -

@Hypex

Quote:

Hypex wrote: @Geri Quote: Right. But it is questionable, if the hardware even supports cable detection. Good point. Even if the chip supports it doesn't mean it is wired in.

Exactly. Maybe the cable detection signals are wired up to some general purpose input on the southbridge, but we don't know which one.

Quote:

Quote: The "boot arguments"! Aha, a merged club mix.

Quote:

Quote: On a normal installed system we could work around it by only putting the ehci_hcd module in the initramfs, which would defer the loading of the ohci/uhci_hcd modules until the root filesystem becomes available. Possible yes. I also tested later kernels and saw same messages.

I somehow managed to install Debian Squeeze with my 2.6.32 kernel here on my new A1SE. But this installation doesn't show the error message. Maybe I should also try to put the ehci/uhci/ohci_hcd modules in the initramfs to see what happens then.

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

- A1 Linux support -

@Geri

Quote:

the PATA (IDE) portion of the VIA6421 is only supported since kernel v2.6.25.

okay. so it looks like i would want to update to a new kernel. It seems I need to anyway to support newer version of Debian. I have no problem with that. The issue is, what is necessary to make a newer Debian kernel for a micro A1? Will the patches for the AmigaOne from the 2.6.18 kernel work or do they need to be completely rewritten? Is this something I could do, or am I woefully ignorant of how involved this is to do?
It seems from a later post that you have gotten later versions of the debian kernel to work on your AmigaOnes. Are these kernels that work on the micro A1 and just need to have the VT6421 driver incorporated? I don't mind trying to do this...i just need to know where to start and how to proceed. I don't expect you to do everything for me....I am willing to do it myself. But obviously its going to take a bit of learning on my part...and i will likely come up with all manner of dumb questions before its all done.
I'll try what you suggested with the dd command using the onboard IDE PIO instead of the SATA and see what happens.
As for the USB2 stuff....yes, I want to know what the difference is between what you have that seems to work and what I have that locks up immediately when a USB device is inserted. It really seems to be the EHCI portion of the handler, since I can comment out the EHCI handler kernel module from kicklayout in OS4 and use the 4 ports on the combo card as USB 1.1 just fine. so OS4 ALMOST works with this card's USB2....again it just being a matter of the EHCI portion. Can it have anything to do with the way the card is initialized (or NOT initialized) because it is done by x86 code that does not run? Can there be some conflict in the I/O memory space being used by the card. I am just trying to imagine what might be wrong.
Anyway, thanks very much for your help! I'll let you know what I find.

Scott

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

@scabit

Quote:

scabit wrote: @Geri okay. so it looks like i would want to update to a new kernel. It seems I need to anyway to support newer version of Debian. I have no problem with that. The issue is, what is necessary to make a newer Debian kernel for a micro A1? Will the patches for the AmigaOne from the 2.6.18 kernel work or do they need to be completely rewritten? Is this something I could do, or am I woefully ignorant of how involved this is to do?

The AmigaOne is officially supported by the Linux kernel since v2.6.30. Unfortunately it still requires some inofficial patches to run on every A1, but they're already available.

Quote:

It seems from a later post that you have gotten later versions of the debian kernel to work on your AmigaOnes. Are these kernels that work on the micro A1 and just need to have the VT6421 driver incorporated?

Yes, these kernels should also run on the uA1. The sata_via driver is probably already included in the 2.6.31 and 2.6.32 kernels I released for use with Debian Squeeze, as they're based on a complete Debian kernel config.

Quote:

I don't mind trying to do this...i just need to know where to start and how to proceed. I don't expect you to do everything for me....I am willing to do it myself. But obviously its going to take a bit of learning on my part...and i will likely come up with all manner of dumb questions before its all done.

That's how it works. I also bored kernel developers with a lot of stupid questions so that they started to ignore me.
But I'll try hard not to do this.

Maybe you could help us with testing Hypex's new Debian Squeeze install CD, which comes with kernel 2.6.31 and thus should include full support for the VT6421. If it is okay for Hypex, then I would upload the image to the sourceforge.net/projects/amigaone-linux website so that it is easily accessible for everyone.

@Hypex: what do you think? Is it ready for some betatesters?

Quote:

I'll try what you suggested with the dd command using the onboard IDE PIO instead of the SATA and see what happens.

Okay. Unfortunately I don't know a lot about how the uA1 behaves. I just acquired an A1SE and learned a lot about its flaws in the last days - especially when it comes to running Linux on it. My A1XE seems to be a completely different machine.

Quote:

As for the USB2 stuff....yes, I want to know what the difference is between what you have that seems to work and what I have that locks up immediately when a USB device is inserted. It really seems to be the EHCI portion of the handler, since I can comment out the EHCI handler kernel module from kicklayout in OS4 and use the 4 ports on the combo card as USB 1.1 just fine. so OS4 ALMOST works with this card's USB2....again it just being a matter of the EHCI portion.

Ouch, I completely forgot about this: you could do the same under Linux by blacklisting the ehci_hcd module. This is done by adding "ehci_hcd.blacklist=yes" to the kernel command line (bootargs). However I don't know for sure, if this works with the 2.6.18 kernel, as the module is linked into the kernel image.

Quote:

Can it have anything to do with the way the card is initialized (or NOT initialized) because it is done by x86 code that does not run?

I don't think so. I never heard of any specific BIOS initialization code for USB host controllers.

Quote:

Can there be some conflict in the I/O memory space being used by the card. I am just trying to imagine what might be wrong.

The main problem is likely the PCI MMIO access mode itself. From what I know the A1 northbridge does not reliably handle concurrent access to a memory mapped (=MMIO) PCI device (for exampe: DMA done by the USB card and the CPU tries to read a control register of the EHCI controller). However the CPUs (G3 or G4) may make a difference, too. Also Hypex sometimes reports problems with Linux on his A1XE G3 that I have never seen on my A1XE G4.

BTW: I currently have a similar problem here with a Sii3114 controller in my A1SE under Linux, which locks up the machine in DMA mode. The same controller however works fine in OS4 and also on my A1XE under Linux. Well, the A1SE is probably not the best machine to compare DMA operations, but I have to find a way to debug this (if this is possible at all ).

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

- A1 Linux support -

@Geri

Quote:

Can you retry it with PIO mode for the onboard IDE controller? (with ide=nodma for kernel 2.6.18).

Hmmm. I tried hooking up this same drive that was connected using the IDE/SATA converter to the internal microA1 IDE bus and booted from the debian boot drive on the sii0680 card. I was sure to add in the ide=nodma to my startup parameters.
When I run gparted, it doesn't see the drive on the internal bus at al, or at least it doesn;t show it as a seperate drive.
When I run blkid though, I get this:

/dev/hda1: UUID="ee0a4d53-8f21-4ae9-a598-ddee37133ee6" SEC_TYPE="ext2" TYPE="ext3"
/dev/hda2: UUID="a4df0530-3e51-4771-acf5-cd4218f35e4b" TYPE="ext3"
/dev/hda4: UUID="bd972560-a4ff-4dcd-ba52-913e30509adb" TYPE="ext2"
/dev/hda21: LABEL="MAC_DHS" TYPE="hfs"
/dev/hda22: LABEL="MAC_DHT" TYPE="hfs"
/dev/sda3: LABEL="MAC_HD" TYPE="hfs"
/dev/sda8: LABEL="MAC_HD2" TYPE="hfs"

The sda3 and sda8 drives are the mac HFS formatted partitions on my drive connected to the internal IDE bus! What is going on here? gparted sees no sda drive at all.....

Scott

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

@Geri

Quote:

Maybe you could help us with testing Hypex's new Debian Squeeze install CD, which comes with kernel 2.6.31

That sounds great!

As for the USB2 stuff, it should be noted as well that even when I use my normal USB 1.1 ports on my micro A1, debian Lenny seems to lock up quit a lot - almost always in fact. I have a USB wireless mouse for example that, if i keep plugged into the rear USB port on my micro, prevents Lenny from fully booting, or at least dies within a minute or two off coming up. Usually though it dies in the middle of starting up -in various random places.
Funny though....I seem to recall there were several days when this USB wireless mouse WAS plugged in and did NOT cause any problems with Lenny.. Hmmm. in fact I remember when I got it on Christmas day thinking "oh oh, now Debian will start locking up" but that didd't start happening until yesterday. Odd. Yes, using the same kernel.

Any ideas on that?

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

@Geri

RE MicroA1 I remember that on OS4 they disabled the driver in Kickstart because there was no apparent way to test for a floppy controller and it crashed. Can floppy be easily disabled?

Quote:

Is it ready for some betatesters?

Hang on a mo, that one is slightly obsolete and may contain bugs. I just finished a new booter I think you'll like with DMA menu. I'll check my bash scripts and then upload it.

@Hypex

What version are you thinking of uploading? just curious. Let us know when it is available, I look forward to it! It would be great to have a compressed install CD ISO that installs the kernel and basic system, then update from the basic skeleton using apt-get install or aptitude as you wish.
The reason i say that is because I have a 5 Gig limit on download per month for ALL of my internet devices. If I go to the library with my netbook or laptop, I can download ISO files (takes forever) without using up all my internet usage for the month. I notice the newer version of Debian come on DVDs! Yikes! I will end up waiting until March to order them by mail (if they re even still available) instead of taking several days of my time sitting doing nothing in the library trying to download them if that's the case.

Scott

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

@scabit

Quote:

scabit wrote: @Geri Hmmm. I tried hooking up this same drive that was connected using the IDE/SATA converter to the internal microA1 IDE bus and booted from the debian boot drive on the sii0680 card. I was sure to add in the ide=nodma to my startup parameters.

Ah, I didn't think about the sii0680 controller. You can also hook up the drive to the sii0680 controller and still run everything in PIO mode with the same kernel option.

Quote:

When I run gparted, it doesn't see the drive on the internal bus at al, or at least it doesn;t show it as a seperate drive. When I run blkid though, I get this: /dev/hda1: UUID="ee0a4d53-8f21-4ae9-a598-ddee37133ee6" SEC_TYPE="ext2" TYPE="ext3" /dev/hda2: UUID="a4df0530-3e51-4771-acf5-cd4218f35e4b" TYPE="ext3" /dev/hda4: UUID="bd972560-a4ff-4dcd-ba52-913e30509adb" TYPE="ext2" /dev/hda21: LABEL="MAC_DHS" TYPE="hfs" /dev/hda22: LABEL="MAC_DHT" TYPE="hfs" /dev/sda3: LABEL="MAC_HD" TYPE="hfs" /dev/sda8: LABEL="MAC_HD2" TYPE="hfs" The sda3 and sda8 drives are the mac HFS formatted partitions on my drive connected to the internal IDE bus! What is going on here? gparted sees no sda drive at all.....

Normally drives only get a /dev/sdxy device node, if they're handled by a scsi or libata driver, which shouldn't be the case for kernel 2.6.18 and the sii0680 or onboard VIA controller. But can you show me the output of lsmod?
Or did you setup the master/slave jumper on the IDE to SATA adapter correctly (if there is one at all)?

edit: Please show me the kernel log (dmesg output), if possible. I'm not sure, where blkid gets its information.

Last edited by Geri on 03-Jan-2013 at 08:35 PM.

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

- A1 Linux support -

@scabit

Quote:

scabit wrote: @Geri As for the USB2 stuff, it should be noted as well that even when I use my normal USB 1.1 ports on my micro A1, debian Lenny seems to lock up quit a lot - almost always in fact. I have a USB wireless mouse for example that, if i keep plugged into the rear USB port on my micro, prevents Lenny from fully booting, or at least dies within a minute or two off coming up. Usually though it dies in the middle of starting up -in various random places.

I have to admit that I don't use the USB ports a lot on my A1, thus I never experienced these problems. I only know that Linux locks up, if you setup U-Boot to use the front instead of the rear USB ports. Some day I have to check which PCI settings U-Boot messes up for this setup.

But does Linux boot normally, if you remove all USB devices. In other words is the behavior predictable in both cases?

Quote:

Funny though....I seem to recall there were several days when this USB wireless mouse WAS plugged in and did NOT cause any problems with Lenny.. Hmmm. in fact I remember when I got it on Christmas day thinking "oh oh, now Debian will start locking up" but that didd't start happening until yesterday. Odd. Yes, using the same kernel.

Which means it may not be directly related to the USB devices and drivers (and thus not predictable). As mentioned above: can you please test without any USB devices connected?

According to some OS4 developer it would also be necessary to work around a bug in the VIA interrupt controller. But I don't know which bug this should be, which negative effects it has and whether it is already fixed or not in Linux. The lockups could be related to this bug, but that is just a wild guess.

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

- A1 Linux support -

@Hypex

Quote:

Hypex wrote: @Geri RE MicroA1 I remember that on OS4 they disabled the driver in Kickstart because there was no apparent way to test for a floppy controller and it crashed. Can floppy be easily disabled?

Uhh, I don't even know, if the floppy drivers works for the 2.6.x kernels. It did with the 2.4.x kernels, but I never ported over the code changes. Anyway I guess you can disable the driver with "floppy.blacklist=yes".

edit: The floppy driver works in virtual DMA mode for kernel 2.6.x, which is activated with the "floppy=nodma" option. For example:

$ modprobe floppy floppy=nodma

Quote:

Quote: Is it ready for some betatesters? Hang on a mo, that one is slightly obsolete and may contain bugs. I just finished a new booter I think you'll like with DMA menu. I'll check my bash scripts and then upload it.

Okay, I didn't have time for testing anyway in the last two days. I'll wait for the new release then.

Last edited by Geri on 04-Jan-2013 at 10:32 PM.

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

- A1 Linux support -

Quote:

Maybe you could help us with testing Hypex's new Debian Squeeze install CD, which comes with kernel 2.6.31 and thus should include full support for the VT6421. If it is okay for Hypex, then I would upload the image to the sourceforge.net/projects/amigaone-linux website so that it is easily accessible for everyone. @Hypex: what do you think? Is it ready for some betatesters?

Here is another kernel beta tester waiting to test!

_________________
--------------------------------------
Search and Destroy
---------------------------------------

@damned

Quote:

damned wrote: Here is another kernel beta tester waiting to test!

It's more an install CD beta test. But give Hypex a little bit time to polish the thing.

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

- A1 Linux support -

@Geri

Quote:

Exactly. Maybe the cable detection signals are wired up to some general purpose input on the southbridge, but we don't know which one.

Either way that would make it hard to detect without a custom driver.

Quote:

Maybe I should also try to put the ehci/uhci/ohci_hcd modules in the initramfs to see what happens then.

You could try that yes. But I wonder if it is because I have a card reader on my internal [front] port? This might cause the UHCI driver to be loaded. Just a thought.

Good job on your installation.

@scabit

My beta installer is based on the 6.0.4 Debian CD. But a later one should work as it really only needs it for the packages. AFAIK in any case but it's here to be tested.

Like the one in this thread, the new installer needs the offical Debian CD to be present. Now, just having the first Debian CD is fine. Do you already have this to hand? If not grab it from here when you go to the library and tell us about it.

http://cdimage.debian.org/debian-cd/current/powerpc/iso-cd/debian-6.0.6-powerpc-CD-1.iso

You could also grab the second CD to cover any extra files needed and avoid using your internet. As it does seem to load stuff off the internet regardless. Possibly updated packages or needed extra files.

Now, to share a little secret, this files should peek your interest:
http://home.lizzy.com.au/dmaen/downloads/Testing/DebianInstaller.lha

A few more secrets come out.

It is an autoboot CD. Just insert and reboot! It has a menu.

If you need to you can disable DMA or IDE bus scanning for up to two controllers.

The install experience is cleaner and should present no errors and minimise user input.

Provided all is well it will install all kernel moduels and a boot menu all reaady for use. Is uses the SLB for this.

If you want to test it out or the above features excited you because they are so good give it a shot!

If you are familar with installing Linux I will leave you to it. After the CD boots and you have made your selection the install image will load and run. It will then detect the CD and load the A1 kernel into memory. It's faster doing this inside Linux.

After this it will do hardware detection and look for the offical CD. You can wait until it looks and tells you about it for one drive, or for two drives put both installer discs in and forget about it.

The install process will then continue as normal. Set up a root partition for the target install and set it as bootable. If you have swap if will find it. If not create a swap partition. Hope you have spare space to test!

After it installs it will finish up and install the kennel modules and boot menu. Or I hope! At this point I am relying on that beta label to pull me through!

I suspect Geri may be testing my installer as we speak. I hope I haven't interrupted the testing process by preempting him and jumping the gun.

Well good luck! And I hope the new installer is an experience we can all more fully enjoy.

@Hypex

Quote:

Hypex wrote: @Geri Either way that would make it hard to detect without a custom driver.

Exactly!

Quote:

Quote: Maybe I should also try to put the ehci/uhci/ohci_hcd modules in the initramfs to see what happens then. You could try that yes. But I wonder if it is because I have a card reader on my internal [front] port? This might cause the UHCI driver to be loaded. Just a thought.

I think it may depend on the order of the PCI devices, as the uhci_hcd driver is loaded a long time before the kernel even knows about your card reader. I currently have the uhci_hcd driver blacklisted, so only ehci_hcd and ohci_hcd drivers are loaded. But I would also have to check, if the OHCI PCI devices are listed before the single EHCI device.

Quote:

Good job on your installation.

Thanks! It was the most complex installation I ever did.

Quote:

Now, to share a little secret, this files should peek your interest: http://home.lizzy.com.au/dmaen/downloads/Testing/DebianInstaller.lha

Good job!

Quote:

A few more secrets come out. It is an autoboot CD. Just insert and reboot! It has a menu.

The menu looks pretty good!

Quote:

After it installs it will finish up and install the kennel modules and boot menu. Or I hope! At this point I am relying on that beta label to pull me through! I suspect Geri may be testing my installer as we speak. I hope I haven't interrupted the testing process by preempting him and jumping the gun.

You did!

I successfully booted the installer on my A1SE without any problem (with DMA off), but didn't have the time yet to start the installation, because I had to move the Linux partitions around first. Unfortunately the slb_v2 still doesn't recognize my Linux installation, which makes it hard to test your installation script. Thus I will probably also try out the installer on my A1XE tomorrow, if all goes well.

Quote:

Well good luck! And I hope the new installer is an experience we can all more fully enjoy.

Thanks again! I hope it makes it easier for users to test new kernel releases. The 2.6.32 kernel is already waiting...

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

- A1 Linux support -

@Geri

Quote:

Anyway I guess you can disable the driver with "floppy.blacklist=yes".

That's good to know if we need it. I can see an advanced options page come up.

Quote:

Exactly!

Just another patch.

Quote:

I think it may depend on the order of the PCI devices, as the uhci_hcd driver is loaded a long time before the kernel even knows about your card reader.

THat's a good point. That could be it!

Quote:

I currently have the uhci_hcd driver blacklisted

Is that for your SE?

Quote:

Thanks! It was the most complex installation I ever did.

And I was so close.

Quote:

Good job!

Thanks!

Quote:

The menu looks pretty good!

That's a good sign. I have plans to tweak it slightly.

Quote:

I successfully booted the installer on my A1SE without any problem (with DMA off), but didn't have the time yet to start the installation

I plan to test a whole install again, but wish you could tell it to procede with "unclean" target and skip the base install. Then I could see if all goes well.

Quote:

Unfortunately the slb_v2 still doesn't recognize my Linux installation,

It is hard to work around this. I wonder if ub2lb works the same as SLB. Then a special debug UB2LB coud be created.

Quote:

The 2.6.32 kernel is already waiting...

Then I can rebuild the installer with that version for the next release. Hopefully a full one!