Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[MacOS Plus]

Here is the telnet log from the last testing nanopico asked me for:

Code: [Select]

Last login: Mon May 16 13:58:30 on ttys000
Hyper-PowerBook-G4:~ administrator$ telnet 192.168.0.250
Trying 192.168.0.250...
Connected to 192.168.0.250.
Escape character is '^]'.
 ok
0 > printenv
-------------- Partition: common -------- Signature: 0x70 ---------------
little-endian?          false                false


Open call - nvram already open
 ok
0 > devalias
pci0                /pci@f0000000
agp                 /pci@f0000000
pci1                /pci@f2000000
pci2                /pci@f4000000
ui2c                /uni-n/i2c
ui2c-serial         /uni-n/i2c/cereal
keyboard            /pseudo-hid/keyboard
mouse               /pseudo-hid/mouse
sound               /pseudo-sound
eject-key           /pseudo-hid/eject-key
nvram               /nvram
enet                /pci@f4000000/ethernet
fw                  /pci@f4000000/firewire
bridge0             /pci@f2000000/@d
pci                 /pci@f2000000/@d
bridge              /pci@f2000000/@d
usb0                /pci@f2000000/@d/usb@8
usb1                /pci@f2000000/@d/usb@9
fwx                 /pci@f2000000/@d/firewire
enetx               /pci@f2000000/@d/ethernet
mac-io              /pci@f2000000/@d/mac-io@7
mpic                /pci@f2000000/@d/mac-io@7/interrupt-controller
hd                  /pci@f2000000/AppleKiwi@15/ata-6@0/disk@0
cd                  /pci@f2000000/@d/mac-io@7/ata-4@1f000/disk@0
zip                 /pci@f2000000/@d/mac-io@7/ata-3@20000/disk@1
ide0                /pci@f2000000/@d/mac-io@7/ata-3@20000/disk@0
ide1                /pci@f2000000/@d/mac-io@7/ata-3@20000/disk@1
ultra0              /pci@f2000000/AppleKiwi@15/ata-6@0/disk@0
ultra1              /pci@f2000000/AppleKiwi@15/ata-6@1/disk@0
scca                /pci@f2000000/@d/mac-io@7/escc/ch-a
sccb                /pci@f2000000/@d/mac-io@7/escc/ch-b
ki2c                /pci@f2000000/@d/mac-io@7/i2c
ki2c-serial         /pci@f2000000/@d/mac-io@7/i2c/cereal
via-pmu             /pci@f2000000/@d/mac-io@7/via-pmu
rtc                 /pci@f2000000/@d/mac-io@7/via-pmu/rtc
pi2c                /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c
ultra2              /pci@f2000000/AppleKiwi@1b/ata-6@0/disk@0
ultra3              /pci@f2000000/AppleKiwi@1b/ata-6@1/disk@0
bridge1             /pci@f2000000/@11
lm87                /uni-n/i2c/lm87
kgpio               /pci@f2000000/@d/mac-io@7/gpio
keyswitch           /pci@f2000000/@d/mac-io@7/gpio/@c
monitor             /pci@f2000000/@d/mac-io@7/gpio/@12
indicator           /pci@f2000000/@d/mac-io@7/gpio/@15
led                 /pci@f2000000/@d/mac-io@7/gpio/@20
9554-0              /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c/@140
9554-1              /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c/@142
9554-2              /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c/@144
9554-3              /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c/@146
9554-M              /pci@f2000000/@d/mac-io@7/via-pmu/pmu-i2c/@148
last-boot           /pci@f4000000/ethernet@f
screen              /pci@f0000000/ATY,Rage128Ps@10
 ok
0 > dev /pci@f2000000/AppleKiwi@15 .properties
vendor-id             

Open call - nvram already open
 ok
0 > dev /pci@f2000000/AppleKiwi@15/ata-6@0 .properties
name                    ata-6

Open call - nvram already open
 ok
0 > dev /pci@f2000000/AppleKiwi@1b .properties
vendor-id             

Open call - nvram already open
 ok
0 > dev /pci@f2000000/AppleKiwi@1b/ata-6@0 .properties
name                    ata-6

Open call - nvram already open
 ok
0 > dev /pci@f2000000/pci-bridge@d/mac-io@7/ata-4@1f000 .properties
name                    ata-4

Open call - nvram already open
 ok
0 > dev /pci@f2000000/pci-bridge@d .properties
vendor-id             

Open call - nvram already open
 ok
0 > dev /pci@f2000000/pci-bridge@11 .properties
vendor-id             

Open call - nvram already open
 ok
0 >

  This is mainly commands listing devices.  Note the legacy presence of an alias for a Zip drive device.  I actually tried once in the past to hook up a Zip drive to the cable in one of the drive caddies.  For some reason this made the ATA controller VERY unhappy!

[MacOS Plus]

  I'd been chatting with nanopico about the situation with the ATA controller(s).  I went back and tested a boot using ELN's original commands, but this time I used an external firewire DVD drive on a rear port instead of the the internal optical drive.

  When I had first tried this configuration months ago I didn't have the benefit of the telnet logging to know exactly what the system was doing while showing only a grey screen.  Now it is clear that a lot more is going on when using the firewire drive to boot the CD versus the internal drive.  So much more that it exceeded the character limit in the messaging window of the forum when I tried to send it to nanopico!  I had to break it up into multiple messages.  As such I won't post it here in whole or part for the moment.  It would be better just to have nanopico comment on it and show relevant log snippets if anything.

  I don't know where that gets us but it certainly was interesting!

[nanopico]

The stuff MacOS Plus generated is daunting to say the least.  But I am going to go over it tonight.
This plus what ELN and I have deciphered with the ROM/trampoline stuff is getting interesting.
I'll update that other post about the trampoline code tonight too.

Oh so many cool things happening.

[nanopico]

If someone has a Sonnet Tempo RAID ATA100 and want's to help please contact me.
We need to locate one for some comparison testing.  If you are willing to sell me or MacOS Plus one awesome.  If not and you are willing to just get us some information from it that would be awesome too.

[nanopico]

Just got a g4 xserve.
It has a few issues that I will have to resolve such as a flakey cd drive (which MacOS Plus was kind enough to point out to me prior to it arriving that this is common), and no hard drives.  The drive sleds are there but I just need to actually install disks.
One of the USB ports is physically destroyed.

This should make things move a little faster. 

[DieHard]

One of the USB ports is physically destroyed.

Before powering up, remember that manged USB ports often create a hard short that will not allow the unit to post,  so pick out all the broken metal pieces with small needle nose pliers and then you can fill the square hole will hot glue from a crafts gun, this stops any accidental use of a bad USB port that could short or damage the MB

[nanopico]

One of the USB ports is physically destroyed.

Before powering up, remember that manged USB ports often create a hard short that will not allow the unit to post,  so pick out all the broken metal pieces with small needle nose pliers and then you can fill the square hole will hot glue from a crafts gun, this stops any accidental use of a bad USB port that could short or damage the MB

Thank you for the insight, that's great advice I will definitely follow.

[MacOS Plus]

Slightly off-topic but in reference to the last couple posts:

  It's really annoying this machine only has two USB ports.  I couldn't understand why they chose not to put one in the front panel to compliment the extra firewire port at least.  I'm considering adding a USB PCI card and some sort of front-bay port panel to my primary OS X Xserve rather than risking dealing with a potentially flaky USB hub.  I'm mostly interested in having another port or two for maintenance purposes and for temporary attachment of USB storage keys or external hard drives.

  While we're talking USB on the Rev.1 Xserve, I've successfully installed a bluetooth USB transceiver and an Apple bluetooth keyboard (the 3-cell edition).  It works just fine in 10.5.  I've never been a big fan of bluetooth mice but I suppose I could add one in order to free-up the other USB port.

  I realize the bluetooth stuff won't carry over to the OS-9-on-Xserve project, but I'm trying to make the best OS X config possible on my first Xserve because it will make for a very useful server which will help a lot with many other projects.  My second Xserve machine is getting the 'OS 9 treatment', or more a 'beat-down' so far.

[nanopico]

Currently working on my xserve.

Got passed the original boot error we were seeing.  Now hitting another one, but may get passed that soon too.

[nanopico]

You can now add the xserve to the list of bootable machines.





Pardon the crappy pictures  Best I could do at the moment.
I got a video too that I will try to get up.

[MacOS Plus]

Holy sh*t!  What do I have to do to try this when I get home tonight?!?

  We spent a while 'battling' the ATA controller - did you finally manage to fully disable that blasted "Kiwi"?  Did you have to disable anything else?  I can hardly wait to open up System Profiler on my machine to see how all the hardware reports itself.

[MacTron]

You can now add the xserve to the list of bootable machines.

WOW!

[nanopico]

Holy sh*t!  What do I have to do to try this when I get home tonight?!?

  We spent a while 'battling' the ATA controller - did you finally manage to fully disable that blasted "Kiwi"?  Did you have to disable anything else?  I can hardly wait to open up System Profiler on my machine to see how all the hardware reports itself.

The damn kiwi.  Yes we were on the right track. After a whole lot of typing I figured it out.
So once it boots you have not hard disks what so ever, after getting over the ata controller it booted with no other issues or changes.

So big step just have to get the controller to work though so it's useful.

On mine it did boot really slow but I equate that to the cd drive being stupid.
So here is what you do.

Code: [Select]

" /pci@F2000000/AppleKiwi@15" find-package drop delete-node
" /pci@F2000000/AppleKiwi@1b" find-package drop delete-node
boot cd:,\\:tbxi
The last line is how I booted off the cd.  How ever you did it before will work too.

this deletes it and takes care of removing all the cross references (which is a lot) to other hardware.

[MacOS Plus]

That's really interesting if you could boot from the internal optical drive over the ATA4 bus because that means my idea to put an SSD in a conversion carrier should work too.  When I get them in the mail I'm going to try that, made easier by the fact I should be able to boot the CD from my external FW DVD drive.

[nanopico]

That's really interesting if you could boot from the internal optical drive over the ATA4 bus because that means my idea to put an SSD in a conversion carrier should work too.  When I get them in the mail I'm going to try that, made easier by the fact I should be able to boot the CD from my external FW DVD drive.

Yup that will work.

I have now added to my list of things to do is to get that controller working in 9. 
I have some ideas, but I've got some research to do first.

[MacOS Plus]

  I was so excited about trying this out I went straight to the machine after work.  I booted the original version of the "Universal 9.2.2" disc from the external FW DVD drive.  In an ironic twist (because of what I'd just been saying about ATI), when it got to the desktop it seemed to get hung up, but pressing "cmd-opt-esc" popped up a force-quit dialog for "ATI Video Accelerator".  After quitting this it quickly finished loading to the desktop successfully!!!!!!!

  The Rage128 Pro AGP video card I'm using wasn't getting along with that one extension, but at least that is an entirely separate and isolated issue.  I can sort that out later.  I'm still getting proper video and full access to resolutions.  The max for this card on VGA output is 1280x1024 at millions of colours.  (It has 16MB VRAM.)

  Next I ran ASR to install to an external SSD on another firewire port.  This ran quite smoothly.  Rebooted to the SSD with extensions off - WORKED!  Cleaned up extensions and control panels (pulled all the ATI ones for now), rebooted - WORKED!!!!!!  IT'S ALIVE!!!!!

  Not only is this machine very fast, even without the video acceleration and a very old video card, but the multiprocessing is WORKING!!!!!!!!  Built-in ethernet works, time server sync works, file sharing enabled (I will test this further later), and internet access WORKS!!!  I managed to download Classilla 9.3.3 using Internet Explorer (obviously I could have just copied it from somewhere else, but that wasn't the point).

  Now that I have a usable and fast browser on this Xserve I thought it would be most appropriate that I post this message directly from the keyboard of this machine.  History, folks!  This is so AWESOME!!!!!!!!!  I feel honored!  This is a huge dream come true for me personally because I originally bought this machine years ago with the hope that some day this would actually work.  I never thought that this long after the official dropping of OS 9 support I would be sitting here running this FOR REAL!  I don't know what else to say - I'm ecstatic!!!!!!!!!

  I will continue testing to see if anything else doesn't work aside from the main ATA controller.  This machine is so simple there isn't really much left that could go wrong.  We'll continue to work at this and keep you all posted!

P.S.  Just before I settle down and go to bed I have one thing to add.  I decided to swap out the AGP Rage128 Pro and AGP riser for a Radeon 7000 PCI 32MB and the PCI riser just to see how a later card would work.  The Radeon plays nice with all the extensions, seems to have acceleration working, and gives me access to 16:9 ratio resolutions for the LCD I have connected.  Now I have a crisp, native 1440x900 and better scolling.  When I locate my Cinebench installer I will give it a run at benchmarking the performance.

[nanopico]

  I was so excited about trying this out I went straight to the machine after work.  I booted the original version of the "Universal 9.2.2" disc from the external FW DVD drive.  In an ironic twist (because of what I'd just been saying about ATI), when it got to the desktop it seemed to get hung up, but pressing "cmd-opt-esc" popped up a force-quit dialog for "ATI Video Accelerator".  After quitting this it quickly finished loading to the desktop successfully!!!!!!!

  The Rage128 Pro AGP video card I'm using wasn't getting along with that one extension, but at least that is an entirely separate and isolated issue.  I can sort that out later.  I'm still getting proper video and full access to resolutions.  The max for this card on VGA output is 1280x1024 at millions of colours.  (It has 16MB VRAM.)

  Next I ran ASR to install to an external SSD on another firewire port.  This ran quite smoothly.  Rebooted to the SSD with extensions off - WORKED!  Cleaned up extensions and control panels (pulled all the ATI ones for now), rebooted - WORKED!!!!!!  IT'S ALIVE!!!!!

  Not only is this machine very fast, even without the video acceleration and a very old video card, but the multiprocessing is WORKING!!!!!!!!  Built-in ethernet works, time server sync works, file sharing enabled (I will test this further later), and internet access WORKS!!!  I managed to download Classilla 9.3.3 using Internet Explorer (obviously I could have just copied it from somewhere else, but that wasn't the point).

  Now that I have a usable and fast browser on this Xserve I thought it would be most appropriate that I post this message directly from the keyboard of this machine.  History, folks!  This is so AWESOME!!!!!!!!!  I feel honored!  This is a huge dream come true for me personally because I originally bought this machine years ago with the hope that some day this would actually work.  I never thought that this long after the official dropping of OS 9 support I would be sitting here running this FOR REAL!  I don't know what else to say - I'm ecstatic!!!!!!!!!

  I will continue testing to see if anything else doesn't work aside from the main ATA controller.  This machine is so simple there isn't really much left that could go wrong.  We'll continue to work at this and keep you all posted!

P.S.  Just before I settle down and go to bed I have one thing to add.  I decided to swap out the AGP Rage128 Pro and AGP riser for a Radeon 7000 PCI 32MB and the PCI riser just to see how a later card would work.  The Radeon plays nice with all the extensions, seems to have acceleration working, and gives me access to 16:9 ratio resolutions for the LCD I have connected.  Now I have a crisp, native 1440x900 and better scolling.  When I locate my Cinebench installer I will give it a run at benchmarking the performance.

For the most part the only issues that should be seen are ones that would have occurred with the MDD's. 
I believe MacTron originally pointed out they are pretty much the same.

[MacOS Plus]

  This morning I installed 1.5GB of RAM and ran the ATI 2005 updater to bring the extensions to their final revisions.  Booting also works from the front-panel firewire port.

  Next I thought I would try seeing just how close this machine really is to an MDD by trying to put it to sleep and wake it.  Selecting sleep brought up a dialog warning that it couldn't proceed until the case-closed switch was engaged.  This is really cool because it recognizes that function through the power management unit and the front panel controls.  Since I was running with the case open I simply held down the switch with my finger and tried again.  It went to sleep properly!  Then I pressed the mouse button to wake it again - it woke!!!  (It also turns out letting go of the switch will wake it too, so the system can respond to numerous inputs for wake.)  What didn't work was that the sleep and wake cycle lost the firewire hard drive creating a condition that required a restart.  At first I thought this was because it had been running on FW bus power and didn't initialize properly when the power came back.  So I tried powering it separately but it still trips up the drive.  I don't think this is an incompatibility issue with the Xserve - it's simply what would have occurred potentially on other machines too.  Not a deal breaker because I normally don't sleep any of my OS9 machines.  Once I get the conversion carrier for the optical drive bus so I can install the SSD there I have a feeling this issue will go away.

  One last note before I leave for the day - Not really a surprise, but the front panel CPU activity guages don't do anything.  I don't know what data source actually drives them.  They may very well never work but we'll see.

[ELN]

Congratulations both of you. Beautiful work!

[MacOS Plus]

  Thanks to member Mat I was able to quickly download Cinebench 2003 for Mac so I could test the functional state of the ATI drivers with the Radeon 7000 PCI card in the OS 9 Xserve.  Running the test proved two things - The acceleration is operational as it appeared to be, and second that the Radeon 7000 kinda sucks (which I knew anyway).  The important thing is it works and I can move on to a better card when I get hold of one.  Here is the report output from Cinebench:

Code: [Select]

CINEBENCH 2003 v1
****************************************************

Tester           : MacOS Plus

Processor        : Xserve G4 Rev.1
MHz              : 1.0GHz 2MB L2
Number of CPUs   : 2
Operating System : Mac OS 9.2.2

Graphics Card    : Radeon 7000 PCI 32MB
Resolution       : 1440x900
Color Depth      : Millions

****************************************************

Rendering (Single   CPU): 94 CB-CPU
Rendering (Multiple CPU): 177 CB-CPU

Multiprocessor Speedup: 1.88

Shading (CINEMA 4D)                : 113 CB-GFX
Shading (OpenGL Software Lighting) : 321 CB-GFX
Shading (OpenGL Hardware Lighting) : 242 CB-GFX

OpenGL Speedup: 2.84

****************************************************
  The super-cool thing is that the test was taking advantage of the multiprocessing and reporting the afforded performance boost!  This was also an important proof of concept on the Xserve platform under OS 9 - that the multiprocessing isn't an non-functional facade.  It reports and works properly!

  Also as part of this test I downloaded Cinebench with the OS X Xserve first so I could test file sharing.  I was able to mount the X-drive on the 9-machine and copy the file over.  Cool - file sharing works!  I was kinda cute really watching two Xserves chatting with one another.  Not something may people would have ever witnessed.  Two roaring monsters they are, but family all the same!

Thanks again, Mat!