E-Maculation

adespoton wrote:Cool... sounds like a bunch of the OpenBIOS changes required have already been completed

Also of note: http://qemu-project.org/Outreachy_2015_MayAugust -- Outreachy attempts to dovetail with GSOC, so any of you looking for more resources on implementing fixes/finding mentors, check there too.

And of course, a number of us on here would be more than happy to clarify odd bits of classic Mac hardware/OS behaviour.

Tracking updates:

• Week 2 (May 12)
  Week 3 (May 20)
  Week 6 (June 9)
  Week 7 (June 17)

Oh yes, and if anyone has the answers to the questions posed in those progress reports, drop one of the participants a line so they can work through that bit a bit faster

More tracking updates:

• Week 9 (July 2)

New here I have been stalking emacucaltion for some years and this thread has made me sign up any way so I was testing the QEMU that can boot OS9 and I found a few things... first of all Mac OS9 wont boot with the -g3beige machine option secondly I found that on both machine options Mac OS 9.04 will crash the emulator (at least the install ISO I have) thirdly when I try to boot OS X (10.2 and 10.4.6 from install disks and 10.4.11 from a pearpc.img) I get a very long Kernel panic of witch ill add some photos this happens on both mac99 and g3beige options (I couldn't find the attachment button I assume I cant do that yet so I tried the /img tag and a URL to it but the image came out very large and broke the website sort of so if some one could help me attach an image it would be appreciated ) from what I can tell the Kernel panic in 10.4.11 mentions Cuda being a thing and the rest mention the IOserial being a thing

LightBulbFun wrote:New here I have been stalking emacucaltion for some years and this thread has made me sign up any way so I was testing the QEMU that can boot OS9 and I found a few things... first of all Mac OS9 wont boot with the -g3beige machine option secondly I found that on both machine options Mac OS 9.04 will crash the emulator (at least the install ISO I have) thirdly when I try to boot OS X (10.2 and 10.4.6 from install disks and 10.4.11 from a pearpc.img) I get a very long Kernel panic of witch ill add some photos this happens on both mac99 and g3beige options (I couldn't find the attachment button I assume I cant do that yet so I tried the /img tag and a URL to it but the image came out very large and broke the website sort of so if some one could help me attach an image it would be appreciated ) from what I can tell the Kernel panic in 10.4.11 mentions Cuda being a thing and the rest mention the IOserial being a thing

Thanks for the feedback! Currently, anything below 10.2 is expected to not boot past the boot screen.

For the 10.2-10.4 issue, it will help if you include the host system you're using and any configuration / build flags you've enabled.

I've had the same long kernel panic on some of my "install from CD" setups I've done with an OS X 10.10 host and default homebrew build, but pre-built images I've made with PearPC work fine -- I'm still attempting to see if I can get my PearPC 10.1 image to actually boot. Sometimes you need to set your memory and screen resolution to the right values to get a proper install working, I've found.

Here's an updated track list; more complete, and easier to follow. Not sure if this should be stickied to the top of the thread or not.

Tracking updates:
Cormac O'Brien's log

Alexander Graf's weekly reports:

• Week 11 (July 15)
  Week 10 (July 8)
  Week 9 (July 2)
  Week 8 (June 24)
  Week 7 (June 17)
  Week 6 (June 9)
  Week 5 (June 3)
  Week 4 (May 27)
  Week 3 (May 20)
  Week 2 (May 12)

my Host system is a OS X 10.10.4 Core i7 4700MQ hackintosh... Like I said in my original post a 10.4.11 PearPC image I made that boots in stock QEMU (Ie QEMU compiled from source altho It would never boot on mac99 only g3beige this is an issue that is known even tho it was said in the 2.2.0 change log that OS X now boots with Mac99... OS X gets stuck at still waiting for root device Ie cant find the HDD its booting from) but when I try and boot that 10.4.11 image on the OS9 qemu I compiled from the web page that was posted I got a Kernel Panic Related from what I can tell to cuda this Panic happens very early on in the boot process and happens on both g3beige and mac99
thats my qemu arguments

from the work log:

- Split 26a9dfe into individual patches
- Check OS X 10.4 conflict described \[here\](http://www.emaculation.com/forum/viewto ... 125#p50822)

So it looks like they're looking into it

Ive got to say this is really exciting stuff!

I tried making a partitioned disk image, and passing that to Sheep to install 9.0.4 and see if Qemu can boot that, but no dice.

But 9.0.4 on ISO boots up and crashes, it's just amazing it even gets that far.

Has anyone tried to make a custom 9 ISO with no extensions or control panels?

Added to this, has anyone tried booting Classic on 10.2/10.3 yet?

adespoton wrote:Added to this, has anyone tried booting Classic on 10.2/10.3 yet?

as I said above... the OS9 QEMU Will not boot any Version of OS X (I have tried 10.1 10.2 and 10.4 all of witch will Kernel panic) feel free to try your self tho heh

And I just had to try...



./bin/qemu-system-ppc -M g3beige -hda 90.disk -cpu G3 -cdrom disk1.iso -boot d -prom-env 'boot-args=-v' -bios openbios-qemu.elf

with the new bios, without g3beige or Mac99 same thing every which way basically... Oh well I'm not complaining, but with all the CUDA messages in the console it is no doubt faults in the fresh new emulation, but once it's fleshed out I imagine the overall MacOS/OS X situation will improve dramatically!

Just did some more testing... and I found that 10.5.8 Does not Kernel Panic on the new OS9 QEMU how ever it does hang on still waiting for root device on the mac99 option like OS X has on stock qemu (10.5.8 wont boot on the g3beige since its missing drivers for that machine model) Neozeed yep thats the same Kernel panic I got... (apart from 10.4.11 where I got a KP that mentioned cuda it self)

Quick question on the ongoing Mac OS 9.X work: there's no need to use ROM files, I assume?

And out of curiosity: I never really read about how it works, but I assume QEMU uses OpenBios and whenever necessary OS9 uses the New World ROM file saved on disk. Is this correct? How did the real machines would boot both OSX and OS9 back in the day?

Thanks! =D

sentient06 wrote:Quick question on the ongoing Mac OS 9.X work: there's no need to use ROM files, I assume?

And out of curiosity: I never really read about how it works, but I assume QEMU uses OpenBios and whenever necessary OS9 uses the New World ROM file saved on disk. Is this correct? How did the real machines would boot both OSX and OS9 back in the day?

Thanks! =D

No, OpenBIOS attempt to find the "Mac OS ROM" file in system folder that is blessed , and runs the code contained, which then boot Mac OS 9.

julialy wrote:No, OpenBIOS attempt to find the "Mac OS ROM" file in system folder that is blessed , and runs the code contained, which then boot Mac OS 9.

Hey, that's what I meant by "New World ROM file saved on disk". xD

That's interesting, is it correct to assume OpenBIOS has some code to run through the HFS tree?

Oh man, it will be really cool when they manage to get this one working! I think I'll just forget about that for a month and come here for the surprise. =D

So...

Any more updates on this? Any more weekly reports? Anything?

DANADAX wrote:So...

Any more updates on this? Any more weekly reports? Anything?

I searched on the mailing list archive, seems there is no more reports.
http://lists.nongnu.org/archive/html/qe ... l/2015-07/

They are working on it, but I one of the mentors seems to be on holidays and progress is slow due to some issues.

Best,
Cat_7

I was passed this fun info on how to enter the nanokernel debugger. I emailed the GSoC dev's but got no reply... SO in case...

Fire off Qemu but don't let it automatically boot, you need to enter some stuff


Namely:
you get the number by adding up the choices in the above screen shot.

If all goes well...





So perhaps it'll help, perhaps not.

hi neozeed, what repo did you build openbios with, and what repo did you get the qemu code with? I keep getting "no valid state has been set by load or init-program" in openbios when trying to use my mac os 9 cd. What version of mac os 9 did you use?

dmbtech wrote:hi neozeed, what repo did you build openbios with, and what repo did you get the qemu code with? I keep getting "no valid state has been set by load or init-program" in openbios when trying to use my mac os 9 cd. What version of mac os 9 did you use?

All the details are here:
http://c-obrien.org/qemu-os9/testing/

He even has an OS 9.2 CD, which works best. I have a 9.04 ISO which doesn't have the full featured nanokernel debug features, 9.2 "works" best.. although it just crashes ultimately.



But you get the idea.

One sec... you've got the networking extension running prior to crash there. Before it wasn't getting that far. Does it get further with extensions disabled?

the keyboard doesn't work so you can disable extensions...

Someone needs to make a custom 9 cd with no extensions

Better for Cormac to make keyboard handling the next task on the list :D

Hi,
The keyboard does react, and one can boot with extensions off from the cd image. But the crash happens anyway.

Best,
Cat_7

Following O'Brien's guide in http://c-obrien.org/qemu-os9/testing/, I've reached the following error:

ERROR: pkg-config binary 'pkg-config' not found

It appears about five seconds after entering

../qemu/configure --target-list=ppc-softmmu --prefix=$HOME/os9/qemu-install

I haven't used brew or anything to install qemu. Might that be the problem? Any advice I can follow? Thanks.

Edit: By the way, shouldn't ../qemu/configure --target-list=ppc-softmmu --prefix=$HOME/os9/qemu-install be expressed as ../qemu/configure --target-list=ppc-softmmu --prefix=$HOME/os9-test/qemu-install? I mean, Cormac O'Brien named the OS 9 test folder as "os9-test", didn't he?