322 replies to this topic

[shoarthing]

Tuesday 3rd November 2009

Guide converted to fully illustrated & updated *.pdf

Please note: this guide assumes you have access to a working OSX installation.

*.pdf is attached to this post together with all required installation files

Please note this guide is *not* for the new ION 330 HD, or Pro versions

Attached Files

•  Workflows.zip   72.84K   2503 downloads
•  ION_330_Install_150d.zip   1.25MB   3075 downloads
•  Asrock_ION_330_OSX86_Leopard_10.5_Guide.pdf   1.44MB   6169 downloads

[DUally]

Wow RESPECT !
All this stuff almost out of the box working on OSX that's awsome !!
what's the issue with audio ? is there no audio at all ?

[DrPoyson]

2 issues for me (instructions followed twice with same result, will try on different disk and osx dmg file later tonight if I get time)

1 - Can't boot without cpus=1
2 - USB drives wont mount.  Nothing happens for a minute or two and then the gfx go all strange (windows move but are transparant, difficult to describe) when I move folders or the machine will not respond to anything (can move mouse though).

I'm certainly not ruling out having cocked something up :-) - hence going to try again from scratch

[shoarthing]

DrPoyson - Hi - speaking as the original King of Pebcak, any hassles are just as likely to be due to me.

Edit 3rd Nov 2009: Present USB1/2 status is sorted w/ BIOS 150d & above guide

[ktfn]

Cool! Does the graphic card support hardware acceleration of HD-playback in OS X? Can you post some screens of activity monitor (cpu load) when playing 720p and 1080p? How does it work?

[UN_WIRED]

First off I have to say EXCELLENT write up shoarthing!!!!!!  Bravo, this is just what the "dr" no pun intended ordered!  

I followed it 100% and before I had read up on this new thread, I was actually doing a similar install, but rebooting my hacked 1000he and just selecting the external drive from the ION, install went similar without issue, but again I have the same issue as before and one of the same ones Drpoyson is having also.  For the record, I am running 1.50 bios

Everything seems to work as it should at this point (video, LAN, USB 2.0 once they are plugged in after boot, are all good) but not unlike Poyson, I still have to use the CPUS=1 flag.  A couple of possibilities of issues we may be having:

1.) I see Shoar, you have 4GB of ram where I only have 2GB installed, will the dsdt file need to be changed to reflect this difference?

2.)  When did you purchase your hardware?  As we all know, vendors can make rev changes to hdwe without notice and there may have been a "Bug fix" in the CPU maybe that changed addressing or something alike?

Checkout the screen shot attached, looks just like yours, except the 4GB of ram.  Again, great write up!

PS-  I have a Belkin F5D7050 also that I have used on hack's before, but once I updated this latest install to 10.5.7, it wont get a DHCP address, so will be working on this issue also.

Attached Files

•  leo_ion_1.tiff   601.65K   354 downloads

[shoarthing]

UN_WIRED, on Aug 7 2009, 08:29 AM, said:

First off I have to say EXCELLENT write up shoarthing!!!!!!  Bravo, this is just what the "dr" no pun intended ordered!  

I followed it 100% and before I had read up on this new thread, I was actually doing a similar install, but rebooting my hacked 1000he and just selecting the external drive from the ION, install went similar without issue, but again I have the same issue as before and one of the same ones Drpoyson is having also.  For the record, I am running 1.50 bios

Everything seems to work as it should at this point (video, LAN, USB 2.0 once they are plugged in after boot, are all good) but not unlike Poyson, I still have to use the CPUS=1 flag.  A couple of possibilities of issues we may be having:

1.) I see Shoar, you have 4GB of ram where I only have 2GB installed, will the dsdt file need to be changed to reflect this difference?

2.)  When did you purchase your hardware?  As we all know, vendors can make rev changes to hdwe without notice and there may have been a "Bug fix" in the CPU maybe that changed addressing or something alike?

Checkout the screen shot attached, looks just like yours, except the 4GB of ram.  Again, great write up!

PS-  I have a Belkin F5D7050 also that I have used on hack's before, but once I updated this latest install to 10.5.7, it wont get a DHCP address, so will be working on this issue also.

. . . hmm: have attached a matched set of BIOS 1.50c [thanks to DrPoyson] & a CPU-alias-edited dsdt.aml made from a Linux dump when running this same BIOS.

*.zip includes the settings I use - plus see from above my:  com.apple.boot.plist

I can't see how the quantity of RAM would affect things. I have a Ralink 2870-chipset 'N' widget which appears to work OK - a neat thing made by Edimax.

I bought my ION 330 in late July: manufacture date is late June. I cannot see a Rev code.

Do you see the same USB issues in OSX 10.6?

[UN_WIRED]

shoarthing, on Aug 7 2009, 02:03 PM, said:

. . . hmm: have attached a matched set of BIOS 1.50c [thanks to DrPoyson] & a CPU-alias-edited dsdt.aml made from a Linux dump when running this same BIOS.

*.zip includes the settings I use - plus see from above my:  com.apple.boot.plist

I can't see how the quantity of RAM would affect things. I have a Ralink 2870-chipset 'N' widget which appears to work OK - a neat thing made by Edimax.

I bought my ION 330 in late July: manufacture date is late June. I cannot see a Rev code.

Do you see the same USB issues in OSX 10.6?

Dont know about the docs purchase timeframe, but mine is dated 7/14/09.  I know these are going out the door as fast as the "Egg" can get them in.

Not much luck with snow last night either, it also imploded when being built on the same 1000he.

When I get  a chance, I am going to try and build the drive today from my Mac Mini and see what I get.  Also interesting, the 10.5.7 combo update went fine last night, but that after I applied the 10.5.8 patch, it imploded and is back to stopping the boot at the headers line again.

[shoarthing]

UN_WIRED, on Aug 7 2009, 03:05 PM, said:

Dont know about the docs purchase timeframe, but mine is dated 7/14/09.  I know these are going out the door as fast as the "Egg" can get them in.

Not much luck with snow last night either, it also imploded when being built on the same 1000he.

When I get  a chance, I am going to try and build the drive today from my Mac Mini and see what I get.  Also interesting, the 10.5.7 combo update went fine last night, but that after I applied the 10.5.8 patch, it imploded and is back to stopping the boot at the headers line again.

. .  the only thing I can think of which will differ between mine and other installs is that the moment I got into OSX I edited a copy of UUID.kext to show the ION 330's en0 MAC address - & then always add this to /Extra/Extensions & within /Extra/Extensions.mkext.

Apple have been increasingly using this UUID - in 10.5.8 it is now shown in under 'Hardware' in Profiler.

[UN_WIRED]

I have to admit that I am as stumped as you are shoar   I just got done reloading the drive with a true vanilla install from my Mac Mini with 10.5.6 retail.  Upon the first boot I got the expected USBEHCI errors, but then it stopped at the mDNSreponder load and did not do any further.  I "Believe" that UUID should not be an issue with the problem Poyson and I are having (Not able to load dsdt.aml). It shows up with a UUID OK during the boot and when it goes to access the root device.

In any case, sounds like Poyson and myself have the same issue with the DSDT file, and would be curious to know if anyone else with this box is having the same issue as well.  Shoar, any chance you still have the old ram lying around and can switch back to 2GB and try it?  If we eliminate that issue. At that point, would it be a safe assumption to say there is a rev difference somewhere on the board maybe?

Is there anyway to capture of the initial load script that runs by, a way to log it maybe via Chameleon?  It flys by so fast, I cannot see what is going on at the tail end of it when the DSDT.aml file should be loading.

Another thought, my curiosity as peeking, what if I restore a vanilla image from my mac mini to the ION?  Will it work?  We shall find out later tonight.  For not, going to have a beer and rest the brain a bit!




WE NEED MORE PEOPLE WITH ASROCK IONS!!!!  If you have one, please test shoarthings  a guide at the top of this post.  Post a reply with your findings, bios rev, ram (2 or 4gb) and anything else you can think of!  Thanks in advance!

QUICK UPDATE:  Just finished installing the 10.5.7 combo from my mini to my ION drive and everything came up as expected (Video, LAN), but still no matter what the  &!%@$  I try, it will not load that DSDT file!  I wish I could look at a Chameleon log to see what it is doing as it flashed by.   Does KEF and the team have that functionality in RC2?

[shoarthing]

. .  well, this dsdt malarkey is fairly simple: if you want to/feel you need to make your own, just d/l a minimal Linux *.iso - because I've used it a lot & it is under 100MB, I generally totter off & get the latest version of Parted Magic.

Burn to CD; use it to boot up the machine you need a dsdt for [use default option - nice & fast], when at the desktop fire up a terminal & type:

cat /proc/acpi/dsdt > dsdt.aml

There's a file manager, & a mounter [both easily to hand]: pop in a USB stick, wait till it mounts, copy dsdt.aml over to it from /

. . . when back in OSX you'll need iasl [PM me if you don't have a modern version]: put a copy in the same folder as a copy of your dsdt.aml, cd an instance of Terminal to the same folder:

./iasl -d dsdt.aml

(decompiles to dsdt.dsl) - this can be opened in any editor.

At the simplest level; you just remove the 4x Processor 'Alias' lines right at the beginning.

(this is what stops you at buffers/headers)

	Scope (_PR)
	{
		Processor (P001, 0x01, 0x00000810, 0x06) {}
		Processor (P002, 0x02, 0x00000000, 0x00) {}
	}

. . . for both physical cores; or

  

 Scope (_PR)
	{
		Processor (P001, 0x01, 0x00000810, 0x06) {}
		Processor (P002, 0x02, 0x00000000, 0x00) {}
		Processor (P003, 0x03, 0x00000000, 0x00) {}
		Processor (P004, 0x04, 0x00000000, 0x00) {}
	}

. .  if you're optimistic about being able to use hyperthreading.

./iasl -tc dsdt.dsl

(recompiles)

. . . & there you are  

[UN_WIRED]

shoarthing, on Aug 7 2009, 11:55 PM, said:

. .  well, this dsdt malarkey is fairly simple: if you want to/feel you need to make your own, just d/l a minimal Linux *.iso - because I've used it a lot & it is under 100MB, I generally totter off & get the latest version of Parted Magic.

Burn to CD; use it to boot up the machine you need a dsdt for [use default option - nice & fast], when at the desktop fire up a terminal & type:

cat /proc/acpi/dsdt > dsdt.aml

There's a file manager, & a mounter [both easily to hand]: pop in a USB stick, wait till it mounts, copy dsdt.aml over to it from /

. . . when back in OSX you'll need iasl [PM me if you don't have a modern version]: put a copy in the same folder as a copy of your dsdt.aml, cd an instance of Terminal to the same folder:

./iasl -d dsdt.aml

(decompiles to dsdt.dsl) - this can be opened in any editor.

At the simplest level; you just remove the 4x Processor 'Alias' lines right at the beginning.

(this is what stops you at buffers/headers)

	Scope (_PR)
	 {
		 Processor (P001, 0x01, 0x00000810, 0x06) {}
		 Processor (P002, 0x02, 0x00000000, 0x00) {}
	 }

. . . for both physical cores; or

  

 Scope (_PR)
	 {
		 Processor (P001, 0x01, 0x00000810, 0x06) {}
		 Processor (P002, 0x02, 0x00000000, 0x00) {}
		 Processor (P003, 0x03, 0x00000000, 0x00) {}
		 Processor (P004, 0x04, 0x00000000, 0x00) {}
	 }

. .  if you're optimistic about being able to use hyperthreading.

./iasl -tc dsdt.dsl

(recompiles)

. .  a there you are.

I am still 100% stumped as to why this works on your machine and not mine!  Any chance you were able to go back your original 2GB of ram?  This COULD (not saying it will) play a factor as the CPU has to be able to initialize all the ram it sees and if you have 4 and I have 2, there would be an issue I believe.  

Still trying some things every which way but loose right now to figure this one out!  We need some other people input as well to figure this one out.

[Huzze]

Thanks! shoarthing for this excellent guide. I followed your guide and have same issue like UN_WIRED (must boot with cpus=1). I notice that my dsdt.aml icon (a gray square) is different than yours. I got BIOS 1.50c from Asrock support but have not uppgraded from BIOS 1.50 yet.

[shoarthing]

Huzze - Hi - I don't think there's a compelling reason to use 1.50c . . .  I've asked Asrock for one with Enable/Disable of "Legacy USB" - ideally one where you can disable 'Low Speed' USB transactions.

I updated the dsdt.aml attached to the first post: it should match 1.50~1.50c - no harm in trying it.

[Huzze]

shoarthing, on Aug 8 2009, 06:26 PM, said:

Huzze - Hi - I don't think there's a compelling reason to use 1.50c . . .  I've asked Asrock for one with Enable/Disable of "Legacy USB" - ideally one where you can disable 'Low Speed' USB transactions.

I updated the dsdt.aml attached to the first post: it should match 1.50~1.50c - no harm in trying it.

Thanks for your advice. I have test your updated dsdt.aml and/or memory upgrade to 4GB. No change. I have to use cpus=1 at boot.

Maybe I´ll try to compile my own dsdt.aml as you suggest in post #11. (If I can find iaml).

Then I have problem with sleep and shutdown (restart is OK!).

Thanks again. I'll hope we can get a updated BIOS from Asrock that solves the USB2 issu.

[UN_WIRED]

Huzze, on Aug 9 2009, 01:42 PM, said:

Thanks for your advice. I have test your updated dsdt.aml and/or memory upgrade to 4GB. No change. I have to use cpus=1 at boot.

Maybe I´ll try to compile my own dsdt.aml as you suggest in post #11. (If I can find iaml).

Then I have problem with sleep and shutdown (restart is OK!).

Thanks again. I'll hope we can get a updated BIOS from Asrock that solves the USB2 issu.

Hey Huzze-

If you need the iaml util, you can find it in one of the packages from the iPC 10.5.6.  It is called DSDT GUI patcher, and I just use the iaml util in terminal to decompile and compile any changes then you can dump the app when done to keep it vanilla.

QUICK ASROCK IMAGE BACKUP/RESTORE...........

THE ISSUE:
I have experimented with this and found that after you boot an image for the first time on the ASROCK, and even create a full image, it will not boot up again (Hangs at the . Maybe it has to do with the UUID changing when you restore it?

THE RESOLUTION:
Although Shoarthings install guide is VERY straight forward and well written, I dont always have the time or patience to wait for a complete reinstall should something blow up. I have use a USB to SATA bridge and as we all know it can take about 50 mins to install even a base leo build to an external dive. I myself do not have another machine with internal SATA in it, cause I use a macinized 1000he.

I have figured out a quick and easy way to reimage your ASROCK ION drive in about 20 mins (Depending on speed of the machine you use). I use my 1000he OSX, or Mac Mini machine to build a drive based on the guide above.

TO BACKUP:
1.) Utilize Shoar's install guide above from a 2nd macinized machine.
2.) BEFORE YOU REBOOT or BOOT the newly built partition for the ASrock, make a FULL image of the partition that you just installed in Disk Utility. If D.U. gives a "Resource busy" error, remember you must unmount the partition you are creating the image from.
3.) Store away your 3.1GB or so image in a safe place for a fast restore.

TO RESTORE:
1.) Format the CORRECT PARTITION or drive for the asrock build. I do more than one partion sometimes, but always use a GUID scheme so I can work on snow leo as well.
2.)After formatting, install the RC2 Chameleon loader and install your kexts and dsdt files accordingly from above.
3.) Use disk utility to restore the "Virgin, never been booted" image to the drive. BE SURE the "Erase Destination Drive" option is not checked or you will blow away Chameleon that you just installed.
4.) You MAY get an error that it cannot find and scan/restore info. If so make sure your image is selected as it appears under your drives, go to "Checksum" on the menu bar, select "CRC-32 Checksum image" This will take 5 mins or so and should be good to go for future installs.
5.) Restore the image, and boot from the ion as normal

In any case, each time you restore and star this image up for the first time, you will get the "Welcome" screen and do the basic setup again. In my case, this saves about 20-30 mins each time I want to wipe a drive, and start over

Hope this helps save you some time as it has me!  Lets get this ASRock, rockin as a new HTPC!

[shoarthing]

UN_WIRED, on Aug 9 2009, 09:02 PM, said:

Hey Huzze-

If you need the iaml util, you can find it in one of the packages from the iPC 10.5.6.  It is called DSDT GUI patcher, and I just use the iaml util in terminal to decompile and compile any changes then you can dump the app when done to keep it vanilla.

. . .  Are you using the option in the 'pcwiz' GUI utility to "Apply to (Volume)"?

[UN_WIRED]

I am in the process of trying to figure out what the $%!$@ is going on with the DSDT.aml file, and why the majority of us cannot load it for 2 cores.  At one point when I used a CPUS=4 variable I got a panic, where the other times it would simply reboot.  My error was that it was expecting 2 cores but only found one on Die 0.

When I checked my working DSDT file on my 1000he (Atom 280), and it had 2 entries for the proc even though this is a single core.  It works fine with HT.

Will post more updates later this evenining once I am done blowing some stuff up here.

[shoarthing]

UN_WIRED, on Aug 10 2009, 01:08 AM, said:

I am in the process of trying to figure out what the $%!$@ is going on with the DSDT.aml file, and why the majority of us cannot load it for 2 cores.  At one point when I used a CPUS=4 variable I got a panic, where the other times it would simply reboot.  My error was that it was expecting 2 cores but only found one on Die 0.

When I checked my working DSDT file on my 1000he (Atom 280), and it had 2 entries for the proc even though this is a single core.  It works fine with HT.

Will post more updates later this evenining once I am done blowing some stuff up here.

. . .  try staying at OSX 10.5.6, or - as I have - after all the updates replacing the 10.5.8 versions with the vanilla 10.5.6 kernel plus matching system & seatbelt kexts (in this case from a 10.5.6 Unibody 17" restore DVD): the 10.5.6 kernel was the last to support HTT for the Atom 330 . . . .

  . . . after making sure that the only copy of dsdt.aml on your ION 330 is the one I've provided, installed to /Extra; with the 10.5.6 kernel you should be able to enable hyperthreading in the BIOS . . . FWIW my Geekbench32 'score' went from 1250~ish with 2 cores (rather worse than a first-gen single-core Intel Mac min) to 1600~ish with 2+2 'cores' (rather better)  

[UN_WIRED]

shoarthing, on Aug 10 2009, 09:23 AM, said:

. . .  try staying at OSX 10.5.6, or - as I have - after all the updates replacing the 10.5.8 versions with the vanilla 10.5.6 kernel plus matching system & seatbelt kexts (in this case from a 10.5.6 Unibody 17" restore DVD): the 10.5.6 kernel was the last to support HTT for the Atom 330 . . . .

  . . . after making sure that the only copy of dsdt.aml on your ION 330 is the one I've provided, installed to /Extra; with the 10.5.6 kernel you should be able to enable hyperthreading in the BIOS & see this:

. .  & this:

. . . FWIW my Geekbench32 'score' went from 1250~ish with 2 cores (rather worse than a first-gen single-core Intel Mac min) to 1600~ish with 2+2 'cores' (rather better)  

I have been working with a full str8t vanilla install of 10.5.6 up to this point.  Only thing that I removed was the cpuintelpowermanagement.kext as I was getting an error from it.  Will be trying again tonight.

Strangely enough though, when I pop up activity monitor, I briefly for a second or so see 2 cpu's before it goes back to one.  Have any idea what that is about?