Jump to content

[Guide] Boot from EFI partition, zero modification installs on Intel SSE2 or better...


munky
 Share

1,404 posts in this topic

Recommended Posts

Hi Munky and all,

I've installed your bootloader on my laptop Asus F3Jc, and booted just fine after install with a Retail DVD.

I'm just having some issues on my Geforce Go 7300 QE/CI, after updating with combo 10.5.6. I've used Nvkush, Nvinject, Nvinjectgo, and Nvdarwin, always with the same result. no QE/CI.

Also, I've got no audio (ALC861). I had sound after installing ALCinject and AppleHDA.kext inside the Extensions folder on the EFI partition, and leaving a copy of the patched AppleHDA.kext on the systems Extensions folder. But it went quiet after two days running. Now I get volume control and listed sound devices, but no sound. After a fresh install, the problem persists.

 

And thanks this was great fun to set up. Next I'm trying it on a P4 and Asus P4V8MX mobo. I hope voodoo works well :(

 

Finally, I miss the grey apple logo :'( on boot. Is there a way to get it back? (I know you took it to save space, but now this thing is so cold and baren).

Thanks!

 

jjduro

 

PS

I translated your guide to Portuguese

Link to comment
Share on other sites

Hi all, hackmates,

 

great guide, munky!!!

 

Everything goes right, for me, but... some questions (I like to know what I'm doing...):

 

1) wich kexts are indispensables to put in /Volumes/EFI/Extra/Extensions?

 

I would to keep only few kexts for my rig, as I wish to keep my hackintosh clean, like a real mac.

2) Or does'n matter how much kext I put in it? May them cause conflicts with the original (with the same name) apple kexts?

 

Thx to all

 

post-338230-1231967729_thumb.png

:(:):D

 

Interesting, isn't it???

 

So, who wants explaine me about differences between various ISOs???

I need to know around indispensables kexts for my rig.... but... very different choices for same P35 chipset... :(

 

Thx to all!!

Link to comment
Share on other sites

Finally, I miss the grey apple logo :'( on boot. Is there a way to get it back? (I know you took it to save space, but now this thing is so cold and baren).

 

Me too...

 

thanks to munky

caiowa

 

EDIT: Because I dont need DSDT I changed bootloader back to v5.1 and the grey apple logo is back :)

Link to comment
Share on other sites

Hey Munky, what about AppleHDA onto EFI partition? Any news? I'm following Cyclonefr, stickpin and The King topics and I know you read there. You say may be a solution for your fantastic method,... but I can't find my way with my AD1988b

At the moment I have to install AZALIA from Taruga for my hackintosh because with the one ad1988b my rig won't shutdown... and this is not vanilla 100% (but it's ok, too)

 

Please, help us poors p5k-e wifi owners ;) !!!

 

Thank you for all

Link to comment
Share on other sites

UP AND RUNNING!

 

Great ! Elegant and effective.

I've had a lot of issues with wrong BOOT-132 image (PATA DVD drive), with the Hybrid

Chameleon DFE installer by kalyway and HD 3850 drivers.

 

But now almost everything is working. Will try to figure out minimal kext needed for P35C-DS3R

rig and 10.5.6.

Link to comment
Share on other sites

OS X Server does not work with this,

The OS X Server software seems to disable the server functions...

Has anyone else seen this ?

 

 

Anyone know what might be happening here ?

I have a fully paid up OS X Licence is the software looking for this kind of install and disabling the server or am I missing something else ?

Thanks...

Link to comment
Share on other sites

G'day guys, and in particular Munky and/or BladeRunner,

I could really do with some help please.

I've successfully installed Leopard 10.5.0 from a retail dvd, using Munky's boot_dfe_modbin.iso CD. I feel I understand what Munky refers to as Phase 0 of this guide, and have worked my way through to Phase 5 Test Boot. Along the way, everything seemed to work, with the occasional tiny delay in Terminal as the commands entered were processed, confirming that something was happening "behind the scenes". My only slight doubt was how to handle the message "Device could not be accessed exclusively..............effect. OK? [n]" after entering "w" at step 3.3. I entered "y", message returns "WritingMBR at offset0", and proceeded onwards. However Phase 5 Test Boot didn't work. My processor is a pre-core Pentium 4, so I assume I need to boot a patched kernel.

Can anyone tell me what I'm doing wrong, either with the above, or with the boot process itself, which after quitting Terminal goes like this:

1 Restart via boot_dfe_modbin.iso CD

2 Press <Enter> at first boot: prompt

3 Enter 81 to select hard drive containing 10.5.0 install

4 At next boot: prompt, enter bt(0,0)/mach_kernel.modbin -v boot-uuid=09f41906-d410..........-658da882748e

Instead of booting, the screen just reverts to the last boot: prompt. So clearly, the magic bt(0,0) command is not doing its stuff!!

Can anyone explain in a bit more detail just what is supposed to happen when the bt(0,0) command is entered. As far as I understand, it should force the booting process to look in the EFI partition from now and for evermore, so that the boot_dfe_modbin.iso CD can be thrown away, and the machine thereafter boots directly from the hard drive/EFI partition.

Is this right? Thanks in anticipation

Rob :D:P

Link to comment
Share on other sites

I think that it can leave the Apple logo if UFS_SUPPORT is excluded like the

Chameleon.

 

i386/libsaio/disk.c

 

// Allow UFS_SUPPORT to be overridden with preprocessor option.

#ifndef UFS_SUPPORT

#define UFS_SUPPORT 0 <--

#endif

Link to comment
Share on other sites

Anyone know what might be happening here ?

I have a fully paid up OS X Licence is the software looking for this kind of install and disabling the server or am I missing something else ?

Thanks...

 

Let me be clearer the Server licence key disappears from the machine and the Server admin tools do not work...

in the applesmbios I find this....

 

<key>com.apple.kpi.usupported<key>

<string>9.0<string>

 

if I get rid of smbios from the volume it makes no difference if I use one from the EFI volume it makes no difference

 

why would the server serial disappear ?

 

Should I use JAs10.4 distro, and if I do what is the actual difference can I just replace the Jas kernal with the real kernal and alter the boot plist and then use the EFI boot method ?

 

It's just that the Jas 10.4 server works but this does not and I would like to use this....

Link to comment
Share on other sites

Let me be clearer the Server licence key disappears from the machine and the Server admin tools do not work...

in the applesmbios I find this....

 

<key>com.apple.kpi.usupported<key>

<string>9.0<string>

 

if I get rid of smbios from the volume it makes no difference if I use one from the EFI volume it makes no difference

 

why would the server serial disappear ?

 

Should I use JAs10.4 distro, and if I do what is the actual difference can I just replace the Jas kernal with the real kernal and alter the boot plist and then use the EFI boot method ?

 

It's just that the Jas 10.4 server works but this does not and I would like to use this....

 

 

OK well it works now...

I put the SMBIOS that was installed in the install on the EFI partition and it just worked, possibly the SMBIOS on the EFI partition was out of date, however I now have "About This Mac" working and that does not even work on a desktop I setup on the same architecture...

 

So in short it looks like and SMBIOS conflict...

Link to comment
Share on other sites

G'day guys, and in particular Munky and/or BladeRunner,

I could really do with some help please.

I've successfully installed Leopard 10.5.0 from a retail dvd, using Munky's boot_dfe_modbin.iso CD. I feel I understand what Munky refers to as Phase 0 of this guide, and have worked my way through to Phase 5 Test Boot. Along the way, everything seemed to work, with the occasional tiny delay in Terminal as the commands entered were processed, confirming that something was happening "behind the scenes". My only slight doubt was how to handle the message "Device could not be accessed exclusively..............effect. OK? [n]" after entering "w" at step 3.3. I entered "y", message returns "WritingMBR at offset0", and proceeded onwards. However Phase 5 Test Boot didn't work. My processor is a pre-core Pentium 4, so I assume I need to boot a patched kernel.

Can anyone tell me what I'm doing wrong, either with the above, or with the boot process itself, which after quitting Terminal goes like this:

1 Restart via boot_dfe_modbin.iso CD

2 Press <Enter> at first boot: prompt

3 Enter 81 to select hard drive containing 10.5.0 install

4 At next boot: prompt, enter bt(0,0)/mach_kernel.modbin -v boot-uuid=09f41906-d410..........-658da882748e

Instead of booting, the screen just reverts to the last boot: prompt. So clearly, the magic bt(0,0) command is not doing its stuff!!

Can anyone explain in a bit more detail just what is supposed to happen when the bt(0,0) command is entered. As far as I understand, it should force the booting process to look in the EFI partition from now and for evermore, so that the boot_dfe_modbin.iso CD can be thrown away, and the machine thereafter boots directly from the hard drive/EFI partition.

Is this right? Thanks in anticipation

Rob :(:)

 

I don't understand why you would use the boot-132 CD after you built and populated the EFI partition. Is it because your Leopard install is not on your primary HDD and the one that is primary does not contain a loader that will point you to the leopard install?

 

If that is the case, try using the boot-132 CD command format of rd(0,1)/mach_kernel.modbin ... to identify the kernel followed by teh boot-uuid for the leopard partition.

Link to comment
Share on other sites

I don't understand why you would use the boot-132 CD after you built and populated the EFI partition. Is it because your Leopard install is not on your primary HDD and the one that is primary does not contain a loader that will point you to the leopard install?

 

If that is the case, try using the boot-132 CD command format of rd(0,1)/mach_kernel.modbin ... to identify the kernel followed by teh boot-uuid for the leopard partition.

hi bladerunner - thanks for quick response. My confusion stems from Munky's guide being extremely precise up to Phase 5 Test Boot, at which point I couldn't work out the exact booting method. The first thing I tried before this plea for help was to restart the machine, go into the machine's (Hewlett Packard) boot menu, and select the Leopard drive as the first boot device. The machine has 2 x SATA drives, the first having only Windows XP on it, the second having only the Leopard install on it. Having selected the Leopard drive as boot, the screen displayed the following:

 

boot0: GPT

boot0: HFS+

boot0: booting

boot0: done

boot1: startupfile - followed by blinking cursor.

 

Then no further progress. No boot: prompt for me to enter the bt(0,0)....etc command. That's when I tried to restart the process with the boot-132 disk, not really understanding the purpose of the bt(0,0)... command, as mentioned in my earlier post. So, do you think my mistake was at Phase 3.3 where after entering "w", I then entered "y" in reply to the message "Device could not be accessed exclusively. A reboot will be needed for changes to take effect. OK? [n]." That returned the message "Writing MBR at offset 0."

 

Seems like I'm pretty close to getting it to work, but your further advice would be appreciated.

 

Many thanks.

EDIT

Well I've just run through the whole process again, this time omitting Phase 3 completely. On boot, I got the same boot0 and boot1 messages as above, then blinking cursor and nothing more.

 

As I worked through the process, I used the ls command to check that the various files had been copied into /Volumes/EFI. Sure enough they were all there - ie all the kexts and the modbin kernel from your bladerunner-2008-09-23.iso download. However I don't know how to confirm that these files were actually on the EFI partition rdisk1s1, although as I said before, the tiny delays between entering a command line in Terminal, and the reappearance of bash-3.2# indicates the inputting process was working. As for your suggestions above, I think I've answered them, if somewhat indirectly!

Link to comment
Share on other sites

I think I am getting to where my problems are I get this in consol so cannot start web services

_CFGetHostUUIDString: unable to determine UUID for host Error 35

 

I see it is a network issue, is there a way to get around this ?

I have seen a few comments about this...

 

So I have now resolved most of my issues most of the error 35 messages I was getting have been resolved with this post

http://fr.pastebin.ca/768829

which resolved my UUID error

then I got an error where the MAC address of the en0 ethernet port could not be picked up and I resolved that here...

 

http://forum.insanelymac.com/lofiversion/i...php/t84783.html

 

However web services still cannot pick up the Mac address...

Jan 18 14:04:14 server servermgrd[59]: *** -[WPWebServices defaultWebServiceWithSession:]: couldn't obtain the MAC address of interface en0.

 

so something is still wrong...

 

Also Both SMBIOSENABLER

and

SMBIOSRESOLVER

 

fail to give me the "About this MAC " without crashing the finder window...

in system profiler however both report the right system buss speed and memory and processor etc...

Link to comment
Share on other sites

hi bladerunner - thanks for quick response. My confusion stems from Munky's guide being extremely precise up to Phase 5 Test Boot, at which point I couldn't work out the exact booting method. The first thing I tried before this plea for help was to restart the machine, go into the machine's (Hewlett Packard) boot menu, and select the Leopard drive as the first boot device. The machine has 2 x SATA drives, the first having only Windows XP on it, the second having only the Leopard install on it. Having selected the Leopard drive as boot, the screen displayed the following:

 

boot0: GPT

boot0: HFS+

boot0: booting

boot0: done

boot1: startupfile - followed by blinking cursor.

 

Then no further progress. No boot: prompt for me to enter the bt(0,0)....etc command. That's when I tried to restart the process with the boot-132 disk, not really understanding the purpose of the bt(0,0)... command, as mentioned in my earlier post. So, do you think my mistake was at Phase 3.3 where after entering "w", I then entered "y" in reply to the message "Device could not be accessed exclusively. A reboot will be needed for changes to take effect. OK? [n]." That returned the message "Writing MBR at offset 0."

 

Seems like I'm pretty close to getting it to work, but your further advice would be appreciated.

 

Many thanks.

EDIT

Well I've just run through the whole process again, this time omitting Phase 3 completely. On boot, I got the same boot0 and boot1 messages as above, then blinking cursor and nothing more.

 

As I worked through the process, I used the ls command to check that the various files had been copied into /Volumes/EFI. Sure enough they were all there - ie all the kexts and the modbin kernel from your bladerunner-2008-09-23.iso download. However I don't know how to confirm that these files were actually on the EFI partition rdisk1s1, although as I said before, the tiny delays between entering a command line in Terminal, and the reappearance of bash-3.2# indicates the inputting process was working. As for your suggestions above, I think I've answered them, if somewhat indirectly!

 

I am not sure I understand your comment "I don't know how to confirm that these files were actually on the EFI partition". If you used diskutil to list the visible drives, determined that the correct partition was mounted as /Volumes/EFI, and used "ls -l /Volumes/EFI " to check the content, what is the issue you question?

 

If you followed Munkys' guide and did Phase 1 & 2 then you need Phase 3 to make the diskXs1 bootable. In phase 3, the "w" -> "y" is correct and required - ignore the error message.

 

You need to be careful that you do phase4 correctly - well, all phases need to be done correctly. In phase4, be sure the mount for the EFI partition worked. Do a "diskutil list" to verify that /dev/diskXs1 is mounted at /Volumes/EFI. If that part fails, you could be copying files to the mount point on your Leopard partition instead.

Link to comment
Share on other sites

I am not sure I understand your comment "I don't know how to confirm that these files were actually on the EFI partition". If you used diskutil to list the visible drives, determined that the correct partition was mounted as /Volumes/EFI, and used "ls -l /Volumes/EFI " to check the content, what is the issue you question?

 

If you followed Munkys' guide and did Phase 1 & 2 then you need Phase 3 to make the diskXs1 bootable. In phase 3, the "w" -> "y" is correct and required - ignore the error message.

 

You need to be careful that you do phase4 correctly - well, all phases need to be done correctly. In phase4, be sure the mount for the EFI partition worked. Do a "diskutil list" to verify that /dev/diskXs1 is mounted at /Volumes/EFI. If that part fails, you could be copying files to the mount point on your Leopard partition instead.

aha! I didn't know the "ls -l" command. I'll check this later this morning (it's now 8.15 am in Melbourne), run through the process again, including Phase 3, and see what happens. Can you confirm that the Phase 5 booting procedure that I 'm following is correct in principle?

 

Thanks

Rob

Link to comment
Share on other sites

aha! I didn't know the "ls -l" command. I'll check this later this morning (it's now 8.15 am in Melbourne), run through the process again, including Phase 3, and see what happens. Can you confirm that the Phase 5 booting procedure that I 'm following is correct in principle?

 

Thanks

Rob

 

I went back and read what you said in an earlier post. If you boot from the boot-132 CD you need to enter rd(0,1)/mach_kernel.modbin to locate the kernel. If you boot directly from the hard disk and EFI partition then the sequence is bt(0,0)/mach_kernel.modbin. Don't ask why they are different. It has something to do with locating the kernel in a ramdisk VS locating it on the hard drive. I think.

Link to comment
Share on other sites

I went back and read what you said in an earlier post. If you boot from the boot-132 CD you need to enter rd(0,1)/mach_kernel.modbin to locate the kernel. If you boot directly from the hard disk and EFI partition then the sequence is bt(0,0)/mach_kernel.modbin. Don't ask why they are different. It has something to do with locating the kernel in a ramdisk VS locating it on the hard drive. I think.

 

Well BladeRunner, I guess we can theorise as much as we like, but clearly it has to do with the mysterious contents of the file boot-turbo-munky.bin! I looked in to this file whilst I was loading this last AND SUCCESSFUL time. There are numerous references hidden in the gobbledegook to rd(xx), bt(yy) etc. Anyway, I'm ashamed to say that I'd made 3 critical and repeated typos during earlier attempts - and they were so subtle as to be hard to find. But after the umpteenth time, I got it right, and have been rewarded by a fully booting Leopard 10.5 OS. It seems though that the bt(0,0) string has to be entered at each start-up?

 

One astonishing outcome of the whole process was that the second main partition on the 500 GB GUID drive was renamed from "Spare" to the password I entered when setting up the install Leopard on the first main partition of the same drive.

 

Curious, eh?

 

I'm now going to attempt to download an upgrade from the Apple site, which I think is 10.5.6. It's 675 MB, so it'll probably take all night to download via this pathetic system we call Broadband down here. Are there any other things of significance in the 54-odd pages of this thread that you'd recommend I look into? Audio and dual-booting are obvious ones but there may be more?

 

Finally, many thanks for taking the trouble to reply and offer your assistance - it's really appreciated. And thanks too to Munky and all the others who actually understand all this at the deepest level, and have made their knowledge available to everyone.

 

:thumbsup_anim:;):D

Link to comment
Share on other sites

Well BladeRunner, I guess we can theorise as much as we like, but clearly it has to do with the mysterious contents of the file boot-turbo-munky.bin! I looked in to this file whilst I was loading this last AND SUCCESSFUL time. There are numerous references hidden in the gobbledegook to rd(xx), bt(yy) etc. Anyway, I'm ashamed to say that I'd made 3 critical and repeated typos during earlier attempts - and they were so subtle as to be hard to find. But after the umpteenth time, I got it right, and have been rewarded by a fully booting Leopard 10.5 OS. It seems though that the bt(0,0) string has to be entered at each start-up?

 

One astonishing outcome of the whole process was that the second main partition on the 500 GB GUID drive was renamed from "Spare" to the password I entered when setting up the install Leopard on the first main partition of the same drive.

 

Curious, eh?

 

I'm now going to attempt to download an upgrade from the Apple site, which I think is 10.5.6. It's 675 MB, so it'll probably take all night to download via this pathetic system we call Broadband down here. Are there any other things of significance in the 54-odd pages of this thread that you'd recommend I look into? Audio and dual-booting are obvious ones but there may be more?

 

Finally, many thanks for taking the trouble to reply and offer your assistance - it's really appreciated. And thanks too to Munky and all the others who actually understand all this at the deepest level, and have made their knowledge available to everyone.

 

:thumbsup_anim: :censored2: :censored2:

 

I am glad you found the problem, and good luck with the upgrade to 10.5.6.

 

re: your comment about needing to enter the bt(0,0)... each time. You can place that data in your Boot.plist file. That can reside in the EFI partition or in /L/P/S. Your choice. My Boot.plist looks like this:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
	<key>Graphics Mode</key>
	<string>"1280x1024x24"</string>
	<key>Kernel</key>
	<string>bt(0,0)/mach_kernel.voodoo</string>
	<key>Kernel Flags</key>
	<string>boot-uuid=B577BA9E-DF3A-3988-859D-86DC513E644C  -v </string>
	<key>Timeout</key>
	<string>7</string>
</dict>
</plist>

Link to comment
Share on other sites

I am glad you found the problem, and good luck with the upgrade to 10.5.6.

 

re: your comment about needing to enter the bt(0,0)... each time. You can place that data in your Boot.plist file. That can reside in the EFI partition or in /L/P/S. Your choice. My Boot.plist looks like this:

<?xml version="1.0" encoding="UTF-8"?>
  <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
  <plist version="1.0">
  <dict>
	   <key>Graphics Mode</key>
	   <string>"1280x1024x24"</string>
	   <key>Kernel</key>
	   <string>bt(0,0)/mach_kernel.voodoo</string>
	   <key>Kernel Flags</key>
	   <string>boot-uuid=B577BA9E-DF3A-3988-859D-86DC513E644C  -v </string>
	   <key>Timeout</key>
	   <string>7</string>
  </dict>
  </plist>

er thanks br, but how to I find the boot.plist?

 

also have successfully updated straight from 10.5.0 to 10.5.6 :) .

 

But I also stuffed up the IONetworkingFamily.kext when I was trying to reinstall my ethernet connection, which incidentally did not "survive" the update. Is there a way to reinstall the kext? I tried copying it from the retail dvd install (from a separate partition on the drive) but with no success. Maybe I have to go right back to the start and reinstall the EFI booting and retail dvd kexts? So at the moment no internet from the hackintosh. :(

Link to comment
Share on other sites

 Share

×
×
  • Create New...