To clarify for users following this thread - this is the latest version compiled against the most up to date source code available found here https://github.com/Piker-Alpha/macosxbootloader ?
FireArse
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Resolved 2006/2007 Mac Pro (1,1/2,1) and OS X Yosemite
- Thread starter Mr. Zarniwoop
- Start date
- Sort by reaction score
To clarify for users following this thread - this is the latest version compiled against the most up to date source code available found here https://github.com/Piker-Alpha/macosxbootloader ?
FireArse
There's only one branch in the project, called "Master". I simply clicked the button "Download ZIP". Then, of course, I unzipped it and compiled "solution" boot.
----------
In my case, it is. I think I built that code after Hennesie. His is supposedly from the same source, but the builds, for unknown reasons, are different.
Yea that's what I did with those two builds. You can how ever click on commits, then click on "<>" to the right of each commit to browse the code at that point in history and then download a zip containing the source from that moment in time.
But since neither of us followed that route, why our builds should have differences remains a mystery, unless we got somewhat different ZIP files. I can't understand the situation. In any case, it would be good if someone with LS 3.4.1 can verify if either of your builds (black/grey) or my later build boot Mavericks and/or Yosemite successfully on an early Mac Pro.
This build also would not boot OS X 10.9.5 with Little Snitch 3.4.1 installed. An interesting distinction does exist between the two (2) aesthetically different Pike Alpha versions however. With Little Snitch 3.4.1 installed, the "black screen white apple" version begins its boot then freezes with the hyphen in the middle of the screen. The "white screen grey apple" version does the continuos boot thing as I described earlier.
I can also confirm that my Dell hackintosh laptop running Public Beta 5 OS X 10.10 booting with Clover has no issues booting with Little Snitch 3.4.1 installed.
Something funky is going on with the kext cache/lzvn compression and the Little Snitch 3.4.1 kext that is not happening with either the Apple stock boot.efi, Tiamo's boot.efi or Clover (or apparently even Chameleon)
Last edited:
I just downloaded a fresh copy of the source and rebuilt it with the exact same results. I am using Windows 7 x64 running in Parallels 9 with MS Visual Studio 2013 Ultimate with Update 3 x86.
I am using a 32-bit edition of Windows 8.1 running in VMWare Fusion 7.0 with MS Visual Studio 2013 Professional with Update 3 x86. I find it odd that there should be such compilation differences.
How you can tell that?
What verbose mode says?
Did you try without kernelcache or with LZSS?
Finally up and running with 2 of my workstations.
Followed Hennesie's latest guide (which is fantastic) but I couldnt get the machine to boot from the USB at all. When trying to select it under System Preferences as the Startup Disk I would get The bless tool was unable to set the current boot disk. Tried a couple more times with that drive and twice with another (one was a cheap USB3 and another was SanDisk Cruzer) but nothing worked.
When I came back to the forum to troubleshoot this somehow the site was down.. I decided to make an 8GB partition on a blank SATA drive and redid the steps again. This time it would only boot if I removed every other drive, but would only get me the language screen before telling me that Yosemite couldnt be installed on this machine. I did make sure the syntax was accurate and not italic but other than that I am unsure of how to check and see where things went wrong.
With the site still down I finally just cloned my MacBook that is running the current build of 10.10 to a blank drive After putting the boot.efi on the clone and updating the PlatformSupport.plist the machine booted up and is running like a champ.
My Apple 5770 card did not have audio via HDMI working by default. I was able to get the audio to work by using kext-dev-mode=1 combined with the guide from this post. https://forums.macrumors.com/threads/1499797/ Audio works great through my off brand Mini DisplayPort to HDMI adapter now.
The first tower then got the 5770 switched out for a non-flashed ATI HD6570 which works flawlessly outside of not having the boot screen.
Finally I repeated the entire process for workstation #2 and had no issues at all (so far). In the system report both workstations are reporting as MacPro 1,1.
I am now onto my final step of doing the upgrade on the main workstation, an upgraded 1,1 that uses a non-flashed GTX 560Ti. It also has a USB3 PCI card as well as an eSATA PCI card but I dont anticipate any issues if I can track down the drivers.
Followed Hennesie's latest guide (which is fantastic) but I couldnt get the machine to boot from the USB at all. When trying to select it under System Preferences as the Startup Disk I would get The bless tool was unable to set the current boot disk. Tried a couple more times with that drive and twice with another (one was a cheap USB3 and another was SanDisk Cruzer) but nothing worked.
When I came back to the forum to troubleshoot this somehow the site was down.. I decided to make an 8GB partition on a blank SATA drive and redid the steps again. This time it would only boot if I removed every other drive, but would only get me the language screen before telling me that Yosemite couldnt be installed on this machine. I did make sure the syntax was accurate and not italic but other than that I am unsure of how to check and see where things went wrong.
With the site still down I finally just cloned my MacBook that is running the current build of 10.10 to a blank drive After putting the boot.efi on the clone and updating the PlatformSupport.plist the machine booted up and is running like a champ.
My Apple 5770 card did not have audio via HDMI working by default. I was able to get the audio to work by using kext-dev-mode=1 combined with the guide from this post. https://forums.macrumors.com/threads/1499797/ Audio works great through my off brand Mini DisplayPort to HDMI adapter now.
The first tower then got the 5770 switched out for a non-flashed ATI HD6570 which works flawlessly outside of not having the boot screen.
Finally I repeated the entire process for workstation #2 and had no issues at all (so far). In the system report both workstations are reporting as MacPro 1,1.
I am now onto my final step of doing the upgrade on the main workstation, an upgraded 1,1 that uses a non-flashed GTX 560Ti. It also has a USB3 PCI card as well as an eSATA PCI card but I dont anticipate any issues if I can track down the drivers.
Attachments
Were you running the previous OS using a bootloader in Legacy mode? You cannot use the bless command when using Clover or Chameleon.
It is not happening with Chameleon bootloader since boot.efi are original and untouched but it was happening with Tiamo's boot under Mavericks for any LS update, it is not a specific Pike's boot problem, it is only that something is missing within these 2 boot files (Tiamo and Pike) assuming that LS is controlling all requests to the Kernel ?
Hennessie, I'm not able to get my MacPro2,1 to boot from an installer disk I built following your guide (actually I built the disk 3 times without success). the machine always crashes very early with a KP. any idea what's going on?
Tried the "mikeboss" first known Pike Alpha complied verbose working version a few pages back in this thread and got the hyphen followed by endless rebooting loop. Returned the Tiamo's boot.efi to my OS X 10.9.5 system with Little Snitch 3.4.1 installed and it returned to booting normally.
----------
I am able to boot normally in OS X 10.9.5 using Tiamo's boot.efi with Little Snitch 3.4.1 installed.
Then it's not the LZVN/kernelcache issue because first verbose version omits kernelcache, so compression as well.
why even try to run Mavericks with the new boot.efi which was built specifically for Yosemite? I don't get it...
Tiamo's for Mavericks and Pike's for Yosemite.
Tiamo's for Mavericks and Pike's for Yosemite.
Hennessie, you can stop re-checking immediately! it seems that my external USB SSD was the problem. I have no idea as of why this was the cause of the KP, but after cloning the install-disk to another (inernal) SSD I'm able to run the installer. very strange, but I learned a long time ago that MacPro1,1 and 2,1 can be very picky with USB devices... especially when it comes to boot OS X off of USB.
thank you for your excellent guide!!
I have found this to be case too. Very unpredictable results associated with 'well thought of' USB drives.
F
Can someone with LS 3.4.1 confirm that Pike's boot.efi can boot Yosemite GM3 on an old Mac Pro? In the meantime, I've explained the situation to Pike, in case he can think of something.
----------
You are absolutely right. But then, if I remember correctly, it simply made it possible for Pike's slightly modified version (at the time) of Tiamo's boot.efi to find the Yosemite kernel and continue booting, so it is highly unlikely that a bug is the culprit. I wonder if earlier iterations of MASM and/or Visual Studio would produce a more compatible boot.efi. It would be great if we could use exactly the same version of both that Tiamo originally used.
Last edited:
Firstly kinda for obvious reasons, I initiated testing on Mavericks to be able to use the variable distinction in determining if it is Little Snitch 3.4.1 that is the problem or just any modified boot.efi or a specific modified boot.efi. As we all know, Tiamo's boot.efi will not boot Yosemite however the Pike Alpha boot.efi version does boot both OS X 10.9 as well as OS X 10.10.
Anyway, due to the fact that my Mac Pro Yosemite drive is a beta, I never bothered installing Little Snitch on it until a few moments ago. I just installed Little Snitch from scratch/for the first time with Build 6c697848ef boot.efi and it boots fine. Therefore I can only conclude that for some reason, something weird occurs with the Pike Alpha boot.efi and updating Little Snitch from a previous version to v3.4.1 which is not the case when booting the same system with either the Apple stock boot.efi, Tiamo's boot.efi or Clover.
Last edited:
MacPro2,1 (flashed from 1,1) with a fresh install of YOSE PB4. installed LS 3.4.1 and all is good. as I said: use Tiamo's boot.efi for Mavericks and Pike's version for Yosemite. I cannot reproduce the behavior others are reporting.
I think just updating period...the previous Little Snitch version on my Mavericks OS X 10.9.5 system before updating to v.3.4.1 was v.3.3.4