Updating a Mac (Pro)s CPU Microcode

[Larsvonhier]

Two notes on recent posts:
* Xraying will show you traces, but not functionality (for which you need the schematics). To reconstruct the so called "net list" from copper traces is possible but involves a huge amount of time and is error prone.
* I´ve also gone so far as to ask several chinese vendors of (Macbook, iPhone, iMac, ...) schematics about Mac Pro schematics. No one has ever heard of them.
* Adding microcodes is what well-known tools such the ones I added here can do as it always has been the official way for Apple to include newer variants of processors to their bootROM and also to release patched versions of microcodes (as a reaction to security vulnerabilities on microcode level!). As to be expected, adding the proper microcodes does not do the trick on its own.

 

[Dayo]

Cool but the impression is that a lot more may be possible with the microcodes and/or the associated memory than appears on the surface. @Petri Krohn posted the item below some time ago (which is why I pinged him):

https://forums.macrumors.com/posts/31855309

 

[Larsvonhier]

Cool but the impression is that a lot more may be possible with the microcodes and/or the associated memory than appears on the surface. @Petri Krohn posted the item below some time ago (which is why I pinged him):

https://forums.macrumors.com/posts/31855309

"Cool, but..."
Yes. Can do a lot more.
Question: What is the goal here? (Updating microcode for what purpose?)
Side project: Get MP1,1 and 2,1 running with newer XEONs that allow macOS up to Ventura (due to SSE4.x support).
Next steps could be then to add SSE4.2 also, but we´re far from the first step. Or to put it another way round: A lot of loose ends, much of the "low hanging fruit" stuff has been tested but does not do the trick.

Have been actively pursueing things with @Petri Krohn a couple of months back, he had good ideas but I think we came to the point where deeper investigation in north bridge init in the EFI boot code is required. Simple trial and error in modifing bits of code did not bring us forward.

btw, can you post a screenshot with details showing your 1,1 running Harpertown Xeon(s) and booted macOS (OSX) after doing the SMC voodoo? Perhaps "about this Mac" and some software like Geekbench showing the actual recognized processor type.
Much appreciated!

 

[Dayo]

I don't run an MP11 but an MP31.
As to the goal, your "side project" of says most of it. SSE42, maybe even AVX Emulation or even more??

People are now able to load stuff into firmware to get bootscreens but I don't think the OP of this thread started in 2017 would have known exactly where the first steps taken there would lead but it can largely be traced back there (at least in terms of stuff out of this forum).

I know next to nothing on the subject btw. Just seems that there might be possibilities there.
I found the "Russian Hackers" video particularly interesting.

 

[umashi1]

Is there a way to get non-mac graphics cards to display before OS on MP 1,1 / 2,1, like patching the EFI or something? Or is that still being worked on?

 

[umashi1]

I came across [this] document for the 5000x chipset.
I don't know if it'll help with getting harpertown CPUs to work with 1,1 / 2,1 mac pro.
I figured i'd post it here just in case.

 

[umashi1]

I just got back the email from sage. The cost to produce a schematic is more than the motherboard is worth.

 

[umashi1]

.

 

[pc297]

Has anyone asked Apple directly?

 

[umashi1]

Has anyone asked Apple directly?

apple most likely wouldn't supply because.
1.Mac Pro 1,1 / 2,1 is considered legacy hardware and therefore not supported anymore.
2. Because they most likely don't have this documentation anymore if by chance they did at one point have it on their servers.
3. Only "apple certified" service technicians have access to most apple documentation and parts.
That was actually one of the contributing factors behind trying to get a right to repair bill passed in new york. Although all the uproar was mostly over iphone repair.
Apple added in a self service repair policy, but i believe that applies to newer hardware. So i'm not sure if apple would supply mac pro schematics to consumers.
You could try contacting them but they'll probably just say it's old out of warranty hardware and not supported or something along those lines.

 

[dald]

Hi, I got my hands on a 1,1 early this year and was looking for upgrading it, of course lost a lot of time as I did not come here in the first place...
But anyhow, there I am, after adding the microcode for a E5440 SLANS (10676), I failed to boot when populated alone, but I managed to get a boot under Debian (Lion is not happy at all) with CPU-A having the E5440 and CPU-B having the original 5150.
Even if the E5440 is not used, what's interesting is to see that in /proc/cpuinfo we see its 4 cores and the correct use of the microcode. Nothing else unfortunately.
I really don't know what I'm doing and just tinkering with the box (might brick it some day if not carefull...)
I know it does not help much, but I wanted to share this test I made.

 

[umashi1]

I came across some documents that talk about the LGA771 CPUs.
i'll share them in case it helps.

Intel 5000X chipset product brief
Intel 5000X (MCH) datasheet may 528 pg
Intel 5000X (MCH) datasheet sep 458 pg
5100 series datasheet
5200 series datasheet
5200 series Thermal/Mechanical Design Guidelines
5300 series datasheet
5400 series specification update
5400 series datasheet
I also came across a datasheet for a similar chipset which goes into more detail
Intel 5000P/5000V/5000Z Chipset (MCH) datasheet
This datasheet talks about the 5000 series chipsets
Intel 5000 Series Chipset (MCH) Thermal/Mechanical Design Guide

In case someone wants to find more useful [datasheets]

 

[dald]

I went into more testing with mix of CPUs in the machine, it boots, but Lion crashes and I'm assuming it's because the systems sees first the E5440 and similar as when under Debian I get some time to time a "illegal instructions" message until I disable the 2 cores of the 5150.
Overall that means to me that the 5000x has no issues with managing the E5440, also I've ran a geekbench after disabling the 5150, but results are very bad :MP1,1-E5440 geekbench5
I tend to think that what's missing to be able to use anything else that natively supported CPUs, is in addition to the microcode in the ROM, part of the drivers from the MP3,1 ROM & not touching the SMC, but I might be completely wrong...
To avoid bricking the box I guess going through the earlier discussion > EFI Shell thingies and start loading some of the MP3,1 drivers for the EFI shell might help to find what's missing in our MP2,1 ROM.
In addition and until a modded ROM works with the E5440, the 5150 cores have to be disabled (not sure how in the EFI Shell) before trying to boot OS X
I'm attaching the /proc/cpuinfo and the output of lscpu after disabling the 5150 cores for curiosity.

 

[umashi1]

If there was a way to get 5400 series cpu's to work on cMp 1,1 / 2,1.
I think it might involve differences in the voltage regulators between [1,1 / 2,1] and [3,1].

I don't have access to an image of the 3,1 board so i can't verify differences.
I have a 1,1 board laying around so i can send off pictures of the chips if someone wants try comparing part differences that might relate to getting 5400 series to boot.

 

[umashi1]

I went into more testing with mix of CPUs in the machine, it boots, but Lion crashes and I'm assuming it's because the systems sees first the E5440 and similar as when under Debian I get some time to time a "illegal instructions" message until I disable the 2 cores of the 5150.
Overall that means to me that the 5000x has no issues with managing the E5440, also I've ran a geekbench after disabling the 5150, but results are very bad :MP1,1-E5440 geekbench5
I tend to think that what's missing to be able to use anything else that natively supported CPUs, is in addition to the microcode in the ROM, part of the drivers from the MP3,1 ROM & not touching the SMC, but I might be completely wrong...
To avoid bricking the box I guess going through the earlier discussion > EFI Shell thingies and start loading some of the MP3,1 drivers for the EFI shell might help to find what's missing in our MP2,1 ROM.
In addition and until a modded ROM works with the E5440, the 5150 cores have to be disabled (not sure how in the EFI Shell) before trying to boot OS X
I'm attaching the /proc/cpuinfo and the output of lscpu after disabling the 5150 cores for curiosity.

If you put two different cpus in, it'll ignore one of them in order to boot.
given that the 1,1 and 2,1 don't support 5400 series, it ignores it.
I wouldn't suggest forcing it, because you may end up making your mac not bootable if you break something in the EFI.

As for the SMC, that has nothing to do with whitelisting CPUs and more to do with the fans and power regulation.

 

[Andrupka]

Alright

I know that I said I would expand on this more today, but I am planning to do it tomorrow instead. I got my 12 E5430's in today and have been testing with them. All of them had a matching cpu/matching pair. 6 of them were made in costa rica, which I expected them all to be, but the other 6 were made in china. From my testing, it seems as though they actually cause different outcomes.

Today I got it to boot with an E5430 in socket b and a 5160 in socket a. Don't get me wrong, it wasn't just as soon as I turned it on, I spent probably 30-45 minutes messing with the smc, but it did boot. There was not a boot screen even with my genuine mac card in it, and it took a while to get into el capitan, but it did it.

I immediately went to dump the rom, not even thinking about checking hardware info, but unfortunately I got a kernel panic and have not gotten it to boot again. Well, I did, but the screen was just white, then went blue, then it shut off, all in a matter of seconds after it first output the white image.

I also have discovered (it might already be known, but not as far as I am aware) how to control which cpu will fail (either only a, only b, or both) and how to get the system to think the cpus are overheating on power up even though they aren't, and when this happens the fans go to 100% but then it will post. All of this was through messing with the smc, nothing more. Same exact components in the exact same locations (except for cpus).

edit: This was on my 1,1 "test" board without the 2,1 firmware, so completely stock firmware and smc.

Is it reproducible with 2 E5430s? This sounds awesome.

 

[Matthew071998]

Has anyone booted a Mac pro 1.1 / 2.1 in verbose mode? This might give some insight in how the efi handles pre-boot phase, cpu detection and cpu configuration.

Enabling verbose mode on a Mac Pro 1.1 (2006) requires modifying the boot arguments using the Open Firmware interface. Here's how you can do it:

1. Start or restart your Mac Pro 1.1.
2. Immediately press and hold the Command (⌘) + Option (⌥) + O + F keys simultaneously.
3. This will launch the Open Firmware interface, indicated by a white text screen with a ">" prompt.
4. At the prompt, type the following command to set the boot-args variable to enable verbose mode:
   
   setenv boot-args -v
5. Press the Return key to execute the command.
6. Type the following command to save the changes:
   
   setenv auto-boot? true
7. Press the Return key to execute the command.
8. Finally, type the command reset-all and press the Return key to restart the Mac Pro.

After the restart, your Mac Pro 1.1 should boot in verbose mode, displaying detailed boot information on the screen.

To revert back to the default boot behavior and disable verbose mode, you can follow the same steps but use the command setenv boot-args without any arguments to clear the boot-args variable:

setenv boot-args

Remember to press the Return key after entering each command and to save the changes using 'setenv auto-boot? true' before restarting the system.

EDIT: 10-5-2023

EFI also has a verbose mode but it won't show pre-boot phase.

 

[tsialex]

Has anyone booted a Mac pro 1.1 / 2.1 in verbose mode? This might give some insight in how the efi handles pre-boot phase, cpu detection and cpu configuration.

Enabling verbose mode on a Mac Pro 1.1 (2006) requires modifying the boot arguments using the Open Firmware interface. Here's how you can do it:

1. Start or restart your Mac Pro 1.1.
2. Immediately press and hold the Command (⌘) + Option (⌥) + O + F keys simultaneously.
3. This will launch the Open Firmware interface, indicated by a white text screen with a ">" prompt.
4. At the prompt, type the following command to set the boot-args variable to enable verbose mode:
   
   setenv boot-args -v
5. Press the Return key to execute the command.
6. Type the following command to save the changes:
   
   setenv auto-boot? true
7. Press the Return key to execute the command.
8. Finally, type the command reset-all and press the Return key to restart the Mac Pro.

After the restart, your Mac Pro 1.1 should boot in verbose mode, displaying detailed boot information on the screen.

To revert back to the default boot behavior and disable verbose mode, you can follow the same steps but use the command setenv boot-args without any arguments to clear the boot-args variable:

setenv boot-args

Remember to press the Return key after entering each command and to save the changes using 'setenv auto-boot? true' before restarting the system.

Intel Mac Pros are EFI, not Open Firmware like PowerPC Macs.

 

[pc297]

That would be a hot lead... if Mr. Piednoël could be reached!

We got this board, its the PSL version and was ready to take the 54xx Xeons.

As for modifying the EFI contents, we have the tools and a modified dual-bank-flash MP2,1 board that allows us to flash anything without risk of bricking the system. The only missing part is detailed knowledge of what to import from 3,1 and what to leave as is in the 2,1 EFI. We started to fiddle around to various different outcomes but never got system control to hand over to PEI phase with correct init of the MCH / RAM. Really helpful would be to hook up a ISP/Debugger or use the intel USB monitoring of the init phase.
Also unknown is whether the SMC interfers or inhibits init when certain CPUs are detected. I managed to dump the contents and even disassemble it (Renesas H8 if I recall right). But debuggers for that are rare and expensive and seem to be tailored to the exact variant of each H8 type. Good thing here is that Apple used the same chip in most Macs, even MacBooks.

Did anyone compare the S5000PSL CPU VRMs with that of previous S5000 boards that didn't take 54xx Xeons?

 

[Larsvonhier]

Did anyone compare the S5000PSL CPU VRMs with that of previous S5000 boards that didn't take 54xx Xeons?

I assume that you´re refering to the voltage regulator modules and their appropriate setting for voltage levels?
We checked this by modifying the onboard VRMs on Mac Pro 1,1 (2,1) boards for the slightly different settings for the 54xx Xeons. That did not show any progress by itself. Seems not to be the (main) reason for no POST success.

 

[pc297]

I assume that you´re refering to the voltage regulator modules and their appropriate setting for voltage levels?
We checked this by modifying the onboard VRMs on Mac Pro 1,1 (2,1) boards for the slightly different settings for the 54xx Xeons. That did not show any progress by itself. Seems not to be the (main) reason for no POST success.

Thanks for the update; sounded like it could have been it or it being an important factor given @ypc0cnz's result with a singular E5430 maybe having reached a "sweet spot" in CPU voltage out of several RAM and GPU configurations.

https://forums.macrumors.com/threads/updating-a-mac-pro-s-cpu-microcode.2114187/post-32004031

So is the no POST result with tweaked CPU voltage with @Petri Krohn's patched 2,1 firmware (with Harpertown microcodes added in)?

Cheers,

 

[Larsvonhier]

Thanks for the update; sounded like it could have been it or it being an important factor given @ypc0cnz's result with a singular E5430 maybe having reached a "sweet spot" in CPU voltage out of several RAM and GPU configurations.

https://forums.macrumors.com/threads/updating-a-mac-pro-s-cpu-microcode.2114187/post-32004031

So is the no POST result with tweaked CPU voltage with @Petri Krohn's patched 2,1 firmware (with Harpertown microcodes added in)?

Cheers,

No POST success so far in any of the investigated combinations of patched FW, HW mods (VRM voltage, processor pin swaps, etc.) and various Xeon types.

 

[pc297]

Silly question, but do Clovertowns and below post successfully on the 3,1?

 

[pc297]

Guys, did you see this post from a year ago about progress on a 10.13 kernel implementing Opemu SSE4.1 emulation on SSE4-less cpus?

Kernel boots but freezes shortly afterwards

TASK FORCE - [AMD] to missing SSE4.1 Instruction set - older CPUs on macOS Sierra (10.12.x)

Hey Coder, Since we have not made a step forwards, with building of working kernel for older AMD CPUs.....and now we hope that some talented Coder would rise to this challenge in order to fix the problem. If any information are desired, then just write a message

www.insanelymac.com

 

[dald]

Arf... My MP1,1 is a huge paper weight at the minute...
I got myself a TL866IIplus and then unsoldered the chip to try modifications, but it appears that the tool is misbehaving while flashing the 2nd hlaf of the chip it breaks the first half, also I did purchase new chips and confirmed the same behaviour (do not seem to be the only one with this problem...M50fw016 is not erased on T56 )
😥
Anyone having experience with the TL866IIplus ?