Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
Just found my other Xeon W3xxx, a W3565 SLBEV that came with my 2010 CPU tray and backplane combo. I'll downgrade to 140.0.0.0.0, install the W3565 into one of my trays, then upgrade the backplane to 142.0.0.0.0.
[doublepost=1552014376][/doublepost]The W3540 was a standard SLBEX, btw.
 
The moral of the story?
Always, always backup your working clean ROM and keep it in a safe place.
Optional: With a little investment you can skip the soldering/de-soldering part with a MATT card in case you brick your cMP.
 
Just "bricked" with W3565 too - it's Xeon processor used in the base single CPU mid-2012 Mac Pro and the model used for the upgraded mid-2010 single CPU. Cheapest mid-2010 came with W3530.

[doublepost=1552018187][/doublepost]Downgraded to 141.0.0.0.0 with X5680:
Screen Shot 2019-03-08 at 00.53.39.png


141.0.0.0.0 with W3565, jet plane mode:
Screen Shot 2019-03-08 at 00.57.48.png


Upgraded to 142.0.0.0.0 via efiupdater and no POST with W3565, as expected.

142.0.0.0.0 "Unbricked" with X5680:
Screen Shot 2019-03-08 at 01.05.08.png
 
Last edited:
A lot of people is asking how to get 140.0.0.0.0. It's not easy to the uninitiated but is doable.

The risk of bricking your Mac Pro is all yours.

1) Download the macOSUpd10.14.3.RecoveryHDUpdate.pkg from Apple SUCatalog (in the future, search the SUCatalog for the most recent RecoveryUpdate) link updated to 10.14.3
2) Open macOSUpd10.14.3.RecoveryHDUpdate.pkg with Pacifist and extract the RecoveryHDMeta.dmg
3) Open RecoveryHDMeta.dmg with Pacifist, go to the Firmware folder and extract it to your desktop.
4) Replace the Firmware folder of your Mojave installer with the one you saved on your desktop.
5) Open the Mojave installer, will ask you to upgrade the firmware.

I'm looking at doing this right now. Tried a standard Mojave install but my standard GFX card isn't metal so it won't proceed. This wasn't using the method above. Will this work on upgrading the firmware if the card is not metal supported (I'll be going back to High Sierra anyway). Cheers. Rob.
 
I'm looking at doing this right now. Tried a standard Mojave install but my standard GFX card isn't metal so it won't proceed. This wasn't using the method above. Will this work on upgrading the firmware if the card is not metal supported (I'll be going back to High Sierra anyway). Cheers. Rob.
Nope, only reconstruction can overcome a no METAL supported GPU.
 
Last edited:
  • Like
Reactions: bsbeamer
So basically I need to buy a metal supported card. I'll start a new thread rather than bombard this one any more. Thank you for your help, Alex.
 
So basically I need to buy a metal supported card. I'll start a new thread rather than bombard this one any more. Thank you for your help, Alex.

This has been discussed numerous times. This is the only info you need. Suggest you buy the bolded.

https://support.apple.com/en-us/HT208898

These specific third-party graphics cards are Metal-capable and compatible with macOS Mojave on Mac Pro (Mid 2010) and Mac Pro (Mid 2012):
  • MSI Gaming Radeon RX 560 128-bit 4GB GDRR5
  • SAPPHIRE Radeon PULSE RX 580 8GB GDDR5
  • SAPPHIRE Radeon HD 7950 Mac Edition
  • NVIDIA Quadro K5000 for Mac
  • NVIDIA GeForce GTX 680 Mac Edition
 
I didn’t had much time to investigate today, but I already found that 142.0.0.0.0 have:

  1. Same empty NVRAM volume as 138.0.0.0.0
  2. Same microcodes as 138.0.0.0.0
  3. Same NVMe module as 140.0.0.0.0
  4. Same APFS module as MP51.0089.B00
Now I need to check the EFI modules that have anything with Xeon configuration and initialization.

I was about to check this when I found a byte immediately after the NVRAM volume headers that has meaning. Initially I thought that was just a byte to make even or odd checksums, but it’s a versioning byte and it’s important for both Fsys and Gaid streams. I went to this rabbit hole and all my free time was spent tracking this versioning byte.
Nice work,
 
I was thinking that the two bytes immediately after the headers of the 3rd/4th NVRAM stream/stores were consecutive versioning, I already knew that it went from 0x0104 to 0x0108, but today I found a mid-2012 RFB backplane dump that have 0x010C.

Screen Shot 2019-03-09 at 13.49.14.png

I don't know if RFB backplanes have it's own versioning, so at the moment I don't know if the versioning is consecutive or anything. EFITool shows this as a unknow field, btw.

Always finding a answer then two or more questions…
[doublepost=1552151990][/doublepost]Found another RFB dump, this time 0x010D:

Screen Shot 2019-03-09 at 14.19.26.png

Maybe this will make sense…
 
Last edited:
Hi Alex,

I am really surprised that there is even another firmware (142.x.x..) in the works. Regardless the outcome and the ongoing thrilled ride to find out the purpose of it, here is a question that came up in our circle.
Will the new Apple display (asuming 30‘ 6K) have a industry standard Display port 1.2 & 1.4 ?
Since it is not a problem to run the LG 5K Display on a cMP Vega card, this should be possible to just plug in the shiny new screen and rest on the good old tower with Mojave and rock another 5 years as a semi pro.
Could the new 142.x firmware be designed to enable this for current cMP users as a last token of pleasure from Apple ?? What limitation could prevent us from this scenario. If the new display is completely proprietary, it would seriously limit its numbers. What’s your take on the purpose for this new firmware, what can the 142.x. do what the 141.x doesn’t ??
 
What’s your take on the purpose for this new firmware, what can the 142.x. do what the 141.x doesn’t ??
All Macs got new BootROMs, when this happens, is a security problem that affects all supported Macs and needs firmware corrections to be implemented. Could be related to what Project Zero published last week, I don't know.

For now, we only found that 142.0.0.0.0 removed support for W3xxx, no improvements made.
 
I made an error before with APFS EFI modules. Apple updated the APFS EFI helper module in the firmware with 142.0.0.0.0.

When I checked if APFS module had any changes with 10.14.4 DP4, I checked just the APFS.efi in the container and not the APFSJumpStart EFI module in the EFI.

APFSJumpStart.ffs from 140 to 142.0.0.0.0:

Screen Shot 2019-03-10 at 12.34.41.png
 
Last edited:
Just found something very interesting:
  • 140.0.0.0.0 has 443 EFI components,
  • 141.0.0.0.0 has 447,
  • 142.0.0.0.0 has 451.
Some modules are always duplicated into the EFI, btw.

Screen Shot 2019-03-10 at 20.11.57.png
Screen Shot 2019-03-10 at 20.26.30.png
Screen Shot 2019-03-10 at 20.12.00.png

[doublepost=1552261232][/doublepost]Added EFI modules between 140 and 142:

Section_Raw_A17B39CE-1916-4CFC-9581-7298A745F7A3_001_body.bin
Section_Raw_A17B39CE-1916-4CFC-9581-7298A745F7A3_002_body.bin
Section_Raw_A17B39CE-1916-4CFC-9581-7298A745F7A3_003_body.bin
Section_Raw_A17B39CE-1916-4CFC-9581-7298A745F7A3_body.bin

Google found nothing about GUID A17B39CE-1916-4CFC-9581-7298A745F7A3 and it's basically an empty module:

Screen Shot 2019-03-10 at 20.46.30.png



[doublepost=1552261342][/doublepost]33 EFI components changed between 140.0.0.0.0 and 142.0.0.0.0, some are duplicated.
 
Last edited:
18 unique ffs that changed between 140.0.0.0 and 142.0.0.0.0.

Sorted by ID:

Code:
7EC21E57-CE8A-4FDE-B147-0F7C0959CD3F - ?? Probably the font/ttf used with pre-boot configuration
8E6BD67D-1872-4932-A39A-FC9DCA4A98CF - ??
AB8DDBAA-9C47-429B-9D12-326F273D9B88 - ??
CFFB32F4-C2A8-48BB-A0EB-6C3CCA3FE847 - APFSJumpStart
18CEC42B-F0FC-4EFA-84AE-27D2184E873A - AppleROMVersion
CC582C73-F48F-4B62-83E8-A586B4C88F84 - AppleFirmwareFeatures.efi
35B898CA-B6A9-49CE-8C72-904735CC49B7 - DxeMain
A17B39CE-1916-4CFC-9581-7298A745F7A3 - IoTrap
DF9A9591-B646-4621-AF0D-18143A73289F - LegacyBootFS.efi
BB1FBD4F-2E30-4793-9BED-74F672BC8FFE - PchResetRuntime
FC1B7640-3466-4C06-B1CC-1C935394B5C2 - PchSerialGpio
E052D8A6-224A-4C32-8D37-2E0AE162364D - PchSmbusDxe
B0D6ED53-B844-43F5-BD2F-61095264E77E - PchSmiDispatcher
C194C6EA-B68C-4981-B64B-9BD271474B20 - PchSpiRuntime.efi
BB65942B-521F-4ec3-BAF9-A92540CF60D2 - SataController
EF0C99B6-B1D3-4025-9405-BF6A560FE0E0 - SmbiosMisc
2FB92EFA-2EE0-4BAE-9EB6-7464125E1EF7 - UhciDxe
240612B7-A063-11D4-9A3A-0090273FC14D - UsbBusDxe

Sorted by GUID:

Code:
18CEC42B-F0FC-4EFA-84AE-27D2184E873A - AppleROMVersion
240612B7-A063-11D4-9A3A-0090273FC14D - UsbBusDxe
2FB92EFA-2EE0-4BAE-9EB6-7464125E1EF7 - UhciDxe
35B898CA-B6A9-49CE-8C72-904735CC49B7 - DxeMain
7EC21E57-CE8A-4FDE-B147-0F7C0959CD3F - ?? Probably the font/ttf used with pre-boot configuration
8E6BD67D-1872-4932-A39A-FC9DCA4A98CF - ??
A17B39CE-1916-4CFC-9581-7298A745F7A3 - IoTrap
AB8DDBAA-9C47-429B-9D12-326F273D9B88 - ??
B0D6ED53-B844-43F5-BD2F-61095264E77E - PchSmiDispatcher
BB1FBD4F-2E30-4793-9BED-74F672BC8FFE - PchResetRuntime
BB65942B-521F-4ec3-BAF9-A92540CF60D2 - SataController
C194C6EA-B68C-4981-B64B-9BD271474B20 - PchSpiRuntime.efi
CC582C73-F48F-4B62-83E8-A586B4C88F84 - AppleFirmwareFeatures.efi
CFFB32F4-C2A8-48BB-A0EB-6C3CCA3FE847 - APFSJumpStart
DF9A9591-B646-4621-AF0D-18143A73289F - LegacyBootFS.efi
E052D8A6-224A-4C32-8D37-2E0AE162364D - PchSmbusDxe
EF0C99B6-B1D3-4025-9405-BF6A560FE0E0 - SmbiosMisc
FC1B7640-3466-4C06-B1CC-1C935394B5C2 - PchSerialGpio
[doublepost=1552263968][/doublepost]So, most are PCH related, some are USB related plus SATAController, APFS and LegacyBootFS.
 
Last edited:
ohh USB changes, wonder if they sorted out the USB detection bug at the boot picker finally?

not sure if its ever been bug reported, but i noticed if I had say a USB installer at the boot picker, and i unplugged it, then plugged it back in, that USB port would no longer show up in the boot picker
 
not sure if its ever been bug reported, but i noticed if I had say a USB installer at the boot picker, and i unplugged it, then plugged it back in, that USB port would no longer show up in the boot picker

If I remember correctly, something similar is one of the problems on the errata for Ibex Peak generation of PCHs.
[doublepost=1552264740][/doublepost]
33 module changed, maybe hundreds of macs bricked but still no boot screen. Lol.

If you are the person responsible for Mac Pro support and Apple never sold a GOP GPU with a Mac Pro, you will change the pre-boot configuration from UGA to GOP after you already declared that mid-2012 Mac Pros are vintage?

Let's be real, GOP support won't ever be implemented. MP5,1 firmware is probably getting corrections that are being sent to all Macs.

Are graphic cards still 5GT/s in 142.0.0.0?

All PCIe cards.
 
Last edited:
  • Like
Reactions: LightBulbFun
Did anyone noticed, that if you exclude the four three two modules that don't have a public ID and we yet don't know the function, not one CPU support related change between 140.0.0.0.0 and 142.0.0.0.0?

Why W3xxx Xeons don't work?
 
Last edited:
  • Like
Reactions: h9826790
It's very weird when you do a web search and one of the first results is something that you wrote. This happened today about one of the updated modules. :p

Btw, now I know how to efficiently check all the modules. I'll probably get everything done maybe in two hours next time Apple releases a firmware.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.