MP 1,1-5,1 MacPro5,1: BootROM thread | 144.0.0.0.0

[tsialex]

Trashcan got a firmware upgrade today too, 129.0.0.0.0:

$IBIOSI$ MP61.88Z.F000.B00.1902141849
Copyright (c) 2005-2019 Apple Inc.  All rights reserved.
Apple ROM Version
  Model:        MP61
  EFI Version:  129.0.0.0.0
  Built by:     root@saumon
  Date:         Thu Feb 14 18:49:50 PST 2019
  Revision:     129 (B&I)
  ROM Version:  F000_B00
  Build Type:   Official Build, Release
  Compiler:     Apple clang version 3.0 (tags/Apple/clang-211.10.1) (based on LLVM 3.0svn)

 

[edgerider]

hi there!
other than bricking the w equiped mac, what are the benefits of 142?
any new functionality ?

 

[TobiasT]

Trashcan got a firmware upgrade today too, 129.0.0.0.0:

$IBIOSI$ MP61.88Z.F000.B00.1902141849
Copyright (c) 2005-2019 Apple Inc.  All rights reserved.
Apple ROM Version
  Model:        MP61
  EFI Version:  129.0.0.0.0
  Built by:     root@saumon
  Date:         Thu Feb 14 18:49:50 PST 2019
  Revision:     129 (B&I)
  ROM Version:  F000_B00
  Build Type:   Official Build, Release
  Compiler:     Apple clang version 3.0 (tags/Apple/clang-211.10.1) (based on LLVM 3.0svn)

This new firmware is not installed if you have a non-apple SSD in MP61, right? I mean, 10.4.4 can be installed, but the firmware is not installed.

 

[Kon_Kipa]

The bricked Mac Pro is a mid2012 and the user has another backplane to replace the bricked one, so in time he will get it back working.

What caused the brick is a mystery for now. That’s why you don’t upgrade to beta BootROMs, leave to the people that have spare backplanes or can desolder/reprogram the SPI flash memory/solder it back.

Be smart, don't risk your Mac Pro with beta BootROMs.

that would be me - i have access to soldering stations & flash programmers

 

[LightBulbFun]

@tsialex have you submitted a bug report to apple about this?

seems pretty serious especially as there are Stock MacPro5,1 configurations with W35xx CPUs

(so even users who have not upgraded their mac pro and completely stuck to apples instructions could end up bricking their MacPro5,1 if this BootROM goes public)

 

[tsialex]

hi there!
other than bricking the w equiped mac, what are the benefits of 142?
any new functionality ?

Apple don't tell anything about BootROM releases before the final release and even after that just sometimes.
[doublepost=1551786805][/doublepost]

This new firmware is not installed if you have a non-apple SSD in MP61, right? I mean, 10.4.4 can be installed, but the firmware is not installed.

MP6,1 don't install firmware updates with other SSD installed, just with the Apple original one. 10.14.4 DP4 can be installed, won't upgrade the firmware.
[doublepost=1551786911][/doublepost]

that would be me - i have access to soldering stations & flash programmers

If you don't know how to find and install beta BootROMs by yourself, it's not for you. Sorry but if you search, you will see that I always say this repeatedly and I won't change it now.
[doublepost=1551786975][/doublepost]

@tsialex have you submitted a bug report to apple about this?

seems pretty serious especially as there are Stock MacPro5,1 configurations with W35xx CPUs

(so even users who have not upgraded their mac pro and completely stuck to apples instructions could end up bricking their MacPro5,1 if this BootROM goes public)

@crjackson2134 already did, I'll will check his radar and add anything if needed.
[doublepost=1551788438][/doublepost]

Would be interesting to see which W intel processor family initialization code was pushed to the 5,1's bootrom... perhaps it was intended for the 7,1?

I bet that Apple made a basic error, like what they did with MP51.0087.B00 forgetting to include microcodes. I'll check the modules when I have time.

Btw, 142.0.0.0.0 has the correct microcodes, exactly the same they used on 138/139/140/141.0.0.0.0.

 

[Zeke D]

So, yes can be just a 3 thing. Someone with an i7 installed needs to confirm it.

Alex, I still have that i7-940 and spare tray if testing on it will help the cause.

 

[tsialex]

Alex, I still have that i7-940 and spare tray if testing on it will help the cause.

Sure. It's the Mac Pro that I have the dump?

 

[Zeke D]

Sure. It's the Mac Pro that I have the dump?

yes, the 2010.

 

[tsialex]

yes, the 2010.

Sent reconstructed 142.0.0.0.0 and 140.0.0.0.0 for your 2010 Mac Pro by PM. Flash 142.0.0.0.0 with your dual Xeon tray, then test if the i7 tray can boot 142.0.0.0.0.

 

[LightBulbFun]

im curious since 142 seems to mess with CPU support code

has anyone shoved a X5687 into a MP on 142 just for the giggles of it/off chance?

 

[tsialex]

Since I don't have any i7 socket 1366, I was curious with the going prices and went to eBay. A i7-990X costs double of the today's asking price of a comparable X5690.

All high clock 1366 i7s cost more, some more than double, than X5xxx equivalent Xeons.

I'll continue without having an i7 to test…

 

[MarkC426]

I currently have Mojave installed on another drive (just for testing).
10.14.1 with fw 140 on a w3xxx xeon.
Can I upgrade to 10.14.2/3 without any upgrade to fw if I use the app store updater?
I assume 140 is the current released fw version.

 

[bsbeamer]

141 & 142 were/are only available in developer preview (beta) testing. You should not encounter any FW update notice from 10.14.2/10.14.3 installer. If you do for some reason, just cancel it and please report here.

 

[handheldgames]

Well, I have single cpu trays with an i7-990x and a w3680 and a dual x5677’s at the house.

 

[Zeke D]

as expected, with 142, my dual X5670 worked just fine. Both my W3530, and the i7-940 are brick.

 

[Zeke D]

I don't know if anyone cares, but injecting 106A4 into 140 allows the i7-940 to boot.

[doublepost=1551838258][/doublepost]Strange things are afoot at the circle K.

Bluetooth is "not available" on sierra running off the i7...

 

[BillyBobBongo]

Strange things are afoot at the circle K.

May I congratulate you on a most Excellent movie reference! [Air Guitar Riff]

Probably going to get told off by @tsialex for going off topic now...it was worth it!

 

[kucharsk]

I already changed the name of the thread to warn everyone that 142.0.0.0.0 bricks with W3xxx Xeons. I bet that people will flash 142.0.0.0.0 anyway…

Where is it listed what kind of CPU you have? All the system report output just says "Xeon."

 

[hwojtek]

Terminal:

sysctl -a | grep machdep.cpu

 

[Macschrauber]

btw. as a repair for bricked W Xeons with 142.0.0.0.0 is to replace the CPU (tray) and downgrade with your rom Backup.

...should not told often enough to backup your rom !!!!

 

[tsialex]

Well, I have single cpu trays with an i7-990x and a w3680 and a dual x5677’s at the house.

I'd like to confirm if hex i7s, the ones that still have support with 206C0/206C2, can work or not.

I forgot that @Zeke D i940, thx for testing, was cutout from microcode support since MP51.0089.B00, so we still don't know if supported i7s can still work with 142.0.0.0.0.
[doublepost=1551887561][/doublepost]Anyway, we already have 4 confirmations of W3xxx not working.

 

[MarkC426]

141 & 142 were/are only available in developer preview (beta) testing. You should not encounter any FW update notice from 10.14.2/10.14.3 installer. If you do for some reason, just cancel it and please report here.

Updated to 10.14.3 without issue......
I was just a bit paranoid with the talk of ‘bricking’ in this thread.

 

[tsialex]

Updated to 10.14.3 without issue......
I was just a bit paranoid with the talk of ‘bricking’ in this thread.

Mac Pro up to 5,1 don't upgrade firmware with updates, only with full installs.

142.0.0.0.0 it's a beta bootROM that you only can get if you are a developer that have access to the Seed channel.

No end user can install it inadvertently.

 

[tsialex]

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. Updated APFSJumpStart EFI Module

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.