Do the reconstructed never-booted BootROM images that you create for users include iCloud and other credentials? Or do they only include hardwareIDs?
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.
MP 1,1-5,1 MacPro5,1: BootROM thread | 144.0.0.0.0
- Thread starter tsialex
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
Do the reconstructed never-booted BootROM images that you create for users include iCloud and other credentials? Or do they only include hardwareIDs?
I have a never-booted image, courtesy of @tsialex, and I'm presuming I could have that flashed to the MATT card before shipping to me, which would be very advantageous.
I'm trying to understand why it would be very advantageous to you?
Just to not have to install the MATT card, boot Recovery, disable SIP, shutdown, power up in Firmware Programming Mode and then run ROMTool? You already do exactly that whenever you need to re-flash the never booted BootROM image.
ah alright... I wasn't sure if I would be able to boot up with the default MATT card as delivered, without first flashing it with something. That makes sense then..thanks for the clarification! And thanks again for the never-booted image...really glad I got that.. I'm much less worried now about what happens if my rom burns out..
Can a never-booted-BootROM image be flashed to a new MXIC MX25L3206E chip instead of the MP51.fd image ?
That is, de-solder the older (potentially end-of-life SPI chip), flash a new MXIC MX25L3206E with your never-booted-BootROM image and solder the new chip onto the logic board ?
Thanks Alex. I did read the first post and it mentions:
- program MP51.fd to the replacement SPI flash memory (Macronix MX25L3205A/MX25L3205D/MX25L3206E, SST 25VF032B),
- program MP51.fd (or a never-booted-BootROM image if you have one) to the replacement SPI flash memory (Macronix MX25L3205A/MX25L3205D/MX25L3206E, SST 25VF032B),
If you have a saved dump, or better yet reconstructed image, use it, no problem. I'll edit the first post to be clear, btw, if you already have a never booted BootROM image, you don't need to take the pictures or contact a firmware engineer for a reconstruction.
I think that I already explained why I do it this way - I always flash the MP51.fd with the external programmer for a series of reasons that probably matter most to me or someone that does this repairs with some frequency:
- easier to boot than a previously saved backup dump (that could have failed GC or other issues)
- always the same checksum
- it's the image I use to test the SPI flash lot when it arrives
- faster and practical, several SPI can be flashed sequentially
- I can pick a SPI from the already flashed stash and install immediately to the backplane.
After the Mac Pro is booting, I flash it's own reconstructed image via ROMTool or flashrom.
Last edited:
I try to upgrade a MP4,1 and I saved a backup dump, looking pretty good:
After SMC reset and NVRAM reset (5 chimes) it looked a bit worse:
and after updating with the latest 4,1 firmware from Apple it was even worse:
Is this normal? I was expecting to see a rather bad one in the beginning and a much better one after the Apple update.
After SMC reset and NVRAM reset (5 chimes) it looked a bit worse:
and after updating with the latest 4,1 firmware from Apple it was even worse:
Is this normal? I was expecting to see a rather bad one in the beginning and a much better one after the Apple update.
MemoryConfig variable format changes between EFI releases, so it's somewhat expected. 21 MemoryConfig entries are a problem and not ok and it's usually a sign of failed garbage collection.
You can force a deep NVRAM reset (with a wired keyboard press continuously CMD-Option-P-R until the 5th chime) and see if garbage collection is really working.
Btw, a lot of MP5,1 knowledge don't directly apply to the MP4,1 NVRAM volume, since the MP4,1 NVRAM is only similar - also MP4,1 firmwares are deprecated since 2010 and it's not really meaningful to actively track any MP4,1 only
issues.
Thank you for your quick reply. I didn't do another deep NVRAM reset, since the first one (after the initial dump) was rather disappointing.
However, the worst is yet to come. After updating to MP51.007F.B03 it was really bad:
After that, I installed High Sierra and it updated the firmware again, which looked a bit better:
My intentions now are to update it to 144.0.0.0.0 and I think this would be a good moment to follow your advice, to force a deep NVRAM reset. Before updating it to 144.0.0.0.0 or after?
What exactly means "if garbage collection is working"? The amount of free space will / should increase?
However, the worst is yet to come. After updating to MP51.007F.B03 it was really bad:
After that, I installed High Sierra and it updated the firmware again, which looked a bit better:
My intentions now are to update it to 144.0.0.0.0 and I think this would be a good moment to follow your advice, to force a deep NVRAM reset. Before updating it to 144.0.0.0.0 or after?
What exactly means "if garbage collection is working"? The amount of free space will / should increase?
Before the upgrade and after it's done.
NVRAM volume and garbage collection work in a very counterintuitive way, to get a better understanding of why/how start reading from the post below:
OpenCore on the Mac Pro
Thank you @tsialex as always! Your knowledge overwhelms me! I have fixed my problem just reflashing BootROM using the reconstructed one that you made it for me. Now everything is ok again
forums.macrumors.com
Also there are several posts, some pages ago on this thread, that I explain all the issues and doubts people have.
Btw, hybrids (early-2009 cross-flashed to MP5,1 frimwares) are extremely prone to bricking, you are seeing the problem right now.
No one should use a MP4,1>5,1 with modern macOS releases, reconstruction is obligatory to anyone with a dual CPU tray.
Thanks. I'll read the entire thread, it will probably take two or three days 
In the end, I'd like to install Open Core and Big Sur on this MP and I try to make sure I understand what I'm doing.
In the end, I'd like to install Open Core and Big Sur on this MP and I try to make sure I understand what I'm doing.
I truly recommend you to obtain a reconstructed ROM from @tsialex. It´s almost a must in 4,1->5,1 Mac Pro.Thanks. I'll read the entire thread, it will probably take two or three days
In the end, I'd like to install Open Core and Big Sur on this MP and I try to make sure I understand what I'm doing.
Last edited:
I understand that, but I suppose those 21 MemoryConfig entries were not real. This dump was right after updating the firmware to the latest 4,1 from Apple, I guess they were included in that image.
The same applies to the 27 MemoryConfig entries that just appeared after updating to MP51.007F.B03
I didn't change anything related to RAM, it has 3 x 8GB and had them in the same slots since the initial dump. Maybe I'm wrong, but I was under the impression that a MemoryConfig entry is added only if something changed about the RAM.
Yes, you are mistaken - learn how the circular log works. Btw, I even explained that the MemoryConfig format changes between EFI releases on post #4,866. See for yourself with UEFITool.
Also the exactly same issue, variable multiplication, also happens with BT related entries/variables - but these variable syntax changes happens between macOS releases while MemoryConfig syntax was changed with EFI releases.
Paging @Macschrauber
First off, thank you for your help in the OC thread! I got my machine back!!!!!
Here are my 3 BootROM entries:
Based on the above, is my BootROM all good?
I haven't yet installed OpenCore -- just running Mojave 10.14.6, but so far everything has been stable.
If I receive yours or @tsialex blessing [basically someone who understands BootROMS] that my BootROM is good to go, then I will put OpenCore 0.7.6, and then just need to figure out Windows.
@KeesMacPro @zzzippp Thank you guys as well for your help/comments!!!
First off, thank you for your help in the OC thread! I got my machine back!!!!!
Here are my 3 BootROM entries:
Based on the above, is my BootROM all good?
I haven't yet installed OpenCore -- just running Mojave 10.14.6, but so far everything has been stable.
If I receive yours or @tsialex blessing [basically someone who understands BootROMS] that my BootROM is good to go, then I will put OpenCore 0.7.6, and then just need to figure out Windows.
@KeesMacPro @zzzippp Thank you guys as well for your help/comments!!!
Static binary analysis is just a tool to help you catch the most common failures. It's not a way to say that a dump is free of problems - at best, it's free of the most common problems. Failures frequently happen in unpredictable ways that you cannot catch just with static binary analysis.
For example, a cross flashed Mac Pro (an early-2009 that had the EFI upgraded to MP5,1 via MacEFIROM tool), fails not just because of the usual NVRAM problems, but also because several deprecated MP4,1 variables continue to be stored inside the VSS store, the circular log have less space to work or is fragmented and garbage collection is forced to work more and with less space (same issue of Windows SecureBoot certificates). This overtime leads to NAND cell exhaustion failure and binary static analysis can't catch it because the variables are correctly stored and valid, but shouldn't be there taking valuable space anymore. Also, the MP4,1 NVRAM volume, and the MP4,1 BootBlock, are slightly incompatible with the MP5,1 EFI and overtime this will result on the corruption of the NVRAM volume. All cross-flashed early-2009 Mac Pro need to be correctly and fully upgraded to the 0x0D firmware standard that Apple sent with the last run of mid-2012s made. This is crucial for early-2009s with dual CPU trays where the NVRAM volume have much less available space (two memory controllers, more DIMMs to have the configuration and the DIMM SPD cached stored = less available VSS store space).
Another common problem where binary static analysis is useless, the BootBlock AAPLEFI1.88Z.0005.I00.1010071430 that Apple installed back in the factory with mid-2010s and most mid-2012s (also installed on cross-flashed early-2009 by the MacEFIROM tool) have a cold boot hang during POST that only presents itself when you have a PCIe switched card, Apple worked for a long time with @crjackson2134 to finally solve this nasty one with 144.0.0.0.0. The problem is, you can't install the updated BootBlock without a BootROM reconstruction, since efiflasher don't upgrade anything besides the EFI firmware inside the BootROM image. So, all Mac Pros that have a BootBlock earlier than AAPLEFI1.88Z.0005.I00.1904121247 (practically all) are subject to this bug, but just people with PCIe switched cards will notice it. Btw, back then @crjackson2134 Mac Pro and mine were practically twins in every way except that his Mac Pro is a mid-2012 and mine was an early-2009 that bricked and was reconstructed with a much newer BootBlock (AAPLEFI1.88Z.0005.I00.1806081704) that was not so prone to the hang and I only had it sometimes while his almost identical config was an edge case where it had the cold boot hang frequently - I would never have bothered to track it since it was very rare to happen with my own Mac Pro and I don't shutdown daily like him.
Anyway, back to the results from @Macschrauber tool, zero MemoryConfigs being present are just impossible - it's like to say you booted without RAM being identified and configured. So, something is seriously wrong with your NVRAM volume/hardware.
Last edited:
Thank you for the analysis!
What do you recommend I do from here on?
Also, is there any way for me to get the 2012 0x0D firmware standard that Apple sent onto my machine?
Last edited: