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

[crjackson2134]

I see it like this …

Even if you go straight to a replacement backplane, you still need the reconstruction to preserve Apple Services, and to cleanup the undoubtedly outdated parts of the Firmware.

One that’s done, a MATT card is an invaluable backup investment (low cost imho), since you’ll be dealing with a used backplane, and have no idea how long the SPI on that board will last.

You could install the replacement backplane, and be in the very same position weeks later. Having the MATT card in hand, and the new Firmware already available, it would be a zero-cost recovery.

 

[48k Productions]

MATT card is very useful for diagnostics, a must for someone that have more than one Mac Pro.

Maybe the cost of a MATT card + BootROM reconstruction service + the possibility of having to replace the backplane doesn't make much economical sense to you, since you seem to do not have software licenses to care.

The only other Mac Pro that I have is an early 2008, which I have no plan of really bothering to boot up again.

as far as purchases go, this is all a tax write off for the business so I don’t mind haha

 

[48k Productions]

I see it like this …

Even if you go straight to a replacement backplane, you still need the reconstruction to preserve Apple Services, and to cleanup the undoubtedly outdated parts of the Firmware.

One that’s done, a MATT card is an invaluable backup investment (low cost imho), since you’ll be dealing with a used backplane, and have no idea how long the SPI on that board will last. You could install the replacement backplane, and be in the very same position weeks later.

Great input/point! Picking up a MATT right now

 

[ToniCH]

Having earlier had my cMP 2009 4,1->5,1 boot rom reconstructed and flashed I now am planning to start using my cMP 2010.

Few questions about the boot rom / SPI Flash memory / MATT card. They have been discussed recently but facts are so scattered in various posts I found it little difficult to be sure so here goes:

1. SPI Flash memory corruption applies to all cMP 5,1's - not just 2009 4,1->5,1 -machines?
2. so, for preventive reasons boot rom reconstruction would be smart with 2010- models too? And at the same time getting the enableGOP injected would be obviously be useful too.

MATT-card:
1. I know that its a good practice to reflash with clean rom every few months but if problems occur...a MATT card prevents the original SPI flash from bricking by being the active SPI and thus sacrificing itself in such incident?
2. can the MATT card be reflashed back to operational status if it corrupts/bricks?
3. can the MATT card be (when operational) be flashed normally in situ like the original? And if necessary outside of the machine with for example CH341A programmer?

Thanks.

 

[tsialex]

1. SPI Flash memory corruption applies to all cMP 5,1's - not just 2009 4,1->5,1 -machines?

Cross-flashed early-2009s are extremely susceptible for reasons that I've explained multiple times here, but happens with all MacPro5,1.

Early made mid-2010s are also very much susceptible.

2. so, for preventive reasons boot rom reconstruction would be smart with 2010- models too? And at the same time getting the enableGOP injected would be obviously be useful too.

Yes to both.

MATT-card:
1. I know that its a good practice to reflash with clean rom every few months but if problems occur...a MATT card prevents the original SPI flash from bricking by being the active SPI and thus sacrificing itself in such incident?
2. can the MATT card be reflashed back to operational status if it corrupts/bricks?
3. can the MATT card be (when operational) be flashed normally in situ like the original?

Yes for all 3 questions. MATT card is a replacement SPI flash and all SPI reads and writes are routed to it when installed to the backplane.

And if necessary outside of the machine with for example CH341A programmer?

Yes, but you will have to buy a quality SOIC-8 clip, won't work with a cheap black Chinese clip or a fake Pomona from AliExpress.

You have to buy a real Pomona 5250 or 3M 923655-08. Both are expensive.

 

[ToniCH]

Yes, but you will have to buy a quality SOIC-8 clip, won't work with a cheap black Chinese clip or a fake Pomona from AliExpress.

You have to buy a real Pomona 5250/3M 923655-08. Both are expensive.

Thanks for the clarifications and tips. I do have 2 of those programmers with the chinese black clip. I've used them successfully few times. But, if you say so then I'll have to order one those better quality clips. The Pomona is less than 20€ (wow, much more expensive in ebay than Mouser) so in larger picture (tool to save the cMP) its not that bad, and its probably useful in other stuff too. But, in ideal case I don't need to do it outside the backplane. 👍

 

[tsialex]

Thanks for the clarifications and tips. I do have 2 of those programmers with the chinese black clip. I've used them successfully few times.

These are practically one use only. I've tried everything to make these to work better, even cleaning it with my ultrasound cleaner prior to using, but is just gigantic waste of time.

I have a box with something like 10 of these junk clips here.

But, if you say so then I'll have to order one those better quality clips. The Pomona is less than 20€ so in larger picture (tool to save the cMP) its not that bad, and its probably useful in other stuff too. But, in ideal case I don't need to do it outside the backplane. 👍

I said expensive relatively to the junk Chinese clips that are less than two dollars. You'll also have to make and solder the cables for the Pomona/3M clips.

 

[KingCornWallis]

Appreciate all the insight of your original post. I arrived here only after seeing a Mac Pro listing on Ebay containing a 'Matt Card'... If this is as serious a problem as I am perceiving, I think the wiki would benefit from a section titled 'The Problem' that basically outlines what it is, why it's happening, and that all cMP's are highly susceptible (ticking timebomb) to this issue and should take preventative action.

(The answers to questions below may also be fitting to add to the wiki, but I suppose 'anyone' can edit)

1. If my Mac Pro is BRICKED, what behavior/activity (if any) from the Mac should I anticipate? How far into the Troubleshooting section of the Apple Technician guide (2009) will I get?
2. You mention 'failed garbage collection' in passing in referring to the core of the problem...but what is this and how often is it supposed to run? Is this akin to SSD wear reduction techniques like TRIM?
3. Where does the manual '3-month' number for re-flashing come from (likely related to previous question).
4. Can I back up my ROM if I already have OC installed? Do I need to undo OC first? Or would it be beneficial to have both a clean 144.0.0.0.0 and the latest modified OC ROM?
5. If I am competent with a flash tool and soldering iron, and am proactive in regularly re-flashing my BootROM, does a Matt Card serve any real purpose? Is it just a redundant BootRom in this case?
6. When is a reconstruction service needed? Just when installing a new chip? You mention it being a better alternative to Cross flashing MP4,1 to 5,1 as well?
7. Is the NVRAM/BootROM equivalent to a Windows BIOS/BIOS chip?
8. Proactively re-flashing the BootROM just helps with garbage collection, but the silicon is still being worn out at an increased rate in newer macOS: What metrics do we have to look at for how often writes to the NVRAM are done in Ventura/Sonoma, to compare against the Data Sheet numbers for a new flash chip to discern longevity?

 

[tsialex]

Appreciate all the insight of your original post. I arrived here only after seeing a Mac Pro listing on Ebay containing a 'Matt Card'... If this is as serious a problem as I am perceiving, I think the wiki would benefit from a section titled 'The Problem' that basically outlines what it is, why it's happening, and that all cMP's are highly susceptible (ticking timebomb) to this issue and should take preventative action.

(The answers to questions below may also be fitting to add to the wiki, but I suppose 'anyone' can edit)

1. If my Mac Pro is BRICKED, what behavior/activity (if any) from the Mac should I anticipate?

More commonly is a power on with no POST and EFI_DONE is off. Also you can get a soft-brick/read-only NVRAM, where you can't change the default boot disk or save anything to the NVRAM, the most common example is the Mac Pro only booting Windows, if you remove the Windows disk you can't boot macOS or anything else.

How far into the Troubleshooting section of the Apple Technician guide (2009) will I get?

Apple Technician Guide diagnostics are not really useful for people that does not have access to Apple (internal, factory and AASPs only) interfaces (for the MacPro5,1 is the LITTLE FRANK one).

If you don't have the interface to follow the steps, you won't get anywhere with minimal config.

Apple Technician Guide was never intended to be read by anyone else than Apple/AASPs technicians.

2. You mention 'failed garbage collection' in passing in referring to the core of the problem...but what is this and how often is it supposed to run?

Garbage collection is automatically executed when the BootROM PEI detects that the VSS store is below 2048 bytes.

The problem is that sometimes the process fails completely, very common when there are a big payload, like the 25KB+ of SSV macOS releases ( Big Sur -> ), or when you have Windows SecureBoot signing of the BootROM.

Is this akin to SSD wear reduction techniques like TRIM?

Nope. TRIM is just a process to erase NAND cells proactively, not related in any way.

Also not related to wear leveling in the way that you think.

3. Where does the manual '3-month' number for re-flashing come from (likely related to previous question).

Empirical evidence after hundreds of dumps/analysis/RI.

Since following the 90day re-flash with the reconstructed BootROM image worked for me since 2018, this is my advice to avoid garbage collection failed runs and lot of people here does the same.

For people that constantly boot different macOS releases, constantly install macOS releases, is better to follow a more on hands look of the VSS store.

Booting between macOS releases newer than Mojave and back to Mojave and earlier cause a lot of writes inside the VSS store because the syntax of BT related variables changed with Catalina. This is one of the often overlooked issues that cause garbage collection to run earlier than expected.

4. Can I back up my ROM if I already have OC installed?

Avoid doing that, sometimes the dump is not realiable when dumped from OC/unsupported macOS releases. Several failures over the years.

Do I need to undo OC first? Or would it be beneficial to have both a clean 144.0.0.0.0 and the latest modified OC ROM?

OC does not modify the Mac Pro firmware, you are mistaking things here, there is no OC ROM.

Since modern macOS releases require a lot of space inside the VSS store compared to 10.6/10.14, now you have a problem with the poor garbage collection of MacPro4,1/5,1. OC/OCLP indirectly also require some considerable space inside the VSS store.

5. If I am competent with a flash tool and soldering iron, and am proactive in regularly re-flashing my BootROM, does a Matt Card serve any real purpose?

Nope, just ease of use.

Is it just a redundant BootRom in this case?

Yes.

6. When is a reconstruction service needed? Just when installing a new chip? You mention it being a better alternative to Cross flashing MP4,1 to 5,1 as well?

When you start to have issues with garbage collection, missing secondary VSS store or when you have a cross-flashed Mac Pro or an old mid-2010 and want to get to mid-2012 0x0d BootROM version, with all the possible upgrades.

Most people do it proactively and to get the most recent components of the BootROM that are not upgraded with the EFI firmware, like having a mid-2012 firmware with an early-2009 or mid-2010.

7. Is the NVRAM/BootROM equivalent to a Windows BIOS/BIOS chip?

Not so much, if you want to make a parallel the Mac Pro BootROM is more equivalent of the UEFI firmware of servers than a normal PC motherboard.

BIOS as we know from the IBM PC is something that died back in 2012ish, even with PCs.

8. Proactively re-flashing the BootROM just helps with garbage collection, but the silicon is still being worn out at an increased rate in newer macOS:

Yes.

What metrics do we have to look at for how often writes to the NVRAM are done in Ventura/Sonoma, to compare against the Data Sheet numbers for a new flash chip to discern longevity?

This generation of flash memories used by Apple with MacPro5,1 does not have any wear management, so, we don't know how current macOS releases really affect durability, but is something that we can see empirically with so much frequent garbage collection runs compared with older macOS and that the VSS store space is a lot smaller.

 

[KingCornWallis]

More commonly is a power on with no POST and EFI_DONE is off. Also you can get a soft-brick/read-only NVRAM, where you can't change the default boot disk or save anything to the NVRAM, the most common example is the Mac Pro only booting Windows, if you remove the Windows disk you can't boot macOS or anything else.



Apple Technician Guide diagnostics are not really useful for people that does not have access to Apple (internal, factory and AASPs only) interfaces (for the MacPro5,1 is the LITTLE FRANK one).

If you don't have the interface to follow the steps, you won't get anywhere with minimal config.

Apple Technician Guide was never intended to be read by anyone else than Apple/AASPs technicians.



Garbage collection is automatically executed when the BootROM PEI detects that the VSS store is below 2048 bytes.

The problem is that sometimes the process fails completely, very common when there are a big payload, like the 25KB+ of SSV macOS releases ( Big Sur -> ), or when you have Windows SecureBoot signing of the BootROM.



Nope. TRIM is just a process to erase NAND cells proactively, not related in any way.

Also not related to wear leveling in the way that you think.



Empirical evidence after hundreds of dumps/analysis/RI.

Since following the 90day re-flash with the reconstructed BootROM image worked for me since 2018, this is my advice to avoid garbage collection failed runs and lot of people here does the same.

For people that constantly boot different macOS releases, constantly install macOS releases, is better to follow a more on hands look of the VSS store.

Booting between macOS releases newer than Mojave and back to Mojave and earlier cause a lot of writes inside the VSS store because the syntax of BT related variables changed with Catalina. This is one of the often overlooked issues that cause garbage collection to run earlier than expected.



Avoid doing that, sometimes the dump is not realiable when dumped from OC/unsupported macOS releases. Several failures over the years.



OC does not modify the Mac Pro firmware, you are mistaking things here, there is no OC ROM.

Since modern macOS releases require a lot of space inside the VSS store compared to 10.6/10.14, now you have a problem with the poor garbage collection of MacPro4,1/5,1. OC/OCLP indirectly also require some considerable space inside the VSS store.



Nope, just ease of use.



Yes.



When you start to have issues with garbage collection, missing secondary VSS store or when you have a cross-flashed Mac Pro or an old mid-2010 and want to get to mid-2012 0x0d BootROM version, with all the possible upgrades.

Most people do it proactively and to get the most recent components of the BootROM that are not upgraded with the EFI firmware, like having a mid-2012 firmware with an early-2009 or mid-2010.



Not so much, if you want to make a parallel the Mac Pro BootROM is more equivalent of the UEFI firmware of servers than a normal PC motherboard.

BIOS as we know from the IBM PC is something that died back in 2012ish, even with PCs.



Yes.



This generation of flash memories used by Apple with MacPro5,1 does not have any wear management, so, we don't know how current macOS releases really affect durability, but is something that we can see empirically with so much frequent garbage collection runs compared with older macOS and that the VSS store space is a lot smaller.

Much appreciated, very thorough. I am aware of the Legacy BIOS vs UEFI discrepancy, common folk still like to say BIOS. I should have remembered we aren't common!

Anyhow, only a few questions remain then.
1. What tangible 'upgrades' are there to be had from a Mid 2012 BootRom? I guess I thought the 2010 was largely comparable to the 2012.
2. I guess I misunderstood what OC is doing exactly to the BootRom. I just know that when booted into Sonoma and I look at the BootRom version, it is not 144.0.0.0.0 (which is the technical limit; I guess Spoofing?) All the magic is in the EFI partition....So I should be able to make a good dump via the guide on here if I boot into Mojave?
3. If you had a Mid-2010 Mac Pro and did NOT have any problems with Garbage Collection or VSS store, would YOU buy a new MX25L3206EM2I-12G chip and follow through to a new reconstruction?

 

[tsialex]

Much appreciated, very thorough. I am aware of the Legacy BIOS vs UEFI discrepancy, common folk still like to say BIOS. I should have remembered we aren't common!

Anyhow, only a few questions remain then.
1. What tangible 'upgrades' are there to be had from a Mid 2012 BootRom?

Hardware descriptor Base_21, instead of Base_20, Fsys version 0x0D, instead of something like 0x04 or 0x05.

I guess I thought the 2010 was largely comparable to the 2012.

Not so simply - Apple made several upgrades over the years, can make this comparison only to mid-2010s made very late in 2011 and in 2012.

The mid-2010s made back in 2010 and most of 2011 are more similar to the refurbished B08 early-2009s than to a mid-2012.

2. I guess I misunderstood what OC is doing exactly to the BootRom. I just know that when booted into Sonoma and I look at the BootRom version, it is not 144.0.0.0.0 (which is the technical limit; I guess Spoofing?) All the magic is in the EFI partition....So I should be able to make a good dump via the guide on here if I boot into Mojave?

Yes, if you boot Mojave natively.

3. If you had a Mid-2010 Mac Pro and did NOT have any problems with Garbage Collection or VSS store, would YOU buy a new MX25L3206EM2I-12G chip and follow through to a new reconstruction?

The SPI flash memory of a mid-2010 is just one year newer than the one from an early-2009. mid-2010s are also entering the age of failed flash memories.

 

[Starkluke]

Hi, ive tried every method i could find online to flash my Mac Pro 4,1 Firmware to the 5,1 Firmware, without success.
I red tsialex offers a boot ROM reconstruction service, which i would like to use.
Im running El Capitan 10.11.6 with the Boot ROM Version MP41.0081.B07
Ive attached an image of my Mac Pros specs

 

[tsialex]

Hi, ive tried every method i could find online to flash my Mac Pro 4,1 Firmware to the 5,1 Firmware, without success.
I red tsialex offers a boot ROM reconstruction service, which i would like to use.
Im running El Capitan 10.11.6 with the Boot ROM Version MP41.0081.B07
Ive attached an image of my Mac Pros specs

Sure, I'll send you a PM with all the info, required files, service fee and turnaround time.

 

[larsba]

Sure, I'll send you a PM with all the info, required files, service fee and turnaround time.

Hello, could you send me a pm regarding reconstruction service too? I recently did a bootrom dump and discovered that my vss store is almost full (26214 free space). Got a 4,1 flashed to 5,1. I am having trouble booting into Big Sur 11.7.4 sometimes.
Thanks a lot!

 

[tsialex]

Hello, could you send me a pm regarding reconstruction service too?

Hi! PM sent.

I recently did a bootrom dump and discovered that my vss store is almost full (26214 free space).

This value is below what you expect with a 8-DIMM dual CPU tray and a problem with software updates when macOS has SSV.

Got a 4,1 flashed to 5,1. I am having trouble booting into Big Sur 11.7.4 sometimes.
Thanks a lot!

Ok, the ancient BootBlock installed by the cross-flashing process could be helping here, but is repairable.

 

[pizzigri]

Hi Alex, could you give me an alternative that will work in the cMP 5.1 for the MX25L3206EM2I-12G that is unavailable for at least another 3-4 months at Mouser? I already have the reconstructed bootrom you made some time ago for me!

 

[tsialex]

Hi Alex, could you give me an alternative that will work in the cMP 5.1 for the MX25L3206EM2I-12G that is unavailable for at least another 3-4 months at Mouser? I already have the reconstructed bootrom you made some time ago for me!

Apple used all the models below, with MacPro4,1 and MacPro5,1:

• SST 25VF032B
• Macronix MX25L3205A
• Macronix MX25L3205D
• Macronix MX25L3206E

Apple's efiflasher code also have support for WinBond W25X32, but I've never saw a backplane with one or installed one to a backplane. Since Apple used WinBond W25X64 with several 2012/2013 Macs, exact same algorithm as W25X32, I suppose that it would also work. If you test one please report the results.

 

[trifero]

Time to reflash my second Mac Pro ROM with the reconstructed one from @tsialex . As usual, all worked like charm.

 

[tsialex]

After a recent SPI flash failure with one of my MATT cards, today I had another unfortunate failure.

I knew for a very long time that the MOLEX connector for the LITTLE FRANK interface was rated for only 20 insertion/removal cycles, today I had my first connector failure ever:

I'll probably try to straighten and put back the damaged contact first, but even if I can do it successfully, I'll need to replace the MOLEX connector in the long run.

 

[trifero]

After a recent SPI flash failure with one of my MATT cards, today I had another unfortunate failure.

I knew for a very long time that the MOLEX connector for the LITTLE FRANK interface was rated for only 20 insertion/removal cycles, today I had my first connector failure ever:

View attachment 2393673

View attachment 2393670

I'll probably try to straighten and put back the damaged contact first, but even if I can do it successfully, I'll need to replace the MOLEX connector in the long run.

I didn´t know Little Frank has a life cycle. In my macs, I have use it twice each. Do you think it will be very hard to replace?

 

[tsialex]

I didn´t know Little Frank has a life cycle. In my macs, I have use it twice each. Do you think it will be very hard to replace?

If you don't remove/re-insert, won't be a problem. The failure happens when you constantly remove/re-insert the MATT card.

For people that have SMD soldering experience, is not difficult to replace it. You will probably need solder paste for temperature sensitive components, a must with these small plastic connectors.

 

[trifero]

If you don't remove/re-insert, won't be a problem. The failure happens when you constantly remove/re-insert the MATT card.

For people that have SMD soldering experience, is not difficult to replace it. You will probably need solder paste for temperature sensitive components, a must with these small plastic connectors.

Let´s hope you can put that connector straight. Usually they broke so easily. Good luck.

 

[crjackson2134]

The failure happens when you constantly remove/re-insert the MATT card.

How many insertion cycles do you estimate you had on this connector.

 

[tsialex]

How many insertion cycles do you estimate you had on this connector.

This one is a mid-2012 Mac Pro bought late last year, I doubt that I had anything near the expected 20 cycles, if much, I’d say half that.

I had a lot more removal/insertion cycles with a my older test mule, a mid-2010, with no problems whatsoever.

I’d say that if you take a lot of care, you will get 15 cycles without any surprises.

 

[joevt]

Does that molex connector have a name? An extension cable could make it last longer - you would only have to replace the extension cable.