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

[tsialex]

Hmmmm. Sounds like users digging for Apple to release a SMC update for the 5,1 that could be applied to all the 4,1's out there. Considering there's never been an update to the unique dual SMC in the 4,1/5,1 platform, perhaps there is a physical / hardware design issue that prevents both SMC's from being updated?

It's not a problem with the Renesas/Hitachi micro-controler, the internal ROM is updatable - can't be dumped but can be updated.

 

[handheldgames]

It's not a problem with the Renesas/Hitachi micro-controler, the internal ROM is updatable - can't be dumped but can be updated.

Ok. So you think that both SMC ROM's are updatable in some manner or form.

Considering the parallel nature of both SMC ROM's, do you think they are updatable under normal system operation, or is the update something that would take place at the factory through the little frank connector on the logic board and from one of the debug pads on the CPU tray?

Considering there have been incorrect PCIe fan speed issues since the GTX680 / HD7950 was introduced, that still persist with the rx580, One might think there would have been an update in some matter or form along the way. While Apple has fixed issues with the EFI Firmware, updated cpu microcodes and added support for NVMe, perhaps they lack a method for reliably updating both SMC's.

 

[tsialex]

Ok. So you think that both SMC ROM's are updatable in some manner or form.

Considering the parallel nature of both SMC ROM's, do you think they are updatable under normal system operation, or is the update something that would take place at the factory through the little frank connector on the logic board and from one of the debug pads on the CPU tray?

The micro- controller used on MP4,1 is a evolution of the one used before on MP1,1 to MP3,1 and Apple software updated the MP2,1 SMC. The MP2,1 SMC update was modified/hacked to update the MP1,1 SMC and the MP1,1 got parity with the MP2,1.

Considering there have been incorrect PCIe fan speed issues since the GTX680 / HD7950 was introduced, that still persist with the rx580, One might think there would have been an update in some matter or form along the way. While Apple has fixed issues with the EFI Firmware, updated cpu microcodes and added support for NVMe, perhaps they lack a method for reliably updating both SMC's.

Speculation mode on - perhaps people who made the MP4,1 and MP5,1 SMC don't work at Apple anymore? I got a word that Apple let go, or changed to iPhone, a lot of Mac firmware engineers some years ago.

 

[Macschrauber]

Dont know if it is discussed that way: The smc upgrade for Mac Pro 1.1 to 2.1 is done in the efi shell via refit / refind.

Could it be that the Efi locks the smc in some way and it is open in the early boot process ?

 

[tsialex]

Dont know if it is discussed that way: The smc upgrade for Mac Pro 1.1 to 2.1 is done in the efi shell via refit / refind.

Could it be that the Efi locks the smc in some way and it is open in the early boot process ?

One of the characteristics of Renesas/NEC/Hitachi H8S embedded micro-controllers is that you have firmware protection. Like some un-dumpable GALs/PALs of years ago.

Since no one got a dump from MP5,1 SMC until now, and the documentation never says about reading the firmware, only writing, I think that the SMC is un-dumpable.

Some ThinkPads have similar H8S embedded micro-controllers, maybe we can get in contact with someone who know this controllers in depth and have some ideas.

 

[tsialex]

Never saw this before, look at this BootROM dump:

View attachment 797091

I have a suspect of why this happened, microcodes again, but in a different way. I'll confirm with the owner and report back.

I got my suspicion confirmed and can tell for sure: the owner tried an i7-940 with MP51.0089.B00, a CPU model that uses 106A4 microcode that Apple removed from the BootROM with all releases newer than MP51.0085.B00.

So the crashes caused the triple IASInstallPhaseList.

 

[Synchro3]

I got my suspicion confirmed and can tell for sure: the owner tried an i7-940 with MP51.0089.B00, a CPU model that Apple removed Intel microcode support from the BootROM with all releases newer than MP51.0085.B00.

So the crashes caused the triple IASInstallPhaseList.

Does that mean that all in the Mac Pro non used CPU's like W3690/X5690 are death row?

 

[tsialex]

Does that mean that all in the Mac Pro non used CPU's like W3690/X5690 are death row?

No, just some i7-920/940/960 that have the old C0/C1 revision and depends on 106A4 microcode that Apple removed. Some very early and rare ES Nehalem Xeons don’t work anymore too.

 

[flowrider]

^^^^Do you know if the added any, like the X5687?

Lou

 

[tsialex]

^^^^Do you know if the added any, like the X5687?

Lou

Support for high multiplier Xeons is not dependent of the microcode, but from the code that enables cores and multipliers. Someone needs to check it, I don't have anything faster than X5677 and X5680.

 

[tu2thepoo]

Does that mean that all in the Mac Pro non used CPU's like W3690/X5690 are death row?

I run a W3690 with no problems, FW138.

[edit]
correction, I'm running FW140 now thanks to tsialex!

 

[flowrider]

Support for high multiplier Xeons is not dependent of the microcode, but from the code that enables cores and multipliers. Someone needs to check it, I don't have anything faster than X5677 and X5680.

Ah, than it's probably still a no! AFAIK, no one has ever gotten a cMP 5,1 to boot with anything faster than an X5677 or an X5690.

Lou

 

[kings79]

HS, X5690's & FW140 no problem.

 

[theone29]

Link is dead. Help please.
Not Found
The requested URL /content/downloads/29/51/041-11143/pph1a58nrlrwgzi8ee5uig1wzr5v42t9gn/macOSUpd10.14.1.RecoveryHDUpdate.pkg was not found on this server.

Additionally, a 404 Not Found error was encountered while trying to use an ErrorDocument to handle the request.

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 macOSUpd10.14.1.RecoveryHDUpdate.pkg from the Apple SUCatalog (in the future, search the SUCatalog for the most recent RecoveryUpdate)
2) Open macOSUpd10.14.1.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 10.14.0 Mojave installer with the one you saved on your desktop.
5) Open the Mojave installer, will ask you to upgrade the firmware.

Please don't ask more about this, if you still can't get updated to 140.0.0.0.0 with this instructions, this BootROM is not for you yet - just wait 10.14.1 release in two or three weeks time.

 

[Sonin]

Thank you it worked for me! But i have a problem

After the Update to 140.0.0.0.0 i inserted my NVME SSD into the PCie Slot and booted into OS. I have chosen the NVME SSD as the Startvolume. After a Restart i got a Blackscreen with my RX 580. I went back to my stock MacPro GPU and nothing changed for me. Still got a Blacksreen....

 

[tsialex]

Thank you it worked for me! But i have a problem

After the Update to 140.0.0.0.0 i inserted my NVME SSD into the PCie Slot and booted into OS. I have chosen the NVME SSD as the Startvolume. After a Restart i got a Blackscreen with my RX 580. I went back to my stock MacPro GPU and nothing changed for me. Still got a Blacksreen....

What is your NVME SSD? 950PRO?

 

[Sonin]

What is your NVME SSD? 950PRO?

Nope a WD Black NVMe SSD 250 GB with Kingshare m.2 Nvme Pcie x4 Adapter. I went back to my HDD with Majove installed on it and have the same Problem with the Blackscreen.

 

[tsialex]

Link is dead. Help please.
Not Found
The requested URL /content/downloads/29/51/041-11143/pph1a58nrlrwgzi8ee5uig1wzr5v42t9gn/macOSUpd10.14.1.RecoveryHDUpdate.pkg was not found on this server.

Additionally, a 404 Not Found error was encountered while trying to use an ErrorDocument to handle the request.

We are in the future now… 10.14.DP3 RecoveryUpdate was succeeded by 10.14.DP4 RecoveryUpdate.

 

[tsialex]

Nope a WD Black NVMe SSD 250 GB with Kingshare m.2 Nvme Pcie x4 Adapter. I went back to my HDD with Majove installed on it and have the same Problem with the Blackscreen.

Some questions to help us troubleshoot your problem:

• Your Mac still chime?
• Can you access it via screen sharing?
• Do you see it on your network?
• With a Mac EFI GPU, can you boot another version of macOS?
  
• Can you boot a createinstallmedia 10.13.6?
• When you press the DIAG button, what leds are on? (page 33 of the Apple Technician Guide)

 

[Sonin]

Thank you for you reply!

I putted back the Original Apple GPU on the Second PCIe x16 Slot and the SSD on the first x16 Slot and it worked for me!

But my RX 580 are still not working. I have got the RX 580 Sapphire Nitro + and the RX 580 Sapphire Pulse. But none of them is working with jet. Neither on the first 16x and on the second 16x Slot...

 

[tsialex]

Thank you for you reply!

I putted back the Original Apple GPU on the Second PCIe x16 Slot and the SSD on the first x16 Slot and it worked for me!

But my RX 580 are still not working. I have got the RX 580 Sapphire Nitro + and the RX 580 Sapphire Pulse. But none of them is working with jet. Neither on the first 16x and on the second 16x Slot...

I'll PM instructions for you do a BootROM dump. I'll check it if you don't have any corruption or trash into it.

 

[tsialex]

Hi!

Just bought Kingston A1000 NVMe ssd.
I have macpro 5.1 and macpro 2009, with firmware updated to 5.1
macOS 10.13.6 on both

In 3 PCIe slots this drive have a speed for both write/reading only 382mb/sec (even in 16x slot) and only in one slot it have 743mb/sec (but not 1300!!) This result same on 2 computers, I have tested different adapters m.2-PCIe with same result. (also I have tested with adapters from AHCI kingston predator ssd, who have 1300mb/sec on my computer)

Any ideas? can kext modifyin magic be better then native Hi Sierra support for NVMe?
may be cause of slow speed is not fresh firmware of ssd?

If I load in windows 8.1, windows see this drive, it appears as a volume. But Kingston SSD Manager v1.1.1.5 program not see it. It see AHCI drives, but not see this NVMe - any ideas?

Yes, it would be great!

Let's continue here, @Vadium. The old MP51.0084.B00 topic needs to rest in peace.

Your Kingston Black NVMe SSD model SA1000M8/960G tops at 1500MB/s read and 1000MB/s write, so your throughput is wrong.

Is it possible to update only ROM, but not install Mojave and still on HiSierra?

Yes, I can reconstruct your BootROM for you with 140.0.0.0.0 and you can use High Sierra, no problem.
[doublepost=1540215622][/doublepost]PM sent.

 

[marcoscc]

Support for high multiplier Xeons is not dependent of the microcode, but from the code that enables cores and multipliers. Someone needs to check it, I don't have anything faster than X5677 and X5680.

I have Pair of Intel Xeon X5690 3.46GHz SLBVX 12MB 6-Core LGA1366 Processor CPU / 128G RAM, with Firmware
138.0.0.0.0 and Mojave 10.14 working perfectly.

 

[tsialex]

I have Pair of Intel Xeon X5690 3.46GHz SLBVX 12MB 6-Core LGA1366 Processor CPU / 128G RAM, with Firmware
138.0.0.0.0 and Mojave 10.14 working perfectly.

X5690 and W3690 are not a problem, fully supported, but faster Xeons like X5687 (4 cores and 8 threads, 3.6GHz) and X5698 (2 cores and 4 threads, 4.4GHz).

 

[VaZ]

I have Pair of Intel Xeon X5690 3.46GHz SLBVX 12MB 6-Core LGA1366 Processor CPU / 128G RAM, with Firmware
138.0.0.0.0 and Mojave 10.14 working perfectly.

X5687 (3.6GHz quad-core) and the X5698 (4.4GHz dual-core) is what we need to test with Mojave and 140.0.0.0