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

tsialex · Feb 18, 2022

From High Sierra, same mid-2010 Mac Pro Server:

Code:

HighSierra:~ tsialex$ ./dmidecode |grep FirmwareFeatures
	FirmwareFeatures: 0xC00C1403
	FirmwareFeaturesMask: 0xFF1FFF3F
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
	<key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
	<data>
	A1QMwA==
	</data>
</dict>
</plist>
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
	<key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
	<data>
	P/8f/w==
	</data>
</dict>
</plist>

From the NVRAM:

FirmwareFeatures = 0xC00C5403
FirmwareFeaturesMask = 0xFF1FFF3F

From the SMBIOS:

FirmwareFeatures: 0xC00C1403
FirmwareFeaturesMask: 0xFF1FFF3F

So, Bit 14 (FW_FEATURE_SUPPORTS_HIGH_SPEED_USB) changes from the NVRAM to the SMBIOS.

Dayo · Feb 18, 2022

Is anyone able to advance an hypothesis on what is happening?
Why Apple seemed to have stopped updating the SMBIOS setting, then apparently actually degraded it in 2012 MP51?

cdf · Feb 18, 2022

Dayo said:
Is anyone able to advance an hypothesis on what is happening?
Why Apple seemed to have stopped updating the SMBIOS setting, then apparently actually degraded it in 2012 MP51?

Well, at least we know that the missing bit (bit 14) in the SMBIOS is not due to cross flashing. The other missing bit in the 2012 model could be explained by an older firmware version in the machine used for the test.

It would be interesting to know now if such a discrepancy also exists in the 3,1.

Dayo · Feb 18, 2022

cdf said:
Well, at least we know that the missing bit (bit 14) in the SMBIOS is not due to cross flashing. The other missing bit in the 2012 model could be explained by an older firmware version in the machine used for the test.

Does seem unrelated to cross flashing indeed.
The tested 2012 model had rebuilt firmware and this presumably would have the latest version.

It would be interesting to know now if such a discrepancy also exists in the 3,1.

I understood that @joevt found them to be the same but no harm in getting additional data points for all the versions looked at.

We are missing output from an unflashed MP41, if one can still be found.
Actually would also love to see from unmodified MP51 as well, just to have certainty.

cdf · Feb 18, 2022

Dayo said:
We are missing output from an unflashed MP41, if one can still be found.

Indeed, these are probably rare at this point. @Bmju would your Mac Pro still have the 4,1 firmware by any chance?

Syncretic · Feb 18, 2022

Another data point, possibly irrelevant - but it differs from the one @Dayo posted earlier:

Code:

sh-3.2# system_profiler SPHardwareDataType
Hardware:

    Hardware Overview:

      Model Name: Mac Pro
      Model Identifier: MacPro3,1
      Processor Name: Quad-Core Intel Xeon
      Processor Speed: 3.2 GHz 
      Number of Processors: 2
      Total Number of Cores: 8
      L2 Cache (per Processor): 12 MB
      Memory: 24 GB
      Boot ROM Version: MP31.006C.B05
      SMC Version (system): 1.25f4
      {snip}

sh-3.2# ./dmidecode | grep FirmwareFeatures
        FirmwareFeatures: 0xC0001403
        FirmwareFeaturesMask: 0xC0003FFF

sh-3.2# nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
        <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
        <data>
        FxQAwA==             // Hex c0 00 14 17
        </data>
</dict>
</plist>

sh-3.2# nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
        <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
        <data>
        /z8AwA==            // Hex c0 00 3f ff
        </data>
</dict>
</plist>
sh-3.2#

FW_FEATURE_DISABLE_BOOTSCRIPT_WORKAROUND      0x80000000U // 31
FW_FEATURE_UNKNOWN_BIT30                      0x40000000U // 30
FW_FEATURE_SUPPORTS_TRB                       0x00001000U // 12
FW_FEATURE_PLATFORM_SECURITY_POLICY_01        0x00000400U // 10
FW_FEATURE_SUPPORTS_NET_BOOT                  0x00000010U // 4 // present in NVRAM, not SMBIOS
FW_FEATURE_SUPPORTS_TARGET_DISK_MODE          0x00000004U // 2 // present in NVRAM, not SMBIOS
FW_FEATURE_SUPPORTS_CD_DRIVE_BOOT             0x00000002U // 1
FW_FEATURE_SUPPORTS_CSM_LEGACY_MODE           0x00000001U // 0

FW_FEATURE_SUPPORTS_HIGH_SPEED_USB            0x00004000U // 14 // ABSENT in both NVRAM and SMBIOS

Macschrauber · Feb 18, 2022

I booted again, Mavericks, this time with 144.0.0.0

dmidecode:

Code:

	FirmwareFeatures: 0xC00C1403
	FirmwareFeaturesMask: 0xFF1FFF3F

first Boot with 0xC00C0403 was with MP51.88Z.0087.B00.1804181525 and Mavericks, unmodified

tsialex · Feb 18, 2022

Mid-2012, from HighSierra, exactly the same install/disk as I used first on the mid-2010 Server, both on 144.0.0.0.0/0x0D (please note that the mid-2010 Server still have the personality server on the BootROM, as sent from factory).

Code:

HighSierra:~ tsialex$ system_profiler SPHardwareDataType
Hardware:

    Hardware Overview:

      Model Name: Mac Pro
      Model Identifier: MacPro5,1
      Processor Name: 6-Core Intel Xeon
      Processor Speed: 3,46 GHz
      Number of Processors: 1
      Total Number of Cores: 6
      L2 Cache (per Core): 256 KB
      L3 Cache: 12 MB
      Memory: 32 GB
      Boot ROM Version: 144.0.0.0.0
      SMC Version (system): 1.39f11
      SMC Version (processor tray): 1.39f11
      Serial Number (system): C0xxxxxxF4MC
   
HighSierra:~ tsialex$ ./dmidecode | grep FirmwareFeatures
    FirmwareFeatures: 0xC00C1403
    FirmwareFeaturesMask: 0xFF1FFF3F
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
    <data>
    F1QMwA==
    </data>
</dict>
</plist>
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
    <data>
    P/8f/w==
    </data>
</dict>
</plist>

FirmwareFeatures from the NVRAM (F1QMwA== / 0xC00C5417) doesn't match the mid-2010 one (A1QMwA== / 0xC00C5403), bits 3 (FW_FEATURE_UNKNOWN_BIT3 ) and 5 (FW_FEATURE_SUPPORTS_SLING_SHOT) are now on. Something is going on.

Mac Pro mid-2012:

Mac Pro Server mid-2010:

Tomorrow I'll check MP41.0081.B03 with my early-2009 B08, it's the oldest working BootROM image I have, with the same HighSierra/HFS install, so, I hope to boot it with just -no_compat_check.

Btw, I'm compiling the data on a spreadsheet.

Bmju · Feb 18, 2022

cdf said:
Indeed, these are probably rare at this point. @Bmju would your Mac Pro still have the 4,1 firmware by any chance?

I think. I haven't had any time to even play around with it at all, apart from using it to notice then check that fix for AudioDxe volume on old macOS which don't produce SystemAudioVolumeDB. PM me and tell me what I need to do to check it!

tsialex · Feb 19, 2022

Early-2009 Mac Pro, model 20H, B08 from factory, with 144.0.0.0.0, same High Sierra disk as last time:

Code:

HighSierra:~ tsialex$ system_profiler SPHardwareDataType
Hardware:

    Hardware Overview:

      Model Name: Mac Pro
      Model Identifier: MacPro5,1
      Processor Name: Quad-Core Intel Xeon
      Processor Speed: 3,2 GHz
      Number of Processors: 1
      Total Number of Cores: 4
      L2 Cache (per Core): 256 KB
      L3 Cache: 8 MB
      Memory: 16 GB
      Boot ROM Version: 144.0.0.0.0
      SMC Version (system): 1.39f5
      SMC Version (processor tray): 1.39f5
      Serial Number (system): xxxxxxxx20H

HighSierra:~ tsialex$ sysctl -n machdep.cpu.brand_string
Intel(R) Xeon(R) CPU           W3565  @ 3.20GHz
HighSierra:~ tsialex$ ./dmidecode | grep FirmwareFeatures
    FirmwareFeatures: 0xC00C1403
    FirmwareFeaturesMask: 0xFF1FFF3F
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
    <data>
    A1QMwA==
    </data>
</dict>
</plist>
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
    <data>
    P/8f/w==
    </data>
</dict>
</plist>
HighSierra:~ tsialex$

NVRAM FirmwareFetaures changed again to A1QMwA==.

I'll downgrade it back to the MP41.88Z.0081.B03.0902231259 BootROM image dump that @Edge sent me, it's the oldest dump I have.

tsialex · Feb 19, 2022

Downgrade successful, I'll boot High Sierra with -no_compat_check now:

cdf · Feb 19, 2022

tsialex said:
NVRAM FirmwareFetaures changed again to A1QMwA==.

And the SMBIOS value is still 0xC00C1403 (AxQMwA==). Seems like the discrepancy is consistent.

tsialex · Feb 19, 2022

Early-2009 Mac Pro, model 20H, B08 from factory, downgraded to @Edge MP41.0081.B03, same High Sierra disk as last time:

Code:

HighSierra:~ tsialex$ system_profiler SPHardwareDataType
Hardware:

    Hardware Overview:

      Model Name: Mac Pro
      Model Identifier: MacPro4,1
      Processor Name: Quad-Core Intel Xeon
      Processor Speed: 3,19 GHz
      Number of Processors: 1
      Total Number of Cores: 4
      L2 Cache (per Core): 256 KB
      L3 Cache: 8 MB
      Memory: 16 GB
      Boot ROM Version: MP41.0081.B03
      SMC Version (system): 1.39f5
      SMC Version (processor tray): 1.39f5
      Serial Number (system): xxxxxxxx4PC
    
 
HighSierra:~ tsialex$ sysctl -n machdep.cpu.brand_string
Intel(R) Xeon(R) CPU           W3565  @ 3.20GHz

HighSierra:~ tsialex$ ./dmidecode | grep FirmwareFeatures
    FirmwareFeatures: 0xC0001403
    FirmwareFeaturesMask: 0xC0007FFF
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
    <data>
    F1QAwA==
    </data>
</dict>
</plist>
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
    <data>
    /38AwA==
    </data>
</dict>
</plist>

NVRAM FirmwareFeatures changed again to F1QAwA== and FirmwareFeaturesMask to /38AwA==. SMBIOS FirmwareFeaturesMask changed to 0xC0007FFF

tsialex · Feb 19, 2022

cdf said:
And the SMBIOS value is still 0xC00C1403 (AxQMwA==). Seems like the discrepancy is consistent.

Yep. Do you wanna check anything else with MP41.0081.B03?

cdf · Feb 19, 2022

tsialex said:
Yep. Do you wanna check anything else with MP41.0081.B03?

No, I can't think of anything to check. And at this point, whether the discrepancy between SMBIOS and NVRAM is actually a bug is even less clear...

tsialex · Feb 19, 2022

cdf said:
No, I can't think of anything to check. And at this point, whether the discrepancy between SMBIOS and NVRAM is actually a bug is even less clear...

Another thing, I have a hunch that different models have different NVRAM firmware features and the firmware features changes between EFI releases. It's not clear what is really going on, I'll probably have to reconstruct different EFI versions for the same model and track this.

I'm going to assemble a Mac Pro from all the spare parts I have and try to track this, I can't do that with my working and test one like I'm doing it right now.

tsialex · Feb 19, 2022

Before going back to the Mac Pro own BootROM image, I've checked Mac Pro EFI Firmware Update 1.4 that upgrades the EFI to MP41.0081.B07 and got half of my hunch confirmed (ping @cdf), NVRAM firmware features changed from F1QAwA== with mask /38AwA== to A1QAwA== with mask still /38AwA==.

Code:

HighSierra:~ tsialex$ system_profiler SPHardwareDataType
Hardware:

    Hardware Overview:

      Model Name: Mac Pro
      Model Identifier: MacPro4,1
      Processor Name: Quad-Core Intel Xeon
      Processor Speed: 3,19 GHz
      Number of Processors: 1
      Total Number of Cores: 4
      L2 Cache (per Core): 256 KB
      L3 Cache: 8 MB
      Memory: 16 GB
      Boot ROM Version: MP41.0081.B07
      SMC Version (system): 1.39f5
      SMC Version (processor tray): 1.39f5
      Serial Number (system): xxxxxxxx4PC

HighSierra:~ tsialex$ sysctl -n machdep.cpu.brand_string
Intel(R) Xeon(R) CPU           W3565  @ 3.20GHz
HighSierra:~ tsialex$ ./dmidecode | grep FirmwareFeatures
    FirmwareFeatures: 0xC0001403
    FirmwareFeaturesMask: 0xC0007FFF
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures</key>
    <data>
    A1QAwA==
    </data>
</dict>
</plist>
HighSierra:~ tsialex$ nvram -x 4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeaturesMask</key>
    <data>
    /38AwA==
    </data>
</dict>
</plist>

Dayo · Feb 19, 2022

tsialex said:
check anything else with MP41.0081.B03?

I wonder whether the MacOS version has an impact as it now seems almost like Apple is essentially "spoofing" these since the outcomes do not appear to represent a hard indication of the CPU features, where one would expect these to be constant.

Basically mean checking the same firmware, doesn't particularly have to be that specific one, with two far apart MacOS versions ... say El Cap and Mojave (for checking impact of APFS)?

If you have RefindPlus, you will be able to boot these regardless the firmware APFS support status.

BTW, my new wild hypothesis is that the mask represents the actual CPU capability and the flags are used to switch these on/off. This may explain why the mask covers more items. Only nit in that is that the mask, while seemingly less prone to change, is not always the same value.

Macschrauber · Feb 20, 2022

Firmware version matter, I guess

my tests were with Mavericks (always use it for tests, flashing and dumping, as it runs on 4.1 and 5.1 and has no SIP)

1.) booted Mavericks in MP51.88Z.0087.B00.1804181525 Firmware and used nvram:

FirmwareFeatures: 0xC00C0403
FirmwareFeaturesMask: 0xFF1FFF3F

2.) flashed a 144 Firmware with clean nvram

3.) booted Mavericks in 144.0.0.0 Firmware:

FirmwareFeatures: 0xC00C1403
FirmwareFeaturesMask: 0xFF1FFF3F

of course I have a firmware dump of Situation 1.) as I reconstructed Firmware for 3.)

startergo · Feb 20, 2022

Macschrauber said:
can give dmidecode output without OC as I run Mavericks atm, machine never seen OC, rebuilt firmware:

FirmwareFeatures: 0xC00C0403
FirmwareFeaturesMask: 0xFF1FFF3F

as I run some long term tests in Mavericks I can give xml output of the FirmwareFeatures tomorrow. (Mavericks nvram cli command does not list xml of given vars)

Interesting. Assuming this system never ran OC the results are different:

Windows suddenly not selectable as startup-disk anymore

Thanks! The deep NVRAM-reset solved it! And it looks somhow better now... ...with the firmware features matching your last post. Im not using OC on this machine. But i made a build for an installer stick for some other cMP. And it looks like this has been enough to corrupt my nvram. I did...

forums.macrumors.com

and here:

Doesn't work if diff Volume type/EFI Mode · Issue #8 · kainjow/BootChamp

I have Windows 8.1 installed in UEFI mode (not standard Boot Camp), on a volume labelled "Windows HD", if I press restart into windows I am just given an error about boot volume not found not sure ...

github.com

cdf · Feb 20, 2022

Summary for MacPro4,1 and MacPro5,1 (so far)

SMBIOS firmware features

4,1: 0xC0001403
5,1: 0xC00C0403 to 0xC00C1403

Speculation: The 5,1 was missing a feature of the 4,1: bit 12 (FW_FEATURE_SUPPORTS_TRB). This feature was added to the 5,1 (at some point between MP51.0089.B00 and 144.0.0.0).

NVRAM firmware features

4,1: 0xC0005417 to 0xC0005403

Observation: the 4,1 originally had extra features: bit 2 (FW_FEATURE_SUPPORTS_TARGET_DISK_MODE) and bit 4 (FW_FEATURE_SUPPORTS_NET_BOOT). For some reason, these features were removed (at some point between MP41.0081.B04 and MP41.0081.B07).

5,1: 0xC00C5403, 0xC00C5417

Observation: For some reason, both values are possible with the same firmware version and machine.

4,1 vs 5,1

All 5,1 data seem to be from firmware versions with APFS support. This explains the presence of certain features not in the 4,1: bit 18 (FW_FEATURE_UNKNOWN_BIT18) and bit 19 (FW_FEATURE_SUPPORTS_APFS).

SMBIOS vs NVRAM

The discrepancy is with bit 2, bit 4, bit 12, and bit 14 (FW_FEATURE_SUPPORTS_HIGH_SPEED_USB). The reason is still unknown.

cdf · Feb 20, 2022

startergo said:
Interesting. Assuming this system never ran OC the results are different:

It seems that these results are either actually consistent with what we have here (after an NVRAM reset) or are from a MacBook Pro.

cdf · Feb 20, 2022

One thing is for sure: the Acidanthera database is off on the firmware features. I wonder how these values were determined...

OpenCorePkg/AppleModels/DataBase/MacPro/MP51.yaml at master · acidanthera/OpenCorePkg

OpenCore bootloader. Contribute to acidanthera/OpenCorePkg development by creating an account on GitHub.

github.com

startergo · Feb 20, 2022

cdf said:
or are from a MacBook Pro

Where is this?

cdf · Feb 20, 2022

startergo said:
Where is this?

Second link. I assumed you were referring to this particular report, which shows very different firmware features: https://github.com/kainjow/BootChamp/issues/8#issuecomment-46055415

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

Contributor

macrumors 68020

macrumors 68020

macrumors 68020

macrumors 68020

macrumors 6502

macrumors 68030

Contributor

macrumors 6502a

Contributor

Contributor

macrumors 68020

Contributor

Contributor

macrumors 68020

Contributor

Contributor

Attachments

macrumors 68020

macrumors 68030

macrumors 603

macrumors 68020

macrumors 68020

macrumors 68020

macrumors 603

macrumors 68020

Our Staff