ASRock H310M-HDV - dual monitor problems

[deeveedee]

@brumas2025 I don't understand why macOS is mapping both displays to con1.  I'm going to need to think about this, because I don't have any guesses to help.  Hopefully others who are watching this can offer suggestions.  I am sorry that I have not been able to help you so far.  I'll revisit this thread if I have any more ideas.  You've been very patient.  Don't give up.

[Asural]

20 minutes ago, brumas2025 said:

I booted up with your example, only one monitor (HDMI) is working, I changed the kext.

I have attached the corrected plist, so please check that it works.

Mod config DVI-D.plist.zip

[brumas2025]

54 minutes ago, deeveedee said:

@brumas2025 I don't understand why macOS is mapping both displays to con1.  I'm going to need to think about this, because I don't have any guesses to help.  Hopefully others who are watching this can offer suggestions.  I am sorry that I have not been able to help you so far.  I'll revisit this thread if I have any more ideas.  You've been very patient.  Don't give up.

Thank you for your time and willingness to help me!

43 minutes ago, Asural said:

I have attached the corrected plist, so please check that it works.

Thanks, I tried your option, but it's loading with two black screens...

[deeveedee]

@Asural As I mentioned here, only the iGPU framebuffer properties that are being changed need to be specified in the config.plist DeviceProperties when using the discrete properties format (vs. using the all-data format).   There is nothing wrong with specifying all iGPU framebuffer properties (both changed and unchanged) when specifying discrete properties (e.g., framebuffer-con0-index, framebuffer-con0-busid ...), but it is unnecessary and re-defining properties that are not changed does not fix anything.  

 

This is a matter of personal preference, but in my opinion, specifying all DeviceProperties, including those that are not changed, just makes the DeviceProperties more difficult to interpret and burdens the reader with figuring out which properties changed and which did not change.  For the reader, it is much easier to specify only the DeviceProperties that require changes.

[brumas2025]

2 hours ago, deeveedee said:

 

I solved my problem with two monitors in an unusual way:

I have a Tanix W2 ultra set-top box, which was connected to the TV.

I saw that after booting up with the HDMI configuration, i could add a wireless display from the set-top box in the display menu.

So I moved it to my second monitor, connected it to the set-top box via an HDMI-VGA adapter, and now my computer has two monitors.

I  achieved what I wanted. 

 

 

 

Edited December 20, 2025 by brumas2025

[deeveedee]

@brumas2025 Very clever way to solve your multi-display problem.  With your solution, you are not limited by the physical port problem.  Well done!

[deeveedee]

@brumas2025 I have been studying Intel iGPU framebuffer patching and am learning that I don't know much (not surprised).  The relationship between the framebuffer pipe and the framebuffer index is still a mystery to me.  macOS may be mapping your motherboard DVI-D and HDMI ports to the same framebuffer index (2) because of your PC's internal iGPU "pipe" architecture.

 

This is a guess for me, so I'm offering this theory with the hope that someone more experienced can comment and help.  I suspect that you might be able to "move" either DVI-D or HDMI to a different logical connector / index by patching the pipe of con0 or con2.  It seems that we're seeing this exact same behavior in this thread.

 

Your chosen framebuffer (3e9b0007) has the following definition from here:

ID: 3E9B0007, STOLEN: 57 MB, FBMEM: 0 bytes, VRAM: 1536 MB, Flags: 0x00801302
TOTAL STOLEN: 58 MB, TOTAL CURSOR: 1 MB (1572864 bytes), MAX STOLEN: 172 MB, MAX OVERALL: 173 MB (181940224 bytes)
Model name: Intel UHD Graphics 630
Camellia: CamelliaDisabled (0), Freq: 0 Hz, FreqMax: 0 Hz
Mobile: 0, PipeCount: 3, PortCount: 3, FBMemoryCount: 3
[1] busId: 0x05, pipe: 9, type: 0x00000400, flags: 0x000003C7 - ConnectorDP
[2] busId: 0x04, pipe: 10, type: 0x00000400, flags: 0x000003C7 - ConnectorDP
[3] busId: 0x06, pipe: 8, type: 0x00000400, flags: 0x000003C7 - ConnectorDP
01050900 00040000 C7030000
02040A00 00040000 C7030000
03060800 00040000 C7030000

The framebuffer pipes defined in this framebuffer are

• con0: pipe 9
• con1: pipe 10 (0x0a)
• con2: pipe 8

You may be able to modify your config.plist by adding DeviceProperty framebuffer-con0-pipe with a value other than 8, 9 or 10 to "move" either your DVI-D or HDMI port to con0 / Index 1.

 

I don't have any suggested pipe values to try at this time, so if you have the patience to experiment, you'll want to try values other than 8, 9 and 10.  I hope that others more experienced can read this theory and comment with more insight.

Edited December 22, 2025 by deeveedee

[brumas2025]

12 hours ago, deeveedee said:

@brumas2025

You may be able to modify your config.plist by adding DeviceProperty framebuffer-con0-pipe with a value other than 8, 9 or 10 to "move" either your DVI-D or HDMI port to con0 / Index 1.

 

I don't have any suggested pipe values to try at this time, so if you have the patience to experiment, you'll want to try values other than 8, 9 and 10.  I hope that others more experienced can read this theory and comment with more insight.

 

Thanks, I'll try that... Or maybe you could make me some trial configs so I can try them out?

Edited December 23, 2025 by brumas2025

[deeveedee]

@brumas2025 Ok.  to help me get started with candidate tests, please test the attached config.plists and report the following for each one:

• Which display is working
• What connector is the display using (e.g., con1)

Thank you.  

 

EDIT: Does your motherboard BIOS allow you to configure DVMT pre-alloc?

config-H31M-test2.plist.zip config-H31M-test3.plist.zip

Edited December 23, 2025 by deeveedee

[brumas2025]

3 hours ago, deeveedee said:

@brumas2025 Ok.  to help me get started with candidate tests, please test the attached config.plists and report the following for each one:

• Which display is working
• What connector is the display using (e.g., con1)

Thank you.  

EDIT: Does your motherboard BIOS allow you to configure DVMT pre-alloc?

Test result:

config-H31M-test2 - successful upload to the HDMI monitor,

config-H31M-test2 - successful upload to the HDMI monitor. Screenshots are attached.

 

The BIOS of my motherboard allows you to set up a DVMT pre-distribution, which one should I install?

Thank you.  

 

Quote

 

 

 

Edited December 23, 2025 by brumas2025

[deeveedee]

@brumas2025 if "Share Memory" is the only graphics memory configuration option in BIOS, you may have an older BIOS that does not allow you to configure DVMT pre-alloc.  Leave the setting unchanged for now.  We can revisit the setting later if we need to.

 

Attached are 4 config.plists as described below.  Please test these and report your results.  All of these config.plists start with config-H31M-test-3.plist from my previous post, so all have 2 ports (framebuffer-portcount = 2, con1 type = HDMI, con1 busid = 4 (default for framebuffer 3e9b0007) ).  Thank you.

• config-H31M-test-4-1.plist: con0 type = HDMI, con0 busid =1
• config-H31M-test-4-2.plist: con0 type = HDMI, con0 busid = 2
• config-H31M-test-4-3.plist: con0 type = DP (default for framebuffer 3e9b0007), con0 busid = 1 (may not be a valid busid for type DP)
• config-H31M-test-4-4.plist: con0 type = DP, con0 busid = 2

 

config-H31M-test-4-1.plist.zip config-H31M-test-4-2.plist.zip config-H31M-test-4-3.plist.zip config-H31M-test-4-4.plist.zip

Edited December 23, 2025 by deeveedee

[brumas2025]

46 minutes ago, deeveedee said:

 

• config-H31M-test-4-1.plist: con0 type = HDMI, con0 busid =1 -successful upload to the HDMI monitor,
• config-H31M-test-4-2.plist: con0 type = HDMI, con0 busid = 2 - couldn't load , panic
• config-H31M-test-4-3.plist: con0 type = DP (default for framebuffer 3e9b0007), con0 busid = 1 (may not be a valid busid for type DP) - successful upload to the HDMI monitor,
• config-H31M-test-4-4.plist: con0 type = DP, con0 busid = 2  - successful upload to the HDMI monitor.

 

Quote

 

Edited December 23, 2025 by brumas2025

[deeveedee]

@brumas2025 I appreciate your patience with my wild guesses.  I tried changing BusIDs first in my previous post, because I have a better understanding of valid BusID values.  I'll be honest - I have no idea what the valid values are for framebuffer-conX-pipe.  I have attached 5 config.plists with different values for framebuffer-con0-pipe.  Please test these and report your findings.  If macOS doesn't boot with a particular config.plist, you don't need to post screenshots.

 

The attached config.plists are based on the previously posted config-H31M-test-3.plist with changes as follows:

• config-H31M-test-5-1.plist: con0 pipe = 0
• config-H31M-test-5-2.plist: con0 pipe = 0x0B
• config-H31M-test-5-3.plist: con0 pipe = 1
• config-H31M-test-5-4.plist: con0 pipe = 0x0C
• config-H31M-test-5-5.plist: con0 pipe = 0x12

config-H31M-test-5-1.plist.zip config-H31M-test-5-2.plist.zip config-H31M-test-5-3.plist.zip config-H31M-test-5-4.plist.zip config-H31M-test-5-5.plist.zip

[brumas2025]

32 minutes ago, deeveedee said:

@brumas2025 I appreciate your patience with my wild guesses.  I tried changing BusIDs first in my previous post, because I have a better understanding of valid BusID values.  I'll be honest - I have no idea what the valid values are for framebuffer-conX-pipe.  I have attached 5 config.plists with different values for framebuffer-con0-pipe.  Please test these and report your findings.  If macOS doesn't boot with a particular config.plist, you don't need to post screenshots.

 

 

• config-H31M-test-5-1.plist: con0 pipe = 0.            successful upload to the HDMI monitor
• config-H31M-test-5-2.plist: con0 pipe = 0x0B.      successful upload to the HDMI monitor
• config-H31M-test-5-3.plist: con0 pipe = 1.             successful upload to the HDMI monitor
• config-H31M-test-5-4.plist: con0 pipe = 0x0C       successful upload to the HDMI monitor
• config-H31M-test-5-5.plist: con0 pipe = 0x12         successful upload to the HDMI monitor

Edited December 23, 2025 by brumas2025

[deeveedee]

@brumas2025 Wow! You're a fast tester!  

 

Attached are the last of my wild guesses for now.  I tried to understand why your original config-DVI-D.plist enabled the DVI-D port to make the attached guesses.

 

Don't worry - I don't have any more guesses after this (for now).  

 

config-H31M-test-5-6.plist.zip config-H31M-test-5-7.plist.zip config-H31M-test-5-8.plist.zip

[brumas2025]

26 минут назад deeveedee сказал:

 

 

Don't worry - I don't have any more guesses after this (for now). 

• config-H31M-test-5-6.plist:             successful upload to the HDMI monitor
• config-H31M-test-5-7.plist:              successful upload to the HDMI monitor
• config-H31M-test-5-8.plist:              successful upload to the HDMI monitor
• 
• 🙂

Edited December 23, 2025 by brumas2025

[deeveedee]

@brumas2025 thank you for testing.  I don't have any more guesses. The person who figures this out will be the one who can examine your config-DVI-D.plist and config-HDMI.plist and figure out why each config.plist enables the DVI-D and HDMI displays respectively.  I am mystified.

 

Thank you again for your patience.  I'm glad you found a working solution with your wireless display.

[brumas2025]

4 минуты назад deeveedee сказал:

@brumas2025спасибо за тестирование.

Thank you for wanting to help!

[Asural]

@brumas2025

So far, we know that index =1 is HDMI, but we don't know index 2 and index 3.

I've attached a test DeviceProperties, so leave the HDMI monitor connected...

To check the index value of the motherboard's VGA connector, connect it to the monitor with a VGA cable and try changing the value set in

framebuffer-con1-busid = <04000000> to 2, 4, 5, or 6 to see which value displays the display.
If no display is obtained with any of the values, index 2 is the built-in VGA,

so set framebuffer-con1-busid = <00000000> framebuffer-con1-enable=<00000000>.

If a display is obtained with any of the values, index 2 is the valid built-in VGA, so set framebuffer-con1-busid to the smallest displayable value.

To check the index value of the motherboard's DVI connector, connect it to a monitor using a DVI to VGA cable.
Try changing the value set in framebuffer-con2-busid = <06000000> to 2, 4, 5, or 6 (avoiding the built-in VGA if it's enabled) to see which value

results in a display.

It's likely that multiple values will work, so record the value and set it to the smallest value that still works.

 

Please let us know the results.
 

H310M_TEST.plist

Edited December 26, 2025 by Asural

[brumas2025]

1 hour ago, Asural said:

@brumas2025    Please let us know the results

Thank you, I'll try your configuration and message a little later.

Edited December 26, 2025 by brumas2025

[brumas2025]

On 12/26/2025 at 8:16 AM, Asural said:

@brumas2025

Please let us know the results.

Good day! I tried your suggestion, but both the HDMI and VGA monitors showed a black screen. I connected via remote access - the schedule was not determined. I don't even know if it's worth continuing the experiments...

[Asural]

14 hours ago, brumas2025 said:

Good day! I tried your suggestion, but both the HDMI and VGA monitors showed a black screen. I connected via remote access - the schedule was not determined. I don't even know if it's worth continuing the experiments...

Please discontinue testing H310M_TEST.plist.

I wanted to test the relationship between index and BusId, but as @deeveedee pointed out here, it appears that both HDMI and DVI are only displayed valid on con1.

I've attached a plist that explicitly specifies index and BusId, so if HDMI display is valid, please perform the following test.
(If HDMI display is not available, this test is not necessary.)

 

Using DVI to VGA and HDMI：
Change framebuffer-con0-busid = <05000000> to 2, 4, 5, or 6 to see which values result in a display.
Change framebuffer-con2-busid = <00000000> to 2, 4, 5, or 6 to see which values result in a display.

 

Append:

My IORegistryExplorer Version 2.1 does not open the ioreg file you attached here, are you using the same version?

H310M_TEST2.plist

Edited December 28, 2025 by Asural

[Asural]

@brumas2025

I've attached the verified configuration for the Gigabyte H310M obtained from this guide.
Please cancel H310M2_TEST.plist and prioritize this.

I'm not sure whether the built-in VGA or DVI to VGA connection will work, so please try them out.
 

GigabyteH310M.plist

[brumas2025]

7 hours ago, Asural said:

@brumas2025

good afternoon, I tried your config, but both monitors have a black screen, it was not possible to connect remotely. I used version 3.0. I did it in version 2.1, and I'm attaching the file.

iMac HDMI 281225.ioreg.zip

Edited December 28, 2025 by brumas2025

[Asural]

44 minutes ago, brumas2025 said:

good afternoon, I tried your config, but both monitors have a black screen, it was not possible to connect remotely. I used version 3.0. I did it in version 2.1, and I'm attaching the file.

iMac HDMI 281225.ioreg.zip 786.13 kB · 0 downloads

I was able to read the ioreg file you attached.
Since con2 is not enabled, I am attaching a version with a different index value. If any monitors are enabled, please attach the ioreg file.
 

GigabyteH310M-2.plist

GigabyteH310M-3.plist

Edited December 28, 2025 by Asural