Jump to content

Marvell (Aquantia) 10 Gb Ethernet support thread


d5aqoep
498 posts in this topic

Recommended Posts

I installed 10.13.2 and updated with the combo updater, but still the same result: device id is 0x0001 :(

As long as I have the Asus card in my hackintosh, some boots crash somewhere around the Aquantia kexts. Next boot seems to be fine.

 

Where can I find the boot log? I can't find the lines from above in my system.log

Link to comment
Share on other sites

I installed 10.13.2 and updated with the combo updater, but still the same result: device id is 0x0001 :(

As long as I have the Asus card in my hackintosh, some boots crash somewhere around the Aquantia kexts. Next boot seems to be fine.

 

Where can I find the boot log? I can't find the lines from above in my system.log

Hello Dogman,

 

Did you tried the "macOS High Sierra 10.13.3 Update for iMac Pro"? https://support.apple.com/kb/DL1952

Or the normal "macOS High Sierra 10.13.3 Combo Update" https://support.apple.com/kb/DL1954

 

But, I'm not sure if this makes a difference...

Link to comment
Share on other sites

I installed 10.13.2 and updated with the combo updater, but still the same result: device id is 0x0001 :(

As long as I have the Asus card in my hackintosh, some boots crash somewhere around the Aquantia kexts. Next boot seems to be fine.

 

Where can I find the boot log? I can't find the lines from above in my system.log

Is your ASUS card on en0 ?

Link to comment
Share on other sites

Windows 7 Looks fine:
PCI\VEN_1D6A&DEV_D107&SUBSYS_87411043&REV_02

Maybe it the second revision that creates the problem?

 

EDIT: now macOS displays

Vendor 0x1d6a

Device 0xd107

SubsystemVendor 0x1043 and

Subsytem id 0x8741

 

 

But my cable is still "unplugged". Will look into this later...

Link to comment
Share on other sites

Windows 7 Looks fine:

PCI\VEN_1D6A&DEV_D107&SUBSYS_87411043&REV_02

 

Maybe it the second revision that creates the problem?

 

EDIT: now macOS displays

Vendor 0x1d6a

Device 0xd107

SubsystemVendor 0x1043 and

Subsytem id 0x8741

 

 

But my cable is still "unplugged". Will look into this later...

As the subsystem vendor ID shows, the update failed. 0x1043 is ASUStek but it should be 0x106b for Apple after a successful firmware update.

 

Mieze

Link to comment
Share on other sites

Windows 7 Looks fine:

PCI\VEN_1D6A&DEV_D107&SUBSYS_87411043&REV_02

 

Maybe it the second revision that creates the problem?

 

EDIT: now macOS displays

Vendor 0x1d6a

Device 0xd107

SubsystemVendor 0x1043 and

Subsytem id 0x8741

 

 

But my cable is still "unplugged". Will look into this later...

Try restarting macOS 2-3 times. Many have had luck once the firmware update is complete.

 

Make sure you have bare minimal clover patches and kexts in hopes that some patch or kext is preventing successful firmware update.

Link to comment
Share on other sites

Do you know what spi is?  These following are used to do firmware update.

 

 

 

AppleEthernetAquantiaAqtion::spi_busy()

 

AppleEthernetAquantiaAqtion::spi_cmd(unsigned char, int, int)

 

AppleEthernetAquantiaAqtion::spi_nvram_cmds_init()

 

AppleEthernetAquantiaAqtion::spi_nvr_wr_enable(int)

 

AppleEthernetAquantiaAqtion::spi_nvr_busy()

 

AppleEthernetAquantiaAqtion::spi_nvr_erase()

 

AppleEthernetAquantiaAqtion::spi_get(unsigned int, unsigned int*, unsigned long)

 

AppleEthernetAquantiaAqtion::spi_put(unsigned int, unsigned int*, unsigned long)

Might there be a way to manually invoke this update rather than relying on the OS to do it at boot time?

Beyond my technical ability to figure out unfortunately but you seem to know your stuff!

Link to comment
Share on other sites

Might there be a way to manually invoke this update rather than relying on the OS to do it at boot time?

Beyond my technical ability to figure out unfortunately but you seem to know your stuff!

It might be possible to patch the driver in a way that it performs the firmware update anyway but this would require you to reverse engineer the driver's start() function in order to find out which conditions must be met for the update to be performed. Unfortunately this is far from being trivial and not done within a weekend.

 

Mieze

  • Like 1
Link to comment
Share on other sites

With regards to Asrock Fatal1ty X299 Professional Gaming i9, we have a confirmation that it works OOTB from 10.13.3 and AQC107 firmware is successfully applied at boot.

 

@ everybody:

Right now in Windows, I am having 3 major issues with my ASUS XG-C100C card.

1. macOS 10.13.3 did a firmware update on the card and now it only accepts Apple's customized bootcamp drivers and refuses to install ASUS official drivers from it's support page.

2. Every 3-4 boots, the card fails to detect a link and network cable unplugged comes up. The solution is to disable and enable it from device manager. Then it works without any issues.

3. Wake on LAN (WOL) simply does not work on any of the drivers. This is pathetic as WOL is a supported feature as per Aquantia's AQC107 specifications page. Link here

Link to comment
Share on other sites

 

With regards to Asrock Fatal1ty X299 Professional Gaming i9, we have a confirmation that it works OOTB from 10.13.3 and AQC107 firmware is successfully applied at boot.
 
@ everybody:
Right now in Windows, I am having 3 major issues with my ASUS XG-C100C card.
1. macOS 10.13.3 did a firmware update on the card and now it only accepts Apple's customized bootcamp drivers and refuses to install ASUS official drivers from it's support page.
2. Every 3-4 boots, the card fails to detect a link and network cable unplugged comes up. The solution is to disable and enable it from device manager. Then it works without any issues.
3. Wake on LAN (WOL) simply does not work on any of the drivers. This is pathetic as WOL is a supported feature as per Aquantia's AQC107 specifications page. Link here

It's obvious that the drivers are still full of bugs which also explains why the Aquantia firmware is applied to cards from different vendors because flashing firmware of a card from a different vendor without explicit consent of the user and in a way that the card won't be recognized by its native drivers anymore is not only stupid but also illegal as it literally causes damage to the card.

 

Another example for Apple's excellent software quality!  :P

 

Mieze

Link to comment
Share on other sites

It's obvious that the drivers are still full of bugs which also explains why the Aquantia firmware is applied to cards from different vendors because flashing firmware of a card from a different vendor without explicit consent of the user and in a way that the card won't be recognized by its native drivers anymore is not only stupid but also illegal as it literally causes damage to the card.

 

Another example for Apple's excellent software quality!  :P

 

Mieze

Aquantia is to blame. The latest driver on ASUS website for XG-C100C is dated around April 2017. That is almost 10 months without any driver update.

It may not be Apple's fault completely beause Apple expects only iMac Pro hardware on their macOS. They don't care about any other possible scenario and totally not concerned about Hackintosh users.

 

Station-drivers have v2.1 driver available dated Dec 2018. But it fails to install on my card with message "This device cannot start."

Link to comment
Share on other sites

Aquantia is to blame. The latest driver on ASUS website for XG-C100C is dated around April 2017. That is almost 10 months without any driver update.

10 month without update? That's nothing for a driver and provided there is no new hardware to support, no significant change in the driver interface or in OS power management, there is no need to update a properly written driver for years.

It may not be Apple's fault completely beause Apple expects only iMac Pro hardware on their macOS. They don't care about any other possible scenario and totally not concerned about Hackintosh users.

That's what the military calls collateral damage and as Macs also support Windows 10 there are several supported scenarios in which this kind of damage may occur (a Aquantia card in a TB to PCIe expansion case attached to a Mac running Win 10 and High Sierra).

 

Mieze

Link to comment
Share on other sites

I purchased ASUS XG-C100C (0x1d6a/0xd107) and applied 10.3.3 update so that its device ID as modified to 0x1d6a/0x07b. and

 

It works ... 

 

It is now recognized as Apple AQC107-AFW.

I did not disable on-board NIC ( 0x8086/0 x15b7 ), did not change machine type to iMac18 either ( still using iMac17,1 ).

 

As I do not have 10G switch, it is connected to a 1G switch and the connection is 100baseTX, I will switch default connection back to my on-board NIC.

 

I also confirmed ASUS Windows drives does not work as the device ID has changed.

 

EDIT: The reason of slow connection was a loose cable. I got 1G speed.

  • Like 2
Link to comment
Share on other sites

Might there be a way to manually invoke this update rather than relying on the OS to do it at boot time?

Beyond my technical ability to figure out unfortunately but you seem to know your stuff!

 

You need make there driver execute the following code.  There are at least three checks, chip version, firmware version and a certain register value.  It is easy to brake a card.  I don't have the hardware to test patches.

 

http://www.insanelymac.com/forum/topic/330614-success-aqc107-10-gbe-native-support-in-high-sierra-10132/?p=2579078

Link to comment
Share on other sites

I booted an ASRock X299 Professional gaming i9 (with built-in AQC107 10GbE) in 10.13.3 with an iMacPro1,1 SMBIOS and it did not work (no cable detected) on first boot, but after 1-2 reboots the 10GbE now works. I did not disable the built-in 1GbE.

 

I assume that if one needs a Windows driver to install then you can just modify the .inf file to include the new deviceID and then use Device Mangers manual driver update to install it.

Link to comment
Share on other sites

I assume that if one needs a Windows driver to install then you can just modify the .inf file to include the new deviceID and then use Device Mangers manual driver update to install it.

I'd suggest to use the Apple provided drivers in Windows after the firmware flash as they match the device now. You'll never know what exactly has changed and the ASUS driver might have problems with the Apple firmware.

 

Mieze

Link to comment
Share on other sites

I booted an ASRock X299 Professional gaming i9 (with built-in AQC107 10GbE) in 10.13.3 with an iMacPro1,1 SMBIOS and it did not work (no cable detected) on first boot, but after 1-2 reboots the 10GbE now works. I did not disable the built-in 1GbE.

 

I assume that if one needs a Windows driver to install then you can just modify the .inf file to include the new deviceID and then use Device Mangers manual driver update to install it.

 

If you modify the inf file, will that break the signature and the driver will fail to install due to it not being WHQL?

Link to comment
Share on other sites

If you modify the inf file, will that break the signature and the driver will fail to install due to it not being WHQL?

 

I was able to modify the inf file for a completely different device not too long ago and the driver still installed on Win8.1. Win10 might be different though. Try and find out?

 

You can also disable the WHQL requirement if need be.

Link to comment
Share on other sites

×
×
  • Create New...