Jump to content

Marvell (Aquantia) 10 Gb Ethernet support thread


d5aqoep
498 posts in this topic

Recommended Posts

I adopted @Shikumo's patch as well, with proper masking which hopefully makes the patch survive longer. I made the masks based on both 11.x and 12.x binaries, so let's see. Thanks @gengik84 for providing the binary from 11.x.

 

Could anyone please test this commit https://github.com/acidanthera/OpenCorePkg/commit/0605cdaa74d7ac6c8c130c2c8513065f2576a5e9?

 

Artifacts can be found at https://github.com/acidanthera/OpenCorePkg/actions/runs/2012498872 once done.

Edited by PMheart
Link to comment
Share on other sites

2 hours ago, PMheart said:

Too lazy to read old posts - could anyone please clarify whether any patch (e.g. built-in or so) is needed before 10.15.4? Thanks!

iirc I didn’t need any patches before Big Sur?

  • Like 1
Link to comment
Share on other sites

Thanks! My AQ107 card works perfectly OC 0.8.0 with the above commit / patch, VT-d enabled and DisableIOMapper=False 

However my Fenvi T919 wifi card will not power on in network preferences anymore so no airdrop, handoff or continuity

Link to comment
Share on other sites

T919 was the one before 12 came up.

 

Since the Changes in BT Stacks 94360ng is one of the most stressless Cards so far.

 

BT and WiFi, but WiFi in 2,4 slow and 5 not that fast. But stable and BT OOB without any additions.

 

 

This AQ Fix is really nice, can’t repeat it often enough.

Edited by kaneske
Link to comment
Share on other sites

On 3/23/2022 at 3:11 PM, Shikumo said:

In theory, but it is currently untested. So it would be nice, if someone with an AQC113 based card could test this and provide some feedback.

 

 

I'll be back in office tomorrow and I'll test it.

Link to comment
Share on other sites

On 3/21/2022 at 12:07 PM, Rankrotten said:

Thanks! My AQ107 card works perfectly OC 0.8.0 with the above commit / patch, VT-d enabled and DisableIOMapper=False 

Anyone get this working with a AMD system that doesn't have VT-d? I added the patch and DisableIOMapper=False. It sees the card but either no IP or freeze when ethernet cable is plugged in. Not sure what to try. I'm going to try OC .80 next but I'm not clear if this is any different than just adding the patch manually.

Edited by italeffect2
Link to comment
Share on other sites

10 hours ago, antuneddu said:

For AMD you have to search in your bios SVM (Secure Virtual Machine) 🤞

Thanks for that. Unfortunately toggling it on didn’t make a difference. I also tried toggling on/off IOMMU and that didn’t matter either. I’m now on OC .80 with everything updated.  It runs fine as long as Ethernet is not plugged in.  Plugging in a cable results in no IP and than in a hard freeze. There’s a new bios update, maybe I’ll try that. 

Link to comment
Share on other sites

10 hours ago, italeffect2 said:

Thanks for that. Unfortunately toggling it on didn’t make a difference. I also tried toggling on/off IOMMU and that didn’t matter either. I’m now on OC .80 with everything updated.  It runs fine as long as Ethernet is not plugged in.  Plugging in a cable results in no IP and than in a hard freeze. There’s a new bios update, maybe I’ll try that. 

Same boat here. I have a Caldigit Connect 10G which is a Thunderbolt 3 device with AQC107. It has always been plug and play on my AMD system (Taichi X570 + Alpine Ridge TB3 AIC). This combination does not require any patching and has always been working from Catalina to Monterey 12.2.1 inclusive. With 12.3, it starts exhibiting the same symptom as above.

 

It looks like whatever IOMMU (on) + DisableIoMapper (false) workaround is not available on AMD system. Interestingly, look at the console of my network switch, the system is sending out a number of DHCP request with random MAC address. It looks as if memory mapping is wrong. Not surprised that the system crashes later on.

 

 

Link to comment
Share on other sites

26 minutes ago, sll1001 said:

Same boat here. I have a Caldigit Connect 10G which is a Thunderbolt 3 device with AQC107. It has always been plug and play on my AMD system (Taichi X570 + Alpine Ridge TB3 AIC). This combination does not require any patching and has always been working from Catalina to Monterey 12.2.1 inclusive. With 12.3, it starts exhibiting the same symptom as above.

 

It looks like whatever IOMMU (on) + DisableIoMapper (false) workaround is not available on AMD system. Interestingly, look at the console of my network switch, the system is sending out a number of DHCP request with random MAC address. It looks as if memory mapping is wrong. Not surprised that the system crashes later on.

 

 

Should also mention that the same device works fine when plugging into an M1 Mac mini running 12.3.

 

Link to comment
Share on other sites

On 3/22/2022 at 11:26 PM, Balamut said:

Will this work on 113 too?

The AQC113 on Intel Z690 works with the patch, but, and the big but is it only works with Monterey not Big Sur.  The old patch works from 12.0 to 12.2.1 and the new patch is needed from 12.3+.  

 

I should also point out that I have two thunderbolt enclosures with the AQC107 inside, one from Caldigit and the other from Cable Matters.  Both work with the Macintosh OS since 10.15 and neither needs a patch... plug them in to the thunderbolt port and they come on without issue.   

 

However if anyone is considering using either of these enclosures with Maple Ridge thunderbolt 4, I would suggest to use the Caldigit, as it has JHL6340 which has robust hotplugging support with Maple RIdge... the Cable Matters is using JHL6240 and it doesn't hotplug with any Maple Ridge NVM greater than 24, it must be connected before the system is turned on.  It also doesn't come back from s3 sleep.  My z690 board is running Maple RIdge NVM 31.  The same JHL6240 Cable Matters  device however works extremely well with M1 Max silicon...it hotplugs just fine, similar how it did on Intel Titan Ridge thunderbolt 3.  So I am not sure what the issue is with Maple Ridge and JHL6240 devices.  Bug!

Edited by dehjomz
Link to comment
Share on other sites

On 3/20/2022 at 10:41 AM, PMheart said:

Too lazy to read old posts - could anyone please clarify whether any patch (e.g. built-in or so) is needed before 10.15.4? Thanks!

 

Aquantia patches were needed starting with Catalina (Mojave worked OOTB with no patches).

  • Like 1
Link to comment
Share on other sites

  • 2 weeks later...

Hi guys,

has anyone been able to use Marvell® AQtion AQC113C 10GbE LAN under Big Sur? The patches and the SSDT make it work under Monterey but I haven't been able to get it to work under Big Sur 11.16.5.

Link to comment
Share on other sites

  • 3 weeks later...

so i took a chance, probably wrong but got this 10gbe card for 60 bucks. 

 

 TEG-10GECTX https://www.amazon.com/dp/B01N5AOWW6/ref=cm_sw_r_apan_i_dl_Z32098JWRGJ41W6HTV11

 

its based on tahuti 4010 and has drivers for mac os windows etc. but the last i see the drivers were supported were 10.13 or so. 

does mac os support OLD NIC cards? 

like if they had mac drivers for 10.13 would they like grandfather-in the drivers/card etc? 

 

 

Edited by dovedescent7
Link to comment
Share on other sites

5 hours ago, dovedescent7 said:

so i took a chance, probably wrong but got this 10gbe card for 60 bucks. 

 

 TEG-10GECTX https://www.amazon.com/dp/B01N5AOWW6/ref=cm_sw_r_apan_i_dl_Z32098JWRGJ41W6HTV11

 

its based on tahuti 4010 and has drivers for mac os windows etc. but the last i see the drivers were supported were 10.13 or so. 

does mac os support OLD NIC cards? 

like if they had mac drivers for 10.13 would they like grandfather-in the drivers/card etc? 

 

 

It's a pity this description doesn't show what chip is in the card. It may be Realtek or Marvell.

Can you look for device-id and vendor-id in Windows for example?

Link to comment
Share on other sites

×
×
  • Create New...