Support for Marvell® AQC113CS-B1-C 10Gbps LAN

U

UserOfNames

I have the MSI Meg X670E Ace. Neither the wifi(Not supported on Linux either) or Ethernet work. The controller it uses is AQC113, which is supported on Linux. However appears to be closed source. Are there any plans of development on getting it to work natively?
 
There is net/aquantia-atlantic-kmod, and it worked for 10Gbps Aquantia NIC (rebranded as Marvell now) on Gigabyte Z590 Master, but it seems to be older AQC107 chip; so the port may or not work for you.

EDIT: Meh, https://github.com/Aquantia/aqtion-freebsd/issues/24:
Code: 
Your device is AQC113 based - its a next generation chip, and its not driver-level compatible with AQC107 family.

Unfortunately this is not supported in freebsd driver. To my knowledge, there will be no official extension of freebsd drivers for this chip. Linux driver although, has all the support in there.

EDIT2: Now thinking more about this, half of the reason for LinuxKPI was porting some network driver(s), so it should not be impossible, someone (tm) just needs to do the work :D
 
cracauer@ said:
not even open source
Click to expand...
github.com

GitHub - Aquantia/AQtion: Marvell (formally Aquantia) AQC multigigabit NIC linux driver (atlantic) - development preview

Marvell (formally Aquantia) AQC multigigabit NIC linux driver (atlantic) - development preview - Aquantia/AQtion
github.com github.com

I don't know why OP says it's closed source.

Also (from that repository):
Code: 
Notice this is an experimental software. For stable driver version please either visit http://www.aquantia.com or take the driver bundled with your linux kernel.

So I guess (not having a linux install to check) it's shipped in linux itself?
 
yuripv79 said:
There is net/aquantia-atlantic-kmod, and it worked for 10Gbps Aquantia NIC (rebranded as Marvell now) on Gigabyte Z590 Master, but it seems to be older AQC107 chip; so the port may or not work for you.

EDIT: Meh, https://github.com/Aquantia/aqtion-freebsd/issues/24:
Code: 
Your device is AQC113 based - its a next generation chip, and its not driver-level compatible with AQC107 family.

Unfortunately this is not supported in freebsd driver. To my knowledge, there will be no official extension of freebsd drivers for this chip. Linux driver although, has all the support in there.

EDIT2: Now thinking more about this, half of the reason for LinuxKPI was porting some network driver(s), so it should not be impossible, someone (tm) just needs to do the work :D
Click to expand...
Yeah unfortunately it doesn't work. After installing the driver there still isn't an interface showing in ifconfig. it does exist in pciconf.

Code: 
none2@pci0:5:0:0:   class=0x020000 rev=0x03 hdr=0x00 vendor=0x1d6a device=0x94c0 subvendor=0x1462 subdevice=0x7d69                         
    vendor     = 'Aquantia Corp.'                                                                                                           
    device     = 'AQC113CS NBase-T/IEEE 802.3bz Ethernet Controller [AQtion]'                                                               
    class      = network                                                                                                                   
    subclass   = ethernet                                                                                                                   
    bar   [10] = type Memory, range 64, base 0xfc100000, size 524288, enabled                                                               
    bar   [18] = type Memory, range 64, base 0xfc180000, size 4096, enabled                                                                 
    bar   [20] = type Memory, range 64, base 0xfbc00000, size 4194304, enabled                                                             
    cap 01[40] = powerspec 3  supports D0 D1 D2 D3  current D0                                                                             
    cap 05[50] = MSI supports 32 messages, 64 bit, vector masks                                                                             
    cap 10[70] = PCI-Express 2 endpoint max data 128(512) FLR RO NS                                                                         
                 max read 512                                                                                                               
                 link x2(x2) speed 8.0(16.0) ClockPM disabled                                                                               
    cap 11[b0] = MSI-X supports 32 messages                                                                                                 
                 Table in map 0x18[0x0], PBA in map 0x18[0x200]                                                                             
    ecap 0001[100] = AER 2 0 fatal 0 non-fatal 0 corrected                                                                                 
    ecap 0002[148] = VC 1 max VC0                                                                                                           
    ecap 0003[168] = Serial 1 0000000000000000                                                                                             
    ecap 0019[178] = PCIe Sec 1 lane errors 0                                                                                               
    ecap 0026[198] = Physical Layer 16.0 GT/s 1                                                                                             
    ecap 0027[1bc] = Lane Margining at Receiver 1                                                                                           
    ecap 0018[1d4] = LTR 1                                                                                                                 
    ecap 001e[1dc] = L1 PM Substates 1                                                                                                     
    ecap 000b[1ec] = Vendor [1] ID 0002 Rev 4 Length 256                                                                                   
    ecap 0025[2ec] = Data Link Feature 1                                                                                                   
    ecap 001f[2f8] = Precision Time Measurement 1                                                                                           
    ecap 000b[304] = Vendor [1] ID 0003 Rev 1 Length 84
 
yuripv79 said:
github.com

GitHub - Aquantia/AQtion: Marvell (formally Aquantia) AQC multigigabit NIC linux driver (atlantic) - development preview

Marvell (formally Aquantia) AQC multigigabit NIC linux driver (atlantic) - development preview - Aquantia/AQtion
github.com github.com

I don't know why OP says it's closed source.

Also (from that repository):
Code: 
Notice this is an experimental software. For stable driver version please either visit http://www.aquantia.com or take the driver bundled with your linux kernel.

So I guess (not having a linux install to check) it's shipped in linux itself?
Click to expand...
My bad. It is under an open-source licence (GPL 2.0) in the aq_common.h header. Got confused because the license in the readme on the top-level directory didn't mention it. Also it seems to be an experimental driver and not the stable one, which is part of the linux kernel.
 
Folks, might this not be an opportunity to further develop net/wifibox to provide integration of AQC113 ?
It uses an Alpine Linux VM which may already have support for this device.
 
Last edited by a moderator:
puretone said:
Folks, might this not be an opportunity to further develop net/wifibox to provide integration of AQC113 ?
It uses an Alpine Linux VM which may already have support for this device.
Click to expand...
Is it not only designed for Wireless NIC(s)?
The github doesn't seem to mention any support for ethernet or LAN on the matter.
 
I've noticed that NetBSD & OpenBSD support the AQC113 driver:

aq(4) - NetBSD Manual Pages

man.netbsd.org man.netbsd.org
github.com

src/sys/dev/pci/pcidevs at master · openbsd/src

Read-only git conversion of OpenBSD's official CVS src repository. Pull requests not accepted - send diffs to the tech@ mailing list. - openbsd/src
github.com github.com
I know FreeBSD and NetBSD have diverged but share the same root. How different or difficult would it be to port over the NetBSD aq driver to FreeBSD?
 
I have this hardware and it runs reasonably well under Linux. I'm not a developer, but if I can help with any testing in FreeBSD I'd be happy to do it.
 
clarknova said:
I have this hardware and it runs reasonably well under Linux. I'm not a developer, but if I can help with any testing in FreeBSD I'd be happy to do it.
Click to expand...
Yeah I had to abandon FreeBSD on my new file server with this card in it for Ubuntu. I am new to both but wanted to focus on learning bsd.
 
clarknova said:
I have this hardware and it runs reasonably well under Linux. I'm not a developer, but if I can help with any testing in FreeBSD I'd be happy to do it.
Click to expand...
I'm wondering if simply adding the device ID's and setting the speed capabilities is enough. That's what I did and it would be interesting if someone could test it on a 14.3-RELEASE (15+ doesn't compile).

The ID's are taken from OpenBSD and if adding the device ID is enough the remaining IDs found in OpenBSD could be added (AQC114/115). There would then be the question of a) how to state the fact that it was taken from OpenBSD and b) if I - without any street credit - can post it (maybe attach to an existing bug/feature request).

Below are the links to the repo/branch and a direct link to the 14.3-RELEASE package. The port net/aquantia-atlantic-kmod does not compile on FreeBSD 15+ (open tickets regarding that ticket).

The standard warning: don't install just any pkg from the internet. Review the code and compile it yourself.

Repo: https://git.sr.ht/~patient0/FreeBSD-ports/refs/AQC113
Direkt link to package: https://git.sr.ht/~patient0/FreeBSD...13/aquantia-atlantic-kmod-0.0.5.1403000_4.pkg
 
patient0 said:
Good idea and thanks for pointing me to the PR!
Click to expand...
/off thread

Curious: is that a Moto Guzzi? It isn't a BMW boxer obviously and I only can remember Honda having a V2 vaguely resembling this, but not at this particular angle of cylinders; lastly the drive is what a Moto Guzzi usually has.
 
I have same problem. If no one's porting it, I was shown these days one can ask a powerful AI (ChatGPT, DeepSeek, Gemini, Grok, K2, Meta, Qwen, etc.) to write drivers but code might need minor corrections (even that for drivers is beyond me for now, though of interest... just don't want to mess up hardware). I'm slightly sceptical because when I asked some Western AIs about an OS package management/building it said use a tool that doesn't exist anywhere (hallucination) and then use a tool only for other OS... if they can't get beginner stuff like this right, I'm sceptical how good code would be,
 
AIs - my take: don't use it unless you have a reliable way of verifying the output (no, asking two different AIs the same questions isn't a reliable way).
 
Erichans said:
/off thread

Curious: is that a Moto Guzzi? It isn't a BMW boxer obviously and I only can remember Honda having a V2 vaguely resembling this, but not at this particular angle of cylinders; lastly the drive is what a Moto Guzzi usually has.
Click to expand...
/off thread: It is indeed a Moto Guzzi Griso 1100 :)
 
You must log in or register to reply here.
Back
Top