Bug 217204

Summary: ASIX AX88179 does not work in 6.X kernel
Product: Drivers Reporter: petr.bahula
Component: USBAssignee: Default virtual assignee for Drivers/USB (drivers_usb)
Status: RESOLVED ANSWERED    
Severity: normal CC: marcan
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 6.1 Subsystem:
Regression: Yes Bisected commit-id:

Description petr.bahula 2023-03-15 16:04:49 UTC
After update from kernel 5.15.80 to 6.1.12 the USB ETH card ASIX AX88179 which is in USB-C dock is not accessible as eth1.

**!!! IMPORTANT !!!
If the HW is booted with good kernel and then rebooted (without losing power) to bad kernel, it will still work. Only if it is booted with bad kernel from cold state (total no power) then it will not work.**

```
Bisecting: 2 revisions left to test after this (roughly 2 steps)
[a41b17ff9dacd22f5f118ee53d82da0f3e52d5e3] dccp: put dccp_qpolicy_full() and dccp_qpolicy_push() in the same lock
Bisecting: 0 revisions left to test after this (roughly 1 step)
[f56530dcdb0684406661ac9f1accf48319d07600] net: usb: make USB_RTL8153_ECM non user configurable
c67cc4315a8e605ec875bd3a1210a549e3562ddc is the first bad commit
commit c67cc4315a8e605ec875bd3a1210a549e3562ddc
Author: Hector Martin <marcan@marcan.st>
Date:   Sun Jul 31 16:22:09 2022 +0900

    net: usb: ax88179_178a: Bind only to vendor-specific interface
    
    The Anker PowerExpand USB-C to Gigabit Ethernet adapter uses this
    chipset, but exposes CDC Ethernet configurations as well as the
    vendor specific one. This driver tries to bind by PID:VID
    unconditionally and ends up picking up the CDC configuration, which
    is supposed to be handled by the class driver. To make things even
    more confusing, it sees both of the CDC class interfaces and tries
    to bind twice, resulting in two broken Ethernet devices.
    
    Change all the ID matches to specifically match the vendor-specific
    interface. By default the device comes up in CDC mode and is bound by
    that driver (which works fine); users may switch it to the vendor
    interface using sysfs to set bConfigurationValue, at which point the
    device actually goes through a reconnect cycle and comes back as a
    vendor specific only device, and then this driver binds and works too.
    
    The affected device uses VID/PID 0b95:1790, but we might as well change
    all of them for good measure, since there is no good reason for this
    driver to bind to standard CDC Ethernet interfaces.
    
    v3: Added VID/PID info to commit message
    
    Signed-off-by: Hector Martin <marcan@marcan.st>
    Link: https://lore.kernel.org/r/20220731072209.45504-1-marcan@marcan.st
    Signed-off-by: Paolo Abeni <pabeni@redhat.com>

 drivers/net/usb/ax88179_178a.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)
```

I've reverted changes made in the c67cc4315a8e605ec875bd3a1210a549e3562ddc commit and gentoo-sources-6.1.12 works correctly now.

Patch which I use for new kernels now:
```
pet@petkub ~ $  cat /etc/portage/patches/sys-kernel/gentoo-sources/ASIX_AX88179.patch 
--- a/drivers/net/usb/ax88179_178a.c    2022-12-11 23:15:18.000000000 +0100
+++ b/drivers/net/usb/ax88179_178a.c    2023-02-23 10:04:47.534060336 +0100
@@ -1844,7 +1844,8 @@
 static const struct usb_device_id products[] = {
 {
        /* ASIX AX88179 10/100/1000 */
-       USB_DEVICE_AND_INTERFACE_INFO(0x0b95, 0x1790, 0xff, 0xff, 0),
+       /*USB_DEVICE_AND_INTERFACE_INFO(0x0b95, 0x1790, 0xff, 0xff, 0),*/
+       USB_DEVICE(0x0b95, 0x1790),
        .driver_info = (unsigned long)&ax88179_info,
 }, {
        /* ASIX AX88178A 10/100/1000 */
```

See also: https://bugs.gentoo.org/895720
Comment 1 Hector Martin 2023-03-17 08:36:21 UTC
I mentioned this by email, but to keep things in the bug tracker: Do you have CONFIG_USB_NET_CDC_NCM enabled? If not, please enable it. That driver is the correct driver to use for standard CDC NCM Ethernet devices.

I would also like the output of `lsusb -v -d 0b95:` (with NCM or in the not-working state, NOT with the old kernel or your patched one) to see exactly what your device is reporting by default.
Comment 2 petr.bahula 2023-03-20 09:00:13 UTC
(In reply to Hector Martin from comment #1)
> I mentioned this by email, but to keep things in the bug tracker: Do you
> have CONFIG_USB_NET_CDC_NCM enabled? If not, please enable it. That driver
> is the correct driver to use for standard CDC NCM Ethernet devices.
> 
Yes, you are true, the CONFIG_USB_NET_CDC_NCM must be enabled. All is working now.
I configured the device according the https://linux-hardware.org/index.php?id=usb:0b95-1790 and the CONFIG_USB_NET_CDC_NCM is not listed there. I know not haw it is generated but I will send info about mistake to info@linux-hardware.org.
Thanks for your help and sorry for lost your time.
Best regards