Re: [PATCH v15 0/13] ax88179_178a: Add support for AX88179A-based chips

[Birger Koblitz <mail@birger-koblitz.de>]

Hi Jianhui,

On 04/10/2026 7:26 am, Jianhui Xu wrote:
> Hi Birger,
> 
> I tested the 13-patch v15 series on an AX88179B adapter (USB 0b95:1790,
> bcdDevice 0x0200) on the declared net-next base
> 071876fd50482a68603a9460d80dd6dd58827ee1.
> 
> Builds, cold DHCP, 1000/100/10 Mbit/s link changes, EEE, pause settings,
> EEPROM reads, autonegotiation restart, module reload, USB reattach, MII
> ioctl reads, and repeated 100/1000 Mbit/s transitions all passed. The build
> only showed the pre-existing unused mask warning in
> asix_ax88772D_get_features(). Deep-S3 resume also passed with both wol
> d and wol g; resume was triggered with QEMU system_wakeup, not a magic
> packet. The guest later hit an init-exit panic during poweroff, which
> appears separate and has not been investigated further.
> 
> However, I did find a reproducible RX VLAN filtering failure. With
> rx-vlan-filter enabled, ip link add for VLAN 123 or 124 returned
> success, but no VLAN interface was created. With the feature disabled,
> VLAN creation worked normally.
> 
> The first read in ax88179a_vlan_rx_add_vid() returns the byte count on
> success, so:
> 
> if (ret)
>      return ret;
> 
> returns 1 before the VLAN bitmap is programmed.
> 
> The patch below fixes both issues and worked correctly when applied
> directly on top of v15.
> 
> Because unmodified v15 has this VLAN failure, please do not carry my
> previous Tested-by for v15.
> 

Thanks so much again for your thorough testing, I know how much work 
this is!
I added your fix and re-sent as v16 without any other changes. Since you 
tested exactly that version successfully, I hope it is OK to keep the 
tested-by for the v16 version.

Best,
   Birger