On 10/13/23 10:36, Arnd Bergmann wrote:
> At the moment, I'd suggest focusing on the drivers that still use wext (git grep
> -w iw_handler_def drivers), if we can show that rtl8192e, rtl8712 or ks7010 have
> been broken for a while, removing those would help with removing wext altogether.
I do not know about the the others, but rtl8712 is still in use according to the
flow of questions I see on a GitHub repo that I maintain.
There has been some recent effort to convert rtl8192e to use cfg/nl/mac80211.
That one is problematic as it has the same PCI_ID as RTL8192SE, but the drivers
differ. Note the special code in the probe routine of
drivers/net/wireless/realtek/rtlwifi/rtl8192se to detect if the wrong one has
triggered the routine. That code should also be in the probe routine for
RTL8192E with the test negated. I have both devices, but it has been a long time
since I actually used either of them. I also have an rtl8712 device, thus I
should be available for testing.
Larry