Short-Lived ASUS Sensors Linux Driver Being Removed In Favor Of Better Alternative

[billyswong]

My Asus B450 board looks still not supported by the new mainline driver. I miss my fan speed reading when the it87 still worked.

And then suddenly today after a kernel update the it87 module gives out fan speed reading again. After years of losing it. What a coincidence.

[imaami]

Sigh, speaking of hwmon stuff that doesn't work, the MSI X570S Carbon Max WiFi has a Nuvoton NCT6687D chip with next to no kernel support. The nct6683 driver should support it but actually doesn't, because it fails to match the X570S Carbon Max's specific customer ID out of the box. And even after adding a macro for the ID the driver still requires an additional hack to make pwm control sort of work. And a bunch of other things are wonky, too.

If anyone else here has this mb and wants to make nct6683 only mostly broken (instead of fully) then have a go: https://github.com/imaami/linux/comm...92b8931eee721f

Btw: my dirty hack doesn't add proper pwm_enable attributes to sysfs. I just made the normal pwm attributes accept the special value -1 in addition to the usual 0-255. Writing -1 switches from manual mode to SmartFan mode. Values 0-255 will automatically switch to manual mode and set the requested pwm value.

[imaami]

MSI X570S Carbon Max WiFi has a Nuvoton NCT6687D chip with next to no kernel support. The nct6683 driver should support it but actually doesn't

Oh and before anyone mentions the out-of-tree nct6687 driver (you can find it on github if you search), yes I'm aware it exists. I just don't find it to my liking. It's very much a modified nct6683 anyway, and has code in it that in all likelihood wouldn't be accepted to upstream hwmon. I'd prefer to see nct6683 get fixed instead.