Re: [PATCH] hwmon: Add driver for MSI Mystic Light controller voltage sensors

From: Guenter Roeck

Date: Sat Nov 07 2026 - 10:11:04 EST


On 11/7/26 00:27, Tomasz Bojanowski wrote:
Your reply did not reach my inbox, because the spam filtering at my mail
provider was set too strictly. I have fixed that, and I read your review
on lore.


There is nothing I can do about that, sorry. No matter which provider I use
to send e-mail, some provider somewhere considers it to be spam.

Thanks for the review, and for trying the driver on your boards.

On 11/6/26 16:36, Guenter Roeck wrote:
Using the USB serial number to detect the device seems fragile. If the
USB vendor/product ID can not be used for some reason, it might make
more sense to implement a DMI match.

Agreed. v2 identifies the board by DMI only and no longer looks at the
serial number. The driver is only registered on supported boards.

This will require additional per-board data. On MAG B650M
MORTAR WIFI (MS-7D76), the USB IDs are 1462:7d76, and
hid_hw_raw_request() with the parameters below returns
a length of 0 bytes. On MAG B650 TOMAHAWK WIFI (MS-7D75),
the USB device ID is 1462:7d75, and on MAG B550M MORTAR
(MS-7C94) it is 1462:7c94.

That makes me wonder if the parameters can be associated
with USB IDs.

On these older boards the controller uses the board code as its
product ID (1462:7dxx, 1462:7cxx). As far as I can tell from the vendor
software, the sensors of these controllers are read with a different
request, not with feature report 0xf0, which would be consistent with
the 0 byte reply. I have not verified this on hardware.


I have the ones mentioned above, so I could test there if you want to
add support for it.

On the 800-series boards the controller uses a common ID, 0db0:0076.
OpenRGB lists more than 30 X870, B850, B840 and Z890 boards behind this
ID and selects its per-board LED configuration by the DMI board name.
The vendor software also handles these boards with one class and tells
them apart by the USB serial number. So the USB ID does not identify
the board there, which is why v2 uses DMI.

For now the driver only covers the 800-series controller on the MAG
B850M MORTAR WIFI (MS-7E61), which is the board I can test on.

v2 also addresses the issues reported by Sashiko and your other
comments: it reads and validates the report in probe, fails probe if
that does not work, only exposes the rails that return a reading, and
no longer returns -ENODATA. I will send it as a new thread.


Thanks a lot for the information.

Guenter