On Sat, 17 Oct 2026, Chris Taraszka wrote:

> On Mon, 12 Oct 2026, Ilpo Järvinen wrote:
> > I've not seen v2 so far, has there been some progress with this?
> >
> > (Not having the generalization seems a blocker adding many models.)
> 
> Hi Ilpo,
> 
> Sorry for the silence. Mingyou's v12 5/6 ("platform/x86:
> bitland-mifs-wmi: Add quirks table") now has the {value, profile}
> pair table I described, plus the no_set flag for GET-only values and
> accept_zero_function for SET replies without a function id. That
> covers what my v2 generic patch would have done, so I won't post a
> second one.

Okay, thanks. I'll go with that series then.

> I'll test v12 on the Xiaomi Book Pro 14 and base the
> model entry on top of it.
> 
> Mingyou, one issue in v12 5/6, in bitland_mifs_wmi_call():
> 
>     memcpy(output->data, payload_data, sizeof(*output->data));
> 
> data is u8[28], so this copies a single byte. laptop_hwmon_read()
> reads res.data[0..7] for the fan speeds, which would then come from
> uninitialized stack. sizeof(output->data) is what you want.
> 
> Kento, you're right that the match has to be on board_name TM2424.
> Since we have the same machine, let's send a single TM2424 entry. If
> you rebase your v2 onto Mingyou's table, I'm fine with it being yours
> and I'll test it here on BIOS XMAPT4B0P0909 with accept_zero_function
> set.
> 
> Matias, sorry for the delay. I think your TM2107 v3 can go on top of
> Mingyou's 5/6 directly.


-- 
 i.