Re: [PATCH] cpufreq: fix using cpufreq-dt as module

[Andreas Kemnade <andreas@kemnade.info>]

Am Mon, 04 Nov 2024 10:35:26 -0800
schrieb Kevin Hilman <khilman@kernel.org>:

> Andreas Kemnade <andreas@kemnade.info> writes:
> 
> > E.g. omap2plus_defconfig compiles cpufreq-dt as module. As there is
> > no module alias nor a module_init(), cpufreq-dt-platdev will not be
> > used and therefore on several omap platforms there is no cpufreq.
> >
> > Enforce builtin compile of cpufreq-dt-platdev to make it effective.
> >
> > Fixes: 3b062a086984 ("cpufreq: dt-platdev: Support building as
> > module") Cc: stable@vger.kernel.org
> > Signed-off-by: Andreas Kemnade <andreas@kemnade.info>  
> 
> I'd much rather see this fixed to work as a module.  You already
> hinted at the right way to do that, so please do that instead.
> 
no clear idea how. What aliases should I add? The cpufreq-dt-platdev is
not a real driver, so I could not create mod_devicetable aliases to
match a given device. It constructs a device under certain conditions
depending on the board compatible, so no simple list of compatibles, it
contains allow and blocklists.

cpufreq-dt then binds to that device and that one can be built as a
module (which then made cpufreq-dt-platdev also a module, causing the
trouble). I do not see any benefit from having cpufreq-dt-platdev as a
module. ti-cpufreq has a similar role and is also just builtin.
It does itself no real work but provides a device cpufreq-dt then binds
to.

Handling module removal would probably need to be added and tested. I
feel not comfortable having such as a regression fix and for stable.

Regards,
Andreas