Re: [PATCH v2 0/2] platform/chrome: Update cros_ec sysfs attribute after sensors are found

[Gwendal Grignou]

The v2 patches I submitted were wrong, as cros_ec sysfs driver is
modifying the attributes of the class driver, not its own. I repost a
patch that takes that into account.
Gwendal.

On Sun, May 30, 2021 at 10:36 PM Dmitry Torokhov <dtor@xxxxxxxxxxxx> wrote:
>
> Hi Gwendal,
>
> On Thu, May 27, 2021 at 2:01 PM Gwendal Grignou <gwendal@xxxxxxxxxxxx> wrote:
> >
> > [+dtor]
> > Is this change acceptable? I was worried it could break when
> > asynchronous probing is used [https://lwn.net/Articles/629895/], but
> > since all probes are deferred in a single thread, it is safe.
>
> I think this is a bit awkward that we need to poke a separate sub-driver.
>
> Have you considered having cros_ec_sensorhub.c create its own
> attribute group (it does not have to have a name and it looks like one
> can register as many unnamed groups as they want) and have wake angle
> show and store methods directly in cros_ec_sensorhub.c?
>
> Thanks,
> Dmitry