Re: [PATCH v4 5/6] at24: Support probing while off
From: Bartosz Golaszewski
Date: Wed Mar 25 2020 - 09:49:03 EST
pon., 23 mar 2020 o 22:31 Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx>
napisaÅ(a):
>
> Bartosz,
>
> On Thu, Mar 12, 2020 at 02:10:32PM +0100, Bartosz Golaszewski wrote:
> > År., 11 mar 2020 o 09:56 Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx> napisaÅ(a):
> > >
> > > Hi Bartosz,
> > >
> > > Thanks for the reply.
> > >
> > > On Wed, Jan 29, 2020 at 02:36:17PM +0100, Bartosz Golaszewski wrote:
> > > > wt., 21 sty 2020 o 14:41 Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx> napisaÅ(a):
> > > > >
> > > > > In certain use cases (where the chip is part of a camera module, and the
> > > > > camera module is wired together with a camera privacy LED), powering on
> > > > > the device during probe is undesirable. Add support for the at24 to
> > > > > execute probe while being powered off. For this to happen, a hint in form
> > > > > of a device property is required from the firmware.
> > > > >
> > > > > Signed-off-by: Sakari Ailus <sakari.ailus@xxxxxxxxxxxxxxx>
> > > > > ---
> > > > > drivers/misc/eeprom/at24.c | 31 +++++++++++++++++++++----------
> >
> > [snip!]
> >
> > > > >
> > > > > static int at24_remove(struct i2c_client *client)
> > > > > {
> > > > > + bool low_power;
> > > > > +
> > > > > pm_runtime_disable(&client->dev);
> > > > > - pm_runtime_set_suspended(&client->dev);
> > > > > + low_power = acpi_dev_state_low_power(&client->dev);
> > > >
> > > > This is inconsistent. You define the low_power field in the context
> > > > structure (BTW the name low_power is a bit vague here - without
> > > > looking at its assignment it would make me think it's about something
> > > > battery-related, how about 'off_at_probe'?) and instead of reusing
> > >
> > > The field was called probe_powered_off in v1, but I changed it to
> > > probe_low_power (and renamed related functions etc.) based on review
> > > comments --- for the device may not be powered off actually.
> > >
> >
> > But is it actually ever low-power? What are the possible logical
> > states of the device? If I understood correctly: it's either off or on
> > at probe - not actually low-power. Am I missing something? In your
> > cover letter you're writing: "These patches enable calling (and
> > finishing) a driver's probe function without powering on the
> > respective device on busses where the practice is to power on the
> > device for probe." To me there's no mention of a low-power state,
> > which makes the name 'probe_low_power' seem completely unrelated.
>
> See <URL:https://patchwork.kernel.org/patch/10938483/>
>
> I've updated the patches according to the comments but did not update the
> cover page accordingly.
>
I see.
Rafael: I think that there are two issues with patch 1/5:
1. It adds a very specific boolean flag to a structure that's meant to
be very general. As I pointed out in the i2c patch: at the very least
this could be made into an int storing flag values, instead of a
boolean field. But rather than that - it looks to me more like a
device (or bus) feature than a driver feature. Is there any ACPI flag
we could use to pass this information to the driver model without
changing the driver structure?
2. The name is still misleading: probe_low_power doesn't correspond
with what it actually does at all (neither did power_off). I'd go with
something like probe_allow_low_power.
> Generally drivers are interested whether a device is powered on so it can
> be accessed, but the actual power state of the device isn't known to the
> driver when it is, well, not in an operational state. A device may be
> powered from a regulator that is always enabled, for instance.
>
> >
> > > > this field here, you call acpi_dev_state_low_power() again. Either
> > > > don't store the context for the life-time of the device if not
> > > > necessary or don't call acpi_dev_state_low_power() at remove, although
> > > > the commit message doesn't describe whether the latter is done on
> > > > purpose.
> > >
> > > Right. probe-low-power property has the same effect on remove for
> > > consistency, i.e. the device can remain in low power state during remove.
> > > This is documented in probe_low_power field documentation in the first
> > > patch.
> > >
> >
> > Just please don't store any state if you're not using it outside of
> > the probe() function.
>
> What exactly are you referring to? The patch adds a local variable to the
> driver's probe and remove functions.
>
Yes, sorry, I looked at the patch and somehow thought it adds a new
field to the data structure and then doesn't reuse it. My bad. Maybe
it was a previous version IDK.
Bartosz
> --
> Kind regards,
>
> Sakari Ailus