Re: [PATCH] i2c: Do not give adapters a default parent

[Kay Sievers <kay.sievers-tD+1rO4QERM@public.gmane.org>]

On Sat, Jul 4, 2009 at 19:14, Jean Delvare<khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org> wrote:
> On Mon, 4 May 2009 14:40:36 +0200, Kay Sievers wrote:

>> (The less difference between classes and buses the better. It is wrong
>> to have two types of subsystems, doing almost the same thing. One
>> could argue that it could be useful inside the kernel, which it isn't,
>> I think, but exporting them to userspace was definitely the wrong
>> thing.)
>
> I finally took a stab at this. The resulting patch is below. I have
> used device_type to differentiate between I2C clients and I2C adapters.
> Is this what you had in mind?

Looks fine, by just looking at the patch.

> It seems to work reasonably well, with the following issues remaining:
>
> * The change breaks at least sensors-detect and libsensors. I can
>  easily modify them so that they work again, but we still have a
>  compatibility issue. Is it possible to have a compatibility option
>  that would add symbolic links from class/i2c-adapter/i2c-* to
>  bus/i2c/devices/i2c-* for a couple years?

Yeah, we can add that. I guess others will need that too, if we
convert things from class to bus. How would that look like? Like a
device_add_class_compat_link(*dev, *class)?

> * Now that i2c-core makes use of device_type, I tried to move the power
>  management handling callbacks there from bus_type, to save a test in
>  each function, however I found that the callback set is different
>  between bus_type and device_type.pm. Why is it so? Is there a document
>  explaining the difference? Is the whole world (including bus_type)
>  eventually moving to dev_pm_ops?

I think this is already removed in the current git tree, and all
should use dev_pm_ops, yes.

> * When i2c-adapters were class devices, virtual ones (for example
>  i2c-stub) appeared in sysfs as devices/virtual/i2c-adapter/i2c-*,
>  which made sense and seemed safe. Now that I have turned them into
>  bus devices, virtual ones appear in sysfs as devices/i2c-* directly,
>  which looks dirty and could result in collisions someday. What should
>  be done about this? I wanted to use virtual_device_parent() but it is
>  internal to the driver core at the moment, and doesn't even exist if
>  CONFIG_SYSFS_DEPRECATED=y.

Yeah, we just need to apply the /sys/devices/virtual logic to bus
devices too, it's currently limited to class devices, because there
was no bus device so far who needed this, but should be an easy
change.

> I would be grateful if you can advise on any of the above points.

If you decide to do it that way, you would need the driver core to be able:
  - to create a link from an otherwise empty "struct class" to an
existing bus-device
  - put bus devices without a parent into the /sys/devices/virtual logic
right? Let me know, I can look into that, if you need that.

Thanks,
Kay