Re: [PATCH 1/6] toshiba_acpi: Convert to use acpi_driver

[Corentin Chary]

On Wed, Sep 21, 2011 at 3:24 PM, Seth Forshee
<seth.forshee@xxxxxxxxxxxxx> wrote:
> On Wed, Sep 21, 2011 at 09:28:30AM +0200, Corentin Chary wrote:
>> On Tue, Sep 20, 2011 at 11:55 PM, Seth Forshee
>> <seth.forshee@xxxxxxxxxxxxx> wrote:
>> > Changes toshiba_acpi to register an acpi driver and eliminates the
>> > platform device it was using.
>>
>> Why to you want to remove the platform device ? If you want to create
>> a sysfs interface later, you'll probably need it. Most of the platform
>> driver I know only use the acpi_device to send proc/netlink events and
>> the platform_device is used everywhere else. (And anyway, it's an x86
>> *platform* driver, not a pure acpi driver).
>
> I removed the platform device because it seems a bit redundant to have
> both, and I don't see what benefit it really provides. I see your point
> that conceptually it makes sense to have it has a platform device,
> although the distinction there is pretty fine.
>
> Anyway, I guess the cost of keeping the platform device in place is
> pretty small, so I can add it back in if that's desirable.

I don't have a strong opinion, so do what you want (or what Matthew
will tell you to), but keeping it would allow easier access to
subdevices (/sys/platform/devices is a better place than
/sys/bus/acpi/devices/ for this kind of device in my opinion).

>> > Also eliminates most global
>> > variables, moving them into toshiba_acpi_dev, along with some
>> > other miscellaneous fixes and cleanup.
>>
>> Good ! Next step would be to deprecate the /proc interface (keeping it
>> for compatibility) and adding a new shinny
>> /sys/platform/device/toshiba-acpi/ interface correctly documented in
>> Documentation/ABI/ :).
>
> I was avoiding chainging any userspace interfaces in this first round of
> patches, but deprecating the proc interface is definitely something I'd
> like to do. I don't know if there's any point to moving it to sysfs
> though. Most of it already has sysfs interfaces via device classes, and
> I don't know that there's any value in the rest of it.

Well, I must admit that I didn't check that, and it's possible that
all you need to do is to deprecate all /proc file if none of them are
actually usefull.

> Thanks,
> Seth
>



-- 
Corentin Chary
http://xf.iksaif.net
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/