Re: [RFC PATCH 2/17] input: RMI4 core bus and sensor drivers.

[Linus Walleij]

On Wed, Sep 26, 2012 at 1:53 AM, Christopher Heiny <cheiny@xxxxxxxxxxxxx> wrote:

>>> >+/* Remove access to raw format string versions */
>>> >+/*#undef simple_show_union_struct
>>> >+#undef show_union_struct_unsigned
>>> >+#undef show_store_union_struct
>>> >+#undef show_repeated_union_struct
>>> >+#undef show_store_repeated_union_struct*/
>>
>> This looks like trying to reimplement ioctl() in sysfs.
>>
>> If what you want is to send big structs in/out of the kernel,
>> use either ioctl() on device nodes (should be trivial since input
>> is using real device nodes) or use configfs.
>
> I'm a little confused.  There's repeated emphasis in the kernel doc that you
> shouldn't use ioctl() anymore - use sysfs instead.  So we've been using
> sysfs, though it seems somewhat klutzy.  If it's actually OK to use ioctl(),
> that could simplify things.  On the other hand, using configfs might be more
> appropriate.

OK yes configfs is said to be ideal for large configuration chunks,
I haven't really used it.

sysfs has this concept of one value per file, and that turns into the
above serialization/marshalling code if followed, so it doesn't look
good. Maybe configfs is the silver bullet.

Yours,
Linus Walleij
--
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/