Re: [PATCH] iio: bme680_i2c: Make bme680_acpi_match depend on CONFIG_ACPI

[Andy Shevchenko]

On Wed, May 5, 2021 at 11:36 AM Jonathan Cameron
<Jonathan.Cameron@xxxxxxxxxx> wrote:
> On Wed, 5 May 2021 09:32:35 +0100
> Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx> wrote:
> > On Tue, 4 May 2021 11:00:52 -0700
> > Guenter Roeck <linux@xxxxxxxxxxxx> wrote:

+Cc: Paul (I hope you are related to coreboot somehow and can
communicate this further), Pavel and Jacek (LED subsystem suffered
with this as well), Hans, Rafael and linux-acpi@

> > Dropping the ones we are fairly sure are spurious is even better!
>
> If I get bored I'll just do a scrub of all the instances of this that
> you haven't already cleaned up.  It's worth noting that we do
> know some highly suspicious looking entries are out there in the wild.

I have counted ~60 users of acpi_device_id in IIO. Brief looking at
the IDs themselves rings an alarm about half of them.

So, here we may have a chicken and egg problem, i.e. somebody has been
using (or used) fake IDs from Linux kernel in the real products. What
I can consider as a course of action is the following:
1. Clean up (by removing as quickly as possible) the IDs that have no
proof to be real from the Linux kernel sources (perhaps marked as
stable material)
2. Notify ASWG / UEFI forum about all IDs that abuse ACPI
specification and are in Linux kernel, so at least we can keep some
kind of "reserved/do not use" list on the official level (Rafael?)
3. Do not accept any IDs without an evidence provided that they are
being in use in the real products (this should be done on Linux
maintainer level in all subsystems that accept drivers


-- 
With Best Regards,
Andy Shevchenko