MIME-Version: 1.0
References: <20211224145903.368999-1-kai.heng.feng@canonical.com>
 <CAHp75Vd3RhVUe_Yoz-fPErzYcV=+gtOVsxNTmn2_52JbSUaMaA@mail.gmail.com> <CAAd53p7bu=+bs5c2Y1LQLC7fPLUihNX_QHw-yOh=fUKUdXWWhg@mail.gmail.com>
In-Reply-To: <CAAd53p7bu=+bs5c2Y1LQLC7fPLUihNX_QHw-yOh=fUKUdXWWhg@mail.gmail.com>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Sat, 25 Dec 2021 18:19:16 +0200
Message-ID: <CAHp75VeVFcaJJeacDCMBEaAVNXUgEF_oaHWXUpHobAYX4EmBuQ@mail.gmail.com>
Subject: Re: [PATCH] iio: humidity: hdc100x: Add ACPI HID table
To:     Kai-Heng Feng <kai.heng.feng@canonical.com>
Cc:     "jic23@kernel.org" <jic23@kernel.org>,
        "lars@metafoo.de" <lars@metafoo.de>,
        Matt Ranostay <matt.ranostay@konsulko.com>,
        Chris Lesiak <chris.lesiak@licor.com>,
        "linux-iio@vger.kernel.org" <linux-iio@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sat, Dec 25, 2021 at 3:05 PM Kai-Heng Feng
<kai.heng.feng@canonical.com> wrote:
> On Sat, Dec 25, 2021 at 7:32 PM Andy Shevchenko
> <andy.shevchenko@gmail.com> wrote:
> > On Friday, December 24, 2021, Kai-Heng Feng <kai.heng.feng@canonical.com> wrote:
> >>
> >> x86 boards may use ACPI HID "HDC1010" to for hdc100x device.
> >>
> >> So add an ACPI match table for that accordingly.
> >
> > No. We do not add abusing IDs blindly.
> > Why this is in use? Is the creative creator of that informed that is an ACPI spec abuse?
>
> Can you please elaborate more on this topic? How is this an ACPI spec abuse?

Yes, https://uefi.org/PNP_ACPI_Registry has links to PNP and ACPI ID registries.
Note, the main differences between them are:
- PNP uses 3-letter vendor ID, ACPI uses 4-letter
- PNP is in maintenance mode and shouldn't be expanded

So, Since above should formally satisfy the PNP ID, the HDC is not a
TI vendor ID in the PNP ID registry.

The section in the spec that refers to it:
https://uefi.org/specs/ACPI/6.4/06_Device_Configuration/Device_Configuration.html#hid-hardware-id

> I did suggest them to use PRP0001, but I also don't think this is an abuse.

The PRP0001 is not an abuse per se, but rather it is highly _not_
recommended for a production use.

> > What devices are those? Is it available on the market? Where is the link to DSDT?
>
> It's not on the market yet.

Thank God!

> Do you need the full DSDT? Or just the
> part of the ACPI device?
> I'll need approve from customer to disclose these info.

This is a requirement only for the devices on the market when nobody
prevented the disaster from becoming... Since it's not yet, no need
for it.

> > Does TI has an ID for that?
>
> I was told by customer this is approved by TI.

How can TI approve that for "HARDCOM ELEKTRONIK & DATATEKNIK"? Is
there evidence that this company has been bought by TI or TI got full
rights on their IPs?

ACPI: TEXAS INSTRUMENTS TXNW
PNP: TEXAS INSTURMENTS TXN
(seems a typo in the company's name, but it should be notified to the
appropriate channels)

So, the ID should start with one of the above, and not HDC as far as I
understood.

PNP: TEXAS MICROSYSTEM TMI
This one probably is not related, but I'm not informed.

I would like you (and your customer) to point to my blog post on the
topic and it explains how the things should be done with the ACPI IDs:
https://andy-shev.dreamwidth.org/151340.html

Hope this will help and be resolved soon.

-- 
With Best Regards,
Andy Shevchenko