From: "Rafael J. Wysocki" <rafael@kernel.org>
To: Raag Jadav <raag.jadav@intel.com>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>,
mika.westerberg@linux.intel.com,
andriy.shevchenko@linux.intel.com, lenb@kernel.org,
robert.moore@intel.com, ardb@kernel.org, will@kernel.org,
mark.rutland@arm.com, linux-acpi@vger.kernel.org,
linux-kernel@vger.kernel.org,
acpica-devel@lists.linuxfoundation.org,
linux-efi@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
mallikarjunappa.sangannavar@intel.com, bala.senthil@intel.com
Subject: Re: [PATCH v2 2/6] ACPI: bus: update acpi_dev_uid_match() to support multiple types
Date: Wed, 22 Nov 2023 12:55:39 +0100 [thread overview]
Message-ID: <CAJZ5v0gyuk-1vfpaRWO1wniYHwMp==Nx9KLVS42=39yXmqKq6Q@mail.gmail.com> (raw)
In-Reply-To: <ZV2KYqah4FHH4tnz@black.fi.intel.com>
On Wed, Nov 22, 2023 at 5:58 AM Raag Jadav <raag.jadav@intel.com> wrote:
>
> On Tue, Nov 21, 2023 at 08:25:20PM +0100, Rafael J. Wysocki wrote:
> > On Tue, Nov 21, 2023 at 11:38 AM Raag Jadav <raag.jadav@intel.com> wrote:
> > >
> > > According to ACPI specification, a _UID object can evaluate to either
> > > a numeric value or a string. Update acpi_dev_uid_match() helper to
> > > support _UID matching for both integer and string types.
> > >
> > > Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> > > Suggested-by: Mika Westerberg <mika.westerberg@linux.intel.com>
> > > Suggested-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > You need to be careful with using this. There are some things below
> > that go beyond what I have suggested.
>
> I think we all suggested some bits and pieces so I included everyone.
> We can drop if there are any objections.
There are, from me and from Andy.
[cut]
> > Up to this point it is all fine IMV.
> >
> > > +/**
> > > + * acpi_dev_uid_match - Match device by supplied UID
> > > + * @adev: ACPI device to match.
> > > + * @uid2: Unique ID of the device.
> > > + *
> > > + * Matches UID in @adev with given @uid2.
> > > + *
> > > + * Returns: %true if matches, %false otherwise.
> > > + */
> > > +
> > > +/* Treat uid as a string for array and pointer types, treat as an integer otherwise */
> > > +#define get_uid_type(x) \
> > > + (__builtin_choose_expr(is_array_or_pointer_type(x), (const char *)0, (u64)0))
> >
> > But I wouldn't use the above.
> >
> > It is far more elaborate than needed for this use case and may not
> > actually work as expected. For instance, why would a pointer to a
> > random struct type be a good candidate for a string?
>
> Such case will not compile, since its data type will not match with
> acpi_str_uid_match() prototype. The compiler does a very good job of
> qualifing only the compatible string types here, which is exactly what
> we want.
>
> error: passing argument 2 of 'acpi_str_uid_match' from incompatible pointer type [-Werror=incompatible-pointer-types]
> if (acpi_dev_uid_match(adev, adev)) {
> ^
> ./include/acpi/acpi_bus.h:870:20: note: expected 'const char *' but argument is of type 'struct acpi_device *'
> static inline bool acpi_str_uid_match(struct acpi_device *adev, const char *uid2)
You are right, it won't compile, but that's not my point. Why would
it be matched with acpi_str_uid_match() in the first place?
> > > +
> > > +#define acpi_dev_uid_match(adev, uid2) \
> > > + _Generic(get_uid_type(uid2), \
> > > + const char *: acpi_str_uid_match, \
> > > + u64: acpi_int_uid_match)(adev, uid2)
> > > +
> >
> > Personally, I would just do something like the following
> >
> > #define acpi_dev_uid_match(adev, uid2) \
> > _Generic((uid2), \
> > const char *: acpi_str_uid_match, \
> > char *: acpi_str_uid_match, \
> > const void *: acpi_str_uid_match, \
> > void *: acpi_str_uid_match, \
> > default: acpi_int_uid_match)(adev, uid2)
> >
> > which doesn't require compiler.h to be fiddled with and is rather
> > straightforward to follow.
> >
> > If I'm to apply the patches, this is about the level of complexity you
> > need to target.
>
> Understood, however this will limit the type support to only a handful
> of types,
Indeed.
> and will not satisfy a few of the existing users, which, for
> example are passing signed or unsigned pointer or an array of u8.
Fair enough, so those types would need to be added to the list.
> Listing every possible type manually for _Generic() looks a bit verbose
> for something that can be simply achieved by __builtin functions in my
> opinion.
But then you don't even need _Generic(), do you?
Wouldn't something like the below work?
#define acpi_dev_uid_match(adev, uid2) \
(__builtin_choose_expr(is_array_or_pointer_type((uid2)),acpi_str_uid_match(adev,
uid2), acpi_int_uid_match(adev, uid2))
In any case, I'm not particularly convinced about the
is_array_or_pointer_type() thing and so I'm not going to apply the
series as is.
next prev parent reply other threads:[~2023-11-22 11:55 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-21 10:38 [PATCH v2 0/6] Support _UID matching for integer types Raag Jadav
2023-11-21 10:38 ` [PATCH v2 1/6] compiler.h: Introduce helpers for identifying array and pointer types Raag Jadav
2023-11-21 10:38 ` [PATCH v2 2/6] ACPI: bus: update acpi_dev_uid_match() to support multiple types Raag Jadav
2023-11-21 16:50 ` Mika Westerberg
2023-11-21 19:25 ` Rafael J. Wysocki
2023-11-22 4:58 ` Raag Jadav
2023-11-22 11:55 ` Rafael J. Wysocki [this message]
2023-11-22 10:12 ` Andy Shevchenko
2023-11-21 10:38 ` [PATCH v2 3/6] ACPI: bus: update acpi_dev_hid_uid_match() " Raag Jadav
2023-11-21 16:52 ` Mika Westerberg
2023-11-21 10:38 ` [PATCH v2 4/6] ACPI: LPSS: use acpi_dev_uid_match() for matching _UID Raag Jadav
2023-11-21 16:52 ` Mika Westerberg
2023-11-21 10:38 ` [PATCH v2 5/6] efi: dev-path-parser: " Raag Jadav
2023-11-21 16:19 ` Ard Biesheuvel
2023-11-21 10:38 ` [PATCH v2 6/6] perf: arm_cspmu: drop redundant acpi_dev_uid_to_integer() Raag Jadav
2023-11-21 15:59 ` Will Deacon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAJZ5v0gyuk-1vfpaRWO1wniYHwMp==Nx9KLVS42=39yXmqKq6Q@mail.gmail.com' \
--to=rafael@kernel.org \
--cc=acpica-devel@lists.linuxfoundation.org \
--cc=andriy.shevchenko@linux.intel.com \
--cc=ardb@kernel.org \
--cc=bala.senthil@intel.com \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-efi@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mallikarjunappa.sangannavar@intel.com \
--cc=mark.rutland@arm.com \
--cc=mika.westerberg@linux.intel.com \
--cc=raag.jadav@intel.com \
--cc=robert.moore@intel.com \
--cc=will@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).