Changes since v7:

* Applied changes from Danilo's and Alex's and reviews (thanks!):
    * Removed a blank line, one each, from the Class and Vendor macros.
    * Moved example code location from struct Vendor, to vendor_id(),
      and introduced it in a later commit, in its final form.
    * Applied Alex's Reviewed-by tag to the series.

Changes since v6:

* Applied changes from Danilo's and Alex's and Elle's reviews (thanks!):
    * Rebased onto driver-core-next, which is here:
          https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git
    * Changed pci::Vendor to be a u16, instead of a u32.
    * Inlined all of the tiniest functions.
    * Changed from Class/Vendor new(), to from_raw().
    * Made from_raw() only accessible to super, which in this case is
      the pci module.
    * Restored infallible operations. That causes Alex's request for the
      following reasonable behavior to work once again:

          from_raw(0x10de).as_raw() == 0x10de

    * Added a new patch, to inline the remaining PCI operations. This
      provides consistent inline choices throughout pci.rs.

Changes since v5:
* Applied changes from Danilo's review (thanks!):
    * Split the nova-core patch into two patches, for nova and pci.
    * Added rust/kernel/pci/ to MAINTAINERS.

Changes since v4:
* Applied changes from Danilo's and Alex's review (thanks!):
    * Reorganized the patches so that the Nova changes consume the
      results of Class and Vendor upgrades, all in one shot.

    * Made Class and Vendor types get constructed infallibly.

    * This was all somewhat disruptive, and also required one more patch
      in order to properly separate the various steps. But I think it is
      all correct now. And CLIPPY=1 builds cleanly too.

* Elle Rhumsaa provided a Reviewed-by for v4 (thanks!), but due to the
  churn in v5 here, I thought it best to not add that tag to v5 yet.
  Instead, I have directly Cc'd Elle on the patches for now.

Changes since v3:
* Applied changes from Danilo's review (thanks!):

    * Moved Class and Vendor to a new pci/id.rs file.
    * Added ClassMask, to constrain callers to use only the two valid
      masks.
    * Removed pci_class_code_raw()
    * Changed Class and Vendor .as_u32() to .as_raw(), because after
      Danilo's comment I looked around rust/kernel and learned that
      .as_raw() is the overwhelmingly used convention.
    * Changed vendor_id() to return a Vendor instance directly.
        * Also, validated Vendor during construction, just as is done
          with Class. Both of these items are expected to match known
          values, even for new devices, so that's a reasonable move.


Changes since v2:

* Applied changes from Danilo's and Alex's review (thanks!):

    * Moved everything possible out of the new define_all_pci_classes!()
      and define_all_pci_vendors!() macros.
    * Used "impl TryFrom<u32> for Class/Vendor", instead of .from_u32().
    * Made the new DeviceId methods infallible.
    * Upgraded DeviceId::from_id() to accept a Vendor struct.

* Changed the names to be a little clearer:
    * class_code_raw() --> pci_class_code_raw()
    * class_enum() --> pci_class()

* Added doctests for the items that are not yet used in real drivers.

v2 is here:
    https://lore.kernel.org/20250818013305.1089446-1-jhubbard@nvidia.com

Changes since v1:

1) Use the pci_device_table for filtering, instead of open-coding
   filters in the .probe() callback.

2) Add PCI Class (class, subclass, implementation) and PCI Vendor to
   Rust for Linux.

3) Rebased onto the latest nova-next branch, which is here:
    https://gitlab.freedesktop.org/drm/nova.git

v1 is here:
    https://lore.kernel.org/20250813232859.224316-1-jhubbard@nvidia.com

Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Alexandre Courbot <acourbot@nvidia.com>
Cc: Elle Rhumsaa <elle@weathered-steel.dev>

John Hubbard (6):
  rust: pci: provide access to PCI Class and Class-related items
  rust: pci: provide access to PCI Vendor values
  rust: pci: add DeviceId::from_class_and_vendor() method
  gpu: nova-core: avoid probing non-display/compute PCI functions
  rust: pci: use pci::Vendor instead of bindings::PCI_VENDOR_ID_*
  rust: pci: inline several tiny functions

 MAINTAINERS                           |   1 +
 drivers/gpu/nova-core/driver.rs       |  33 +-
 rust/kernel/pci.rs                    |  72 +++-
 rust/kernel/pci/id.rs                 | 558 ++++++++++++++++++++++++++
 samples/rust/rust_dma.rs              |   6 +-
 samples/rust/rust_driver_auxiliary.rs |  12 +-
 samples/rust/rust_driver_pci.rs       |   9 +-
 7 files changed, 663 insertions(+), 28 deletions(-)
 create mode 100644 rust/kernel/pci/id.rs


base-commit: 8f5ae30d69d7543eee0d70083daf4de8fe15d585
-- 
2.51.0

On Sat Aug 30, 2025 at 12:36 AM CEST, John Hubbard wrote:
> Changes since v7:
>
> * Applied changes from Danilo's and Alex's and reviews (thanks!):
>     * Removed a blank line, one each, from the Class and Vendor macros.
>     * Moved example code location from struct Vendor, to vendor_id(),
>       and introduced it in a later commit, in its final form.
>     * Applied Alex's Reviewed-by tag to the series.

I think you forgot to align Debug and Display, i.e. Debug still prints decimal
values.

Is this intentional? If not, no worries, I can fix it up on apply (which a few
minor doc-comment nits):

diff --git a/rust/kernel/pci/id.rs b/rust/kernel/pci/id.rs
index f6ce8f8a2a4d..f534133aed3d 100644
--- a/rust/kernel/pci/id.rs
+++ b/rust/kernel/pci/id.rs
@@ -26,7 +26,7 @@
 ///     Ok(())
 /// }
 /// ```
-#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+#[derive(Clone, Copy, PartialEq, Eq)]
 #[repr(transparent)]
 pub struct Class(u32);

@@ -81,12 +81,18 @@ const fn to_24bit_class(val: u32) -> u32 {
     }
 }

-impl fmt::Display for Class {
+impl fmt::Debug for Class {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "0x{:06x}", self.0)
     }
 }

+impl fmt::Display for Class {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        <Self as fmt::Debug>::fmt(self, f)
+    }
+}
+
 impl ClassMask {
     /// Get the raw mask value.
     #[inline]

On 9/1/25 7:58 AM, Danilo Krummrich wrote:
> On Sat Aug 30, 2025 at 12:36 AM CEST, John Hubbard wrote:
>> Changes since v7:
>>
>> * Applied changes from Danilo's and Alex's and reviews (thanks!):
>>      * Removed a blank line, one each, from the Class and Vendor macros.
>>      * Moved example code location from struct Vendor, to vendor_id(),
>>        and introduced it in a later commit, in its final form.
>>      * Applied Alex's Reviewed-by tag to the series.
> 
> I think you forgot to align Debug and Display, i.e. Debug still prints decimal
> values.
> 
> Is this intentional? If not, no worries, I can fix it up on apply (which a few
> minor doc-comment nits):

Yes, I missed that one. Fixup on apply would be great, yes.


thanks,
John Hubbard

> 
> diff --git a/rust/kernel/pci/id.rs b/rust/kernel/pci/id.rs
> index f6ce8f8a2a4d..f534133aed3d 100644
> --- a/rust/kernel/pci/id.rs
> +++ b/rust/kernel/pci/id.rs
> @@ -26,7 +26,7 @@
>   ///     Ok(())
>   /// }
>   /// ```
> -#[derive(Debug, Clone, Copy, PartialEq, Eq)]
> +#[derive(Clone, Copy, PartialEq, Eq)]
>   #[repr(transparent)]
>   pub struct Class(u32);
> 
> @@ -81,12 +81,18 @@ const fn to_24bit_class(val: u32) -> u32 {
>       }
>   }
> 
> -impl fmt::Display for Class {
> +impl fmt::Debug for Class {
>       fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
>           write!(f, "0x{:06x}", self.0)
>       }
>   }
> 
> +impl fmt::Display for Class {
> +    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
> +        <Self as fmt::Debug>::fmt(self, f)
> +    }
> +}
> +
>   impl ClassMask {
>       /// Get the raw mask value.
>       #[inline]

On Sat Aug 30, 2025 at 12:36 AM CEST, John Hubbard wrote:
> John Hubbard (6):

Applied to driver-core-testing, thanks!

>   rust: pci: provide access to PCI Class and Class-related items

    [ Minor doc-comment improvements, align Debug and Display. - Danilo ]

>   rust: pci: provide access to PCI Vendor values

    [ Minor doc-comment improvements, align Debug and Display. - Danilo ]

>   rust: pci: add DeviceId::from_class_and_vendor() method

    [ Minor doc-comment improvements. - Danilo ]

>   gpu: nova-core: avoid probing non-display/compute PCI functions
>   rust: pci: use pci::Vendor instead of bindings::PCI_VENDOR_ID_*

    [ Replace "as a validated vendor" with "as [`Vendor`]". - Danilo ]

>   rust: pci: inline several tiny functions