Nouveau - Aug 2025 - [PATCH v4 2/3] gpu: nova-core: avoid probing non-display/compute PCI functions

NovaCore has so far been too imprecise about figuring out if .probe()
has found a supported PCI PF (Physical Function). By that I mean:
.probe() sets up BAR0 (which involves a lot of very careful devres and
Device<Bound> details behind the scenes). And then if it is dealing with
a non-supported device such as the .1 audio PF on many GPUs, it fails
out due to an unexpected BAR0 size. We have been fortunate that the BAR0
sizes are different.

Really, we should be filtering on PCI class ID instead. These days I
think we can confidently pick out Nova's supported PF's via PCI class
ID. And if not, then we'll revisit.

The approach here is to filter on "Display VGA" or "Display
3D", which
is how PCI class IDs express "this is a modern GPU's PF".

Cc: Danilo Krummrich <dakr at kernel.org>
Cc: Alexandre Courbot <acourbot at nvidia.com>
Signed-off-by: John Hubbard <jhubbard at nvidia.com>
---
 drivers/gpu/nova-core/driver.rs | 28 +++++++++++++++++++++++-----
 1 file changed, 23 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/nova-core/driver.rs b/drivers/gpu/nova-core/driver.rs
index 274989ea1fb4..b60c9defa9d1 100644
--- a/drivers/gpu/nova-core/driver.rs
+++ b/drivers/gpu/nova-core/driver.rs
@@ -1,6 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0
 
-use kernel::{auxiliary, bindings, c_str, device::Core, pci, prelude::*,
sizes::SZ_16M, sync::Arc};
+use kernel::{
+    auxiliary, bindings, c_str, device::Core, pci, pci::Class, pci::ClassMask,
prelude::*,
+    sizes::SZ_16M, sync::Arc,
+};
 
 use crate::gpu::Gpu;
 
@@ -18,10 +21,25 @@ pub(crate) struct NovaCore {
     PCI_TABLE,
     MODULE_PCI_TABLE,
     <NovaCore as pci::Driver>::IdInfo,
-    [(
-        pci::DeviceId::from_id(bindings::PCI_VENDOR_ID_NVIDIA,
bindings::PCI_ANY_ID as u32),
-        ()
-    )]
+    [
+        // Modern NVIDIA GPUs will show up as either VGA or 3D controllers.
+        (
+            pci::DeviceId::from_class_and_vendor(
+                Class::DISPLAY_VGA,
+                ClassMask::ClassSubclass,
+                bindings::PCI_VENDOR_ID_NVIDIA
+            ),
+            ()
+        ),
+        (
+            pci::DeviceId::from_class_and_vendor(
+                Class::DISPLAY_3D,
+                ClassMask::ClassSubclass,
+                bindings::PCI_VENDOR_ID_NVIDIA
+            ),
+            ()
+        ),
+    ]
 );
 
 impl pci::Driver for NovaCore {
-- 
2.50.1

On Wed Aug 20, 2025 at 12:08 PM JST, John Hubbard wrote:
> NovaCore has so far been too imprecise about figuring out if .probe()
> has found a supported PCI PF (Physical Function). By that I mean:
> .probe() sets up BAR0 (which involves a lot of very careful devres and
> Device<Bound> details behind the scenes). And then if it is dealing
with
> a non-supported device such as the .1 audio PF on many GPUs, it fails
> out due to an unexpected BAR0 size. We have been fortunate that the BAR0
> sizes are different.
>
> Really, we should be filtering on PCI class ID instead. These days I
> think we can confidently pick out Nova's supported PF's via PCI
class
> ID. And if not, then we'll revisit.
>
> The approach here is to filter on "Display VGA" or "Display
3D", which
> is how PCI class IDs express "this is a modern GPU's PF".
>
> Cc: Danilo Krummrich <dakr at kernel.org>
> Cc: Alexandre Courbot <acourbot at nvidia.com>
> Signed-off-by: John Hubbard <jhubbard at nvidia.com>
> ---
>  drivers/gpu/nova-core/driver.rs | 28 +++++++++++++++++++++++-----
>  1 file changed, 23 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/nova-core/driver.rs
b/drivers/gpu/nova-core/driver.rs
> index 274989ea1fb4..b60c9defa9d1 100644
> --- a/drivers/gpu/nova-core/driver.rs
> +++ b/drivers/gpu/nova-core/driver.rs
> @@ -1,6 +1,9 @@
>  // SPDX-License-Identifier: GPL-2.0
>  
> -use kernel::{auxiliary, bindings, c_str, device::Core, pci, prelude::*,
sizes::SZ_16M, sync::Arc};
> +use kernel::{
> +    auxiliary, bindings, c_str, device::Core, pci, pci::Class,
pci::ClassMask, prelude::*,
> +    sizes::SZ_16M, sync::Arc,
> +};
>  
>  use crate::gpu::Gpu;
>  
> @@ -18,10 +21,25 @@ pub(crate) struct NovaCore {
>      PCI_TABLE,
>      MODULE_PCI_TABLE,
>      <NovaCore as pci::Driver>::IdInfo,
> -    [(
> -        pci::DeviceId::from_id(bindings::PCI_VENDOR_ID_NVIDIA,
bindings::PCI_ANY_ID as u32),
> -        ()
> -    )]
> +    [
> +        // Modern NVIDIA GPUs will show up as either VGA or 3D
controllers.
> +        (
> +            pci::DeviceId::from_class_and_vendor(
> +                Class::DISPLAY_VGA,
> +                ClassMask::ClassSubclass,
> +                bindings::PCI_VENDOR_ID_NVIDIA
> +            ),
> +            ()
> +        ),
> +        (
> +            pci::DeviceId::from_class_and_vendor(
> +                Class::DISPLAY_3D,
> +                ClassMask::ClassSubclass,
> +                bindings::PCI_VENDOR_ID_NVIDIA
> +            ),
This is making use of `from_class_and_vendor`, which is modified in the
next patch, requiring to modify this part of the file again. How about
switching this patch with 3/3 so we only modify the nova-core code once?

I also wonder if we want to merge 1/3 and (the current) 3/3, since 1/3
alone leaves `from_class_and_vendor` into some intermediate state that
nobody will ever get a chance to use anyway, and one doesn't really make
sense without the other. WDYT?