Re: [PATCH v6 07/34] gpu: nova-core: move GPU init and DMA mask setup into Gpu::new()

[Alexandre Courbot]

On Tue Mar 10, 2026 at 11:10 AM JST, John Hubbard wrote:
> Move Spec creation, the dev_info log, and DMA mask setup from the
> driver's probe() into Gpu::new(), so that all GPU-specific
> initialization lives in the Gpu constructor.
>
> This restructures Gpu::new() to use pin_init_scope wrapping
> try_pin_init!, which allows running fallible setup code (Spec::new,
> dma_set_mask_and_coherent) before the pin-initializer. The parameter
> type changes from pci::Device<device::Bound> to pci::Device<device::Core>
> because the DMA call requires the Core device state.
>
> Also makes Chipset::arch() const, adds Spec::chipset() accessor, and
> makes Spec::new() pub(crate) for use by later patches.
>
> No functional change: the same 47-bit DMA mask is applied.
>
> Cc: Danilo Krummrich <dakr@xxxxxxxxxx>
> Cc: Gary Guo <gary@xxxxxxxxxxx>
> Signed-off-by: John Hubbard <jhubbard@xxxxxxxxxx>
> ---
>  drivers/gpu/nova-core/driver.rs | 15 --------
>  drivers/gpu/nova-core/gpu.rs    | 64 ++++++++++++++++++++++-----------
>  2 files changed, 43 insertions(+), 36 deletions(-)
>
> diff --git a/drivers/gpu/nova-core/driver.rs b/drivers/gpu/nova-core/driver.rs
> index 84b0e1703150..e07f7122b35c 100644
> --- a/drivers/gpu/nova-core/driver.rs
> +++ b/drivers/gpu/nova-core/driver.rs
> @@ -4,8 +4,6 @@
>      auxiliary,
>      device::Core,
>      devres::Devres,
> -    dma::Device,
> -    dma::DmaMask,
>      pci,
>      pci::{
>          Class,
> @@ -38,14 +36,6 @@ pub(crate) struct NovaCore {
>  
>  const BAR0_SIZE: usize = SZ_16M;
>  
> -// For now we only support Ampere which can use up to 47-bit DMA addresses.
> -//
> -// TODO: Add an abstraction for this to support newer GPUs which may support
> -// larger DMA addresses. Limiting these GPUs to smaller address widths won't
> -// have any adverse affects, unless installed on systems which require larger
> -// DMA addresses. These systems should be quite rare.
> -const GPU_DMA_BITS: u32 = 47;
> -
>  pub(crate) type Bar0 = pci::Bar<BAR0_SIZE>;
>  
>  kernel::pci_device_table!(
> @@ -84,11 +74,6 @@ fn probe(pdev: &pci::Device<Core>, _info: &Self::IdInfo) -> impl PinInit<Self, E
>              pdev.enable_device_mem()?;
>              pdev.set_master();
>  
> -            // SAFETY: No concurrent DMA allocations or mappings can be made because
> -            // the device is still being probed and therefore isn't being used by
> -            // other threads of execution.
> -            unsafe { pdev.dma_set_mask_and_coherent(DmaMask::new::<GPU_DMA_BITS>())? };

I think the reason why this was in driver.rs is that the method we are
in (`probe`) is enough to justify the safety argument.

Whereas `Gpu::new` can technically be called at a different time, and
the safety requirement becomes more difficult to justify - we would
actually need to make `Gpu::new` unsafe and set its own `SAFETY` section
to cover the safety requirement of `dma_set_mask_and_coherent`.

So I think it is preferable to keep this part as-is.

> -
>              let bar = Arc::pin_init(
>                  pdev.iomap_region_sized::<BAR0_SIZE>(0, c"nova-core/bar0"),
>                  GFP_KERNEL,
> diff --git a/drivers/gpu/nova-core/gpu.rs b/drivers/gpu/nova-core/gpu.rs
> index 3b4ccc3d18b9..a7f1957880ff 100644
> --- a/drivers/gpu/nova-core/gpu.rs
> +++ b/drivers/gpu/nova-core/gpu.rs
> @@ -3,6 +3,10 @@
>  use kernel::{
>      device,
>      devres::Devres,
> +    dma::{
> +        Device,
> +        DmaMask, //
> +    },
>      fmt,
>      pci,
>      prelude::*,
> @@ -162,6 +166,10 @@ pub(crate) enum Architecture {
>      Blackwell = 0x1b,
>  }
>  
> +// TODO: Set the DMA mask per-architecture. Hopper and Blackwell support 52-bit
> +// DMA addresses. For now, use 47-bit which is correct for Turing, Ampere, and Ada.
> +const GPU_DMA_BITS: u32 = 47;
> +
>  impl TryFrom<u8> for Architecture {
>      type Error = Error;
>  
> @@ -211,7 +219,7 @@ pub(crate) struct Spec {
>  }
>  
>  impl Spec {
> -    fn new(dev: &device::Device, bar: &Bar0) -> Result<Spec> {
> +    pub(crate) fn new(dev: &device::Device, bar: &Bar0) -> Result<Spec> {

The visibility of this methods doesn't need to be changed with the
current code - although it will per the next patch if we keep the DMA
mask setup in `driver.rs`. If we opt for that direction, let's merge
that part into the next patch.