Message-Id: <DFUE9X1HQG9S.2YDWQ6GDLN36G@garyguo.net>
Date: Wed, 21 Jan 2026 16:06:37 +0000
From: "Gary Guo" <gary@...yguo.net>
To: "John Hubbard" <jhubbard@...dia.com>, "Danilo Krummrich"
 <dakr@...nel.org>
Cc: "Alexandre Courbot" <acourbot@...dia.com>, "Joel Fernandes"
 <joelagnelf@...dia.com>, "Timur Tabi" <ttabi@...dia.com>, "Alistair Popple"
 <apopple@...dia.com>, "Edwin Peer" <epeer@...dia.com>, "Zhi Wang"
 <zhiw@...dia.com>, "David Airlie" <airlied@...il.com>, "Simona Vetter"
 <simona@...ll.ch>, "Bjorn Helgaas" <bhelgaas@...gle.com>, "Miguel Ojeda"
 <ojeda@...nel.org>, "Alex Gaynor" <alex.gaynor@...il.com>, "Boqun Feng"
 <boqun.feng@...il.com>, "Gary Guo" <gary@...yguo.net>,
 Björn Roy Baron <bjorn3_gh@...tonmail.com>, "Benno Lossin"
 <lossin@...nel.org>, "Andreas Hindborg" <a.hindborg@...nel.org>, "Alice
 Ryhl" <aliceryhl@...gle.com>, "Trevor Gross" <tmgross@...ch.edu>,
 <nouveau@...ts.freedesktop.org>, <rust-for-linux@...r.kernel.org>, "LKML"
 <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 15/31] gpu: nova-core: Hopper/Blackwell: add FSP falcon
 EMEM operations

On Wed Dec 3, 2025 at 5:59 AM GMT, John Hubbard wrote:
> Add external memory (EMEM) read/write operations to the GPU's FSP falcon
> engine. These operations use Falcon PIO (Programmed I/O) to communicate
> with the FSP through indirect memory access.
>
> Signed-off-by: John Hubbard <jhubbard@...dia.com>
> ---
>  drivers/gpu/nova-core/falcon/fsp.rs | 60 ++++++++++++++++++++++++++++-
>  drivers/gpu/nova-core/regs.rs       | 10 +++++
>  2 files changed, 69 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/nova-core/falcon/fsp.rs b/drivers/gpu/nova-core/falcon/fsp.rs
> index 7323ae2f2302..9e796e82e556 100644
> --- a/drivers/gpu/nova-core/falcon/fsp.rs
> +++ b/drivers/gpu/nova-core/falcon/fsp.rs
> @@ -5,15 +5,27 @@
>  //! The FSP falcon handles secure boot and Chain of Trust operations
>  //! on Hopper and Blackwell architectures, replacing SEC2's role.
>  
> +use kernel::prelude::*;
> +
>  use crate::{
> +    driver::Bar0,
>      falcon::{
> +        Falcon,
>          FalconEngine,
>          PFalcon2Base,
>          PFalconBase, //
>      },
> -    regs::macros::RegisterBase,
> +    regs::{
> +        self,
> +        macros::RegisterBase, //
> +    },
>  };
>  
> +/// EMEM control register bit 24: write mode.
> +const EMEM_CTL_WRITE: u32 = 1 << 24;
> +/// EMEM control register bit 25: read mode.
> +const EMEM_CTL_READ: u32 = 1 << 25;
> +
>  /// Type specifying the `Fsp` falcon engine. Cannot be instantiated.
>  #[allow(dead_code)]
>  pub(crate) struct Fsp(());
> @@ -30,3 +42,49 @@ impl RegisterBase<PFalcon2Base> for Fsp {
>  impl FalconEngine for Fsp {
>      const ID: Self = Fsp(());
>  }
> +
> +impl Falcon<Fsp> {
> +    /// Writes `data` to FSP external memory at byte `offset` using Falcon PIO.
> +    ///
> +    /// Returns `EINVAL` if offset or data length is not 4-byte aligned.
> +    #[allow(dead_code)]
> +    pub(crate) fn write_emem(&self, bar: &Bar0, offset: u32, data: &[u8]) -> Result {
> +        if offset % 4 != 0 || data.len() % 4 != 0 {

I was about to suggest `is_multiple_of`, but then realize that it's only
available in Rust 1.82+...

> +            return Err(EINVAL);
> +        }
> +
> +        regs::NV_PFALCON_FALCON_EMEM_CTL::default()
> +            .set_value(EMEM_CTL_WRITE | offset)
> +            .write(bar, &Fsp::ID);
> +
> +        for chunk in data.chunks_exact(4) {
> +            let word = u32::from_le_bytes([chunk[0], chunk[1], chunk[2], chunk[3]]);

Use `as_chunks()` can save you from doing this.

Also, a typical pattern is `chunk.try_into().unwrap()` -- yes it has an unwrap
as `[T]` -> `[T; N]` can fail in general, but not with `chunks_exact`.

Best,
Gary

> +            regs::NV_PFALCON_FALCON_EMEM_DATA::default()
> +                .set_data(word)
> +                .write(bar, &Fsp::ID);
> +        }
> +
> +        Ok(())
> +    }
> +
> +    /// Reads FSP external memory at byte `offset` into `data` using Falcon PIO.
> +    ///
> +    /// Returns `EINVAL` if offset or data length is not 4-byte aligned.
> +    #[allow(dead_code)]
> +    pub(crate) fn read_emem(&self, bar: &Bar0, offset: u32, data: &mut [u8]) -> Result {
> +        if offset % 4 != 0 || data.len() % 4 != 0 {
> +            return Err(EINVAL);
> +        }
> +
> +        regs::NV_PFALCON_FALCON_EMEM_CTL::default()
> +            .set_value(EMEM_CTL_READ | offset)
> +            .write(bar, &Fsp::ID);
> +
> +        for chunk in data.chunks_exact_mut(4) {
> +            let word = regs::NV_PFALCON_FALCON_EMEM_DATA::read(bar, &Fsp::ID).data();
> +            chunk.copy_from_slice(&word.to_le_bytes());
> +        }
> +
> +        Ok(())
> +    }
> +}
> diff --git a/drivers/gpu/nova-core/regs.rs b/drivers/gpu/nova-core/regs.rs
> index 82cc6c0790e5..b642cee9611d 100644
> --- a/drivers/gpu/nova-core/regs.rs
> +++ b/drivers/gpu/nova-core/regs.rs
> @@ -391,6 +391,16 @@ pub(crate) fn mem_scrubbing_done(self) -> bool {
>      8:8     br_fetch as bool;
>  });
>  
> +// GP102 EMEM PIO registers (used by FSP for Hopper/Blackwell)
> +// These registers provide falcon external memory communication interface
> +register!(NV_PFALCON_FALCON_EMEM_CTL @ PFalconBase[0x00000ac0] {
> +    31:0    value as u32;       // EMEM control register
> +});
> +
> +register!(NV_PFALCON_FALCON_EMEM_DATA @ PFalconBase[0x00000ac4] {
> +    31:0    data as u32;        // EMEM data register
> +});
> +
>  // The modules below provide registers that are not identical on all supported chips. They should
>  // only be used in HAL modules.
>  

Powered by blists - more mailing lists