From: "Gary Guo" <gary@garyguo.net>
To: "John Hubbard" <jhubbard@nvidia.com>,
"Danilo Krummrich" <dakr@kernel.org>
Cc: "Alexandre Courbot" <acourbot@nvidia.com>,
"Joel Fernandes" <joelagnelf@nvidia.com>,
"Alistair Popple" <apopple@nvidia.com>,
"Zhi Wang" <zhiw@nvidia.com>, "Simona Vetter" <simona@ffwll.ch>,
"Bjorn Helgaas" <bhelgaas@google.com>,
"Miguel Ojeda" <ojeda@kernel.org>,
"Alex Gaynor" <alex.gaynor@gmail.com>,
"Boqun Feng" <boqun.feng@gmail.com>,
"Gary Guo" <gary@garyguo.net>,
"Björn Roy Baron" <bjorn3_gh@protonmail.com>,
"Benno Lossin" <lossin@kernel.org>,
"Andreas Hindborg" <a.hindborg@kernel.org>,
"Alice Ryhl" <aliceryhl@google.com>,
"Trevor Gross" <tmgross@umich.edu>,
nouveau@lists.freedesktop.org, rust-for-linux@vger.kernel.org,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 29/31] gpu: nova-core: Hopper/Blackwell: add FSP Chain of Trust boot path
Date: Wed, 21 Jan 2026 16:35:53 +0000 [thread overview]
Message-ID: <DFUEWBKOLFQ1.3NSTFO4L97503@garyguo.net> (raw)
In-Reply-To: <20251203055923.1247681-30-jhubbard@nvidia.com>
On Wed Dec 3, 2025 at 5:59 AM GMT, John Hubbard wrote:
> Add the FSP-based boot path for Hopper and Blackwell GPUs. Unlike
> Turing/Ampere/Ada which use SEC2 to load the booter firmware, Hopper
> and Blackwell use FSP (Firmware System Processor) with FMC firmware
> to establish a Chain of Trust and boot GSP directly.
>
> The boot() function now dispatches to either run_booter() (SEC2 path)
> or run_fsp() (FSP path) based on the GPU architecture. The cmdq
> commands are moved to after GSP boot, and the GSP sequencer is only
> run for SEC2-based architectures.
>
> Signed-off-by: John Hubbard <jhubbard@nvidia.com>
> ---
> drivers/gpu/nova-core/firmware/fsp.rs | 6 +-
> drivers/gpu/nova-core/fsp.rs | 6 +-
> drivers/gpu/nova-core/gsp/boot.rs | 159 ++++++++++++++++++++------
> 3 files changed, 126 insertions(+), 45 deletions(-)
>
> diff --git a/drivers/gpu/nova-core/firmware/fsp.rs b/drivers/gpu/nova-core/firmware/fsp.rs
> index 80401b964488..d88c7a91e2bc 100644
> --- a/drivers/gpu/nova-core/firmware/fsp.rs
> +++ b/drivers/gpu/nova-core/firmware/fsp.rs
> @@ -13,16 +13,14 @@
> gpu::Chipset, //
> };
>
> -#[expect(unused)]
> pub(crate) struct FspFirmware {
> /// FMC firmware image data (only the .image section)
> - fmc_image: DmaObject,
> + pub(crate) fmc_image: DmaObject,
> /// Full FMC ELF data (for signature extraction)
> - fmc_full: DmaObject,
> + pub(crate) fmc_full: DmaObject,
> }
>
> impl FspFirmware {
> - #[expect(unused)]
> pub(crate) fn new(
> dev: &device::Device<device::Bound>,
> chipset: Chipset,
> diff --git a/drivers/gpu/nova-core/fsp.rs b/drivers/gpu/nova-core/fsp.rs
> index 7d46fbcc7abd..9c11ceb6ab4d 100644
> --- a/drivers/gpu/nova-core/fsp.rs
> +++ b/drivers/gpu/nova-core/fsp.rs
> @@ -1,8 +1,5 @@
> // SPDX-License-Identifier: GPL-2.0
>
> -// TODO: remove this once the code is fully functional
> -#![expect(dead_code)]
> -
> //! FSP (Firmware System Processor) interface for Hopper/Blackwell GPUs.
> //!
> //! Hopper/Blackwell use a simplified firmware boot sequence: FMC --> FSP --> GSP.
> @@ -11,6 +8,7 @@
>
> use kernel::{
> device,
> + dma::CoherentAllocation,
> io::poll::read_poll_timeout,
> prelude::*,
> ptr::{
> @@ -381,8 +379,6 @@ pub(crate) fn create_fmc_boot_params(
> wpr_meta_size: u32,
> libos_addr: u64,
> ) -> Result<kernel::dma::CoherentAllocation<GspFmcBootParams>> {
> - use kernel::dma::CoherentAllocation;
> -
> const GSP_DMA_TARGET_COHERENT_SYSTEM: u32 = 1;
> const GSP_DMA_TARGET_NONCOHERENT_SYSTEM: u32 = 2;
>
> diff --git a/drivers/gpu/nova-core/gsp/boot.rs b/drivers/gpu/nova-core/gsp/boot.rs
> index 4d04135a700e..0fbaa73eb55c 100644
> --- a/drivers/gpu/nova-core/gsp/boot.rs
> +++ b/drivers/gpu/nova-core/gsp/boot.rs
> @@ -13,6 +13,7 @@
> use crate::{
> driver::Bar0,
> falcon::{
> + fsp::Fsp as FspEngine,
> gsp::Gsp,
> sec2::Sec2,
> Falcon,
> @@ -24,6 +25,7 @@
> BooterFirmware,
> BooterKind, //
> },
> + fsp::FspFirmware,
> fwsec::{
> FwsecCommand,
> FwsecFirmware, //
> @@ -31,9 +33,11 @@
> gsp::GspFirmware,
> FIRMWARE_VERSION, //
> },
> - gpu::Chipset,
> + fsp::Fsp,
> + gpu::{Architecture, Chipset},
> gsp::{
> commands,
> + fw::LibosMemoryRegionInitArgument,
> sequencer::{
> GspSequencer,
> GspSequencerParams, //
> @@ -155,6 +159,59 @@ fn run_booter(
> Ok(())
> }
>
> + fn run_fsp(
> + dev: &device::Device<device::Bound>,
> + bar: &Bar0,
> + chipset: Chipset,
> + gsp_falcon: &Falcon<Gsp>,
> + wpr_meta: &CoherentAllocation<GspFwWprMeta>,
> + libos: &CoherentAllocation<LibosMemoryRegionInitArgument>,
> + fb_layout: &FbLayout,
> + ) -> Result {
> + let fsp_falcon = Falcon::<FspEngine>::new(dev, chipset)?;
> +
> + Fsp::wait_secure_boot(dev, bar, chipset.arch())?;
> +
> + let fsp_fw = FspFirmware::new(dev, chipset, FIRMWARE_VERSION)?;
> +
> + // SAFETY: fmc_full is a valid DmaObject with a contiguous allocation of size() bytes
> + // starting at start_ptr(). The slice is only used for signature extraction within this
> + // function scope while fsp_fw remains valid.
> + let fmc_full_data = unsafe {
> + core::slice::from_raw_parts(fsp_fw.fmc_full.start_ptr(), fsp_fw.fmc_full.size())
> + };
The justification is week because it does not mention about the non-race nature
of this, which need to be justified for a DMA allocation. If you use
`CoherentAllocation::as_slice`, then this requirement would be obvious.
For example:
// SAFETY: the dma buffer is not yet submitted too hardware and we are the
// unique owner at this point.
let fmc_full_data = unsafe { fsp_fw.fmc_full.as_slice(0, fsp_fw.fmc_full.size()) };
Best,
Gary
> + let signatures = Fsp::extract_fmc_signatures_static(dev, fmc_full_data)?;
> +
> + // Create FMC boot parameters
> + let fmc_boot_params = Fsp::create_fmc_boot_params(
> + dev,
> + wpr_meta.dma_handle(),
> + core::mem::size_of::<GspFwWprMeta>() as u32,
> + libos.dma_handle(),
> + )?;
> +
> + // Execute FSP Chain of Trust
> + // NOTE: FSP Chain of Trust handles GSP boot internally - we do NOT reset or boot GSP
> + Fsp::boot_gsp_fmc_with_signatures(
> + dev,
> + bar,
> + chipset,
> + &fsp_fw.fmc_image,
> + &fmc_boot_params,
> + u64::from(fb_layout.total_reserved_size),
> + false, // not resuming
> + &fsp_falcon,
> + &signatures,
> + )?;
> +
> + // Wait for GSP lockdown to be released
> + let fmc_boot_params_addr = fmc_boot_params.dma_handle();
> + let _mbox0 =
> + Self::wait_for_gsp_lockdown_release(dev, bar, gsp_falcon, fmc_boot_params_addr)?;
> +
> + Ok(())
> + }
> +
> /// Check if GSP lockdown has been released after FSP Chain of Trust
> fn gsp_lockdown_released(
> dev: &device::Device,
> @@ -192,7 +249,6 @@ fn gsp_lockdown_released(
> }
>
> /// Wait for GSP lockdown to be released after FSP Chain of Trust
> - #[expect(dead_code)]
> fn wait_for_gsp_lockdown_release(
> dev: &device::Device,
> bar: &Bar0,
> @@ -255,8 +311,6 @@ pub(crate) fn boot(
> ) -> Result {
> let dev = pdev.as_ref();
>
> - let bios = Vbios::new(dev, bar)?;
> -
> let gsp_fw = KBox::pin_init(
> GspFirmware::new(dev, chipset, FIRMWARE_VERSION)?,
> GFP_KERNEL,
> @@ -265,36 +319,58 @@ pub(crate) fn boot(
> let fb_layout = FbLayout::new(chipset, bar, &gsp_fw)?;
> dev_dbg!(dev, "{:#x?}\n", fb_layout);
>
> - Self::run_fwsec_frts(dev, gsp_falcon, bar, &bios, &fb_layout)?;
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + let bios = Vbios::new(dev, bar)?;
> + Self::run_fwsec_frts(dev, gsp_falcon, bar, &bios, &fb_layout)?;
> + }
>
> let wpr_meta =
> CoherentAllocation::<GspFwWprMeta>::alloc_coherent(dev, 1, GFP_KERNEL | __GFP_ZERO)?;
> dma_write!(wpr_meta[0] = GspFwWprMeta::new(&gsp_fw, &fb_layout))?;
>
> - self.cmdq
> - .send_command(bar, commands::SetSystemInfo::new(pdev))?;
> - self.cmdq.send_command(bar, commands::SetRegistry::new())?;
> + // For SEC2-based architectures, reset GSP and boot it before SEC2
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + gsp_falcon.reset(bar)?;
> + let libos_handle = self.libos.dma_handle();
> + let (mbox0, mbox1) = gsp_falcon.boot(
> + bar,
> + Some(libos_handle as u32),
> + Some((libos_handle >> 32) as u32),
> + )?;
> + dev_dbg!(
> + pdev.as_ref(),
> + "GSP MBOX0: {:#x}, MBOX1: {:#x}\n",
> + mbox0,
> + mbox1
> + );
>
> - gsp_falcon.reset(bar)?;
> - let libos_handle = self.libos.dma_handle();
> - let (mbox0, mbox1) = gsp_falcon.boot(
> - bar,
> - Some(libos_handle as u32),
> - Some((libos_handle >> 32) as u32),
> - )?;
> - dev_dbg!(
> - pdev.as_ref(),
> - "GSP MBOX0: {:#x}, MBOX1: {:#x}\n",
> - mbox0,
> - mbox1
> - );
> + dev_dbg!(
> + pdev.as_ref(),
> + "Using SEC2 to load and run the booter_load firmware...\n"
> + );
> + }
>
> - dev_dbg!(
> - pdev.as_ref(),
> - "Using SEC2 to load and run the booter_load firmware...\n"
> - );
> + match chipset.arch() {
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada => {
> + Self::run_booter(dev, bar, chipset, sec2_falcon, &wpr_meta)?
> + }
>
> - Self::run_booter(dev, bar, chipset, sec2_falcon, &wpr_meta)?;
> + Architecture::Hopper | Architecture::Blackwell => Self::run_fsp(
> + dev,
> + bar,
> + chipset,
> + gsp_falcon,
> + &wpr_meta,
> + &self.libos,
> + &fb_layout,
> + )?,
> + }
>
> gsp_falcon.write_os_version(bar, gsp_fw.bootloader.app_version);
>
> @@ -312,16 +388,27 @@ pub(crate) fn boot(
> gsp_falcon.is_riscv_active(bar),
> );
>
> - // Create and run the GSP sequencer.
> - let seq_params = GspSequencerParams {
> - bootloader_app_version: gsp_fw.bootloader.app_version,
> - libos_dma_handle: libos_handle,
> - gsp_falcon,
> - sec2_falcon,
> - dev: pdev.as_ref().into(),
> - bar,
> - };
> - GspSequencer::run(&mut self.cmdq, seq_params)?;
> + // Now that GSP is active, send system info and registry
> + self.cmdq
> + .send_command(bar, commands::SetSystemInfo::new(pdev))?;
> + self.cmdq.send_command(bar, commands::SetRegistry::new())?;
> +
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + let libos_handle = self.libos.dma_handle();
> + // Create and run the GSP sequencer.
> + let seq_params = GspSequencerParams {
> + bootloader_app_version: gsp_fw.bootloader.app_version,
> + libos_dma_handle: libos_handle,
> + gsp_falcon,
> + sec2_falcon,
> + dev: pdev.as_ref().into(),
> + bar,
> + };
> + GspSequencer::run(&mut self.cmdq, seq_params)?;
> + }
>
> // Wait until GSP is fully initialized.
> commands::wait_gsp_init_done(&mut self.cmdq)?;
WARNING: multiple messages have this Message-ID (diff)
From: "Gary Guo" <gary@garyguo.net>
To: "John Hubbard" <jhubbard@nvidia.com>,
"Danilo Krummrich" <dakr@kernel.org>
Cc: "Alexandre Courbot" <acourbot@nvidia.com>,
"Joel Fernandes" <joelagnelf@nvidia.com>,
"Timur Tabi" <ttabi@nvidia.com>,
"Alistair Popple" <apopple@nvidia.com>,
"Edwin Peer" <epeer@nvidia.com>, "Zhi Wang" <zhiw@nvidia.com>,
"David Airlie" <airlied@gmail.com>,
"Simona Vetter" <simona@ffwll.ch>,
"Bjorn Helgaas" <bhelgaas@google.com>,
"Miguel Ojeda" <ojeda@kernel.org>,
"Alex Gaynor" <alex.gaynor@gmail.com>,
"Boqun Feng" <boqun.feng@gmail.com>,
"Gary Guo" <gary@garyguo.net>,
"Björn Roy Baron" <bjorn3_gh@protonmail.com>,
"Benno Lossin" <lossin@kernel.org>,
"Andreas Hindborg" <a.hindborg@kernel.org>,
"Alice Ryhl" <aliceryhl@google.com>,
"Trevor Gross" <tmgross@umich.edu>,
nouveau@lists.freedesktop.org, rust-for-linux@vger.kernel.org,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 29/31] gpu: nova-core: Hopper/Blackwell: add FSP Chain of Trust boot path
Date: Wed, 21 Jan 2026 16:35:53 +0000 [thread overview]
Message-ID: <DFUEWBKOLFQ1.3NSTFO4L97503@garyguo.net> (raw)
In-Reply-To: <20251203055923.1247681-30-jhubbard@nvidia.com>
On Wed Dec 3, 2025 at 5:59 AM GMT, John Hubbard wrote:
> Add the FSP-based boot path for Hopper and Blackwell GPUs. Unlike
> Turing/Ampere/Ada which use SEC2 to load the booter firmware, Hopper
> and Blackwell use FSP (Firmware System Processor) with FMC firmware
> to establish a Chain of Trust and boot GSP directly.
>
> The boot() function now dispatches to either run_booter() (SEC2 path)
> or run_fsp() (FSP path) based on the GPU architecture. The cmdq
> commands are moved to after GSP boot, and the GSP sequencer is only
> run for SEC2-based architectures.
>
> Signed-off-by: John Hubbard <jhubbard@nvidia.com>
> ---
> drivers/gpu/nova-core/firmware/fsp.rs | 6 +-
> drivers/gpu/nova-core/fsp.rs | 6 +-
> drivers/gpu/nova-core/gsp/boot.rs | 159 ++++++++++++++++++++------
> 3 files changed, 126 insertions(+), 45 deletions(-)
>
> diff --git a/drivers/gpu/nova-core/firmware/fsp.rs b/drivers/gpu/nova-core/firmware/fsp.rs
> index 80401b964488..d88c7a91e2bc 100644
> --- a/drivers/gpu/nova-core/firmware/fsp.rs
> +++ b/drivers/gpu/nova-core/firmware/fsp.rs
> @@ -13,16 +13,14 @@
> gpu::Chipset, //
> };
>
> -#[expect(unused)]
> pub(crate) struct FspFirmware {
> /// FMC firmware image data (only the .image section)
> - fmc_image: DmaObject,
> + pub(crate) fmc_image: DmaObject,
> /// Full FMC ELF data (for signature extraction)
> - fmc_full: DmaObject,
> + pub(crate) fmc_full: DmaObject,
> }
>
> impl FspFirmware {
> - #[expect(unused)]
> pub(crate) fn new(
> dev: &device::Device<device::Bound>,
> chipset: Chipset,
> diff --git a/drivers/gpu/nova-core/fsp.rs b/drivers/gpu/nova-core/fsp.rs
> index 7d46fbcc7abd..9c11ceb6ab4d 100644
> --- a/drivers/gpu/nova-core/fsp.rs
> +++ b/drivers/gpu/nova-core/fsp.rs
> @@ -1,8 +1,5 @@
> // SPDX-License-Identifier: GPL-2.0
>
> -// TODO: remove this once the code is fully functional
> -#![expect(dead_code)]
> -
> //! FSP (Firmware System Processor) interface for Hopper/Blackwell GPUs.
> //!
> //! Hopper/Blackwell use a simplified firmware boot sequence: FMC --> FSP --> GSP.
> @@ -11,6 +8,7 @@
>
> use kernel::{
> device,
> + dma::CoherentAllocation,
> io::poll::read_poll_timeout,
> prelude::*,
> ptr::{
> @@ -381,8 +379,6 @@ pub(crate) fn create_fmc_boot_params(
> wpr_meta_size: u32,
> libos_addr: u64,
> ) -> Result<kernel::dma::CoherentAllocation<GspFmcBootParams>> {
> - use kernel::dma::CoherentAllocation;
> -
> const GSP_DMA_TARGET_COHERENT_SYSTEM: u32 = 1;
> const GSP_DMA_TARGET_NONCOHERENT_SYSTEM: u32 = 2;
>
> diff --git a/drivers/gpu/nova-core/gsp/boot.rs b/drivers/gpu/nova-core/gsp/boot.rs
> index 4d04135a700e..0fbaa73eb55c 100644
> --- a/drivers/gpu/nova-core/gsp/boot.rs
> +++ b/drivers/gpu/nova-core/gsp/boot.rs
> @@ -13,6 +13,7 @@
> use crate::{
> driver::Bar0,
> falcon::{
> + fsp::Fsp as FspEngine,
> gsp::Gsp,
> sec2::Sec2,
> Falcon,
> @@ -24,6 +25,7 @@
> BooterFirmware,
> BooterKind, //
> },
> + fsp::FspFirmware,
> fwsec::{
> FwsecCommand,
> FwsecFirmware, //
> @@ -31,9 +33,11 @@
> gsp::GspFirmware,
> FIRMWARE_VERSION, //
> },
> - gpu::Chipset,
> + fsp::Fsp,
> + gpu::{Architecture, Chipset},
> gsp::{
> commands,
> + fw::LibosMemoryRegionInitArgument,
> sequencer::{
> GspSequencer,
> GspSequencerParams, //
> @@ -155,6 +159,59 @@ fn run_booter(
> Ok(())
> }
>
> + fn run_fsp(
> + dev: &device::Device<device::Bound>,
> + bar: &Bar0,
> + chipset: Chipset,
> + gsp_falcon: &Falcon<Gsp>,
> + wpr_meta: &CoherentAllocation<GspFwWprMeta>,
> + libos: &CoherentAllocation<LibosMemoryRegionInitArgument>,
> + fb_layout: &FbLayout,
> + ) -> Result {
> + let fsp_falcon = Falcon::<FspEngine>::new(dev, chipset)?;
> +
> + Fsp::wait_secure_boot(dev, bar, chipset.arch())?;
> +
> + let fsp_fw = FspFirmware::new(dev, chipset, FIRMWARE_VERSION)?;
> +
> + // SAFETY: fmc_full is a valid DmaObject with a contiguous allocation of size() bytes
> + // starting at start_ptr(). The slice is only used for signature extraction within this
> + // function scope while fsp_fw remains valid.
> + let fmc_full_data = unsafe {
> + core::slice::from_raw_parts(fsp_fw.fmc_full.start_ptr(), fsp_fw.fmc_full.size())
> + };
The justification is week because it does not mention about the non-race nature
of this, which need to be justified for a DMA allocation. If you use
`CoherentAllocation::as_slice`, then this requirement would be obvious.
For example:
// SAFETY: the dma buffer is not yet submitted too hardware and we are the
// unique owner at this point.
let fmc_full_data = unsafe { fsp_fw.fmc_full.as_slice(0, fsp_fw.fmc_full.size()) };
Best,
Gary
> + let signatures = Fsp::extract_fmc_signatures_static(dev, fmc_full_data)?;
> +
> + // Create FMC boot parameters
> + let fmc_boot_params = Fsp::create_fmc_boot_params(
> + dev,
> + wpr_meta.dma_handle(),
> + core::mem::size_of::<GspFwWprMeta>() as u32,
> + libos.dma_handle(),
> + )?;
> +
> + // Execute FSP Chain of Trust
> + // NOTE: FSP Chain of Trust handles GSP boot internally - we do NOT reset or boot GSP
> + Fsp::boot_gsp_fmc_with_signatures(
> + dev,
> + bar,
> + chipset,
> + &fsp_fw.fmc_image,
> + &fmc_boot_params,
> + u64::from(fb_layout.total_reserved_size),
> + false, // not resuming
> + &fsp_falcon,
> + &signatures,
> + )?;
> +
> + // Wait for GSP lockdown to be released
> + let fmc_boot_params_addr = fmc_boot_params.dma_handle();
> + let _mbox0 =
> + Self::wait_for_gsp_lockdown_release(dev, bar, gsp_falcon, fmc_boot_params_addr)?;
> +
> + Ok(())
> + }
> +
> /// Check if GSP lockdown has been released after FSP Chain of Trust
> fn gsp_lockdown_released(
> dev: &device::Device,
> @@ -192,7 +249,6 @@ fn gsp_lockdown_released(
> }
>
> /// Wait for GSP lockdown to be released after FSP Chain of Trust
> - #[expect(dead_code)]
> fn wait_for_gsp_lockdown_release(
> dev: &device::Device,
> bar: &Bar0,
> @@ -255,8 +311,6 @@ pub(crate) fn boot(
> ) -> Result {
> let dev = pdev.as_ref();
>
> - let bios = Vbios::new(dev, bar)?;
> -
> let gsp_fw = KBox::pin_init(
> GspFirmware::new(dev, chipset, FIRMWARE_VERSION)?,
> GFP_KERNEL,
> @@ -265,36 +319,58 @@ pub(crate) fn boot(
> let fb_layout = FbLayout::new(chipset, bar, &gsp_fw)?;
> dev_dbg!(dev, "{:#x?}\n", fb_layout);
>
> - Self::run_fwsec_frts(dev, gsp_falcon, bar, &bios, &fb_layout)?;
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + let bios = Vbios::new(dev, bar)?;
> + Self::run_fwsec_frts(dev, gsp_falcon, bar, &bios, &fb_layout)?;
> + }
>
> let wpr_meta =
> CoherentAllocation::<GspFwWprMeta>::alloc_coherent(dev, 1, GFP_KERNEL | __GFP_ZERO)?;
> dma_write!(wpr_meta[0] = GspFwWprMeta::new(&gsp_fw, &fb_layout))?;
>
> - self.cmdq
> - .send_command(bar, commands::SetSystemInfo::new(pdev))?;
> - self.cmdq.send_command(bar, commands::SetRegistry::new())?;
> + // For SEC2-based architectures, reset GSP and boot it before SEC2
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + gsp_falcon.reset(bar)?;
> + let libos_handle = self.libos.dma_handle();
> + let (mbox0, mbox1) = gsp_falcon.boot(
> + bar,
> + Some(libos_handle as u32),
> + Some((libos_handle >> 32) as u32),
> + )?;
> + dev_dbg!(
> + pdev.as_ref(),
> + "GSP MBOX0: {:#x}, MBOX1: {:#x}\n",
> + mbox0,
> + mbox1
> + );
>
> - gsp_falcon.reset(bar)?;
> - let libos_handle = self.libos.dma_handle();
> - let (mbox0, mbox1) = gsp_falcon.boot(
> - bar,
> - Some(libos_handle as u32),
> - Some((libos_handle >> 32) as u32),
> - )?;
> - dev_dbg!(
> - pdev.as_ref(),
> - "GSP MBOX0: {:#x}, MBOX1: {:#x}\n",
> - mbox0,
> - mbox1
> - );
> + dev_dbg!(
> + pdev.as_ref(),
> + "Using SEC2 to load and run the booter_load firmware...\n"
> + );
> + }
>
> - dev_dbg!(
> - pdev.as_ref(),
> - "Using SEC2 to load and run the booter_load firmware...\n"
> - );
> + match chipset.arch() {
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada => {
> + Self::run_booter(dev, bar, chipset, sec2_falcon, &wpr_meta)?
> + }
>
> - Self::run_booter(dev, bar, chipset, sec2_falcon, &wpr_meta)?;
> + Architecture::Hopper | Architecture::Blackwell => Self::run_fsp(
> + dev,
> + bar,
> + chipset,
> + gsp_falcon,
> + &wpr_meta,
> + &self.libos,
> + &fb_layout,
> + )?,
> + }
>
> gsp_falcon.write_os_version(bar, gsp_fw.bootloader.app_version);
>
> @@ -312,16 +388,27 @@ pub(crate) fn boot(
> gsp_falcon.is_riscv_active(bar),
> );
>
> - // Create and run the GSP sequencer.
> - let seq_params = GspSequencerParams {
> - bootloader_app_version: gsp_fw.bootloader.app_version,
> - libos_dma_handle: libos_handle,
> - gsp_falcon,
> - sec2_falcon,
> - dev: pdev.as_ref().into(),
> - bar,
> - };
> - GspSequencer::run(&mut self.cmdq, seq_params)?;
> + // Now that GSP is active, send system info and registry
> + self.cmdq
> + .send_command(bar, commands::SetSystemInfo::new(pdev))?;
> + self.cmdq.send_command(bar, commands::SetRegistry::new())?;
> +
> + if matches!(
> + chipset.arch(),
> + Architecture::Turing | Architecture::Ampere | Architecture::Ada
> + ) {
> + let libos_handle = self.libos.dma_handle();
> + // Create and run the GSP sequencer.
> + let seq_params = GspSequencerParams {
> + bootloader_app_version: gsp_fw.bootloader.app_version,
> + libos_dma_handle: libos_handle,
> + gsp_falcon,
> + sec2_falcon,
> + dev: pdev.as_ref().into(),
> + bar,
> + };
> + GspSequencer::run(&mut self.cmdq, seq_params)?;
> + }
>
> // Wait until GSP is fully initialized.
> commands::wait_gsp_init_done(&mut self.cmdq)?;
next prev parent reply other threads:[~2026-01-21 16:36 UTC|newest]
Thread overview: 115+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-03 5:58 [PATCH 00/31] gpu: nova-core: firmware: Hopper/Blackwell support John Hubbard
2025-12-03 5:58 ` [PATCH 01/31] gpu: nova-core: print FB sizes, along with ranges John Hubbard
2025-12-03 19:35 ` Timur Tabi
2025-12-04 7:27 ` John Hubbard
2026-01-13 13:28 ` Gary Guo
2026-01-13 13:28 ` Gary Guo
2026-01-13 13:42 ` Miguel Ojeda
2026-01-13 13:42 ` Miguel Ojeda
2026-01-14 2:23 ` John Hubbard
2026-01-14 2:23 ` John Hubbard
2026-01-23 3:09 ` John Hubbard
2026-01-23 18:04 ` Gary Guo
2025-12-03 5:58 ` [PATCH 02/31] gpu: nova-core: add FbRange.len() and use it in boot.rs John Hubbard
2026-01-13 13:29 ` Gary Guo
2026-01-13 13:29 ` Gary Guo
2025-12-03 5:58 ` [PATCH 03/31] gpu: nova-core: Hopper/Blackwell: basic GPU identification John Hubbard
2025-12-03 5:58 ` [PATCH 04/31] nova-core: factor .fwsignature* selection into a new get_gsp_sigs_section() John Hubbard
2026-01-13 13:33 ` Gary Guo
2026-01-13 13:33 ` Gary Guo
2026-01-14 2:24 ` John Hubbard
2026-01-14 2:24 ` John Hubbard
2025-12-03 5:58 ` [PATCH 05/31] gpu: nova-core: use GPU Architecture to simplify HAL selections John Hubbard
2025-12-03 19:38 ` Timur Tabi
2025-12-04 7:28 ` John Hubbard
2025-12-03 5:58 ` [PATCH 06/31] gpu: nova-core: apply the one "use" item per line policy to commands.rs John Hubbard
2026-01-13 13:35 ` Gary Guo
2026-01-13 13:35 ` Gary Guo
2025-12-03 5:58 ` [PATCH 07/31] gpu: nova-core: set DMA mask width based on GPU architecture John Hubbard
2026-01-13 13:43 ` Gary Guo
2026-01-13 13:43 ` Gary Guo
2026-01-14 3:03 ` John Hubbard
2026-01-14 3:03 ` John Hubbard
2025-12-03 5:59 ` [PATCH 08/31] gpu: nova-core: move firmware image parsing code to firmware.rs John Hubbard
2026-01-13 13:44 ` Gary Guo
2026-01-13 13:44 ` Gary Guo
2025-12-03 5:59 ` [PATCH 09/31] gpu: nova-core: factor out a section_name_eq() function John Hubbard
2026-01-13 13:57 ` Gary Guo
2026-01-13 13:57 ` Gary Guo
2026-01-14 3:18 ` John Hubbard
2026-01-14 3:18 ` John Hubbard
2026-01-14 14:40 ` Gary Guo
2026-01-14 14:40 ` Gary Guo
2025-12-03 5:59 ` [PATCH 10/31] gpu: nova-core: don't assume 64-bit firmware images John Hubbard
2025-12-03 5:59 ` [PATCH 11/31] gpu: nova-core: add support for 32-bit " John Hubbard
2025-12-03 5:59 ` [PATCH 12/31] gpu: nova-core: add auto-detection of 32-bit, 64-bit " John Hubbard
2025-12-03 5:59 ` [PATCH 13/31] gpu: nova-core: Hopper/Blackwell: add FMC firmware image, in support of FSP John Hubbard
2025-12-03 5:59 ` [PATCH 14/31] gpu: nova-core: Hopper/Blackwell: add FSP falcon engine stub John Hubbard
2025-12-03 5:59 ` [PATCH 15/31] gpu: nova-core: Hopper/Blackwell: add FSP falcon EMEM operations John Hubbard
2025-12-03 6:04 ` Timur Tabi
2025-12-03 6:07 ` John Hubbard
2026-01-21 16:06 ` Gary Guo
2026-01-21 16:06 ` Gary Guo
2026-01-21 16:17 ` Miguel Ojeda
2026-01-21 16:17 ` Miguel Ojeda
2026-01-23 23:48 ` John Hubbard
2026-01-23 23:48 ` John Hubbard
2025-12-03 5:59 ` [PATCH 16/31] gpu: nova-core: Hopper/Blackwell: add FSP message infrastructure John Hubbard
2025-12-05 16:47 ` Joel Fernandes
2026-01-03 2:15 ` John Hubbard
2026-01-03 2:15 ` John Hubbard
2025-12-03 5:59 ` [PATCH 17/31] gpu: nova-core: Hopper/Blackwell: calculate reserved FB heap size John Hubbard
2025-12-03 20:48 ` Timur Tabi
2025-12-04 7:34 ` John Hubbard
2026-01-21 16:10 ` Gary Guo
2026-01-21 16:10 ` Gary Guo
2026-01-23 23:56 ` John Hubbard
2026-01-23 23:56 ` John Hubbard
2025-12-03 5:59 ` [PATCH 18/31] gpu: nova-core: Hopper/Blackwell: add needs_large_reserved_mem() John Hubbard
2025-12-03 20:51 ` Timur Tabi
2025-12-04 7:36 ` John Hubbard
2025-12-03 5:59 ` [PATCH 19/31] gpu: nova-core: Hopper/Blackwell: add FSP secure boot completion waiting John Hubbard
2025-12-03 5:59 ` [PATCH 20/31] gpu: nova-core: Hopper/Blackwell: add FSP message structures John Hubbard
2025-12-03 5:59 ` [PATCH 21/31] gpu: nova-core: Hopper/Blackwell: add FMC signature extraction John Hubbard
2025-12-03 15:45 ` Joel Fernandes
2025-12-04 7:55 ` John Hubbard
2026-01-21 16:15 ` Gary Guo
2026-01-21 16:15 ` Gary Guo
2026-01-24 0:45 ` John Hubbard
2026-01-24 0:45 ` John Hubbard
2025-12-03 5:59 ` [PATCH 22/31] gpu: nova-core: Hopper/Blackwell: add FSP send/receive messaging John Hubbard
2025-12-03 5:59 ` [PATCH 23/31] gpu: nova-core: Hopper/Blackwell: add FSP Chain of Trust boot John Hubbard
2025-12-05 17:15 ` Joel Fernandes
2025-12-08 6:00 ` John Hubbard
2025-12-06 21:36 ` Joel Fernandes
2025-12-08 6:09 ` John Hubbard
2025-12-03 5:59 ` [PATCH 24/31] gpu: nova-core: Hopper/Blackwell: larger non-WPR heap John Hubbard
2026-01-21 16:18 ` Gary Guo
2026-01-21 16:18 ` Gary Guo
2026-01-24 1:50 ` John Hubbard
2026-01-24 1:50 ` John Hubbard
2025-12-03 5:59 ` [PATCH 25/31] gpu: nova-core: Hopper/Blackwell: larger WPR2 (GSP) heap John Hubbard
2025-12-03 5:59 ` [PATCH 26/31] gpu: nova-core: refactor SEC2 booter loading into run_booter() helper John Hubbard
2025-12-03 20:53 ` Timur Tabi
2025-12-04 7:37 ` John Hubbard
2025-12-03 5:59 ` [PATCH 27/31] gpu: nova-core: Hopper/Blackwell: skip GFW boot waiting John Hubbard
2025-12-03 5:59 ` [PATCH 28/31] gpu: nova-core: Hopper/Blackwell: add GSP lockdown release polling John Hubbard
2025-12-03 20:59 ` Timur Tabi
2025-12-04 7:49 ` John Hubbard
2026-01-21 16:20 ` Gary Guo
2026-01-21 16:20 ` Gary Guo
2026-01-24 1:10 ` John Hubbard
2026-01-24 1:10 ` John Hubbard
2025-12-03 5:59 ` [PATCH 29/31] gpu: nova-core: Hopper/Blackwell: add FSP Chain of Trust boot path John Hubbard
2026-01-21 16:35 ` Gary Guo [this message]
2026-01-21 16:35 ` Gary Guo
2026-01-24 1:38 ` John Hubbard
2026-01-24 1:38 ` John Hubbard
2026-01-24 1:42 ` John Hubbard
2026-01-24 1:42 ` John Hubbard
2026-01-26 13:08 ` Gary Guo
2026-01-26 13:08 ` Gary Guo
2026-01-26 19:53 ` John Hubbard
2026-01-26 19:53 ` John Hubbard
2025-12-03 5:59 ` [PATCH 30/31] gpu: nova-core: Hopper/Blackwell: new location for PCI config mirror John Hubbard
2025-12-03 5:59 ` [PATCH 31/31] gpu: nova-core: clarify the GPU firmware boot steps John Hubbard
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=DFUEWBKOLFQ1.3NSTFO4L97503@garyguo.net \
--to=gary@garyguo.net \
--cc=a.hindborg@kernel.org \
--cc=acourbot@nvidia.com \
--cc=alex.gaynor@gmail.com \
--cc=aliceryhl@google.com \
--cc=apopple@nvidia.com \
--cc=bhelgaas@google.com \
--cc=bjorn3_gh@protonmail.com \
--cc=boqun.feng@gmail.com \
--cc=dakr@kernel.org \
--cc=jhubbard@nvidia.com \
--cc=joelagnelf@nvidia.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lossin@kernel.org \
--cc=nouveau@lists.freedesktop.org \
--cc=ojeda@kernel.org \
--cc=rust-for-linux@vger.kernel.org \
--cc=simona@ffwll.ch \
--cc=tmgross@umich.edu \
--cc=zhiw@nvidia.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.