public inbox for rust-for-linux@vger.kernel.org
 help / color / mirror / Atom feed
From: John Hubbard <jhubbard@nvidia.com>
To: Danilo Krummrich <dakr@kernel.org>,
	Alexandre Courbot <acourbot@nvidia.com>
Cc: "Joel Fernandes" <joelagnelf@nvidia.com>,
	"Timur Tabi" <ttabi@nvidia.com>,
	"Alistair Popple" <apopple@nvidia.com>,
	"Eliot Courtney" <ecourtney@nvidia.com>,
	"Shashank Sharma" <shashanks@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>,
	rust-for-linux@vger.kernel.org,
	LKML <linux-kernel@vger.kernel.org>,
	"John Hubbard" <jhubbard@nvidia.com>
Subject: [PATCH v9 27/31] gpu: nova-core: Hopper/Blackwell: larger WPR2 (GSP) heap
Date: Wed, 25 Mar 2026 18:38:58 -0700	[thread overview]
Message-ID: <20260326013902.588242-28-jhubbard@nvidia.com> (raw)
In-Reply-To: <20260326013902.588242-1-jhubbard@nvidia.com>

Hopper, Blackwell and later GPUs require a larger heap for WPR2.

Signed-off-by: John Hubbard <jhubbard@nvidia.com>
---
 drivers/gpu/nova-core/gsp/fw.rs | 63 +++++++++++++++++++++++++--------
 1 file changed, 49 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/nova-core/gsp/fw.rs b/drivers/gpu/nova-core/gsp/fw.rs
index c510a9cea3ba..f1531b71ac62 100644
--- a/drivers/gpu/nova-core/gsp/fw.rs
+++ b/drivers/gpu/nova-core/gsp/fw.rs
@@ -106,21 +106,41 @@ enum GspFwHeapParams {}
 /// Minimum required alignment for the GSP heap.
 const GSP_HEAP_ALIGNMENT: Alignment = Alignment::new::<{ 1 << 20 }>();
 
+// These constants override the generated bindings for architecture-specific heap sizing.
+// See Open RM: kgspCalculateGspFwHeapSize and related functions.
+//
+// 14MB for Hopper/Blackwell+.
+const GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100: u64 = 14 * num::usize_as_u64(SZ_1M);
+// 142MB client alloc for ~188MB total.
+const GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE_GH100: u64 = 142 * num::usize_as_u64(SZ_1M);
+// Hopper/Blackwell+ minimum heap size: 170MB (88 + 12 + 70).
+// See Open RM: GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB for the base 88MB,
+// plus Hopper+ additions in kgspCalculateGspFwHeapSize_GH100.
+const GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB_HOPPER: u64 = 170;
+
 impl GspFwHeapParams {
     /// Returns the amount of GSP-RM heap memory used during GSP-RM boot and initialization (up to
     /// and including the first client subdevice allocation).
-    fn base_rm_size(_chipset: Chipset) -> u64 {
-        // TODO: this needs to be updated to return the correct value for Hopper+ once support for
-        // them is added:
-        // u64::from(bindings::GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100)
-        u64::from(bindings::GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X)
+    fn base_rm_size(chipset: Chipset) -> u64 {
+        use crate::gpu::Architecture;
+        match chipset.arch() {
+            Architecture::Hopper | Architecture::BlackwellGB10x | Architecture::BlackwellGB20x => {
+                GSP_FW_HEAP_PARAM_BASE_RM_SIZE_GH100
+            }
+            _ => u64::from(bindings::GSP_FW_HEAP_PARAM_BASE_RM_SIZE_TU10X),
+        }
     }
 
     /// Returns the amount of heap memory required to support a single channel allocation.
-    fn client_alloc_size() -> u64 {
-        u64::from(bindings::GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE)
-            .align_up(GSP_HEAP_ALIGNMENT)
-            .unwrap_or(u64::MAX)
+    fn client_alloc_size(chipset: Chipset) -> Result<u64> {
+        use crate::gpu::Architecture;
+        let size = match chipset.arch() {
+            Architecture::Hopper | Architecture::BlackwellGB10x | Architecture::BlackwellGB20x => {
+                GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE_GH100
+            }
+            _ => u64::from(bindings::GSP_FW_HEAP_PARAM_CLIENT_ALLOC_SIZE),
+        };
+        size.align_up(GSP_HEAP_ALIGNMENT).ok_or(EINVAL)
     }
 
     /// Returns the amount of memory to reserve for management purposes for a framebuffer of size
@@ -164,12 +184,27 @@ impl LibosParams {
                 * num::usize_as_u64(SZ_1M),
     };
 
+    /// Hopper/Blackwell+ GPUs need a larger minimum heap size than the bindings specify.
+    /// The r570 bindings set LIBOS3_BAREMETAL_MIN_MB to 88MB, but Hopper/Blackwell+ actually
+    /// requires 170MB (88 + 12 + 70).
+    const LIBOS_HOPPER: LibosParams = LibosParams {
+        carveout_size: num::u32_as_u64(bindings::GSP_FW_HEAP_PARAM_OS_SIZE_LIBOS3_BAREMETAL),
+        allowed_heap_size: GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MIN_MB_HOPPER
+            * num::usize_as_u64(SZ_1M)
+            ..num::u32_as_u64(bindings::GSP_FW_HEAP_SIZE_OVERRIDE_LIBOS3_BAREMETAL_MAX_MB)
+                * num::usize_as_u64(SZ_1M),
+    };
+
     /// Returns the libos parameters corresponding to `chipset`.
     pub(crate) fn from_chipset(chipset: Chipset) -> &'static LibosParams {
-        if chipset < Chipset::GA102 {
-            &Self::LIBOS2
-        } else {
-            &Self::LIBOS3
+        use crate::gpu::Architecture;
+        match chipset.arch() {
+            Architecture::Turing => &Self::LIBOS2,
+            Architecture::Ampere if chipset == Chipset::GA100 => &Self::LIBOS2,
+            Architecture::Ampere | Architecture::Ada => &Self::LIBOS3,
+            Architecture::Hopper | Architecture::BlackwellGB10x | Architecture::BlackwellGB20x => {
+                &Self::LIBOS_HOPPER
+            }
         }
     }
 
@@ -183,7 +218,7 @@ pub(crate) fn wpr_heap_size(&self, chipset: Chipset, fb_size: u64) -> Result<u64
             // RM boot working memory,
             .saturating_add(GspFwHeapParams::base_rm_size(chipset))
             // One RM client,
-            .saturating_add(GspFwHeapParams::client_alloc_size())
+            .saturating_add(GspFwHeapParams::client_alloc_size(chipset)?)
             // Overhead for memory management.
             .saturating_add(GspFwHeapParams::management_overhead(fb_size)?)
             // Clamp to the supported heap sizes.
-- 
2.53.0


  parent reply	other threads:[~2026-03-26  1:39 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-26  1:38 [PATCH v9 00/31] gpu: nova-core: firmware: Hopper/Blackwell support John Hubbard
2026-03-26  1:38 ` [PATCH v9 01/31] gpu: nova-core: Hopper/Blackwell: basic GPU identification John Hubbard
2026-03-26  1:38 ` [PATCH v9 02/31] gpu: nova-core: factor .fwsignature* selection into a new find_gsp_sigs_section() John Hubbard
2026-03-30 14:29   ` Alexandre Courbot
2026-03-30 17:51     ` John Hubbard
2026-03-26  1:38 ` [PATCH v9 03/31] gpu: nova-core: use GPU Architecture to simplify HAL selections John Hubbard
2026-03-26  1:38 ` [PATCH v9 04/31] gpu: nova-core: add Copy/Clone to Spec and Revision, add chipset() accessor John Hubbard
2026-03-26  1:38 ` [PATCH v9 05/31] gpu: nova-core: set DMA mask width based on GPU architecture John Hubbard
2026-03-30 14:32   ` Alexandre Courbot
2026-03-30 21:31     ` John Hubbard
2026-03-26  1:38 ` [PATCH v9 06/31] gpu: nova-core: Hopper/Blackwell: skip GFW boot waiting John Hubbard
2026-03-30 14:52   ` Alexandre Courbot
2026-03-30 15:20     ` Gary Guo
2026-03-30 18:33     ` Joel Fernandes
2026-03-30 19:15       ` John Hubbard
2026-03-31  0:18       ` Alexandre Courbot
2026-03-31 16:17         ` Joel Fernandes
2026-03-26  1:38 ` [PATCH v9 07/31] gpu: nova-core: move firmware image parsing code to firmware.rs John Hubbard
2026-03-26  1:38 ` [PATCH v9 08/31] gpu: nova-core: factor out an elf_str() function John Hubbard
2026-03-26  1:38 ` [PATCH v9 09/31] gpu: nova-core: don't assume 64-bit firmware images John Hubbard
2026-03-26  1:38 ` [PATCH v9 10/31] gpu: nova-core: add support for 32-bit " John Hubbard
2026-03-26  1:38 ` [PATCH v9 11/31] gpu: nova-core: add auto-detection of 32-bit, 64-bit " John Hubbard
2026-03-26  1:38 ` [PATCH v9 12/31] gpu: nova-core: Hopper/Blackwell: add FMC firmware image, in support of FSP John Hubbard
2026-03-26  1:38 ` [PATCH v9 13/31] gpu: nova-core: Hopper/Blackwell: add FSP falcon engine stub John Hubbard
2026-03-26  1:38 ` [PATCH v9 14/31] gpu: nova-core: Hopper/Blackwell: add FSP falcon EMEM operations John Hubbard
2026-03-26  1:38 ` [PATCH v9 15/31] gpu: nova-core: Hopper/Blackwell: add FSP message infrastructure John Hubbard
2026-03-26  1:38 ` [PATCH v9 16/31] rust: ptr: add const_align_up() John Hubbard
2026-03-27  9:33   ` Miguel Ojeda
2026-03-30 21:41     ` John Hubbard
2026-03-31  0:03       ` Miguel Ojeda
2026-03-31  2:23         ` Alexandre Courbot
2026-03-31 10:26           ` Miguel Ojeda
2026-03-31  2:21   ` Alexandre Courbot
2026-03-31  2:36     ` John Hubbard
2026-03-31 10:24       ` Miguel Ojeda
2026-03-31 11:53         ` Danilo Krummrich
2026-04-03 10:01           ` Miguel Ojeda
2026-04-03 10:02   ` Miguel Ojeda
2026-03-26  1:38 ` [PATCH v9 17/31] gpu: nova-core: Hopper/Blackwell: calculate reserved FB heap size John Hubbard
2026-03-26  1:38 ` [PATCH v9 18/31] gpu: nova-core: add MCTP/NVDM protocol types for firmware communication John Hubbard
2026-03-26  1:38 ` [PATCH v9 19/31] gpu: nova-core: Hopper/Blackwell: add FSP secure boot completion waiting John Hubbard
2026-03-26  1:38 ` [PATCH v9 20/31] gpu: nova-core: Hopper/Blackwell: add FMC signature extraction John Hubbard
2026-03-26  1:38 ` [PATCH v9 21/31] gpu: nova-core: Hopper/Blackwell: add FSP send/receive messaging John Hubbard
2026-03-26  1:38 ` [PATCH v9 22/31] gpu: nova-core: Hopper/Blackwell: add FspCotVersion type John Hubbard
2026-03-26  1:38 ` [PATCH v9 23/31] gpu: nova-core: Hopper/Blackwell: larger non-WPR heap John Hubbard
2026-03-26  1:38 ` [PATCH v9 24/31] gpu: nova-core: Hopper/Blackwell: add FSP Chain of Trust boot John Hubbard
2026-03-30 15:11   ` Alexandre Courbot
2026-03-30 22:54     ` John Hubbard
2026-03-26  1:38 ` [PATCH v9 25/31] gpu: nova-core: Blackwell: use correct sysmem flush registers John Hubbard
2026-03-26  1:38 ` [PATCH v9 26/31] gpu: nova-core: make WPR heap sizing fallible John Hubbard
2026-03-26  1:38 ` John Hubbard [this message]
2026-03-26  1:38 ` [PATCH v9 28/31] gpu: nova-core: refactor SEC2 booter loading into BooterFirmware::run() John Hubbard
2026-03-26  1:39 ` [PATCH v9 29/31] gpu: nova-core: Hopper/Blackwell: add GSP lockdown release polling John Hubbard
2026-03-26  1:39 ` [PATCH v9 30/31] gpu: nova-core: Hopper/Blackwell: new location for PCI config mirror John Hubbard
2026-03-26  1:39 ` [PATCH v9 31/31] gpu: nova-core: Hopper/Blackwell: integrate FSP boot path into boot() John Hubbard
2026-03-30  5:10 ` [PATCH v9 00/31] gpu: nova-core: firmware: Hopper/Blackwell support Alexandre Courbot
2026-03-30 22:47   ` 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=20260326013902.588242-28-jhubbard@nvidia.com \
    --to=jhubbard@nvidia.com \
    --cc=a.hindborg@kernel.org \
    --cc=acourbot@nvidia.com \
    --cc=airlied@gmail.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=ecourtney@nvidia.com \
    --cc=gary@garyguo.net \
    --cc=joelagnelf@nvidia.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lossin@kernel.org \
    --cc=ojeda@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=shashanks@nvidia.com \
    --cc=simona@ffwll.ch \
    --cc=tmgross@umich.edu \
    --cc=ttabi@nvidia.com \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox