From: John Hubbard <jhubbard@nvidia.com>
To: Danilo Krummrich <dakr@kernel.org>,
Alexandre Courbot <acourbot@nvidia.com>
Cc: "Timur Tabi" <ttabi@nvidia.com>,
"Alistair Popple" <apopple@nvidia.com>,
"Eliot Courtney" <ecourtney@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>,
nova-gpu@lists.linux.dev, LKML <linux-kernel@vger.kernel.org>,
"John Hubbard" <jhubbard@nvidia.com>
Subject: [PATCH v2 18/31] gpu: nova-core: gsp: add GMC dispatch on receive
Date: Fri, 21 Aug 2026 18:54:35 -0700 [thread overview]
Message-ID: <20260822015448.238214-19-jhubbard@nvidia.com> (raw)
In-Reply-To: <20260822015448.238214-1-jhubbard@nvidia.com>
The r000 boot path delivers the load-and-execute steps as GMC events
keyed by command id, where the RPC path keys on a function code, so the
boot loop needs its own dispatch.
Add a GMC receive that dispatches on the command id. A GMC element and
an RPC element share every field through the NVDM header, so this checks
the NVDM type before reading the GMC header and drops an element of the
other kind rather than handing the handler a command id decoded from an
RPC header. It also drops an element whose GMC payload size disagrees
with the length the transport header declares, since GSP-RM sends the
two as one length and nothing says which to believe. An element nobody
claims yields Ok(None), and the read pointer advances on every path.
Assisted-by: Cursor:claude-opus-5
Reviewed-by: Timur Tabi <ttabi@nvidia.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
---
drivers/gpu/nova-core/gsp/cmdq.rs | 79 ++++++++++++++++++++++++++++++-
drivers/gpu/nova-core/gsp/fw.rs | 20 ++++++++
2 files changed, 97 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/nova-core/gsp/cmdq.rs b/drivers/gpu/nova-core/gsp/cmdq.rs
index 96128c418dc9..98d749674df1 100644
--- a/drivers/gpu/nova-core/gsp/cmdq.rs
+++ b/drivers/gpu/nova-core/gsp/cmdq.rs
@@ -641,7 +641,6 @@ struct GspMessage<'a> {
/// A GMC message ready to be processed from the message queue.
///
/// This is the type returned by [`CmdqInner::wait_for_gmc_msg`].
-#[expect(dead_code)]
struct GmcMessage<'a> {
// Reference to the header of the message.
header: &'a GspGmcMsgElement,
@@ -791,6 +790,22 @@ fn receive_msg<M: MessageFromGsp>(&self, timeout: Delta) -> Result<M>
self.inner.lock().receive_msg(timeout, None)
}
+ /// Receives one GMC event from the GSP and passes its command id and raw payload slices to
+ /// `handler`.
+ ///
+ /// This method may sleep while waiting. The [`CmdqInner`] mutex stays locked across the wait
+ /// and across the `handler` call, so `handler` must not call back into this [`Cmdq`].
+ ///
+ /// See [`CmdqInner::receive_gmc_and_dispatch`] for return values, queue state, and errors.
+ #[expect(dead_code)]
+ pub(crate) fn receive_gmc_and_dispatch<R>(
+ &self,
+ timeout: Delta,
+ handler: impl FnOnce(u32, &[u8], &[u8]) -> Option<R>,
+ ) -> Result<Option<R>> {
+ self.inner.lock().receive_gmc_and_dispatch(timeout, handler)
+ }
+
/// Waits for an unsolicited GSP event of type `M`, dispatching any other event that arrives
/// first.
///
@@ -1314,7 +1329,6 @@ fn drain(&mut self) -> Result {
/// - `ETIMEDOUT` if `timeout` has elapsed before any message becomes available.
/// - `EIO` if the framing is invalid, or the queue was already poisoned by an earlier such
/// failure. Either failure poisons the queue, so recovery requires a reset.
- #[expect(dead_code)]
fn wait_for_gmc_msg(&self, timeout: Delta) -> Result<GmcMessage<'_>> {
if self.poisoned.get() {
return Err(EIO);
@@ -1370,4 +1384,65 @@ fn wait_for_gmc_msg(&self, timeout: Delta) -> Result<GmcMessage<'_>> {
contents: (slice_1, slice_2),
})
}
+
+ /// Receive the next GMC event from the GSP and dispatch it through a handler.
+ ///
+ /// The handler receives the GMC command id and the raw payload slices that follow the
+ /// [`super::fw::GmcApiHeader`] (two slices because the circular buffer may wrap). It returns
+ /// `None` for an event it does not handle.
+ ///
+ /// Where [`Self::receive_msg`] keys on [`MsgFunction`], this keys on the GMC command id,
+ /// which is the form the r000 firmware uses for boot events.
+ ///
+ /// Returns `Ok(None)` when nothing claimed the element, either because it is not a GMC
+ /// element or because the handler declined it. The read pointer is advanced on every path.
+ ///
+ /// # Errors
+ ///
+ /// - `ETIMEDOUT` if `timeout` has elapsed before any message becomes available.
+ /// - `EIO` if the queue is poisoned or the element fails framing validation (see
+ /// [`Self::wait_for_gmc_msg`]).
+ fn receive_gmc_and_dispatch<R>(
+ &mut self,
+ timeout: Delta,
+ handler: impl FnOnce(u32, &[u8], &[u8]) -> Option<R>,
+ ) -> Result<Option<R>> {
+ let message = self.wait_for_gmc_msg(timeout)?;
+ let header = message.header;
+ let length = header.length();
+
+ // The RPC and GMC elements share every field through `nvdm_header`, so `gmc` holds an
+ // RPC header rather than a GMC one unless the NVDM type says otherwise.
+ let result = if !header.is_gmc_api() {
+ dev_warn!(&self.dev, "GSP GMC: dropping non-GMC queue element\n");
+ None
+ } else if num::u32_as_usize(header.gmc.size) != header.payload_length() {
+ // GSP-RM sends the element as the GMC header plus `size` bytes, so the two lengths
+ // describe the same payload and a handler cannot tell which one to believe.
+ dev_err!(
+ &self.dev,
+ "GSP GMC: payload is {} bytes, transport declares {}\n",
+ header.gmc.size,
+ header.payload_length(),
+ );
+ None
+ } else {
+ let command_id = header.gmc.command_id();
+
+ dev_dbg!(
+ &self.dev,
+ "GSP GMC: event: seq# {}, command_id=0x{:x}, length=0x{:x}\n",
+ header.gmc.sequence,
+ command_id,
+ length,
+ );
+
+ handler(command_id, message.contents.0, message.contents.1)
+ };
+
+ self.gsp_mem
+ .advance_cpu_read_ptr(u32::try_from(length.div_ceil(GSP_PAGE_SIZE))?);
+
+ Ok(result)
+ }
}
diff --git a/drivers/gpu/nova-core/gsp/fw.rs b/drivers/gpu/nova-core/gsp/fw.rs
index bbcc2e2b8401..31c0e6351fdc 100644
--- a/drivers/gpu/nova-core/gsp/fw.rs
+++ b/drivers/gpu/nova-core/gsp/fw.rs
@@ -1008,6 +1008,10 @@ pub(crate) struct GmcApiHeader {
reserved: [u32; 5],
}
+/// Command identifier bits of [`GmcApiHeader::command`], matching Open RM's
+/// `GMCAPI_HEADER_COMMAND_ID_MASK`. The remaining byte carries flags.
+const GMCAPI_COMMAND_ID_MASK: u32 = 0x00ff_ffff;
+
static_assert!(size_of::<GmcApiHeader>() == size_of::<r000_00::GMCAPI_HEADER>());
static_assert!(
core::mem::offset_of!(GmcApiHeader, command)
@@ -1030,6 +1034,13 @@ pub(crate) struct GmcApiHeader {
== core::mem::offset_of!(r000_00::GMCAPI_HEADER, reserved)
);
+impl GmcApiHeader {
+ /// Returns the command identifier, without the flag byte.
+ pub(crate) fn command_id(&self) -> u32 {
+ self.command & GMCAPI_COMMAND_ID_MASK
+ }
+}
+
// SAFETY: All fields are integer types with no uninitialized padding bytes.
unsafe impl AsBytes for GmcApiHeader {}
@@ -1134,6 +1145,15 @@ pub(crate) fn validate_framing(&self) -> Result {
)
}
+ /// Returns `true` if the NVDM header routes this element to the GSP's GMC dispatch.
+ ///
+ /// A [`GspMsgElement`] and a [`GspGmcMsgElement`] share every field through `nvdm_header`,
+ /// so the NVDM type is what distinguishes the two on the queue, and `gmc` holds an RPC
+ /// header rather than a GMC one when this returns `false`.
+ pub(crate) fn is_gmc_api(&self) -> bool {
+ self.nvdm_header.validate(NvdmType::GmcApi)
+ }
+
/// Returns the number of elements (i.e. memory pages) used by this message.
pub(crate) fn element_count(&self) -> u32 {
self.mctp_payload_size
--
2.55.0
next prev parent reply other threads:[~2026-08-22 1:55 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-22 1:54 [PATCH v2 00/31] gpu: nova-core: boot on the r000 GSP firmware John Hubbard
2026-08-22 1:54 ` [PATCH v2 01/31] rust: pci: add domain_nr() accessor John Hubbard
2026-08-22 7:57 ` Miguel Ojeda
2026-08-22 20:11 ` John Hubbard
2026-08-23 0:08 ` Alexandre Courbot
2026-08-22 1:54 ` [PATCH v2 02/31] gpu: nova-core: firmware: add r000 bindings John Hubbard
2026-08-22 1:54 ` [PATCH v2 03/31] gpu: nova-core: extract radix3 page table into its own module John Hubbard
2026-08-22 1:54 ` [PATCH v2 04/31] gpu: nova-core: set MCTP transport header version to 1 John Hubbard
2026-08-22 1:54 ` [PATCH v2 05/31] gpu: nova-core: add Falcon helpers for r000 LOAD_EXEC events John Hubbard
2026-08-22 1:54 ` [PATCH v2 06/31] gpu: nova-core: zero-pad radix3 page table levels to page boundary John Hubbard
2026-08-22 1:54 ` [PATCH v2 07/31] gpu: nova-core: distinguish async GSP RPC traffic in debug logs John Hubbard
2026-08-22 1:54 ` [PATCH v2 08/31] gpu: nova-core: add optional ucodes firmware loading John Hubbard
2026-08-23 16:26 ` M Henning
2026-08-23 19:51 ` John Hubbard
2026-08-22 1:54 ` [PATCH v2 09/31] gpu: nova-core: add LIBOS3 log buffers and state monitor buffer John Hubbard
2026-08-22 1:54 ` [PATCH v2 10/31] gpu: nova-core: add build ID headers to debugfs log buffer dumps John Hubbard
2026-08-22 1:54 ` [PATCH v2 11/31] gpu: nova-core: rename the FbRanges elf field to fw_image John Hubbard
2026-08-22 1:54 ` [PATCH v2 12/31] gpu: nova-core: regs: add msgq v2 BAR0 register declarations John Hubbard
2026-08-22 1:54 ` [PATCH v2 13/31] gpu: nova-core: gsp: add msgq v2 internals John Hubbard
2026-08-22 1:54 ` [PATCH v2 14/31] gpu: nova-core: generalize allocate_command() for variable headers John Hubbard
2026-08-22 1:54 ` [PATCH v2 15/31] gpu: nova-core: add GMC API message types John Hubbard
2026-08-22 1:54 ` [PATCH v2 16/31] gpu: nova-core: add GMC send path John Hubbard
2026-08-22 1:54 ` [PATCH v2 17/31] gpu: nova-core: add GMC transport receive path John Hubbard
2026-08-22 1:54 ` John Hubbard [this message]
2026-08-22 1:54 ` [PATCH v2 19/31] gpu: nova-core: separate the generic falcon bootloader from FWSEC John Hubbard
2026-08-22 1:54 ` [PATCH v2 20/31] gpu: nova-core: handle the r000 load-and-execute HS binary event John Hubbard
2026-08-22 1:54 ` [PATCH v2 21/31] gpu: nova-core: handle the r000 load-and-execute bootloader event John Hubbard
2026-08-22 1:54 ` [PATCH v2 22/31] gpu: nova-core: gsp: add the GMC boot event dispatcher John Hubbard
2026-08-22 1:54 ` [PATCH v2 23/31] gpu: nova-core: gsp: add the GSP_INIT request builder John Hubbard
2026-08-22 1:54 ` [PATCH v2 24/31] gpu: nova-core: gsp: send GSP_INIT and decode its reply John Hubbard
2026-08-22 1:54 ` [PATCH v2 25/31] gpu: nova-core: gsp: pass the remaining log buffers to GSP-RM John Hubbard
2026-08-22 1:54 ` [PATCH v2 26/31] gpu: nova-core: switch to the r000 GSP firmware John Hubbard
2026-08-22 1:54 ` [PATCH v2 27/31] gpu: nova-core: gsp: validate RPC element framing on receive John Hubbard
2026-08-22 1:54 ` [PATCH v2 28/31] gpu: nova-core: gsp: remove the RPCs that GSP_INIT replaced John Hubbard
2026-08-22 1:54 ` [PATCH v2 29/31] gpu: nova-core: firmware: delete the r570 bindings John Hubbard
2026-08-22 1:54 ` [PATCH v2 30/31] gpu: nova-core: print GMC command names in debug logs John Hubbard
2026-08-22 1:54 ` [PATCH v2 31/31] gpu: nova-core: distinguish GMC event and response " 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=20260822015448.238214-19-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=linux-kernel@vger.kernel.org \
--cc=lossin@kernel.org \
--cc=nova-gpu@lists.linux.dev \
--cc=ojeda@kernel.org \
--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
all inboxes | Powered by JetHome®