[PATCH v2] gpu: nova-core: gsp: cmdq: reduce sleep delta in wait_for_msg

[PATCH v2] gpu: nova-core: gsp: cmdq: reduce sleep delta in wait_for_msg

[Alexandre Courbot]

If a message is not immediately available on the queue, `wait_for_msg`
sleeps for 1 millisecond before checking again. This is 3 orders of
magnitude more than Nouveau (which sleeps for 1 microsecond) or OpenRM
(which does not sleep at all and busy-polls) and is way too much, as GSP
commands are expected to be processed much faster than that.

Reduce the sleep time to 1 microsecond. This effectively makes `fsleep`
busy-loop with a call to `cpu_relax` afterwards, keeping latency to a
minimum.

Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
---
Changes in v2:
- Wait for 1 microsecond. (Eliot)
- Link to v1: https://patch.msgid.link/20261030-cmdq_wait_msg_timeout-v1-1-a1e628d2732c@nvidia.com
---
 drivers/gpu/nova-core/gsp/cmdq.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/nova-core/gsp/cmdq.rs b/drivers/gpu/nova-core/gsp/cmdq.rs
index 936406004a57..8ee8a99e4d7e 100644
--- a/drivers/gpu/nova-core/gsp/cmdq.rs
+++ b/drivers/gpu/nova-core/gsp/cmdq.rs
@@ -603,7 +603,7 @@ fn wait_for_msg(&self, timeout: Delta) -> Result<GspMessage<'_>> {
             read_poll_timeout(
                 || Ok(self.gsp_mem.driver_read_area()),
                 |driver_area| !driver_area.0.is_empty(),
-                Delta::from_millis(1),
+                Delta::from_micros(1),
                 timeout,
             )
         }

---
base-commit: 9fe580fd90e20c8432832c5f9a7e92b049ae9af6
change-id: 20261030-cmdq_wait_msg_timeout-26df11234d10

Best regards,
--  
Alexandre Courbot <acourbot@nvidia.com>

Re: [PATCH v2] gpu: nova-core: gsp: cmdq: reduce sleep delta in wait_for_msg

[Eliot Courtney]

On Mon Nov 2, 2026 at 9:30 PM JST, Alexandre Courbot wrote:
> If a message is not immediately available on the queue, `wait_for_msg`
> sleeps for 1 millisecond before checking again. This is 3 orders of
> magnitude more than Nouveau (which sleeps for 1 microsecond) or OpenRM
> (which does not sleep at all and busy-polls) and is way too much, as GSP
> commands are expected to be processed much faster than that.
>
> Reduce the sleep time to 1 microsecond. This effectively makes `fsleep`
> busy-loop with a call to `cpu_relax` afterwards, keeping latency to a
> minimum.
>
> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
> ---
> Changes in v2:
> - Wait for 1 microsecond. (Eliot)
> - Link to v1: https://patch.msgid.link/20261030-cmdq_wait_msg_timeout-v1-1-a1e628d2732c@nvidia.com
> ---

Reviewed-by: Eliot Courtney <ecourtney@nvidia.com>

Re: [PATCH v2] gpu: nova-core: gsp: cmdq: reduce sleep delta in wait_for_msg

[Danilo Krummrich]

On Mon, Nov 2, 2026, at 1:30 PM, Alexandre Courbot wrote:
> If a message is not immediately available on the queue, `wait_for_msg`
> sleeps for 1 millisecond before checking again. This is 3 orders of
> magnitude more than Nouveau (which sleeps for 1 microsecond) or OpenRM
> (which does not sleep at all and busy-polls) and is way too much, as GSP
> commands are expected to be processed much faster than that.
>
> Reduce the sleep time to 1 microsecond. This effectively makes `fsleep`
> busy-loop with a call to `cpu_relax` afterwards, keeping latency to a
> minimum.
>
> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>

Acked-by: Danilo Krummrich <dakr@kernel.org>

Re: [PATCH v2] gpu: nova-core: gsp: cmdq: reduce sleep delta in wait_for_msg

[Alexandre Courbot]

On Mon Nov 2, 2026 at 9:30 PM JST, Alexandre Courbot wrote:
> If a message is not immediately available on the queue, `wait_for_msg`
> sleeps for 1 millisecond before checking again. This is 3 orders of
> magnitude more than Nouveau (which sleeps for 1 microsecond) or OpenRM
> (which does not sleep at all and busy-polls) and is way too much, as GSP
> commands are expected to be processed much faster than that.
>
> Reduce the sleep time to 1 microsecond. This effectively makes `fsleep`
> busy-loop with a call to `cpu_relax` afterwards, keeping latency to a
> minimum.
>
> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>

Pushed to drm-rust-next, thanks!