[PATCH] accel/ivpu: Reorder doorbell unregister and command queue destruction

[PATCH] accel/ivpu: Reorder doorbell unregister and command queue destruction

[Maciej Falkowski]

From: Karol Wachowski <karol.wachowski@intel.com>

Refactor ivpu_cmdq_unregister() to ensure the doorbell is unregistered
before destroying the command queue. The NPU firmware requires doorbells
to be unregistered prior to command queue destruction.

If doorbell remains registered when command queue destroy command is sent
firmware will automatically unregister the doorbell, making subsequent
unregister attempts no-operations (NOPs).

Ensure compliance with firmware expectations by moving the doorbell
unregister call ahead of the command queue destruction logic,
thus preventing unnecessary NOP operation.

Fixes: 2a18ceff9482 ("accel/ivpu: Implement support for hardware scheduler")
Cc: <stable@vger.kernel.org> # v6.11+
Signed-off-by: Karol Wachowski <karol.wachowski@intel.com>
Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>
---
 drivers/accel/ivpu/ivpu_job.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/accel/ivpu/ivpu_job.c b/drivers/accel/ivpu/ivpu_job.c
index bd351dd46d6b..060f1fc031d3 100644
--- a/drivers/accel/ivpu/ivpu_job.c
+++ b/drivers/accel/ivpu/ivpu_job.c
@@ -271,6 +271,10 @@ static int ivpu_cmdq_unregister(struct ivpu_file_priv *file_priv, struct ivpu_cm
 	if (!cmdq->db_id)
 		return 0;
 
+	ret = ivpu_jsm_unregister_db(vdev, cmdq->db_id);
+	if (!ret)
+		ivpu_dbg(vdev, JOB, "DB %d unregistered\n", cmdq->db_id);
+
 	if (vdev->fw->sched_mode == VPU_SCHEDULING_MODE_HW) {
 		ret = ivpu_jsm_hws_destroy_cmdq(vdev, file_priv->ctx.id, cmdq->id);
 		if (!ret)
@@ -278,10 +282,6 @@ static int ivpu_cmdq_unregister(struct ivpu_file_priv *file_priv, struct ivpu_cm
 				 cmdq->id, file_priv->ctx.id);
 	}
 
-	ret = ivpu_jsm_unregister_db(vdev, cmdq->db_id);
-	if (!ret)
-		ivpu_dbg(vdev, JOB, "DB %d unregistered\n", cmdq->db_id);
-
 	xa_erase(&file_priv->vdev->db_xa, cmdq->db_id);
 	cmdq->db_id = 0;
 
-- 
2.43.0

Re: [PATCH] accel/ivpu: Reorder doorbell unregister and command queue destruction

[Jeff Hugo]

On 6/4/2025 9:44 AM, Maciej Falkowski wrote:
> From: Karol Wachowski <karol.wachowski@intel.com>
> 
> Refactor ivpu_cmdq_unregister() to ensure the doorbell is unregistered
> before destroying the command queue. The NPU firmware requires doorbells
> to be unregistered prior to command queue destruction.
> 
> If doorbell remains registered when command queue destroy command is sent
> firmware will automatically unregister the doorbell, making subsequent
> unregister attempts no-operations (NOPs).
> 
> Ensure compliance with firmware expectations by moving the doorbell
> unregister call ahead of the command queue destruction logic,
> thus preventing unnecessary NOP operation.
> 
> Fixes: 2a18ceff9482 ("accel/ivpu: Implement support for hardware scheduler")
> Cc: <stable@vger.kernel.org> # v6.11+
> Signed-off-by: Karol Wachowski <karol.wachowski@intel.com>
> Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>

Huh?  This was posted to the list on May 15th, and Jacek applied it to 
drm-misc-fixes on May 28th.

-Jeff

Re: [PATCH] accel/ivpu: Reorder doorbell unregister and command queue destruction

[Falkowski, Maciej]

On 6/4/2025 6:18 PM, Jeff Hugo wrote:

> On 6/4/2025 9:44 AM, Maciej Falkowski wrote:
>> From: Karol Wachowski <karol.wachowski@intel.com>
>>
>> Refactor ivpu_cmdq_unregister() to ensure the doorbell is unregistered
>> before destroying the command queue. The NPU firmware requires doorbells
>> to be unregistered prior to command queue destruction.
>>
>> If doorbell remains registered when command queue destroy command is 
>> sent
>> firmware will automatically unregister the doorbell, making subsequent
>> unregister attempts no-operations (NOPs).
>>
>> Ensure compliance with firmware expectations by moving the doorbell
>> unregister call ahead of the command queue destruction logic,
>> thus preventing unnecessary NOP operation.
>>
>> Fixes: 2a18ceff9482 ("accel/ivpu: Implement support for hardware 
>> scheduler")
>> Cc: <stable@vger.kernel.org> # v6.11+
>> Signed-off-by: Karol Wachowski <karol.wachowski@intel.com>
>> Signed-off-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>
>
> Huh?  This was posted to the list on May 15th, and Jacek applied it to 
> drm-misc-fixes on May 28th.
My apologies, you are right. I accidentally included it while checking 
latest changes.

Best regards,
Maciej
>
> -Jeff