All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sunil Khatri <sunil.khatri@amd.com>
To: "Alex Deucher" <alexander.deucher@amd.com>,
	"Christian König" <christian.koenig@amd.com>
Cc: amd-gfx@lists.freedesktop.org, Sunil Khatri <sunil.khatri@amd.com>
Subject: [PATCH V1 1/2] drm/amdgpu/userq: avoid uneccessary locking in amdgpu_userq_create
Date: Wed,  8 Apr 2026 11:06:02 +0530	[thread overview]
Message-ID: <20260408053603.2894742-1-sunil.khatri@amd.com> (raw)

Reorganise code to avoid holding mutex userq_mutex while
also trying to grab exec lock ww_mutex where its not needed
for function amdgpu_userq_input_va_validate

Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c | 33 +++++++++++------------
 1 file changed, 15 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
index 3a6e7a569c78..3f502c18879a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
@@ -737,28 +737,17 @@ amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
 		return r;
 	}
 
-	/*
-	 * There could be a situation that we are creating a new queue while
-	 * the other queues under this UQ_mgr are suspended. So if there is any
-	 * resume work pending, wait for it to get done.
-	 *
-	 * This will also make sure we have a valid eviction fence ready to be used.
-	 */
-	amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr);
-
 	uq_funcs = adev->userq_funcs[args->in.ip_type];
 	if (!uq_funcs) {
 		drm_file_err(uq_mgr->file, "Usermode queue is not supported for this IP (%u)\n",
 			     args->in.ip_type);
-		r = -EINVAL;
-		goto unlock;
+		return -EINVAL;
 	}
 
 	queue = kzalloc(sizeof(struct amdgpu_usermode_queue), GFP_KERNEL);
 	if (!queue) {
 		drm_file_err(uq_mgr->file, "Failed to allocate memory for queue\n");
-		r = -ENOMEM;
-		goto unlock;
+		return -ENOMEM;
 	}
 
 	INIT_LIST_HEAD(&queue->userq_va_list);
@@ -781,12 +770,21 @@ amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
 		goto free_queue;
 	}
 
+	/*
+	 * There could be a situation that we are creating a new queue while
+	 * the other queues under this UQ_mgr are suspended. So if there is any
+	 * resume work pending, wait for it to get done.
+	 *
+	 * This will also make sure we have a valid eviction fence ready to be used.
+	 */
+	amdgpu_userq_ensure_ev_fence(&fpriv->userq_mgr, &fpriv->evf_mgr);
+
 	/* Convert relative doorbell offset into absolute doorbell index */
 	index = amdgpu_userq_get_doorbell_index(uq_mgr, &db_info, filp);
 	if (index == (uint64_t)-EINVAL) {
 		drm_file_err(uq_mgr->file, "Failed to get doorbell for queue\n");
 		r = -EINVAL;
-		goto free_queue;
+		goto unlock;
 	}
 
 	queue->doorbell_index = index;
@@ -794,7 +792,7 @@ amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
 	r = amdgpu_userq_fence_driver_alloc(adev, &queue->fence_drv);
 	if (r) {
 		drm_file_err(uq_mgr->file, "Failed to alloc fence driver\n");
-		goto free_queue;
+		goto unlock;
 	}
 
 	r = uq_funcs->mqd_create(queue, &args->in);
@@ -858,11 +856,10 @@ amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
 	up_read(&adev->reset_domain->sem);
 clean_fence_driver:
 	amdgpu_userq_fence_driver_free(queue);
-free_queue:
-	kfree(queue);
 unlock:
 	mutex_unlock(&uq_mgr->userq_mutex);
-
+free_queue:
+	kfree(queue);
 	return r;
 }
 
-- 
2.34.1


             reply	other threads:[~2026-04-08  5:36 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-08  5:36 Sunil Khatri [this message]
2026-04-08  5:36 ` [PATCH V1 2/2] drm/amdgpu/userq: clean the VA mapping list for failed queue creation Sunil Khatri
2026-04-08  8:23 ` [PATCH V1 1/2] drm/amdgpu/userq: avoid uneccessary locking in amdgpu_userq_create Christian König
2026-04-08  8:36   ` Khatri, Sunil
2026-04-08  8:53     ` Christian König

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=20260408053603.2894742-1-sunil.khatri@amd.com \
    --to=sunil.khatri@amd.com \
    --cc=alexander.deucher@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    --cc=christian.koenig@amd.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.