[PATCH v2] drm/amdkfd: Fix an eviction fence leak

Lang Yu lang.yu at amd.com
Mon Sep 30 06:38:12 UTC 2024


Only creating a new reference for each process instead of each VM.

Fixes: 9a1c1339abf9 ("drm/amdkfd: Run restore_workers on freezable WQs")

Suggested-by: Felix Kuehling <felix.kuehling at amd.com>
Signed-off-by: Lang Yu <lang.yu at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 4 ++--
 drivers/gpu/drm/amd/amdkfd/kfd_process.c         | 7 +++++--
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index ce5ca304dba9..fa572ba7f9fc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1439,8 +1439,8 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 	list_add_tail(&vm->vm_list_node,
 			&(vm->process_info->vm_list_head));
 	vm->process_info->n_vms++;
-
-	*ef = dma_fence_get(&vm->process_info->eviction_fence->base);
+	if (ef)
+		*ef = dma_fence_get(&vm->process_info->eviction_fence->base);
 	mutex_unlock(&vm->process_info->lock);
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index d07acf1b2f93..d665ecdcd12f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1702,12 +1702,15 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 
 	ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(dev->adev, avm,
 						     &p->kgd_process_info,
-						     &ef);
+						     p->ef ? NULL : &ef);
 	if (ret) {
 		dev_err(dev->adev->dev, "Failed to create process VM object\n");
 		return ret;
 	}
-	RCU_INIT_POINTER(p->ef, ef);
+
+	if (!p->ef)
+		RCU_INIT_POINTER(p->ef, ef);
+
 	pdd->drm_priv = drm_file->private_data;
 
 	ret = kfd_process_device_reserve_ib_mem(pdd);
-- 
2.25.1



More information about the amd-gfx mailing list