drm/amdkfd: Acquire cwsr buffer during CRIU resume

In normal kfd queue create, a reference to the cwsr svm range
is acquired so that this buffer is not freed from underneath the
queue. We weren't doing that.

After svm ranges are restored in RESUME phase of CRIU, go through
all the queues and acquire a reference to the cwsr.

Signed-off-by: David Francis <David.Francis@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 5a00bdd267a5d87307f9aaddabf550f2410c9d40)
Cc: stable@vger.kernel.org

4 files changed, 39 insertions, 2 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index ed5a6c3263bf..988f4cf6da90 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -2861,6 +2861,8 @@ static int criu_resume(struct file *filep,
 		goto exit;
 	}
 
+	kfd_criu_resume_queues(target);
+
 	ret =  amdgpu_amdkfd_criu_resume(target->kgd_process_info);
 	if (ret)
 		pr_err("amdgpu_amdkfd_criu_resume failed for %i\n", args->pid);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index 13852c2fd868..436160549386 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -1376,6 +1376,8 @@ int kfd_criu_restore_event(struct file *devkfd,
 			   uint8_t __user *user_priv_data,
 			   uint64_t *priv_data_offset,
 			   uint64_t max_priv_data_size);
+
+int kfd_criu_resume_queues(struct kfd_process *p);
 /* CRIU - End */
 
 /* Queue Context Management */
@@ -1392,6 +1394,7 @@ void kfd_queue_unref_bo_va(struct amdgpu_vm *vm, struct amdgpu_bo **bo);
 int kfd_queue_unref_bo_vas(struct kfd_process_device *pdd,
 			   struct queue_properties *properties);
 void kfd_queue_ctx_save_restore_size(struct kfd_topology_device *dev);
+int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size);
 
 struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type,
 		struct kfd_node *dev);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index 5741ed95c381..b68e884b1c37 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -1124,6 +1124,38 @@ exit:
 	return ret;
 }
 
+int kfd_criu_resume_queues(struct kfd_process *p)
+{
+	int ret, i;
+	u64 total_cwsr_size;
+
+	for (i = 0; i < p->n_pdds; i++) {
+		struct kfd_process_device *pdd = p->pdds[i];
+		struct kfd_topology_device *topo_dev;
+		struct queue *q;
+
+		topo_dev = kfd_topology_device_by_id(pdd->dev->id);
+
+		list_for_each_entry(q, &pdd->qpd.queues_list, list) {
+			total_cwsr_size = q->properties.ctx_save_restore_area_size +
+					  topo_dev->node_props.debug_memory_size;
+			if (check_mul_overflow(total_cwsr_size,
+					       NUM_XCC(pdd->dev->xcc_mask),
+					       &total_cwsr_size))
+				return -EINVAL;
+			total_cwsr_size = ALIGN(total_cwsr_size, PAGE_SIZE);
+
+			ret = kfd_queue_buffer_svm_get(pdd, q->properties.ctx_save_restore_area_address,
+						       total_cwsr_size);
+			if (ret)
+				return ret;
+		}
+
+	}
+
+	return 0;
+}
+
 int pqm_get_queue_checkpoint_info(struct process_queue_manager *pqm,
 				  unsigned int qid,
 				  uint32_t *mqd_size,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
index 9dcf0134f35b..cd41b10e0873 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_queue.c
@@ -88,7 +88,7 @@ void uninit_queue(struct queue *q)
 
 #if IS_ENABLED(CONFIG_HSA_AMD_SVM)
 
-static int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
+int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
 {
 	struct kfd_process *p = pdd->process;
 	struct list_head update_list;
@@ -182,7 +182,7 @@ static void kfd_queue_buffer_svm_put(struct kfd_process_device *pdd, u64 addr, u
 }
 #else
 
-static int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
+int kfd_queue_buffer_svm_get(struct kfd_process_device *pdd, u64 addr, u64 size)
 {
 	return -EINVAL;
 }