drm/amdkfd: Implement KFD process eviction/restore

When the TTM memory manager in KGD evicts BOs, all user mode queues
potentially accessing these BOs must be evicted temporarily. Once
user mode queues are evicted, the eviction fence is signaled,
allowing the migration of the BO to proceed.

A delayed worker is scheduled to restore all the BOs belonging to
the evicted process and restart its queues.

During suspend/resume of the GPU we also evict all processes to allow
KGD to save BOs in system memory, since VRAM will be lost.

v2:
* Account for eviction when updating of q->is_active in MQD manager

Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
This commit is contained in:
Felix Kuehling
2018-02-06 20:32:45 -05:00
committed by Oded Gabbay
parent 403575c44e
commit 26103436da
8 changed files with 547 additions and 8 deletions

View File

@@ -79,6 +79,10 @@ struct device_process_node {
*
* @process_termination: Clears all process queues belongs to that device.
*
* @evict_process_queues: Evict all active queues of a process
*
* @restore_process_queues: Restore all evicted queues queues of a process
*
*/
struct device_queue_manager_ops {
@@ -129,6 +133,11 @@ struct device_queue_manager_ops {
int (*process_termination)(struct device_queue_manager *dqm,
struct qcm_process_device *qpd);
int (*evict_process_queues)(struct device_queue_manager *dqm,
struct qcm_process_device *qpd);
int (*restore_process_queues)(struct device_queue_manager *dqm,
struct qcm_process_device *qpd);
};
struct device_queue_manager_asic_ops {