video: tegra: host: dla: use 0x92 preaction

- PREACTION_SEM_GE i.e. 0x92 blocks processing of further task until
  value at given pointer becomes greater than equal to updating value
- use 0x92 (PREACTION_SEM_GE) as default action for processing syncpoint
  as to make inline with fence wait
- Use MACROS from interface file

Jira DLA-198

Change-Id: Ibc5dbf171fd9a62cfd1a74341753a094d2caeddd
Signed-off-by: Shridhar Rasal <srasal@nvidia.com>
Signed-off-by: Prashant Gaikwad <pgaikwad@nvidia.com>
Reviewed-on: http://git-master/r/1257860
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
Shridhar Rasal
2016-11-22 12:26:22 +05:30
committed by Laxman Dewangan
parent 25eec664e7
commit e018a06107
2 changed files with 4 additions and 9 deletions

View File

@@ -36,11 +36,6 @@
#define MAX_NUM_ACTION_LIST 1
/* TODO: Below should come from firmware interface */
#define ACTION_OPCODE_TERMINATE 0x00
#define ACTION_OPCODE_WRITE_SEM 0x80
#define ACTION_OPCODE_READ_SEM 0x90
/**
* Maximum number of queue's per engine
*/

View File

@@ -576,12 +576,12 @@ struct nvdla_task *nvdla_task_alloc(struct nvhost_queue *queue,
/* update end of action list */
if (i == user_task->num_prefences) {
opcode->value = ACTION_OPCODE_TERMINATE;
opcode->value = PREACTION_TERMINATE;
break;
}
/* set action type */
opcode->value = ACTION_OPCODE_READ_SEM;
opcode->value = PREACTION_SEM_GE;
/* get actual preaction address */
preaction = (struct dla_action_semaphore *)
@@ -606,12 +606,12 @@ struct nvdla_task *nvdla_task_alloc(struct nvhost_queue *queue,
/* update end of list */
if (i == user_task->num_postfences) {
opcode->value = ACTION_OPCODE_TERMINATE;
opcode->value = POSTACTION_TERMINATE;
break;
}
/* set action type */
opcode->value = ACTION_OPCODE_WRITE_SEM;
opcode->value = POSTACTION_SEM;
/* get actual post action mem */
postaction = (struct dla_action_semaphore *)