*/
panfrost_flush_batches_accessing_bo(ctx, bo, PAN_BO_ACCESS_WRITE);
- panfrost_bo_wait(bo, INT64_MAX, PAN_BO_ACCESS_WRITE);
+ panfrost_bo_wait(bo, INT64_MAX, false);
create_new_bo = true;
copy_resource = true;
* batches), we try to allocate a new one to avoid waiting.
*/
if (panfrost_pending_batches_access_bo(ctx, bo) ||
- !panfrost_bo_wait(bo, 0, PAN_BO_ACCESS_RW)) {
+ !panfrost_bo_wait(bo, 0, true)) {
/* We want the BO to be MMAPed. */
uint32_t flags = bo->flags & ~PAN_BO_DELAY_MMAP;
struct panfrost_bo *newbo = NULL;
} else if (!(usage & PIPE_TRANSFER_UNSYNCHRONIZED)) {
if (usage & PIPE_TRANSFER_WRITE) {
panfrost_flush_batches_accessing_bo(ctx, bo, true);
- panfrost_bo_wait(bo, INT64_MAX, PAN_BO_ACCESS_RW);
+ panfrost_bo_wait(bo, INT64_MAX, true);
} else if (usage & PIPE_TRANSFER_READ) {
panfrost_flush_batches_accessing_bo(ctx, bo, false);
- panfrost_bo_wait(bo, INT64_MAX, PAN_BO_ACCESS_WRITE);
+ panfrost_bo_wait(bo, INT64_MAX, false);
}
}