gpu: nvgpu: remove os_fence dependency from channel_sync

Move the wait_cmd_buffer programming for channel_sync->wait_fd to
channel_sync.c.  nvgpu_os_fence->program_waits
interface is now removed. channel_sync can directly retrieve
syncpt/semaphore from the interfaces of struct nvgpu_os_fence_syncpt
and struct nvgpu_os_fence_sema and use it for the wait programming.

Also, change int to u32 for some variables such as num_fences,
max_wait_size and wait_cmd_size.

Jira NVGPU-1093

Change-Id: I19c1b10d676caff49ce57861091f7f0ea65e7676
Signed-off-by: Debarshi Dutta <ddutta@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/1829719
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: Konsta Holtta <kholtta@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Vijayakumar Subbu <vsubbu@nvidia.com>
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>
This commit is contained in:
ddutta
2018-09-18 11:16:31 +05:30
committed by mobile promotions
parent 9f948ed07f
commit 80b5e2b8d6
6 changed files with 89 additions and 127 deletions

View File

@@ -80,7 +80,7 @@ static int nvgpu_submit_prepare_syncs(struct channel_gk20a *c,
* submission when user requested and the wait hasn't expired.
*/
if (flags & NVGPU_SUBMIT_FLAGS_FENCE_WAIT) {
int max_wait_cmds = c->deterministic ? 1 : 0;
u32 max_wait_cmds = c->deterministic ? 1U : 0U;
if (!pre_alloc_enabled) {
job->wait_cmd = nvgpu_kzalloc(g,