nv-tegra-mirror/linux-nvgpu
1
0
Fork 0
mirror of git://nv-tegra.nvidia.com/linux-nvgpu.git synced 2025-12-25 02:52:51 +03:00
Code Issues Packages Projects Releases Wiki Activity

gpu: nvgpu: use nvgpu list for pending semaphore waits

Browse Source 
Use nvgpu list APIs instead of linux list APIs
to store pending semaphore waits

Jira NVGPU-13

Change-Id: I42fc6c6233e39f475a939ddd6a81c0cda851b6bf
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1454693
Reviewed-by: svccoveritychecker <svccoveritychecker@nvidia.com>
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>
This commit is contained in:
Deepak Nibade
2017-04-04 14:30:15 +05:30
committed by mobile promotions
parent 692b56aa4b
commit 92ab6c3bbc
3 changed files with 20 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
drivers/gpu/nvgpu/gk20a/channel_sync_gk20a.c
View File

@@ -399,7 +399,14 @@ struct wait_fence_work {
struct channel_gk20a *ch;
struct nvgpu_semaphore *sema;
struct gk20a *g;
struct list_head entry;
struct nvgpu_list_node entry;
};
static inline struct wait_fence_work *
wait_fence_work_from_entry(struct nvgpu_list_node *node)
{
return (struct wait_fence_work *)
((uintptr_t)node - offsetof(struct wait_fence_work, entry));
};
/*
@@ -413,7 +420,7 @@ static void gk20a_add_pending_sema_wait(struct gk20a *g,
struct wait_fence_work *work)
{
nvgpu_raw_spinlock_acquire(&g->pending_sema_waits_lock);
list_add(&work->entry, &g->pending_sema_waits);
nvgpu_list_add(&work->entry, &g->pending_sema_waits);
nvgpu_raw_spinlock_release(&g->pending_sema_waits_lock);
}
@@ -422,10 +429,10 @@ static void gk20a_add_pending_sema_wait(struct gk20a *g,
* then delete the entire pending list.
*/
static void gk20a_start_sema_wait_cancel(struct gk20a *g,
struct list_head *list)
struct nvgpu_list_node *list)
{
nvgpu_raw_spinlock_acquire(&g->pending_sema_waits_lock);
list_replace_init(&g->pending_sema_waits, list);
nvgpu_list_replace_init(&g->pending_sema_waits, list);
nvgpu_raw_spinlock_release(&g->pending_sema_waits_lock);
}
@@ -448,18 +455,18 @@ static void gk20a_start_sema_wait_cancel(struct gk20a *g,
void gk20a_channel_cancel_pending_sema_waits(struct gk20a *g)
{
struct wait_fence_work *work;
struct list_head local_pending_sema_waits;
struct nvgpu_list_node local_pending_sema_waits;
gk20a_start_sema_wait_cancel(g, &local_pending_sema_waits);
while (!list_empty(&local_pending_sema_waits)) {
while (!nvgpu_list_empty(&local_pending_sema_waits)) {
int ret;
work = list_first_entry(&local_pending_sema_waits,
struct wait_fence_work,
work = nvgpu_list_first_entry(&local_pending_sema_waits,
wait_fence_work,
entry);
list_del_init(&work->entry);
nvgpu_list_del(&work->entry);
/*
* Only nvgpu_kfree() work if the cancel is successful.
@@ -486,8 +493,8 @@ static void gk20a_channel_semaphore_launcher(
* otherwise it's possible that the deterministic submit path suffers.
*/
nvgpu_raw_spinlock_acquire(&g->pending_sema_waits_lock);
if (!list_empty(&g->pending_sema_waits))
list_del_init(&w->entry);
if (!nvgpu_list_empty(&g->pending_sema_waits))
nvgpu_list_del(&w->entry);
nvgpu_raw_spinlock_release(&g->pending_sema_waits_lock);
gk20a_dbg_info("waiting for pre fence %p '%s'",