@@ -570,23 +570,19 @@ static void engines_idle_release(struct i915_gem_context *ctx,
engines->ctx = i915_gem_context_get(ctx);
for_each_gem_engine(ce, engines, it) {
- struct dma_fence *fence;
- int err = 0;
+ int err;
/* serialises with execbuf */
set_bit(CONTEXT_CLOSED_BIT, &ce->flags);
if (!intel_context_pin_if_active(ce))
continue;
- fence = i915_active_fence_get(&ce->timeline->last_request);
- if (fence) {
- err = i915_sw_fence_await_dma_fence(&engines->fence,
- fence, 0,
- GFP_KERNEL);
- dma_fence_put(fence);
- }
+ /* Wait until context is finally scheduled out and retired */
+ err = i915_sw_fence_await_active(&engines->fence,
+ &ce->active,
+ I915_ACTIVE_AWAIT_BARRIER);
intel_context_unpin(ce);
- if (err < 0)
+ if (err)
goto kill;
}
@@ -542,6 +542,54 @@ static int __await_active(struct i915_active_fence *active,
return 0;
}
+struct wait_barrier {
+ struct wait_queue_entry base;
+ struct i915_active *ref;
+};
+
+static int
+barrier_wake(wait_queue_entry_t *wq, unsigned mode, int flags, void *key)
+{
+ struct wait_barrier *wb = container_of(wq, typeof(*wb), base);
+
+ if (i915_active_is_idle(wb->ref)) { /* shared waitqueue, must check! */
+ list_del(&wq->entry);
+ i915_sw_fence_complete(wq->private);
+ kfree(wq);
+ }
+
+ return 0;
+}
+
+static int __await_barrier(struct i915_active *ref, struct i915_sw_fence *fence)
+{
+ struct wait_barrier *wb;
+
+ wb = kmalloc(sizeof(*wb), GFP_NOWAIT | __GFP_NOWARN);
+ if (unlikely(!wb))
+ return -ENOMEM;
+
+ if (!i915_active_acquire_if_busy(ref)) {
+ kfree(wb);
+ return 0;
+ }
+
+ if (!i915_sw_fence_await(fence)) {
+ kfree(wb);
+ i915_active_release(ref);
+ return -EINVAL;
+ }
+
+ wb->base.flags = 0;
+ wb->base.func = barrier_wake;
+ wb->base.private = fence;
+
+ add_wait_queue(__var_waitqueue(ref), &wb->base);
+
+ i915_active_release(ref);
+ return 0;
+}
+
static int await_active(struct i915_active *ref,
unsigned int flags,
int (*fn)(void *arg, struct dma_fence *fence),
@@ -556,7 +604,8 @@ static int await_active(struct i915_active *ref,
return err;
}
- if (flags & I915_ACTIVE_AWAIT_ALL && i915_active_acquire_if_busy(ref)) {
+ if (flags & I915_ACTIVE_AWAIT_ACTIVE &&
+ i915_active_acquire_if_busy(ref)) {
struct active_node *it, *n;
rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) {
@@ -569,6 +618,12 @@ static int await_active(struct i915_active *ref,
return err;
}
+ if (flags & I915_ACTIVE_AWAIT_BARRIER) {
+ err = __await_barrier(ref, arg);
+ if (err)
+ return err;
+ }
+
return 0;
}
@@ -193,7 +193,8 @@ int i915_sw_fence_await_active(struct i915_sw_fence *fence,
int i915_request_await_active(struct i915_request *rq,
struct i915_active *ref,
unsigned int flags);
-#define I915_ACTIVE_AWAIT_ALL BIT(0)
+#define I915_ACTIVE_AWAIT_ACTIVE BIT(0)
+#define I915_ACTIVE_AWAIT_BARRIER BIT(1)
int i915_active_acquire(struct i915_active *ref);
bool i915_active_acquire_if_busy(struct i915_active *ref);
If we want to percolate information back from the HW, up through the GEM context, we need to wait until the intel_context is scheduled out for the last time. This is handled by the retirement of the intel_context's barrier, i.e. by listening to the pulse after the notional unpin. So wait until the intel_context is finally retired before releasing the engine, so that we can inspect the final context state and pass it on. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> --- drivers/gpu/drm/i915/gem/i915_gem_context.c | 16 +++--- drivers/gpu/drm/i915/i915_active.c | 57 ++++++++++++++++++++- drivers/gpu/drm/i915/i915_active.h | 3 +- 3 files changed, 64 insertions(+), 12 deletions(-)