Re: [Intel-gfx] [RFC 04/12] drm/i915: Use explicit flag to mark unreachable intel_context

[Chris Wilson]

Quoting Tvrtko Ursulin (2020-03-09 18:31:21)
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c 
> b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> index 0893ce781a84..0302757396d5 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
> @@ -2547,8 +2547,8 @@ static void eb_request_add(struct i915_execbuffer *eb)
> prev = __i915_request_commit(rq);
>  
> /* Check that the context wasn't destroyed before submission */
> -   if (likely(rcu_access_pointer(eb->context->gem_context))) {
> -   attr = eb->gem_context->sched;
> +   if (likely(!READ_ONCE(eb->context->closed))) {
> +   attr = rcu_dereference(eb->gem_context)->sched;

That's the warn. We don't have a rcu_read_lock here so it complains.

eb->gem_context is a strong ref, no rcu markup require.
(it's the eb->context->gem_context that needs annotation)
-Chris
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[Intel-gfx] [RFC 04/12] drm/i915: Use explicit flag to mark unreachable intel_context

[Tvrtko Ursulin]

From: Tvrtko Ursulin 

I need to keep the GEM context around a bit longer so adding an explicit
flag for syncing execbuf with closed/abandonded contexts.

Signed-off-by: Tvrtko Ursulin 
---
 drivers/gpu/drm/i915/gem/i915_gem_context.c| 3 ++-
 drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 4 ++--
 drivers/gpu/drm/i915/gt/intel_context_types.h  | 1 +
 3 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c 
b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index 0f4150c8d7fe..abc3a3e2fcf1 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -568,7 +568,8 @@ static void engines_idle_release(struct i915_gem_context 
*ctx,
int err = 0;
 
/* serialises with execbuf */
-   RCU_INIT_POINTER(ce->gem_context, NULL);
+   smp_store_mb(ce->closed, true);
+
if (!intel_context_pin_if_active(ce))
continue;
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c 
b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index 0893ce781a84..0302757396d5 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -2547,8 +2547,8 @@ static void eb_request_add(struct i915_execbuffer *eb)
prev = __i915_request_commit(rq);
 
/* Check that the context wasn't destroyed before submission */
-   if (likely(rcu_access_pointer(eb->context->gem_context))) {
-   attr = eb->gem_context->sched;
+   if (likely(!READ_ONCE(eb->context->closed))) {
+   attr = rcu_dereference(eb->gem_context)->sched;
 
/*
 * Boost actual workloads past semaphores!
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h 
b/drivers/gpu/drm/i915/gt/intel_context_types.h
index 11278343b9b5..c60490e756f9 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -50,6 +50,7 @@ struct intel_context {
 
struct i915_address_space *vm;
struct i915_gem_context __rcu *gem_context;
+   bool closed;
 
struct list_head signal_link;
struct list_head signals;
-- 
2.20.1

___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx