From e85ade1f50aae464ce196672faa7a099fd1721ed Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris@chris-wilson.co.uk>
Date: Wed, 18 Dec 2019 10:40:43 +0000
Subject: [PATCH] drm/i915: Hold reference to intel_frontbuffer as we track activity
Git-commit: e85ade1f50aae464ce196672faa7a099fd1721ed
Patch-mainline: v5.5-rc4
No-fix: da42104f589d979bbe402703fd836cec60befae1
References: git-fixes
Since obj->frontbuffer is no longer protected by the struct_mutex, as we
are processing the execbuf, it may be removed. Mark the
intel_frontbuffer as rcu protected, and so acquire a reference to
the struct as we track activity upon it.
Closes: https://gitlab.freedesktop.org/drm/intel/issues/827
Fixes: 8e7cb1799b4f ("drm/i915: Extract intel_frontbuffer active tracking")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: <stable@vger.kernel.org> # v5.4+
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20191218104043.3539458-1-chris@chris-wilson.co.uk
(cherry picked from commit da42104f589d979bbe402703fd836cec60befae1)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Acked-by: Takashi Iwai <tiwai@suse.de>
---
drivers/gpu/drm/i915/display/intel_display.c | 2 -
drivers/gpu/drm/i915/display/intel_frontbuffer.c | 14 +++------
drivers/gpu/drm/i915/display/intel_frontbuffer.h | 34 ++++++++++++++++++++---
drivers/gpu/drm/i915/display/intel_overlay.c | 17 ++++++++---
drivers/gpu/drm/i915/gem/i915_gem_clflush.c | 3 +-
drivers/gpu/drm/i915/gem/i915_gem_domain.c | 4 +-
drivers/gpu/drm/i915/gem/i915_gem_object.c | 26 ++++++++++++++++-
drivers/gpu/drm/i915/gem/i915_gem_object.h | 23 ++++++++++++++-
drivers/gpu/drm/i915/gem/i915_gem_object_types.h | 2 -
drivers/gpu/drm/i915/i915_gem.c | 10 +++---
drivers/gpu/drm/i915/i915_vma.c | 10 +++++-
11 files changed, 115 insertions(+), 30 deletions(-)
--- a/drivers/gpu/drm/i915/display/intel_display.c
+++ b/drivers/gpu/drm/i915/display/intel_display.c
@@ -14429,7 +14429,7 @@ intel_prepare_plane_fb(struct drm_plane
return ret;
fb_obj_bump_render_priority(obj);
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_DIRTYFB);
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_DIRTYFB);
if (!new_state->fence) { /* implicit fencing */
struct dma_fence *fence;
--- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c
+++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c
@@ -225,7 +225,7 @@ static void frontbuffer_release(struct k
spin_unlock(&to_i915(front->obj->base.dev)->fb_tracking.lock);
i915_gem_object_put(front->obj);
- kfree(front);
+ kfree_rcu(front, rcu);
}
struct intel_frontbuffer *
@@ -234,11 +234,7 @@ intel_frontbuffer_get(struct drm_i915_ge
struct drm_i915_private *i915 = to_i915(obj->base.dev);
struct intel_frontbuffer *front;
- spin_lock(&i915->fb_tracking.lock);
- front = obj->frontbuffer;
- if (front)
- kref_get(&front->ref);
- spin_unlock(&i915->fb_tracking.lock);
+ front = __intel_frontbuffer_get(obj);
if (front)
return front;
@@ -253,13 +249,13 @@ intel_frontbuffer_get(struct drm_i915_ge
frontbuffer_active, frontbuffer_retire);
spin_lock(&i915->fb_tracking.lock);
- if (obj->frontbuffer) {
+ if (rcu_access_pointer(obj->frontbuffer)) {
kfree(front);
- front = obj->frontbuffer;
+ front = rcu_dereference_protected(obj->frontbuffer, true);
kref_get(&front->ref);
} else {
i915_gem_object_get(obj);
- obj->frontbuffer = front;
+ rcu_assign_pointer(obj->frontbuffer, front);
}
spin_unlock(&i915->fb_tracking.lock);
--- a/drivers/gpu/drm/i915/display/intel_frontbuffer.h
+++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.h
@@ -27,10 +27,10 @@
#include <linux/atomic.h>
#include <linux/kref.h>
+#include "gem/i915_gem_object_types.h"
#include "i915_active.h"
struct drm_i915_private;
-struct drm_i915_gem_object;
enum fb_op_origin {
ORIGIN_GTT,
@@ -45,6 +45,7 @@ struct intel_frontbuffer {
atomic_t bits;
struct i915_active write;
struct drm_i915_gem_object *obj;
+ struct rcu_head rcu;
};
void intel_frontbuffer_flip_prepare(struct drm_i915_private *i915,
@@ -54,6 +55,35 @@ void intel_frontbuffer_flip_complete(str
void intel_frontbuffer_flip(struct drm_i915_private *i915,
unsigned frontbuffer_bits);
+void intel_frontbuffer_put(struct intel_frontbuffer *front);
+
+static inline struct intel_frontbuffer *
+__intel_frontbuffer_get(const struct drm_i915_gem_object *obj)
+{
+ struct intel_frontbuffer *front;
+
+ if (likely(!rcu_access_pointer(obj->frontbuffer)))
+ return NULL;
+
+ rcu_read_lock();
+ do {
+ front = rcu_dereference(obj->frontbuffer);
+ if (!front)
+ break;
+
+ if (unlikely(!kref_get_unless_zero(&front->ref)))
+ continue;
+
+ if (likely(front == rcu_access_pointer(obj->frontbuffer)))
+ break;
+
+ intel_frontbuffer_put(front);
+ } while (1);
+ rcu_read_unlock();
+
+ return front;
+}
+
struct intel_frontbuffer *
intel_frontbuffer_get(struct drm_i915_gem_object *obj);
@@ -119,6 +149,4 @@ void intel_frontbuffer_track(struct inte
struct intel_frontbuffer *new,
unsigned int frontbuffer_bits);
-void intel_frontbuffer_put(struct intel_frontbuffer *front);
-
#endif /* __INTEL_FRONTBUFFER_H__ */
--- a/drivers/gpu/drm/i915/display/intel_overlay.c
+++ b/drivers/gpu/drm/i915/display/intel_overlay.c
@@ -278,12 +278,21 @@ static void intel_overlay_flip_prepare(s
struct i915_vma *vma)
{
enum pipe pipe = overlay->crtc->pipe;
+ struct intel_frontbuffer *from = NULL, *to = NULL;
WARN_ON(overlay->old_vma);
- intel_frontbuffer_track(overlay->vma ? overlay->vma->obj->frontbuffer : NULL,
- vma ? vma->obj->frontbuffer : NULL,
- INTEL_FRONTBUFFER_OVERLAY(pipe));
+ if (overlay->vma)
+ from = intel_frontbuffer_get(overlay->vma->obj);
+ if (vma)
+ to = intel_frontbuffer_get(vma->obj);
+
+ intel_frontbuffer_track(from, to, INTEL_FRONTBUFFER_OVERLAY(pipe));
+
+ if (to)
+ intel_frontbuffer_put(to);
+ if (from)
+ intel_frontbuffer_put(from);
intel_frontbuffer_flip_prepare(overlay->i915,
INTEL_FRONTBUFFER_OVERLAY(pipe));
@@ -768,7 +777,7 @@ static int intel_overlay_do_put_image(st
ret = PTR_ERR(vma);
goto out_pin_section;
}
- intel_frontbuffer_flush(new_bo->frontbuffer, ORIGIN_DIRTYFB);
+ i915_gem_object_flush_frontbuffer(new_bo, ORIGIN_DIRTYFB);
if (!overlay->active) {
u32 oconfig;
--- a/drivers/gpu/drm/i915/gem/i915_gem_clflush.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_clflush.c
@@ -20,7 +20,8 @@ static void __do_clflush(struct drm_i915
{
GEM_BUG_ON(!i915_gem_object_has_pages(obj));
drm_clflush_sg(obj->mm.pages);
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_CPU);
+
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
}
static int clflush_work(struct dma_fence_work *base)
--- a/drivers/gpu/drm/i915/gem/i915_gem_domain.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_domain.c
@@ -674,7 +674,7 @@ i915_gem_set_domain_ioctl(struct drm_dev
i915_gem_object_unlock(obj);
if (write_domain)
- intel_frontbuffer_invalidate(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
out_unpin:
i915_gem_object_unpin_pages(obj);
@@ -794,7 +794,7 @@ int i915_gem_object_prepare_write(struct
}
out:
- intel_frontbuffer_invalidate(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
obj->mm.dirty = true;
/* return with the pages pinned */
return 0;
--- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
@@ -270,7 +270,7 @@ i915_gem_object_flush_write_domain(struc
for_each_ggtt_vma(vma, obj)
intel_gt_flush_ggtt_writes(vma->vm->gt);
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
for_each_ggtt_vma(vma, obj) {
if (vma->iomap)
@@ -298,6 +298,30 @@ i915_gem_object_flush_write_domain(struc
obj->write_domain = 0;
}
+void __i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin)
+{
+ struct intel_frontbuffer *front;
+
+ front = __intel_frontbuffer_get(obj);
+ if (front) {
+ intel_frontbuffer_flush(front, origin);
+ intel_frontbuffer_put(front);
+ }
+}
+
+void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin)
+{
+ struct intel_frontbuffer *front;
+
+ front = __intel_frontbuffer_get(obj);
+ if (front) {
+ intel_frontbuffer_invalidate(front, origin);
+ intel_frontbuffer_put(front);
+ }
+}
+
void i915_gem_init__objects(struct drm_i915_private *i915)
{
INIT_WORK(&i915->mm.free_work, __i915_gem_free_work);
--- a/drivers/gpu/drm/i915/gem/i915_gem_object.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h
@@ -13,8 +13,8 @@
#include <drm/i915_drm.h>
+#include "display/intel_frontbuffer.h"
#include "i915_gem_object_types.h"
-
#include "i915_gem_gtt.h"
void i915_gem_init__objects(struct drm_i915_private *i915);
@@ -431,4 +431,25 @@ int i915_gem_object_wait_priority(struct
const struct i915_sched_attr *attr);
#define I915_PRIORITY_DISPLAY I915_USER_PRIORITY(I915_PRIORITY_MAX)
+void __i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin);
+void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin);
+
+static inline void
+i915_gem_object_flush_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin)
+{
+ if (unlikely(rcu_access_pointer(obj->frontbuffer)))
+ __i915_gem_object_flush_frontbuffer(obj, origin);
+}
+
+static inline void
+i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
+ enum fb_op_origin origin)
+{
+ if (unlikely(rcu_access_pointer(obj->frontbuffer)))
+ __i915_gem_object_invalidate_frontbuffer(obj, origin);
+}
+
#endif
--- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h
@@ -143,7 +143,7 @@ struct drm_i915_gem_object {
*/
u16 write_domain;
- struct intel_frontbuffer *frontbuffer;
+ struct intel_frontbuffer __rcu *frontbuffer;
/** Current tiling stride for the object, if it's tiled. */
unsigned int tiling_and_stride;
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -143,7 +143,7 @@ i915_gem_phys_pwrite(struct drm_i915_gem
* We manually control the domain here and pretend that it
* remains coherent i.e. in the GTT domain, like shmem_pwrite.
*/
- intel_frontbuffer_invalidate(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
if (copy_from_user(vaddr, user_data, args->size))
return -EFAULT;
@@ -151,7 +151,7 @@ i915_gem_phys_pwrite(struct drm_i915_gem
drm_clflush_virt_range(vaddr, args->size);
intel_gt_chipset_flush(&to_i915(obj->base.dev)->gt);
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
return 0;
}
@@ -588,7 +588,7 @@ i915_gem_gtt_pwrite_fast(struct drm_i915
goto out_unpin;
}
- intel_frontbuffer_invalidate(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_invalidate_frontbuffer(obj, ORIGIN_CPU);
user_data = u64_to_user_ptr(args->data_ptr);
offset = args->offset;
@@ -630,7 +630,7 @@ i915_gem_gtt_pwrite_fast(struct drm_i915
user_data += page_length;
offset += page_length;
}
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
i915_gem_object_unlock_fence(obj, fence);
out_unpin:
@@ -723,7 +723,7 @@ i915_gem_shmem_pwrite(struct drm_i915_ge
offset = 0;
}
- intel_frontbuffer_flush(obj->frontbuffer, ORIGIN_CPU);
+ i915_gem_object_flush_frontbuffer(obj, ORIGIN_CPU);
i915_gem_object_unlock_fence(obj, fence);
return ret;
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -907,8 +907,14 @@ int i915_vma_move_to_active(struct i915_
return err;
if (flags & EXEC_OBJECT_WRITE) {
- if (intel_frontbuffer_invalidate(obj->frontbuffer, ORIGIN_CS))
- i915_active_add_request(&obj->frontbuffer->write, rq);
+ struct intel_frontbuffer *front;
+
+ front = __intel_frontbuffer_get(obj);
+ if (unlikely(front)) {
+ if (intel_frontbuffer_invalidate(front, ORIGIN_CS))
+ i915_active_add_request(&front->write, rq);
+ intel_frontbuffer_put(front);
+ }
reservation_object_add_excl_fence(vma->resv, &rq->fence);
obj->write_domain = I915_GEM_DOMAIN_RENDER;