drm/i915/ttm: handle blitter failure on DG2

If the move or clear operation somehow fails, and the memory underneath
is not cleared, like when moving to lmem, then we currently fallback to
memcpy or memset. However with small-BAR systems this fallback might no
longer be possible. For now we use the set_wedged sledgehammer if we
ever encounter such a scenario, and mark the object as borked to plug
any holes where access to the memory underneath can happen. Add some
basic selftests to exercise this.

v2:
  - In the selftests make sure we grab the runtime pm around the reset.
    Also make sure we grab the reset lock before checking if the device
    is wedged, since the wedge might still be in-progress and hence the
    bit might not be set yet.
  - Don't wedge or put the object into an unknown state, if the request
    construction fails (or similar). Just returning an error and
    skipping the fallback should be safe here.
  - Make sure we wedge each gt. (Thomas)
  - Peek at the unknown_state in io_reserve, that way we don't have to
    export or hand roll the fault_wait_for_idle. (Thomas)
  - Add the missing read-side barriers for the unknown_state. (Thomas)
  - Some kernel-doc fixes. (Thomas)
v3:
  - Tweak the ordering of the set_wedged, also add FIXME.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Jon Bloomfield <jon.bloomfield@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jordan Justen <jordan.l.justen@intel.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220629174350.384910-11-matthew.auld@intel.com

1 files changed, 13 insertions, 12 deletions

diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index 5d5828b9a242..43339ecabd73 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -310,7 +310,7 @@ struct i915_vma_work {
 	struct i915_address_space *vm;
 	struct i915_vm_pt_stash stash;
 	struct i915_vma_resource *vma_res;
-	struct drm_i915_gem_object *pinned;
+	struct drm_i915_gem_object *obj;
 	struct i915_sw_dma_fence_cb cb;
 	enum i915_cache_level cache_level;
 	unsigned int flags;
@@ -321,17 +321,25 @@ static void __vma_bind(struct dma_fence_work *work)
 	struct i915_vma_work *vw = container_of(work, typeof(*vw), base);
 	struct i915_vma_resource *vma_res = vw->vma_res;
 
+	/*
+	 * We are about the bind the object, which must mean we have already
+	 * signaled the work to potentially clear/move the pages underneath. If
+	 * something went wrong at that stage then the object should have
+	 * unknown_state set, in which case we need to skip the bind.
+	 */
+	if (i915_gem_object_has_unknown_state(vw->obj))
+		return;
+
 	vma_res->ops->bind_vma(vma_res->vm, &vw->stash,
 			       vma_res, vw->cache_level, vw->flags);
-
 }
 
 static void __vma_release(struct dma_fence_work *work)
 {
 	struct i915_vma_work *vw = container_of(work, typeof(*vw), base);
 
-	if (vw->pinned)
-		i915_gem_object_put(vw->pinned);
+	if (vw->obj)
+		i915_gem_object_put(vw->obj);
 
 	i915_vm_free_pt_stash(vw->vm, &vw->stash);
 	if (vw->vma_res)
@@ -517,14 +525,7 @@ int i915_vma_bind(struct i915_vma *vma,
 		}
 
 		work->base.dma.error = 0; /* enable the queue_work() */
-
-		/*
-		 * If we don't have the refcounted pages list, keep a reference
-		 * on the object to avoid waiting for the async bind to
-		 * complete in the object destruction path.
-		 */
-		if (!work->vma_res->bi.pages_rsgt)
-			work->pinned = i915_gem_object_get(vma->obj);
+		work->obj = i915_gem_object_get(vma->obj);
 	} else {
 		ret = i915_gem_object_wait_moving_fence(vma->obj, true);
 		if (ret) {