[PATCH v4 0/4] drm/i915: Asynchronous vma unbinding part1

[PATCH v4 0/4] drm/i915: Asynchronous vma unbinding part1

[Thomas Hellström]

This is the first three already reviewed patches from the patch series titled
"Asynchronous vma unbinding", with an additional cleanup patch from Christian,
which would otherwise conflict heavily with this series.

Christian König (1):
  drm/i915: remove questionable fence optimization during copy

Thomas Hellström (3):
  drm/i915: Avoid using the i915_fence_array when collecting
    dependencies
  drm/i915: Break out the i915_deps utility
  drm/i915: Require the vm mutex for i915_vma_bind()

 drivers/gpu/drm/i915/Makefile                 |   1 +
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    |  50 ++-
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c  | 323 +++---------------
 drivers/gpu/drm/i915/gt/intel_migrate.c       |  24 +-
 drivers/gpu/drm/i915/gt/intel_migrate.h       |   9 +-
 drivers/gpu/drm/i915/i915_deps.c              | 237 +++++++++++++
 drivers/gpu/drm/i915/i915_deps.h              |  45 +++
 drivers/gpu/drm/i915/i915_request.c           |  22 ++
 drivers/gpu/drm/i915/i915_request.h           |   2 +
 drivers/gpu/drm/i915/i915_vma.c               |   1 +
 10 files changed, 412 insertions(+), 302 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/i915_deps.c
 create mode 100644 drivers/gpu/drm/i915/i915_deps.h

-- 
2.31.1

[Intel-gfx] [PATCH v4 0/4] drm/i915: Asynchronous vma unbinding part1

[Thomas Hellström]

This is the first three already reviewed patches from the patch series titled
"Asynchronous vma unbinding", with an additional cleanup patch from Christian,
which would otherwise conflict heavily with this series.

Christian König (1):
  drm/i915: remove questionable fence optimization during copy

Thomas Hellström (3):
  drm/i915: Avoid using the i915_fence_array when collecting
    dependencies
  drm/i915: Break out the i915_deps utility
  drm/i915: Require the vm mutex for i915_vma_bind()

 drivers/gpu/drm/i915/Makefile                 |   1 +
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    |  50 ++-
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c  | 323 +++---------------
 drivers/gpu/drm/i915/gt/intel_migrate.c       |  24 +-
 drivers/gpu/drm/i915/gt/intel_migrate.h       |   9 +-
 drivers/gpu/drm/i915/i915_deps.c              | 237 +++++++++++++
 drivers/gpu/drm/i915/i915_deps.h              |  45 +++
 drivers/gpu/drm/i915/i915_request.c           |  22 ++
 drivers/gpu/drm/i915/i915_request.h           |   2 +
 drivers/gpu/drm/i915/i915_vma.c               |   1 +
 10 files changed, 412 insertions(+), 302 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/i915_deps.c
 create mode 100644 drivers/gpu/drm/i915/i915_deps.h

-- 
2.31.1

[PATCH v4 1/4] drm/i915: Avoid using the i915_fence_array when collecting dependencies

[Thomas Hellström]

Since the gt migration code was using only a single fence for
dependencies, these were collected in a dma_fence_array. However, it
turns out that it's illegal to use some dma_fences in a dma_fence_array,
in particular other dma_fence_arrays and dma_fence_chains, and this
causes trouble for us moving forward.

Have the gt migration code instead take a const struct i915_deps for
dependencies. This means we can skip the dma_fence_array creation
and instead pass the struct i915_deps instead to circumvent the
problem.

v2:
- Make the prev_deps() function static. (kernel test robot <lkp@intel.com>)
- Update the struct i915_deps kerneldoc.
v4:
- Rebase.

Fixes: 5652df829b3c ("drm/i915/ttm: Update i915_gem_obj_copy_ttm() to be asynchronous")
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 129 +++++--------------
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h |  17 +++
 drivers/gpu/drm/i915/gt/intel_migrate.c      |  24 ++--
 drivers/gpu/drm/i915/gt/intel_migrate.h      |   9 +-
 drivers/gpu/drm/i915/i915_request.c          |  22 ++++
 drivers/gpu/drm/i915/i915_request.h          |   2 +
 6 files changed, 91 insertions(+), 112 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 8ad09fcf3698..62a6fd2d127d 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -3,8 +3,6 @@
  * Copyright © 2021 Intel Corporation
  */
 
-#include <linux/dma-fence-array.h>
-
 #include <drm/ttm/ttm_bo_driver.h>
 
 #include "i915_drv.h"
@@ -44,17 +42,12 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
 #endif
 
 /**
- * DOC: Set of utilities to dynamically collect dependencies and
- * eventually coalesce them into a single fence which is fed into
- * the GT migration code, since it only accepts a single dependency
- * fence.
- * The single fence returned from these utilities, in the case of
- * dependencies from multiple fence contexts, a struct dma_fence_array,
- * since the i915 request code can break that up and await the individual
- * fences.
+ * DOC: Set of utilities to dynamically collect dependencies into a
+ * structure which is fed into the GT migration code.
  *
  * Once we can do async unbinding, this is also needed to coalesce
- * the migration fence with the unbind fences.
+ * the migration fence with the unbind fences if these are coalesced
+ * post-migration.
  *
  * While collecting the individual dependencies, we store the refcounted
  * struct dma_fence pointers in a realloc-managed pointer array, since
@@ -65,32 +58,13 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
  * A struct i915_deps need to be initialized using i915_deps_init().
  * If i915_deps_add_dependency() or i915_deps_add_resv() return an
  * error code they will internally call i915_deps_fini(), which frees
- * all internal references and allocations. After a call to
- * i915_deps_to_fence(), or i915_deps_sync(), the struct should similarly
- * be viewed as uninitialized.
+ * all internal references and allocations.
  *
  * We might want to break this out into a separate file as a utility.
  */
 
 #define I915_DEPS_MIN_ALLOC_CHUNK 8U
 
-/**
- * struct i915_deps - Collect dependencies into a single dma-fence
- * @single: Storage for pointer if the collection is a single fence.
- * @fence: Allocated array of fence pointers if more than a single fence;
- * otherwise points to the address of @single.
- * @num_deps: Current number of dependency fences.
- * @fences_size: Size of the @fences array in number of pointers.
- * @gfp: Allocation mode.
- */
-struct i915_deps {
-	struct dma_fence *single;
-	struct dma_fence **fences;
-	unsigned int num_deps;
-	unsigned int fences_size;
-	gfp_t gfp;
-};
-
 static void i915_deps_reset_fences(struct i915_deps *deps)
 {
 	if (deps->fences != &deps->single)
@@ -163,7 +137,7 @@ static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
 	return ret;
 }
 
-static int i915_deps_sync(struct i915_deps *deps,
+static int i915_deps_sync(const struct i915_deps *deps,
 			  const struct ttm_operation_ctx *ctx)
 {
 	struct dma_fence **fences = deps->fences;
@@ -183,7 +157,6 @@ static int i915_deps_sync(struct i915_deps *deps,
 			break;
 	}
 
-	i915_deps_fini(deps);
 	return ret;
 }
 
@@ -221,34 +194,6 @@ static int i915_deps_add_dependency(struct i915_deps *deps,
 	return i915_deps_grow(deps, fence, ctx);
 }
 
-static struct dma_fence *i915_deps_to_fence(struct i915_deps *deps,
-					    const struct ttm_operation_ctx *ctx)
-{
-	struct dma_fence_array *array;
-
-	if (deps->num_deps == 0)
-		return NULL;
-
-	if (deps->num_deps == 1) {
-		deps->num_deps = 0;
-		return deps->fences[0];
-	}
-
-	/*
-	 * TODO: Alter the allocation mode here to not try too hard to
-	 * make things async.
-	 */
-	array = dma_fence_array_create(deps->num_deps, deps->fences, 0, 0,
-				       false);
-	if (!array)
-		return ERR_PTR(i915_deps_sync(deps, ctx));
-
-	deps->fences = NULL;
-	i915_deps_reset_fences(deps);
-
-	return &array->base;
-}
-
 static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
 			      bool all, const bool no_excl,
 			      const struct ttm_operation_ctx *ctx)
@@ -387,7 +332,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 					     struct ttm_resource *dst_mem,
 					     struct ttm_tt *dst_ttm,
 					     struct sg_table *dst_st,
-					     struct dma_fence *dep)
+					     const struct i915_deps *deps)
 {
 	struct drm_i915_private *i915 = container_of(bo->bdev, typeof(*i915),
 						     bdev);
@@ -411,7 +356,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 			return ERR_PTR(-EINVAL);
 
 		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
-		ret = intel_context_migrate_clear(to_gt(i915)->migrate.context, dep,
+		ret = intel_context_migrate_clear(to_gt(i915)->migrate.context, deps,
 						  dst_st->sgl, dst_level,
 						  i915_ttm_gtt_binds_lmem(dst_mem),
 						  0, &rq);
@@ -425,7 +370,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 		src_level = i915_ttm_cache_level(i915, bo->resource, src_ttm);
 		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
 		ret = intel_context_migrate_copy(to_gt(i915)->migrate.context,
-						 dep, src_rsgt->table.sgl,
+						 deps, src_rsgt->table.sgl,
 						 src_level,
 						 i915_ttm_gtt_binds_lmem(bo->resource),
 						 dst_st->sgl, dst_level,
@@ -610,10 +555,11 @@ i915_ttm_memcpy_work_arm(struct i915_ttm_memcpy_work *work,
 }
 
 static struct dma_fence *
-__i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
+__i915_ttm_move(struct ttm_buffer_object *bo,
+		const struct ttm_operation_ctx *ctx, bool clear,
 		struct ttm_resource *dst_mem, struct ttm_tt *dst_ttm,
 		struct i915_refct_sgt *dst_rsgt, bool allow_accel,
-		struct dma_fence *move_dep)
+		const struct i915_deps *move_deps)
 {
 	struct i915_ttm_memcpy_work *copy_work = NULL;
 	struct i915_ttm_memcpy_arg _arg, *arg = &_arg;
@@ -621,7 +567,7 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 
 	if (allow_accel) {
 		fence = i915_ttm_accel_move(bo, clear, dst_mem, dst_ttm,
-					    &dst_rsgt->table, move_dep);
+					    &dst_rsgt->table, move_deps);
 
 		/*
 		 * We only need to intercept the error when moving to lmem.
@@ -655,8 +601,8 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 
 		if (!IS_ERR(fence))
 			goto out;
-	} else if (move_dep) {
-		int err = dma_fence_wait(move_dep, true);
+	} else if (move_deps) {
+		int err = i915_deps_sync(move_deps, ctx);
 
 		if (err)
 			return ERR_PTR(err);
@@ -680,29 +626,21 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 	return fence;
 }
 
-static struct dma_fence *prev_fence(struct ttm_buffer_object *bo,
-				    struct ttm_operation_ctx *ctx)
+static int
+prev_deps(struct ttm_buffer_object *bo, struct ttm_operation_ctx *ctx,
+	  struct i915_deps *deps)
 {
-	struct i915_deps deps;
 	int ret;
 
-	/*
-	 * Instead of trying hard with GFP_KERNEL to allocate memory,
-	 * the dependency collection will just sync if it doesn't
-	 * succeed.
-	 */
-	i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
-	ret = i915_deps_add_dependency(&deps, bo->moving, ctx);
+	ret = i915_deps_add_dependency(deps, bo->moving, ctx);
 	if (!ret)
 		/*
 		 * TODO: Only await excl fence here, and shared fences before
 		 * signaling the migration fence.
 		 */
-		ret = i915_deps_add_resv(&deps, bo->base.resv, true, false, ctx);
-	if (ret)
-		return ERR_PTR(ret);
+		ret = i915_deps_add_resv(deps, bo->base.resv, true, false, ctx);
 
-	return i915_deps_to_fence(&deps, ctx);
+	return ret;
 }
 
 /**
@@ -756,16 +694,18 @@ int i915_ttm_move(struct ttm_buffer_object *bo, bool evict,
 
 	clear = !i915_ttm_cpu_maps_iomem(bo->resource) && (!ttm || !ttm_tt_is_populated(ttm));
 	if (!(clear && ttm && !(ttm->page_flags & TTM_TT_FLAG_ZERO_ALLOC))) {
-		struct dma_fence *dep = prev_fence(bo, ctx);
+		struct i915_deps deps;
 
-		if (IS_ERR(dep)) {
+		i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
+		ret = prev_deps(bo, ctx, &deps);
+		if (ret) {
 			i915_refct_sgt_put(dst_rsgt);
-			return PTR_ERR(dep);
+			return ret;
 		}
 
-		migration_fence = __i915_ttm_move(bo, clear, dst_mem, bo->ttm,
-						  dst_rsgt, true, dep);
-		dma_fence_put(dep);
+		migration_fence = __i915_ttm_move(bo, ctx, clear, dst_mem, bo->ttm,
+						  dst_rsgt, true, &deps);
+		i915_deps_fini(&deps);
 	}
 
 	/* We can possibly get an -ERESTARTSYS here */
@@ -826,7 +766,7 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 		.interruptible = intr,
 	};
 	struct i915_refct_sgt *dst_rsgt;
-	struct dma_fence *copy_fence, *dep_fence;
+	struct dma_fence *copy_fence;
 	struct i915_deps deps;
 	int ret, shared_err;
 
@@ -847,15 +787,12 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 	if (ret)
 		return ret;
 
-	dep_fence = i915_deps_to_fence(&deps, &ctx);
-	if (IS_ERR(dep_fence))
-		return PTR_ERR(dep_fence);
-
 	dst_rsgt = i915_ttm_resource_get_st(dst, dst_bo->resource);
-	copy_fence = __i915_ttm_move(src_bo, false, dst_bo->resource,
+	copy_fence = __i915_ttm_move(src_bo, &ctx, false, dst_bo->resource,
 				     dst_bo->ttm, dst_rsgt, allow_accel,
-				     dep_fence);
+				     &deps);
 
+	i915_deps_fini(&deps);
 	i915_refct_sgt_put(dst_rsgt);
 	if (IS_ERR_OR_NULL(copy_fence))
 		return PTR_ERR_OR_ZERO(copy_fence);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
index d2e7f149e05c..138b7647a558 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
@@ -18,6 +18,23 @@ struct ttm_tt;
 struct drm_i915_gem_object;
 struct i915_refct_sgt;
 
+/**
+ * struct i915_deps - Collect dependencies into a single dma-fence
+ * @single: Storage for pointer if the collection is a single fence.
+ * @fences: Allocated array of fence pointers if more than a single fence;
+ * otherwise points to the address of @single.
+ * @num_deps: Current number of dependency fences.
+ * @fences_size: Size of the @fences array in number of pointers.
+ * @gfp: Allocation mode.
+ */
+struct i915_deps {
+	struct dma_fence *single;
+	struct dma_fence **fences;
+	unsigned int num_deps;
+	unsigned int fences_size;
+	gfp_t gfp;
+};
+
 int i915_ttm_move_notify(struct ttm_buffer_object *bo);
 
 I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c
index 19a01878fee3..18b44af56969 100644
--- a/drivers/gpu/drm/i915/gt/intel_migrate.c
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.c
@@ -404,7 +404,7 @@ static int emit_copy(struct i915_request *rq, int size)
 
 int
 intel_context_migrate_copy(struct intel_context *ce,
-			   struct dma_fence *await,
+			   const struct i915_deps *deps,
 			   struct scatterlist *src,
 			   enum i915_cache_level src_cache_level,
 			   bool src_is_lmem,
@@ -431,8 +431,8 @@ intel_context_migrate_copy(struct intel_context *ce,
 			goto out_ce;
 		}
 
-		if (await) {
-			err = i915_request_await_dma_fence(rq, await);
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
 			if (err)
 				goto out_rq;
 
@@ -442,7 +442,7 @@ intel_context_migrate_copy(struct intel_context *ce,
 					goto out_rq;
 			}
 
-			await = NULL;
+			deps = NULL;
 		}
 
 		/* The PTE updates + copy must not be interrupted. */
@@ -525,7 +525,7 @@ static int emit_clear(struct i915_request *rq, int size, u32 value)
 
 int
 intel_context_migrate_clear(struct intel_context *ce,
-			    struct dma_fence *await,
+			    const struct i915_deps *deps,
 			    struct scatterlist *sg,
 			    enum i915_cache_level cache_level,
 			    bool is_lmem,
@@ -550,8 +550,8 @@ intel_context_migrate_clear(struct intel_context *ce,
 			goto out_ce;
 		}
 
-		if (await) {
-			err = i915_request_await_dma_fence(rq, await);
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
 			if (err)
 				goto out_rq;
 
@@ -561,7 +561,7 @@ intel_context_migrate_clear(struct intel_context *ce,
 					goto out_rq;
 			}
 
-			await = NULL;
+			deps = NULL;
 		}
 
 		/* The PTE updates + clear must not be interrupted. */
@@ -599,7 +599,7 @@ intel_context_migrate_clear(struct intel_context *ce,
 
 int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_gem_ww_ctx *ww,
-		       struct dma_fence *await,
+		       const struct i915_deps *deps,
 		       struct scatterlist *src,
 		       enum i915_cache_level src_cache_level,
 		       bool src_is_lmem,
@@ -624,7 +624,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 	if (err)
 		goto out;
 
-	err = intel_context_migrate_copy(ce, await,
+	err = intel_context_migrate_copy(ce, deps,
 					 src, src_cache_level, src_is_lmem,
 					 dst, dst_cache_level, dst_is_lmem,
 					 out);
@@ -638,7 +638,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 int
 intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_gem_ww_ctx *ww,
-		    struct dma_fence *await,
+		    const struct i915_deps *deps,
 		    struct scatterlist *sg,
 		    enum i915_cache_level cache_level,
 		    bool is_lmem,
@@ -661,7 +661,7 @@ intel_migrate_clear(struct intel_migrate *m,
 	if (err)
 		goto out;
 
-	err = intel_context_migrate_clear(ce, await, sg, cache_level,
+	err = intel_context_migrate_clear(ce, deps, sg, cache_level,
 					  is_lmem, value, out);
 
 	intel_context_unpin(ce);
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.h b/drivers/gpu/drm/i915/gt/intel_migrate.h
index 4e18e755a00b..ccc677ec4aa3 100644
--- a/drivers/gpu/drm/i915/gt/intel_migrate.h
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.h
@@ -11,6 +11,7 @@
 #include "intel_migrate_types.h"
 
 struct dma_fence;
+struct i915_deps;
 struct i915_request;
 struct i915_gem_ww_ctx;
 struct intel_gt;
@@ -23,7 +24,7 @@ struct intel_context *intel_migrate_create_context(struct intel_migrate *m);
 
 int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_gem_ww_ctx *ww,
-		       struct dma_fence *await,
+		       const struct i915_deps *deps,
 		       struct scatterlist *src,
 		       enum i915_cache_level src_cache_level,
 		       bool src_is_lmem,
@@ -33,7 +34,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_request **out);
 
 int intel_context_migrate_copy(struct intel_context *ce,
-			       struct dma_fence *await,
+			       const struct i915_deps *deps,
 			       struct scatterlist *src,
 			       enum i915_cache_level src_cache_level,
 			       bool src_is_lmem,
@@ -45,7 +46,7 @@ int intel_context_migrate_copy(struct intel_context *ce,
 int
 intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_gem_ww_ctx *ww,
-		    struct dma_fence *await,
+		    const struct i915_deps *deps,
 		    struct scatterlist *sg,
 		    enum i915_cache_level cache_level,
 		    bool is_lmem,
@@ -53,7 +54,7 @@ intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_request **out);
 int
 intel_context_migrate_clear(struct intel_context *ce,
-			    struct dma_fence *await,
+			    const struct i915_deps *deps,
 			    struct scatterlist *sg,
 			    enum i915_cache_level cache_level,
 			    bool is_lmem,
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 40cf8ec0b9fc..7d804df27546 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -32,6 +32,7 @@
 #include <linux/sched/mm.h>
 
 #include "gem/i915_gem_context.h"
+#include "gem/i915_gem_ttm_move.h"
 #include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine.h"
@@ -1543,6 +1544,27 @@ i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
 	return 0;
 }
 
+/**
+ * i915_request_await_deps - set this request to (async) wait upon a struct
+ * i915_deps dma_fence collection
+ * @rq: request we are wishing to use
+ * @deps: The struct i915_deps containing the dependencies.
+ *
+ * Returns 0 if successful, negative error code on error.
+ */
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps)
+{
+	int i, err;
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		err = i915_request_await_dma_fence(rq, deps->fences[i]);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
 /**
  * i915_request_await_object - set this request to (async) wait upon a bo
  * @to: request we are wishing to use
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
index ce7714210697..170ee78c2858 100644
--- a/drivers/gpu/drm/i915/i915_request.h
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -47,6 +47,7 @@
 struct drm_file;
 struct drm_i915_gem_object;
 struct drm_printer;
+struct i915_deps;
 struct i915_request;
 
 #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
@@ -411,6 +412,7 @@ int i915_request_await_object(struct i915_request *to,
 			      bool write);
 int i915_request_await_dma_fence(struct i915_request *rq,
 				 struct dma_fence *fence);
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps);
 int i915_request_await_execution(struct i915_request *rq,
 				 struct dma_fence *fence);
 
-- 
2.31.1

[Intel-gfx] [PATCH v4 1/4] drm/i915: Avoid using the i915_fence_array when collecting dependencies

[Thomas Hellström]

Since the gt migration code was using only a single fence for
dependencies, these were collected in a dma_fence_array. However, it
turns out that it's illegal to use some dma_fences in a dma_fence_array,
in particular other dma_fence_arrays and dma_fence_chains, and this
causes trouble for us moving forward.

Have the gt migration code instead take a const struct i915_deps for
dependencies. This means we can skip the dma_fence_array creation
and instead pass the struct i915_deps instead to circumvent the
problem.

v2:
- Make the prev_deps() function static. (kernel test robot <lkp@intel.com>)
- Update the struct i915_deps kerneldoc.
v4:
- Rebase.

Fixes: 5652df829b3c ("drm/i915/ttm: Update i915_gem_obj_copy_ttm() to be asynchronous")
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 129 +++++--------------
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h |  17 +++
 drivers/gpu/drm/i915/gt/intel_migrate.c      |  24 ++--
 drivers/gpu/drm/i915/gt/intel_migrate.h      |   9 +-
 drivers/gpu/drm/i915/i915_request.c          |  22 ++++
 drivers/gpu/drm/i915/i915_request.h          |   2 +
 6 files changed, 91 insertions(+), 112 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 8ad09fcf3698..62a6fd2d127d 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -3,8 +3,6 @@
  * Copyright © 2021 Intel Corporation
  */
 
-#include <linux/dma-fence-array.h>
-
 #include <drm/ttm/ttm_bo_driver.h>
 
 #include "i915_drv.h"
@@ -44,17 +42,12 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
 #endif
 
 /**
- * DOC: Set of utilities to dynamically collect dependencies and
- * eventually coalesce them into a single fence which is fed into
- * the GT migration code, since it only accepts a single dependency
- * fence.
- * The single fence returned from these utilities, in the case of
- * dependencies from multiple fence contexts, a struct dma_fence_array,
- * since the i915 request code can break that up and await the individual
- * fences.
+ * DOC: Set of utilities to dynamically collect dependencies into a
+ * structure which is fed into the GT migration code.
  *
  * Once we can do async unbinding, this is also needed to coalesce
- * the migration fence with the unbind fences.
+ * the migration fence with the unbind fences if these are coalesced
+ * post-migration.
  *
  * While collecting the individual dependencies, we store the refcounted
  * struct dma_fence pointers in a realloc-managed pointer array, since
@@ -65,32 +58,13 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
  * A struct i915_deps need to be initialized using i915_deps_init().
  * If i915_deps_add_dependency() or i915_deps_add_resv() return an
  * error code they will internally call i915_deps_fini(), which frees
- * all internal references and allocations. After a call to
- * i915_deps_to_fence(), or i915_deps_sync(), the struct should similarly
- * be viewed as uninitialized.
+ * all internal references and allocations.
  *
  * We might want to break this out into a separate file as a utility.
  */
 
 #define I915_DEPS_MIN_ALLOC_CHUNK 8U
 
-/**
- * struct i915_deps - Collect dependencies into a single dma-fence
- * @single: Storage for pointer if the collection is a single fence.
- * @fence: Allocated array of fence pointers if more than a single fence;
- * otherwise points to the address of @single.
- * @num_deps: Current number of dependency fences.
- * @fences_size: Size of the @fences array in number of pointers.
- * @gfp: Allocation mode.
- */
-struct i915_deps {
-	struct dma_fence *single;
-	struct dma_fence **fences;
-	unsigned int num_deps;
-	unsigned int fences_size;
-	gfp_t gfp;
-};
-
 static void i915_deps_reset_fences(struct i915_deps *deps)
 {
 	if (deps->fences != &deps->single)
@@ -163,7 +137,7 @@ static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
 	return ret;
 }
 
-static int i915_deps_sync(struct i915_deps *deps,
+static int i915_deps_sync(const struct i915_deps *deps,
 			  const struct ttm_operation_ctx *ctx)
 {
 	struct dma_fence **fences = deps->fences;
@@ -183,7 +157,6 @@ static int i915_deps_sync(struct i915_deps *deps,
 			break;
 	}
 
-	i915_deps_fini(deps);
 	return ret;
 }
 
@@ -221,34 +194,6 @@ static int i915_deps_add_dependency(struct i915_deps *deps,
 	return i915_deps_grow(deps, fence, ctx);
 }
 
-static struct dma_fence *i915_deps_to_fence(struct i915_deps *deps,
-					    const struct ttm_operation_ctx *ctx)
-{
-	struct dma_fence_array *array;
-
-	if (deps->num_deps == 0)
-		return NULL;
-
-	if (deps->num_deps == 1) {
-		deps->num_deps = 0;
-		return deps->fences[0];
-	}
-
-	/*
-	 * TODO: Alter the allocation mode here to not try too hard to
-	 * make things async.
-	 */
-	array = dma_fence_array_create(deps->num_deps, deps->fences, 0, 0,
-				       false);
-	if (!array)
-		return ERR_PTR(i915_deps_sync(deps, ctx));
-
-	deps->fences = NULL;
-	i915_deps_reset_fences(deps);
-
-	return &array->base;
-}
-
 static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
 			      bool all, const bool no_excl,
 			      const struct ttm_operation_ctx *ctx)
@@ -387,7 +332,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 					     struct ttm_resource *dst_mem,
 					     struct ttm_tt *dst_ttm,
 					     struct sg_table *dst_st,
-					     struct dma_fence *dep)
+					     const struct i915_deps *deps)
 {
 	struct drm_i915_private *i915 = container_of(bo->bdev, typeof(*i915),
 						     bdev);
@@ -411,7 +356,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 			return ERR_PTR(-EINVAL);
 
 		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
-		ret = intel_context_migrate_clear(to_gt(i915)->migrate.context, dep,
+		ret = intel_context_migrate_clear(to_gt(i915)->migrate.context, deps,
 						  dst_st->sgl, dst_level,
 						  i915_ttm_gtt_binds_lmem(dst_mem),
 						  0, &rq);
@@ -425,7 +370,7 @@ static struct dma_fence *i915_ttm_accel_move(struct ttm_buffer_object *bo,
 		src_level = i915_ttm_cache_level(i915, bo->resource, src_ttm);
 		intel_engine_pm_get(to_gt(i915)->migrate.context->engine);
 		ret = intel_context_migrate_copy(to_gt(i915)->migrate.context,
-						 dep, src_rsgt->table.sgl,
+						 deps, src_rsgt->table.sgl,
 						 src_level,
 						 i915_ttm_gtt_binds_lmem(bo->resource),
 						 dst_st->sgl, dst_level,
@@ -610,10 +555,11 @@ i915_ttm_memcpy_work_arm(struct i915_ttm_memcpy_work *work,
 }
 
 static struct dma_fence *
-__i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
+__i915_ttm_move(struct ttm_buffer_object *bo,
+		const struct ttm_operation_ctx *ctx, bool clear,
 		struct ttm_resource *dst_mem, struct ttm_tt *dst_ttm,
 		struct i915_refct_sgt *dst_rsgt, bool allow_accel,
-		struct dma_fence *move_dep)
+		const struct i915_deps *move_deps)
 {
 	struct i915_ttm_memcpy_work *copy_work = NULL;
 	struct i915_ttm_memcpy_arg _arg, *arg = &_arg;
@@ -621,7 +567,7 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 
 	if (allow_accel) {
 		fence = i915_ttm_accel_move(bo, clear, dst_mem, dst_ttm,
-					    &dst_rsgt->table, move_dep);
+					    &dst_rsgt->table, move_deps);
 
 		/*
 		 * We only need to intercept the error when moving to lmem.
@@ -655,8 +601,8 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 
 		if (!IS_ERR(fence))
 			goto out;
-	} else if (move_dep) {
-		int err = dma_fence_wait(move_dep, true);
+	} else if (move_deps) {
+		int err = i915_deps_sync(move_deps, ctx);
 
 		if (err)
 			return ERR_PTR(err);
@@ -680,29 +626,21 @@ __i915_ttm_move(struct ttm_buffer_object *bo, bool clear,
 	return fence;
 }
 
-static struct dma_fence *prev_fence(struct ttm_buffer_object *bo,
-				    struct ttm_operation_ctx *ctx)
+static int
+prev_deps(struct ttm_buffer_object *bo, struct ttm_operation_ctx *ctx,
+	  struct i915_deps *deps)
 {
-	struct i915_deps deps;
 	int ret;
 
-	/*
-	 * Instead of trying hard with GFP_KERNEL to allocate memory,
-	 * the dependency collection will just sync if it doesn't
-	 * succeed.
-	 */
-	i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
-	ret = i915_deps_add_dependency(&deps, bo->moving, ctx);
+	ret = i915_deps_add_dependency(deps, bo->moving, ctx);
 	if (!ret)
 		/*
 		 * TODO: Only await excl fence here, and shared fences before
 		 * signaling the migration fence.
 		 */
-		ret = i915_deps_add_resv(&deps, bo->base.resv, true, false, ctx);
-	if (ret)
-		return ERR_PTR(ret);
+		ret = i915_deps_add_resv(deps, bo->base.resv, true, false, ctx);
 
-	return i915_deps_to_fence(&deps, ctx);
+	return ret;
 }
 
 /**
@@ -756,16 +694,18 @@ int i915_ttm_move(struct ttm_buffer_object *bo, bool evict,
 
 	clear = !i915_ttm_cpu_maps_iomem(bo->resource) && (!ttm || !ttm_tt_is_populated(ttm));
 	if (!(clear && ttm && !(ttm->page_flags & TTM_TT_FLAG_ZERO_ALLOC))) {
-		struct dma_fence *dep = prev_fence(bo, ctx);
+		struct i915_deps deps;
 
-		if (IS_ERR(dep)) {
+		i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
+		ret = prev_deps(bo, ctx, &deps);
+		if (ret) {
 			i915_refct_sgt_put(dst_rsgt);
-			return PTR_ERR(dep);
+			return ret;
 		}
 
-		migration_fence = __i915_ttm_move(bo, clear, dst_mem, bo->ttm,
-						  dst_rsgt, true, dep);
-		dma_fence_put(dep);
+		migration_fence = __i915_ttm_move(bo, ctx, clear, dst_mem, bo->ttm,
+						  dst_rsgt, true, &deps);
+		i915_deps_fini(&deps);
 	}
 
 	/* We can possibly get an -ERESTARTSYS here */
@@ -826,7 +766,7 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 		.interruptible = intr,
 	};
 	struct i915_refct_sgt *dst_rsgt;
-	struct dma_fence *copy_fence, *dep_fence;
+	struct dma_fence *copy_fence;
 	struct i915_deps deps;
 	int ret, shared_err;
 
@@ -847,15 +787,12 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 	if (ret)
 		return ret;
 
-	dep_fence = i915_deps_to_fence(&deps, &ctx);
-	if (IS_ERR(dep_fence))
-		return PTR_ERR(dep_fence);
-
 	dst_rsgt = i915_ttm_resource_get_st(dst, dst_bo->resource);
-	copy_fence = __i915_ttm_move(src_bo, false, dst_bo->resource,
+	copy_fence = __i915_ttm_move(src_bo, &ctx, false, dst_bo->resource,
 				     dst_bo->ttm, dst_rsgt, allow_accel,
-				     dep_fence);
+				     &deps);
 
+	i915_deps_fini(&deps);
 	i915_refct_sgt_put(dst_rsgt);
 	if (IS_ERR_OR_NULL(copy_fence))
 		return PTR_ERR_OR_ZERO(copy_fence);
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
index d2e7f149e05c..138b7647a558 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
@@ -18,6 +18,23 @@ struct ttm_tt;
 struct drm_i915_gem_object;
 struct i915_refct_sgt;
 
+/**
+ * struct i915_deps - Collect dependencies into a single dma-fence
+ * @single: Storage for pointer if the collection is a single fence.
+ * @fences: Allocated array of fence pointers if more than a single fence;
+ * otherwise points to the address of @single.
+ * @num_deps: Current number of dependency fences.
+ * @fences_size: Size of the @fences array in number of pointers.
+ * @gfp: Allocation mode.
+ */
+struct i915_deps {
+	struct dma_fence *single;
+	struct dma_fence **fences;
+	unsigned int num_deps;
+	unsigned int fences_size;
+	gfp_t gfp;
+};
+
 int i915_ttm_move_notify(struct ttm_buffer_object *bo);
 
 I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c
index 19a01878fee3..18b44af56969 100644
--- a/drivers/gpu/drm/i915/gt/intel_migrate.c
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.c
@@ -404,7 +404,7 @@ static int emit_copy(struct i915_request *rq, int size)
 
 int
 intel_context_migrate_copy(struct intel_context *ce,
-			   struct dma_fence *await,
+			   const struct i915_deps *deps,
 			   struct scatterlist *src,
 			   enum i915_cache_level src_cache_level,
 			   bool src_is_lmem,
@@ -431,8 +431,8 @@ intel_context_migrate_copy(struct intel_context *ce,
 			goto out_ce;
 		}
 
-		if (await) {
-			err = i915_request_await_dma_fence(rq, await);
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
 			if (err)
 				goto out_rq;
 
@@ -442,7 +442,7 @@ intel_context_migrate_copy(struct intel_context *ce,
 					goto out_rq;
 			}
 
-			await = NULL;
+			deps = NULL;
 		}
 
 		/* The PTE updates + copy must not be interrupted. */
@@ -525,7 +525,7 @@ static int emit_clear(struct i915_request *rq, int size, u32 value)
 
 int
 intel_context_migrate_clear(struct intel_context *ce,
-			    struct dma_fence *await,
+			    const struct i915_deps *deps,
 			    struct scatterlist *sg,
 			    enum i915_cache_level cache_level,
 			    bool is_lmem,
@@ -550,8 +550,8 @@ intel_context_migrate_clear(struct intel_context *ce,
 			goto out_ce;
 		}
 
-		if (await) {
-			err = i915_request_await_dma_fence(rq, await);
+		if (deps) {
+			err = i915_request_await_deps(rq, deps);
 			if (err)
 				goto out_rq;
 
@@ -561,7 +561,7 @@ intel_context_migrate_clear(struct intel_context *ce,
 					goto out_rq;
 			}
 
-			await = NULL;
+			deps = NULL;
 		}
 
 		/* The PTE updates + clear must not be interrupted. */
@@ -599,7 +599,7 @@ intel_context_migrate_clear(struct intel_context *ce,
 
 int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_gem_ww_ctx *ww,
-		       struct dma_fence *await,
+		       const struct i915_deps *deps,
 		       struct scatterlist *src,
 		       enum i915_cache_level src_cache_level,
 		       bool src_is_lmem,
@@ -624,7 +624,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 	if (err)
 		goto out;
 
-	err = intel_context_migrate_copy(ce, await,
+	err = intel_context_migrate_copy(ce, deps,
 					 src, src_cache_level, src_is_lmem,
 					 dst, dst_cache_level, dst_is_lmem,
 					 out);
@@ -638,7 +638,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 int
 intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_gem_ww_ctx *ww,
-		    struct dma_fence *await,
+		    const struct i915_deps *deps,
 		    struct scatterlist *sg,
 		    enum i915_cache_level cache_level,
 		    bool is_lmem,
@@ -661,7 +661,7 @@ intel_migrate_clear(struct intel_migrate *m,
 	if (err)
 		goto out;
 
-	err = intel_context_migrate_clear(ce, await, sg, cache_level,
+	err = intel_context_migrate_clear(ce, deps, sg, cache_level,
 					  is_lmem, value, out);
 
 	intel_context_unpin(ce);
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.h b/drivers/gpu/drm/i915/gt/intel_migrate.h
index 4e18e755a00b..ccc677ec4aa3 100644
--- a/drivers/gpu/drm/i915/gt/intel_migrate.h
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.h
@@ -11,6 +11,7 @@
 #include "intel_migrate_types.h"
 
 struct dma_fence;
+struct i915_deps;
 struct i915_request;
 struct i915_gem_ww_ctx;
 struct intel_gt;
@@ -23,7 +24,7 @@ struct intel_context *intel_migrate_create_context(struct intel_migrate *m);
 
 int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_gem_ww_ctx *ww,
-		       struct dma_fence *await,
+		       const struct i915_deps *deps,
 		       struct scatterlist *src,
 		       enum i915_cache_level src_cache_level,
 		       bool src_is_lmem,
@@ -33,7 +34,7 @@ int intel_migrate_copy(struct intel_migrate *m,
 		       struct i915_request **out);
 
 int intel_context_migrate_copy(struct intel_context *ce,
-			       struct dma_fence *await,
+			       const struct i915_deps *deps,
 			       struct scatterlist *src,
 			       enum i915_cache_level src_cache_level,
 			       bool src_is_lmem,
@@ -45,7 +46,7 @@ int intel_context_migrate_copy(struct intel_context *ce,
 int
 intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_gem_ww_ctx *ww,
-		    struct dma_fence *await,
+		    const struct i915_deps *deps,
 		    struct scatterlist *sg,
 		    enum i915_cache_level cache_level,
 		    bool is_lmem,
@@ -53,7 +54,7 @@ intel_migrate_clear(struct intel_migrate *m,
 		    struct i915_request **out);
 int
 intel_context_migrate_clear(struct intel_context *ce,
-			    struct dma_fence *await,
+			    const struct i915_deps *deps,
 			    struct scatterlist *sg,
 			    enum i915_cache_level cache_level,
 			    bool is_lmem,
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 40cf8ec0b9fc..7d804df27546 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -32,6 +32,7 @@
 #include <linux/sched/mm.h>
 
 #include "gem/i915_gem_context.h"
+#include "gem/i915_gem_ttm_move.h"
 #include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine.h"
@@ -1543,6 +1544,27 @@ i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence)
 	return 0;
 }
 
+/**
+ * i915_request_await_deps - set this request to (async) wait upon a struct
+ * i915_deps dma_fence collection
+ * @rq: request we are wishing to use
+ * @deps: The struct i915_deps containing the dependencies.
+ *
+ * Returns 0 if successful, negative error code on error.
+ */
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps)
+{
+	int i, err;
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		err = i915_request_await_dma_fence(rq, deps->fences[i]);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
 /**
  * i915_request_await_object - set this request to (async) wait upon a bo
  * @to: request we are wishing to use
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
index ce7714210697..170ee78c2858 100644
--- a/drivers/gpu/drm/i915/i915_request.h
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -47,6 +47,7 @@
 struct drm_file;
 struct drm_i915_gem_object;
 struct drm_printer;
+struct i915_deps;
 struct i915_request;
 
 #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
@@ -411,6 +412,7 @@ int i915_request_await_object(struct i915_request *to,
 			      bool write);
 int i915_request_await_dma_fence(struct i915_request *rq,
 				 struct dma_fence *fence);
+int i915_request_await_deps(struct i915_request *rq, const struct i915_deps *deps);
 int i915_request_await_execution(struct i915_request *rq,
 				 struct dma_fence *fence);
 
-- 
2.31.1

[PATCH v4 2/4] drm/i915: remove questionable fence optimization during copy

[Thomas Hellström]

From: Christian König <christian.koenig@amd.com>

First of all as discussed multiple times now kernel copies *must* always
wait for all fences in a BO before actually doing the copy. This is
mandatory.

Additional to that drop the handling when there can't be a shared slot
allocated on the source BO and just properly return an error code.
Otherwise this code path would only be tested under out of memory
conditions.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 43 +++++++-------------
 1 file changed, 14 insertions(+), 29 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 62a6fd2d127d..5dbaccf3f201 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -195,19 +195,14 @@ static int i915_deps_add_dependency(struct i915_deps *deps,
 }
 
 static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
-			      bool all, const bool no_excl,
 			      const struct ttm_operation_ctx *ctx)
 {
 	struct dma_resv_iter iter;
 	struct dma_fence *fence;
+	int ret;
 
 	dma_resv_assert_held(resv);
-	dma_resv_for_each_fence(&iter, resv, all, fence) {
-		int ret;
-
-		if (no_excl && dma_resv_iter_is_exclusive(&iter))
-			continue;
-
+	dma_resv_for_each_fence(&iter, resv, true, fence) {
 		ret = i915_deps_add_dependency(deps, fence, ctx);
 		if (ret)
 			return ret;
@@ -634,11 +629,7 @@ prev_deps(struct ttm_buffer_object *bo, struct ttm_operation_ctx *ctx,
 
 	ret = i915_deps_add_dependency(deps, bo->moving, ctx);
 	if (!ret)
-		/*
-		 * TODO: Only await excl fence here, and shared fences before
-		 * signaling the migration fence.
-		 */
-		ret = i915_deps_add_resv(deps, bo->base.resv, true, false, ctx);
+		ret = i915_deps_add_resv(deps, bo->base.resv, ctx);
 
 	return ret;
 }
@@ -768,22 +759,21 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 	struct i915_refct_sgt *dst_rsgt;
 	struct dma_fence *copy_fence;
 	struct i915_deps deps;
-	int ret, shared_err;
+	int ret;
 
 	assert_object_held(dst);
 	assert_object_held(src);
 	i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
 
-	/*
-	 * We plan to add a shared fence only for the source. If that
-	 * fails, we await all source fences before commencing
-	 * the copy instead of only the exclusive.
-	 */
-	shared_err = dma_resv_reserve_shared(src_bo->base.resv, 1);
-	ret = i915_deps_add_resv(&deps, dst_bo->base.resv, true, false, &ctx);
-	if (!ret)
-		ret = i915_deps_add_resv(&deps, src_bo->base.resv,
-					 !!shared_err, false, &ctx);
+	ret = dma_resv_reserve_shared(src_bo->base.resv, 1);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, dst_bo->base.resv, &ctx);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, src_bo->base.resv, &ctx);
 	if (ret)
 		return ret;
 
@@ -798,12 +788,7 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 		return PTR_ERR_OR_ZERO(copy_fence);
 
 	dma_resv_add_excl_fence(dst_bo->base.resv, copy_fence);
-
-	/* If we failed to reserve a shared slot, add an exclusive fence */
-	if (shared_err)
-		dma_resv_add_excl_fence(src_bo->base.resv, copy_fence);
-	else
-		dma_resv_add_shared_fence(src_bo->base.resv, copy_fence);
+	dma_resv_add_shared_fence(src_bo->base.resv, copy_fence);
 
 	dma_fence_put(copy_fence);
 
-- 
2.31.1

[Intel-gfx] [PATCH v4 2/4] drm/i915: remove questionable fence optimization during copy

[Thomas Hellström]

From: Christian König <christian.koenig@amd.com>

First of all as discussed multiple times now kernel copies *must* always
wait for all fences in a BO before actually doing the copy. This is
mandatory.

Additional to that drop the handling when there can't be a shared slot
allocated on the source BO and just properly return an error code.
Otherwise this code path would only be tested under out of memory
conditions.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 43 +++++++-------------
 1 file changed, 14 insertions(+), 29 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 62a6fd2d127d..5dbaccf3f201 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -195,19 +195,14 @@ static int i915_deps_add_dependency(struct i915_deps *deps,
 }
 
 static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
-			      bool all, const bool no_excl,
 			      const struct ttm_operation_ctx *ctx)
 {
 	struct dma_resv_iter iter;
 	struct dma_fence *fence;
+	int ret;
 
 	dma_resv_assert_held(resv);
-	dma_resv_for_each_fence(&iter, resv, all, fence) {
-		int ret;
-
-		if (no_excl && dma_resv_iter_is_exclusive(&iter))
-			continue;
-
+	dma_resv_for_each_fence(&iter, resv, true, fence) {
 		ret = i915_deps_add_dependency(deps, fence, ctx);
 		if (ret)
 			return ret;
@@ -634,11 +629,7 @@ prev_deps(struct ttm_buffer_object *bo, struct ttm_operation_ctx *ctx,
 
 	ret = i915_deps_add_dependency(deps, bo->moving, ctx);
 	if (!ret)
-		/*
-		 * TODO: Only await excl fence here, and shared fences before
-		 * signaling the migration fence.
-		 */
-		ret = i915_deps_add_resv(deps, bo->base.resv, true, false, ctx);
+		ret = i915_deps_add_resv(deps, bo->base.resv, ctx);
 
 	return ret;
 }
@@ -768,22 +759,21 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 	struct i915_refct_sgt *dst_rsgt;
 	struct dma_fence *copy_fence;
 	struct i915_deps deps;
-	int ret, shared_err;
+	int ret;
 
 	assert_object_held(dst);
 	assert_object_held(src);
 	i915_deps_init(&deps, GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN);
 
-	/*
-	 * We plan to add a shared fence only for the source. If that
-	 * fails, we await all source fences before commencing
-	 * the copy instead of only the exclusive.
-	 */
-	shared_err = dma_resv_reserve_shared(src_bo->base.resv, 1);
-	ret = i915_deps_add_resv(&deps, dst_bo->base.resv, true, false, &ctx);
-	if (!ret)
-		ret = i915_deps_add_resv(&deps, src_bo->base.resv,
-					 !!shared_err, false, &ctx);
+	ret = dma_resv_reserve_shared(src_bo->base.resv, 1);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, dst_bo->base.resv, &ctx);
+	if (ret)
+		return ret;
+
+	ret = i915_deps_add_resv(&deps, src_bo->base.resv, &ctx);
 	if (ret)
 		return ret;
 
@@ -798,12 +788,7 @@ int i915_gem_obj_copy_ttm(struct drm_i915_gem_object *dst,
 		return PTR_ERR_OR_ZERO(copy_fence);
 
 	dma_resv_add_excl_fence(dst_bo->base.resv, copy_fence);
-
-	/* If we failed to reserve a shared slot, add an exclusive fence */
-	if (shared_err)
-		dma_resv_add_excl_fence(src_bo->base.resv, copy_fence);
-	else
-		dma_resv_add_shared_fence(src_bo->base.resv, copy_fence);
+	dma_resv_add_shared_fence(src_bo->base.resv, copy_fence);
 
 	dma_fence_put(copy_fence);
 
-- 
2.31.1

[PATCH v4 3/4] drm/i915: Break out the i915_deps utility

[Thomas Hellström]

Since it's starting to be used outside the i915 TTM move code, move it
to a separate set of files.

v2:
- Update the documentation.
v4:
- Rebase.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/i915/Makefile                |   1 +
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 171 +------------
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h |  17 --
 drivers/gpu/drm/i915/i915_deps.c             | 237 +++++++++++++++++++
 drivers/gpu/drm/i915/i915_deps.h             |  45 ++++
 drivers/gpu/drm/i915/i915_request.c          |   2 +-
 6 files changed, 285 insertions(+), 188 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/i915_deps.c
 create mode 100644 drivers/gpu/drm/i915/i915_deps.h

diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 6ddd2d2bbaaf..1b62b9f65196 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -163,6 +163,7 @@ i915-y += \
 	  i915_active.o \
 	  i915_buddy.o \
 	  i915_cmd_parser.o \
+	  i915_deps.o \
 	  i915_gem_evict.o \
 	  i915_gem_gtt.o \
 	  i915_gem_ww.o \
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 5dbaccf3f201..ee9612a3ee5e 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -5,6 +5,7 @@
 
 #include <drm/ttm/ttm_bo_driver.h>
 
+#include "i915_deps.h"
 #include "i915_drv.h"
 #include "intel_memory_region.h"
 #include "intel_region_ttm.h"
@@ -41,176 +42,6 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
 }
 #endif
 
-/**
- * DOC: Set of utilities to dynamically collect dependencies into a
- * structure which is fed into the GT migration code.
- *
- * Once we can do async unbinding, this is also needed to coalesce
- * the migration fence with the unbind fences if these are coalesced
- * post-migration.
- *
- * While collecting the individual dependencies, we store the refcounted
- * struct dma_fence pointers in a realloc-managed pointer array, since
- * that can be easily fed into a dma_fence_array. Other options are
- * available, like for example an xarray for similarity with drm/sched.
- * Can be changed easily if needed.
- *
- * A struct i915_deps need to be initialized using i915_deps_init().
- * If i915_deps_add_dependency() or i915_deps_add_resv() return an
- * error code they will internally call i915_deps_fini(), which frees
- * all internal references and allocations.
- *
- * We might want to break this out into a separate file as a utility.
- */
-
-#define I915_DEPS_MIN_ALLOC_CHUNK 8U
-
-static void i915_deps_reset_fences(struct i915_deps *deps)
-{
-	if (deps->fences != &deps->single)
-		kfree(deps->fences);
-	deps->num_deps = 0;
-	deps->fences_size = 1;
-	deps->fences = &deps->single;
-}
-
-static void i915_deps_init(struct i915_deps *deps, gfp_t gfp)
-{
-	deps->fences = NULL;
-	deps->gfp = gfp;
-	i915_deps_reset_fences(deps);
-}
-
-static void i915_deps_fini(struct i915_deps *deps)
-{
-	unsigned int i;
-
-	for (i = 0; i < deps->num_deps; ++i)
-		dma_fence_put(deps->fences[i]);
-
-	if (deps->fences != &deps->single)
-		kfree(deps->fences);
-}
-
-static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
-			  const struct ttm_operation_ctx *ctx)
-{
-	int ret;
-
-	if (deps->num_deps >= deps->fences_size) {
-		unsigned int new_size = 2 * deps->fences_size;
-		struct dma_fence **new_fences;
-
-		new_size = max(new_size, I915_DEPS_MIN_ALLOC_CHUNK);
-		new_fences = kmalloc_array(new_size, sizeof(*new_fences), deps->gfp);
-		if (!new_fences)
-			goto sync;
-
-		memcpy(new_fences, deps->fences,
-		       deps->fences_size * sizeof(*new_fences));
-		swap(new_fences, deps->fences);
-		if (new_fences != &deps->single)
-			kfree(new_fences);
-		deps->fences_size = new_size;
-	}
-	deps->fences[deps->num_deps++] = dma_fence_get(fence);
-	return 0;
-
-sync:
-	if (ctx->no_wait_gpu && !dma_fence_is_signaled(fence)) {
-		ret = -EBUSY;
-		goto unref;
-	}
-
-	ret = dma_fence_wait(fence, ctx->interruptible);
-	if (ret)
-		goto unref;
-
-	ret = fence->error;
-	if (ret)
-		goto unref;
-
-	return 0;
-
-unref:
-	i915_deps_fini(deps);
-	return ret;
-}
-
-static int i915_deps_sync(const struct i915_deps *deps,
-			  const struct ttm_operation_ctx *ctx)
-{
-	struct dma_fence **fences = deps->fences;
-	unsigned int i;
-	int ret = 0;
-
-	for (i = 0; i < deps->num_deps; ++i, ++fences) {
-		if (ctx->no_wait_gpu && !dma_fence_is_signaled(*fences)) {
-			ret = -EBUSY;
-			break;
-		}
-
-		ret = dma_fence_wait(*fences, ctx->interruptible);
-		if (!ret)
-			ret = (*fences)->error;
-		if (ret)
-			break;
-	}
-
-	return ret;
-}
-
-static int i915_deps_add_dependency(struct i915_deps *deps,
-				    struct dma_fence *fence,
-				    const struct ttm_operation_ctx *ctx)
-{
-	unsigned int i;
-	int ret;
-
-	if (!fence)
-		return 0;
-
-	if (dma_fence_is_signaled(fence)) {
-		ret = fence->error;
-		if (ret)
-			i915_deps_fini(deps);
-		return ret;
-	}
-
-	for (i = 0; i < deps->num_deps; ++i) {
-		struct dma_fence *entry = deps->fences[i];
-
-		if (!entry->context || entry->context != fence->context)
-			continue;
-
-		if (dma_fence_is_later(fence, entry)) {
-			dma_fence_put(entry);
-			deps->fences[i] = dma_fence_get(fence);
-		}
-
-		return 0;
-	}
-
-	return i915_deps_grow(deps, fence, ctx);
-}
-
-static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
-			      const struct ttm_operation_ctx *ctx)
-{
-	struct dma_resv_iter iter;
-	struct dma_fence *fence;
-	int ret;
-
-	dma_resv_assert_held(resv);
-	dma_resv_for_each_fence(&iter, resv, true, fence) {
-		ret = i915_deps_add_dependency(deps, fence, ctx);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
-}
-
 static enum i915_cache_level
 i915_ttm_cache_level(struct drm_i915_private *i915, struct ttm_resource *res,
 		     struct ttm_tt *ttm)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
index 138b7647a558..d2e7f149e05c 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
@@ -18,23 +18,6 @@ struct ttm_tt;
 struct drm_i915_gem_object;
 struct i915_refct_sgt;
 
-/**
- * struct i915_deps - Collect dependencies into a single dma-fence
- * @single: Storage for pointer if the collection is a single fence.
- * @fences: Allocated array of fence pointers if more than a single fence;
- * otherwise points to the address of @single.
- * @num_deps: Current number of dependency fences.
- * @fences_size: Size of the @fences array in number of pointers.
- * @gfp: Allocation mode.
- */
-struct i915_deps {
-	struct dma_fence *single;
-	struct dma_fence **fences;
-	unsigned int num_deps;
-	unsigned int fences_size;
-	gfp_t gfp;
-};
-
 int i915_ttm_move_notify(struct ttm_buffer_object *bo);
 
 I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
diff --git a/drivers/gpu/drm/i915/i915_deps.c b/drivers/gpu/drm/i915/i915_deps.c
new file mode 100644
index 000000000000..999210b37325
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.c
@@ -0,0 +1,237 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/dma-fence.h>
+#include <linux/slab.h>
+
+#include <drm/ttm/ttm_bo_api.h>
+
+#include "i915_deps.h"
+
+/**
+ * DOC: Set of utilities to dynamically collect dependencies into a
+ * structure which is fed into the GT migration code.
+ *
+ * Once we can do async unbinding, this is also needed to coalesce
+ * the migration fence with the unbind fences if these are coalesced
+ * post-migration.
+ *
+ * While collecting the individual dependencies, we store the refcounted
+ * struct dma_fence pointers in a realloc-managed pointer array, since
+ * that can be easily fed into a dma_fence_array. Other options are
+ * available, like for example an xarray for similarity with drm/sched.
+ * Can be changed easily if needed.
+ *
+ * A struct i915_deps need to be initialized using i915_deps_init().
+ * If i915_deps_add_dependency() or i915_deps_add_resv() return an
+ * error code they will internally call i915_deps_fini(), which frees
+ * all internal references and allocations.
+ */
+
+/* Min number of fence pointers in the array when an allocation occurs. */
+#define I915_DEPS_MIN_ALLOC_CHUNK 8U
+
+static void i915_deps_reset_fences(struct i915_deps *deps)
+{
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+	deps->num_deps = 0;
+	deps->fences_size = 1;
+	deps->fences = &deps->single;
+}
+
+/**
+ * i915_deps_init - Initialize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to initialize.
+ * @gfp: The allocation mode for subsequenst allocations.
+ */
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp)
+{
+	deps->fences = NULL;
+	deps->gfp = gfp;
+	i915_deps_reset_fences(deps);
+}
+
+/**
+ * i915_deps_fini - Finalize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to finalize.
+ *
+ * This function drops all fence references taken, conditionally frees and
+ * then resets the fences array.
+ */
+void i915_deps_fini(struct i915_deps *deps)
+{
+	unsigned int i;
+
+	for (i = 0; i < deps->num_deps; ++i)
+		dma_fence_put(deps->fences[i]);
+
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+}
+
+static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
+			  const struct ttm_operation_ctx *ctx)
+{
+	int ret;
+
+	if (deps->num_deps >= deps->fences_size) {
+		unsigned int new_size = 2 * deps->fences_size;
+		struct dma_fence **new_fences;
+
+		new_size = max(new_size, I915_DEPS_MIN_ALLOC_CHUNK);
+		new_fences = kmalloc_array(new_size, sizeof(*new_fences), deps->gfp);
+		if (!new_fences)
+			goto sync;
+
+		memcpy(new_fences, deps->fences,
+		       deps->fences_size * sizeof(*new_fences));
+		swap(new_fences, deps->fences);
+		if (new_fences != &deps->single)
+			kfree(new_fences);
+		deps->fences_size = new_size;
+	}
+	deps->fences[deps->num_deps++] = dma_fence_get(fence);
+	return 0;
+
+sync:
+	if (ctx->no_wait_gpu && !dma_fence_is_signaled(fence)) {
+		ret = -EBUSY;
+		goto unref;
+	}
+
+	ret = dma_fence_wait(fence, ctx->interruptible);
+	if (ret)
+		goto unref;
+
+	ret = fence->error;
+	if (ret)
+		goto unref;
+
+	return 0;
+
+unref:
+	i915_deps_fini(deps);
+	return ret;
+}
+
+/**
+ * i915_deps_sync - Wait for all the fences in the dependency collection
+ * @deps: Pointer to the i915_deps structure the fences of which to wait for.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how the waits
+ * should be performed.
+ *
+ * This function waits for fences in the dependency collection. If it
+ * encounters an error during the wait or a fence error, the wait for
+ * further fences is aborted and the error returned.
+ *
+ * Return: Zero if successful, Negative error code on error.
+ */
+int i915_deps_sync(const struct i915_deps *deps, const struct ttm_operation_ctx *ctx)
+{
+	struct dma_fence **fences = deps->fences;
+	unsigned int i;
+	int ret = 0;
+
+	for (i = 0; i < deps->num_deps; ++i, ++fences) {
+		if (ctx->no_wait_gpu && !dma_fence_is_signaled(*fences)) {
+			ret = -EBUSY;
+			break;
+		}
+
+		ret = dma_fence_wait(*fences, ctx->interruptible);
+		if (!ret)
+			ret = (*fences)->error;
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+/**
+ * i915_deps_add_dependency - Add a fence to the dependency collection
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @fence: The fence to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Adds a fence to the dependency collection, and takes a reference on it.
+ * If the fence context is not zero and there was a later fence from the
+ * same fence context already added, then the fence is not added to the
+ * dependency collection. If the fence context is not zero and there was
+ * an earlier fence already added, then the fence will replace the older
+ * fence from the same context and the reference on the earlier fence will
+ * be dropped.
+ * If there is a failure to allocate memory to accommodate the new fence to
+ * be added, the new fence will instead be waited for and an error may
+ * be returned; depending on the value of @ctx, or if there was a fence
+ * error. If an error was returned, the dependency collection will be
+ * finalized and all fence reference dropped.
+ *
+ * Return: 0 if success. Negative error code on error.
+ */
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx)
+{
+	unsigned int i;
+	int ret;
+
+	if (!fence)
+		return 0;
+
+	if (dma_fence_is_signaled(fence)) {
+		ret = fence->error;
+		if (ret)
+			i915_deps_fini(deps);
+		return ret;
+	}
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		struct dma_fence *entry = deps->fences[i];
+
+		if (!entry->context || entry->context != fence->context)
+			continue;
+
+		if (dma_fence_is_later(fence, entry)) {
+			dma_fence_put(entry);
+			deps->fences[i] = dma_fence_get(fence);
+		}
+
+		return 0;
+	}
+
+	return i915_deps_grow(deps, fence, ctx);
+}
+
+/**
+ * i915_deps_add_resv - Add the fences of a reservation object to a dependency
+ * collection.
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @resv: The reservation object, then fences of which to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Calls i915_deps_add_depencency() on the indicated fences of @resv.
+ *
+ * Return: Zero on success. Negative error code on error.
+ */
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx)
+{
+	struct dma_resv_iter iter;
+	struct dma_fence *fence;
+
+	dma_resv_assert_held(resv);
+	dma_resv_for_each_fence(&iter, resv, true, fence) {
+		int ret = i915_deps_add_dependency(deps, fence, ctx);
+
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_deps.h b/drivers/gpu/drm/i915/i915_deps.h
new file mode 100644
index 000000000000..d76c0106c910
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _I915_DEPS_H_
+#define _I915_DEPS_H_
+
+#include <linux/types.h>
+
+struct ttm_operation_ctx;
+struct dma_fence;
+struct dma_resv;
+
+/**
+ * struct i915_deps - Collect dependencies into a single dma-fence
+ * @single: Storage for pointer if the collection is a single fence.
+ * @fences: Allocated array of fence pointers if more than a single fence;
+ * otherwise points to the address of @single.
+ * @num_deps: Current number of dependency fences.
+ * @fences_size: Size of the @fences array in number of pointers.
+ * @gfp: Allocation mode.
+ */
+struct i915_deps {
+	struct dma_fence *single;
+	struct dma_fence **fences;
+	unsigned int num_deps;
+	unsigned int fences_size;
+	gfp_t gfp;
+};
+
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp);
+
+void i915_deps_fini(struct i915_deps *deps);
+
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx);
+
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx);
+
+int i915_deps_sync(const struct i915_deps *deps,
+		   const struct ttm_operation_ctx *ctx);
+#endif
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 7d804df27546..76cf5ac91e94 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -32,7 +32,6 @@
 #include <linux/sched/mm.h>
 
 #include "gem/i915_gem_context.h"
-#include "gem/i915_gem_ttm_move.h"
 #include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine.h"
@@ -43,6 +42,7 @@
 #include "gt/intel_rps.h"
 
 #include "i915_active.h"
+#include "i915_deps.h"
 #include "i915_drv.h"
 #include "i915_trace.h"
 #include "intel_pm.h"
-- 
2.31.1

[Intel-gfx] [PATCH v4 3/4] drm/i915: Break out the i915_deps utility

[Thomas Hellström]

Since it's starting to be used outside the i915 TTM move code, move it
to a separate set of files.

v2:
- Update the documentation.
v4:
- Rebase.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/i915/Makefile                |   1 +
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c | 171 +------------
 drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h |  17 --
 drivers/gpu/drm/i915/i915_deps.c             | 237 +++++++++++++++++++
 drivers/gpu/drm/i915/i915_deps.h             |  45 ++++
 drivers/gpu/drm/i915/i915_request.c          |   2 +-
 6 files changed, 285 insertions(+), 188 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/i915_deps.c
 create mode 100644 drivers/gpu/drm/i915/i915_deps.h

diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 6ddd2d2bbaaf..1b62b9f65196 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -163,6 +163,7 @@ i915-y += \
 	  i915_active.o \
 	  i915_buddy.o \
 	  i915_cmd_parser.o \
+	  i915_deps.o \
 	  i915_gem_evict.o \
 	  i915_gem_gtt.o \
 	  i915_gem_ww.o \
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
index 5dbaccf3f201..ee9612a3ee5e 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.c
@@ -5,6 +5,7 @@
 
 #include <drm/ttm/ttm_bo_driver.h>
 
+#include "i915_deps.h"
 #include "i915_drv.h"
 #include "intel_memory_region.h"
 #include "intel_region_ttm.h"
@@ -41,176 +42,6 @@ void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
 }
 #endif
 
-/**
- * DOC: Set of utilities to dynamically collect dependencies into a
- * structure which is fed into the GT migration code.
- *
- * Once we can do async unbinding, this is also needed to coalesce
- * the migration fence with the unbind fences if these are coalesced
- * post-migration.
- *
- * While collecting the individual dependencies, we store the refcounted
- * struct dma_fence pointers in a realloc-managed pointer array, since
- * that can be easily fed into a dma_fence_array. Other options are
- * available, like for example an xarray for similarity with drm/sched.
- * Can be changed easily if needed.
- *
- * A struct i915_deps need to be initialized using i915_deps_init().
- * If i915_deps_add_dependency() or i915_deps_add_resv() return an
- * error code they will internally call i915_deps_fini(), which frees
- * all internal references and allocations.
- *
- * We might want to break this out into a separate file as a utility.
- */
-
-#define I915_DEPS_MIN_ALLOC_CHUNK 8U
-
-static void i915_deps_reset_fences(struct i915_deps *deps)
-{
-	if (deps->fences != &deps->single)
-		kfree(deps->fences);
-	deps->num_deps = 0;
-	deps->fences_size = 1;
-	deps->fences = &deps->single;
-}
-
-static void i915_deps_init(struct i915_deps *deps, gfp_t gfp)
-{
-	deps->fences = NULL;
-	deps->gfp = gfp;
-	i915_deps_reset_fences(deps);
-}
-
-static void i915_deps_fini(struct i915_deps *deps)
-{
-	unsigned int i;
-
-	for (i = 0; i < deps->num_deps; ++i)
-		dma_fence_put(deps->fences[i]);
-
-	if (deps->fences != &deps->single)
-		kfree(deps->fences);
-}
-
-static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
-			  const struct ttm_operation_ctx *ctx)
-{
-	int ret;
-
-	if (deps->num_deps >= deps->fences_size) {
-		unsigned int new_size = 2 * deps->fences_size;
-		struct dma_fence **new_fences;
-
-		new_size = max(new_size, I915_DEPS_MIN_ALLOC_CHUNK);
-		new_fences = kmalloc_array(new_size, sizeof(*new_fences), deps->gfp);
-		if (!new_fences)
-			goto sync;
-
-		memcpy(new_fences, deps->fences,
-		       deps->fences_size * sizeof(*new_fences));
-		swap(new_fences, deps->fences);
-		if (new_fences != &deps->single)
-			kfree(new_fences);
-		deps->fences_size = new_size;
-	}
-	deps->fences[deps->num_deps++] = dma_fence_get(fence);
-	return 0;
-
-sync:
-	if (ctx->no_wait_gpu && !dma_fence_is_signaled(fence)) {
-		ret = -EBUSY;
-		goto unref;
-	}
-
-	ret = dma_fence_wait(fence, ctx->interruptible);
-	if (ret)
-		goto unref;
-
-	ret = fence->error;
-	if (ret)
-		goto unref;
-
-	return 0;
-
-unref:
-	i915_deps_fini(deps);
-	return ret;
-}
-
-static int i915_deps_sync(const struct i915_deps *deps,
-			  const struct ttm_operation_ctx *ctx)
-{
-	struct dma_fence **fences = deps->fences;
-	unsigned int i;
-	int ret = 0;
-
-	for (i = 0; i < deps->num_deps; ++i, ++fences) {
-		if (ctx->no_wait_gpu && !dma_fence_is_signaled(*fences)) {
-			ret = -EBUSY;
-			break;
-		}
-
-		ret = dma_fence_wait(*fences, ctx->interruptible);
-		if (!ret)
-			ret = (*fences)->error;
-		if (ret)
-			break;
-	}
-
-	return ret;
-}
-
-static int i915_deps_add_dependency(struct i915_deps *deps,
-				    struct dma_fence *fence,
-				    const struct ttm_operation_ctx *ctx)
-{
-	unsigned int i;
-	int ret;
-
-	if (!fence)
-		return 0;
-
-	if (dma_fence_is_signaled(fence)) {
-		ret = fence->error;
-		if (ret)
-			i915_deps_fini(deps);
-		return ret;
-	}
-
-	for (i = 0; i < deps->num_deps; ++i) {
-		struct dma_fence *entry = deps->fences[i];
-
-		if (!entry->context || entry->context != fence->context)
-			continue;
-
-		if (dma_fence_is_later(fence, entry)) {
-			dma_fence_put(entry);
-			deps->fences[i] = dma_fence_get(fence);
-		}
-
-		return 0;
-	}
-
-	return i915_deps_grow(deps, fence, ctx);
-}
-
-static int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
-			      const struct ttm_operation_ctx *ctx)
-{
-	struct dma_resv_iter iter;
-	struct dma_fence *fence;
-	int ret;
-
-	dma_resv_assert_held(resv);
-	dma_resv_for_each_fence(&iter, resv, true, fence) {
-		ret = i915_deps_add_dependency(deps, fence, ctx);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
-}
-
 static enum i915_cache_level
 i915_ttm_cache_level(struct drm_i915_private *i915, struct ttm_resource *res,
 		     struct ttm_tt *ttm)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
index 138b7647a558..d2e7f149e05c 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_move.h
@@ -18,23 +18,6 @@ struct ttm_tt;
 struct drm_i915_gem_object;
 struct i915_refct_sgt;
 
-/**
- * struct i915_deps - Collect dependencies into a single dma-fence
- * @single: Storage for pointer if the collection is a single fence.
- * @fences: Allocated array of fence pointers if more than a single fence;
- * otherwise points to the address of @single.
- * @num_deps: Current number of dependency fences.
- * @fences_size: Size of the @fences array in number of pointers.
- * @gfp: Allocation mode.
- */
-struct i915_deps {
-	struct dma_fence *single;
-	struct dma_fence **fences;
-	unsigned int num_deps;
-	unsigned int fences_size;
-	gfp_t gfp;
-};
-
 int i915_ttm_move_notify(struct ttm_buffer_object *bo);
 
 I915_SELFTEST_DECLARE(void i915_ttm_migrate_set_failure_modes(bool gpu_migration,
diff --git a/drivers/gpu/drm/i915/i915_deps.c b/drivers/gpu/drm/i915/i915_deps.c
new file mode 100644
index 000000000000..999210b37325
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.c
@@ -0,0 +1,237 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#include <linux/dma-fence.h>
+#include <linux/slab.h>
+
+#include <drm/ttm/ttm_bo_api.h>
+
+#include "i915_deps.h"
+
+/**
+ * DOC: Set of utilities to dynamically collect dependencies into a
+ * structure which is fed into the GT migration code.
+ *
+ * Once we can do async unbinding, this is also needed to coalesce
+ * the migration fence with the unbind fences if these are coalesced
+ * post-migration.
+ *
+ * While collecting the individual dependencies, we store the refcounted
+ * struct dma_fence pointers in a realloc-managed pointer array, since
+ * that can be easily fed into a dma_fence_array. Other options are
+ * available, like for example an xarray for similarity with drm/sched.
+ * Can be changed easily if needed.
+ *
+ * A struct i915_deps need to be initialized using i915_deps_init().
+ * If i915_deps_add_dependency() or i915_deps_add_resv() return an
+ * error code they will internally call i915_deps_fini(), which frees
+ * all internal references and allocations.
+ */
+
+/* Min number of fence pointers in the array when an allocation occurs. */
+#define I915_DEPS_MIN_ALLOC_CHUNK 8U
+
+static void i915_deps_reset_fences(struct i915_deps *deps)
+{
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+	deps->num_deps = 0;
+	deps->fences_size = 1;
+	deps->fences = &deps->single;
+}
+
+/**
+ * i915_deps_init - Initialize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to initialize.
+ * @gfp: The allocation mode for subsequenst allocations.
+ */
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp)
+{
+	deps->fences = NULL;
+	deps->gfp = gfp;
+	i915_deps_reset_fences(deps);
+}
+
+/**
+ * i915_deps_fini - Finalize an i915_deps structure
+ * @deps: Pointer to the i915_deps structure to finalize.
+ *
+ * This function drops all fence references taken, conditionally frees and
+ * then resets the fences array.
+ */
+void i915_deps_fini(struct i915_deps *deps)
+{
+	unsigned int i;
+
+	for (i = 0; i < deps->num_deps; ++i)
+		dma_fence_put(deps->fences[i]);
+
+	if (deps->fences != &deps->single)
+		kfree(deps->fences);
+}
+
+static int i915_deps_grow(struct i915_deps *deps, struct dma_fence *fence,
+			  const struct ttm_operation_ctx *ctx)
+{
+	int ret;
+
+	if (deps->num_deps >= deps->fences_size) {
+		unsigned int new_size = 2 * deps->fences_size;
+		struct dma_fence **new_fences;
+
+		new_size = max(new_size, I915_DEPS_MIN_ALLOC_CHUNK);
+		new_fences = kmalloc_array(new_size, sizeof(*new_fences), deps->gfp);
+		if (!new_fences)
+			goto sync;
+
+		memcpy(new_fences, deps->fences,
+		       deps->fences_size * sizeof(*new_fences));
+		swap(new_fences, deps->fences);
+		if (new_fences != &deps->single)
+			kfree(new_fences);
+		deps->fences_size = new_size;
+	}
+	deps->fences[deps->num_deps++] = dma_fence_get(fence);
+	return 0;
+
+sync:
+	if (ctx->no_wait_gpu && !dma_fence_is_signaled(fence)) {
+		ret = -EBUSY;
+		goto unref;
+	}
+
+	ret = dma_fence_wait(fence, ctx->interruptible);
+	if (ret)
+		goto unref;
+
+	ret = fence->error;
+	if (ret)
+		goto unref;
+
+	return 0;
+
+unref:
+	i915_deps_fini(deps);
+	return ret;
+}
+
+/**
+ * i915_deps_sync - Wait for all the fences in the dependency collection
+ * @deps: Pointer to the i915_deps structure the fences of which to wait for.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how the waits
+ * should be performed.
+ *
+ * This function waits for fences in the dependency collection. If it
+ * encounters an error during the wait or a fence error, the wait for
+ * further fences is aborted and the error returned.
+ *
+ * Return: Zero if successful, Negative error code on error.
+ */
+int i915_deps_sync(const struct i915_deps *deps, const struct ttm_operation_ctx *ctx)
+{
+	struct dma_fence **fences = deps->fences;
+	unsigned int i;
+	int ret = 0;
+
+	for (i = 0; i < deps->num_deps; ++i, ++fences) {
+		if (ctx->no_wait_gpu && !dma_fence_is_signaled(*fences)) {
+			ret = -EBUSY;
+			break;
+		}
+
+		ret = dma_fence_wait(*fences, ctx->interruptible);
+		if (!ret)
+			ret = (*fences)->error;
+		if (ret)
+			break;
+	}
+
+	return ret;
+}
+
+/**
+ * i915_deps_add_dependency - Add a fence to the dependency collection
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @fence: The fence to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Adds a fence to the dependency collection, and takes a reference on it.
+ * If the fence context is not zero and there was a later fence from the
+ * same fence context already added, then the fence is not added to the
+ * dependency collection. If the fence context is not zero and there was
+ * an earlier fence already added, then the fence will replace the older
+ * fence from the same context and the reference on the earlier fence will
+ * be dropped.
+ * If there is a failure to allocate memory to accommodate the new fence to
+ * be added, the new fence will instead be waited for and an error may
+ * be returned; depending on the value of @ctx, or if there was a fence
+ * error. If an error was returned, the dependency collection will be
+ * finalized and all fence reference dropped.
+ *
+ * Return: 0 if success. Negative error code on error.
+ */
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx)
+{
+	unsigned int i;
+	int ret;
+
+	if (!fence)
+		return 0;
+
+	if (dma_fence_is_signaled(fence)) {
+		ret = fence->error;
+		if (ret)
+			i915_deps_fini(deps);
+		return ret;
+	}
+
+	for (i = 0; i < deps->num_deps; ++i) {
+		struct dma_fence *entry = deps->fences[i];
+
+		if (!entry->context || entry->context != fence->context)
+			continue;
+
+		if (dma_fence_is_later(fence, entry)) {
+			dma_fence_put(entry);
+			deps->fences[i] = dma_fence_get(fence);
+		}
+
+		return 0;
+	}
+
+	return i915_deps_grow(deps, fence, ctx);
+}
+
+/**
+ * i915_deps_add_resv - Add the fences of a reservation object to a dependency
+ * collection.
+ * @deps: Pointer to the i915_deps structure a fence is to be added to.
+ * @resv: The reservation object, then fences of which to add.
+ * @ctx: Pointer to a struct ttm_operation_ctx indicating how waits are to
+ * be performed if waiting.
+ *
+ * Calls i915_deps_add_depencency() on the indicated fences of @resv.
+ *
+ * Return: Zero on success. Negative error code on error.
+ */
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx)
+{
+	struct dma_resv_iter iter;
+	struct dma_fence *fence;
+
+	dma_resv_assert_held(resv);
+	dma_resv_for_each_fence(&iter, resv, true, fence) {
+		int ret = i915_deps_add_dependency(deps, fence, ctx);
+
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/i915_deps.h b/drivers/gpu/drm/i915/i915_deps.h
new file mode 100644
index 000000000000..d76c0106c910
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_deps.h
@@ -0,0 +1,45 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef _I915_DEPS_H_
+#define _I915_DEPS_H_
+
+#include <linux/types.h>
+
+struct ttm_operation_ctx;
+struct dma_fence;
+struct dma_resv;
+
+/**
+ * struct i915_deps - Collect dependencies into a single dma-fence
+ * @single: Storage for pointer if the collection is a single fence.
+ * @fences: Allocated array of fence pointers if more than a single fence;
+ * otherwise points to the address of @single.
+ * @num_deps: Current number of dependency fences.
+ * @fences_size: Size of the @fences array in number of pointers.
+ * @gfp: Allocation mode.
+ */
+struct i915_deps {
+	struct dma_fence *single;
+	struct dma_fence **fences;
+	unsigned int num_deps;
+	unsigned int fences_size;
+	gfp_t gfp;
+};
+
+void i915_deps_init(struct i915_deps *deps, gfp_t gfp);
+
+void i915_deps_fini(struct i915_deps *deps);
+
+int i915_deps_add_dependency(struct i915_deps *deps,
+			     struct dma_fence *fence,
+			     const struct ttm_operation_ctx *ctx);
+
+int i915_deps_add_resv(struct i915_deps *deps, struct dma_resv *resv,
+		       const struct ttm_operation_ctx *ctx);
+
+int i915_deps_sync(const struct i915_deps *deps,
+		   const struct ttm_operation_ctx *ctx);
+#endif
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 7d804df27546..76cf5ac91e94 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -32,7 +32,6 @@
 #include <linux/sched/mm.h>
 
 #include "gem/i915_gem_context.h"
-#include "gem/i915_gem_ttm_move.h"
 #include "gt/intel_breadcrumbs.h"
 #include "gt/intel_context.h"
 #include "gt/intel_engine.h"
@@ -43,6 +42,7 @@
 #include "gt/intel_rps.h"
 
 #include "i915_active.h"
+#include "i915_deps.h"
 #include "i915_drv.h"
 #include "i915_trace.h"
 #include "intel_pm.h"
-- 
2.31.1

[PATCH v4 4/4] drm/i915: Require the vm mutex for i915_vma_bind()

[Thomas Hellström]

Protect updates of struct i915_vma flags and async binding / unbinding
with the vm::mutex. This means that i915_vma_bind() needs to assert
vm::mutex held. In order to make that possible drop the caching of
kmap_atomic() maps around i915_vma_bind().

An alternative would be to use kmap_local() but since we block cpu
unplugging during sleeps inside kmap_local() sections this may have
unwanted side-effects. Particularly since we might wait for gpu while
holding the vm mutex.

This change may theoretically increase execbuf cpu-usage on snb, but
at least on non-highmem systems that increase should be very small.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    | 50 ++++++++++++++++++-
 drivers/gpu/drm/i915/i915_vma.c               |  1 +
 2 files changed, 50 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index ec7c4a29a720..b8330f0bf652 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -1109,6 +1109,47 @@ static inline struct i915_ggtt *cache_to_ggtt(struct reloc_cache *cache)
 	return &i915->ggtt;
 }
 
+static void reloc_cache_unmap(struct reloc_cache *cache)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	vaddr = unmask_page(cache->vaddr);
+	if (cache->vaddr & KMAP)
+		kunmap_atomic(vaddr);
+	else
+		io_mapping_unmap_atomic((void __iomem *)vaddr);
+}
+
+static void reloc_cache_remap(struct reloc_cache *cache,
+			      struct drm_i915_gem_object *obj)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	if (cache->vaddr & KMAP) {
+		struct page *page = i915_gem_object_get_page(obj, cache->page);
+
+		vaddr = kmap_atomic(page);
+		cache->vaddr = unmask_flags(cache->vaddr) |
+			(unsigned long)vaddr;
+	} else {
+		struct i915_ggtt *ggtt = cache_to_ggtt(cache);
+		unsigned long offset;
+
+		offset = cache->node.start;
+		if (!drm_mm_node_allocated(&cache->node))
+			offset += cache->page << PAGE_SHIFT;
+
+		cache->vaddr = (unsigned long)
+			io_mapping_map_atomic_wc(&ggtt->iomap, offset);
+	}
+}
+
 static void reloc_cache_reset(struct reloc_cache *cache, struct i915_execbuffer *eb)
 {
 	void *vaddr;
@@ -1373,10 +1414,17 @@ eb_relocate_entry(struct i915_execbuffer *eb,
 		 * batchbuffers.
 		 */
 		if (reloc->write_domain == I915_GEM_DOMAIN_INSTRUCTION &&
-		    GRAPHICS_VER(eb->i915) == 6) {
+		    GRAPHICS_VER(eb->i915) == 6 &&
+		    !i915_vma_is_bound(target->vma, I915_VMA_GLOBAL_BIND)) {
+			struct i915_vma *vma = target->vma;
+
+			reloc_cache_unmap(&eb->reloc_cache);
+			mutex_lock(&vma->vm->mutex);
 			err = i915_vma_bind(target->vma,
 					    target->vma->obj->cache_level,
 					    PIN_GLOBAL, NULL);
+			mutex_unlock(&vma->vm->mutex);
+			reloc_cache_remap(&eb->reloc_cache, ev->vma->obj);
 			if (err)
 				return err;
 		}
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index de24e4b3b19b..be208a8f1ed0 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -393,6 +393,7 @@ int i915_vma_bind(struct i915_vma *vma,
 	u32 bind_flags;
 	u32 vma_flags;
 
+	lockdep_assert_held(&vma->vm->mutex);
 	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
 	GEM_BUG_ON(vma->size > vma->node.size);
 
-- 
2.31.1

[Intel-gfx] [PATCH v4 4/4] drm/i915: Require the vm mutex for i915_vma_bind()

[Thomas Hellström]

Protect updates of struct i915_vma flags and async binding / unbinding
with the vm::mutex. This means that i915_vma_bind() needs to assert
vm::mutex held. In order to make that possible drop the caching of
kmap_atomic() maps around i915_vma_bind().

An alternative would be to use kmap_local() but since we block cpu
unplugging during sleeps inside kmap_local() sections this may have
unwanted side-effects. Particularly since we might wait for gpu while
holding the vm mutex.

This change may theoretically increase execbuf cpu-usage on snb, but
at least on non-highmem systems that increase should be very small.

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 .../gpu/drm/i915/gem/i915_gem_execbuffer.c    | 50 ++++++++++++++++++-
 drivers/gpu/drm/i915/i915_vma.c               |  1 +
 2 files changed, 50 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index ec7c4a29a720..b8330f0bf652 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -1109,6 +1109,47 @@ static inline struct i915_ggtt *cache_to_ggtt(struct reloc_cache *cache)
 	return &i915->ggtt;
 }
 
+static void reloc_cache_unmap(struct reloc_cache *cache)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	vaddr = unmask_page(cache->vaddr);
+	if (cache->vaddr & KMAP)
+		kunmap_atomic(vaddr);
+	else
+		io_mapping_unmap_atomic((void __iomem *)vaddr);
+}
+
+static void reloc_cache_remap(struct reloc_cache *cache,
+			      struct drm_i915_gem_object *obj)
+{
+	void *vaddr;
+
+	if (!cache->vaddr)
+		return;
+
+	if (cache->vaddr & KMAP) {
+		struct page *page = i915_gem_object_get_page(obj, cache->page);
+
+		vaddr = kmap_atomic(page);
+		cache->vaddr = unmask_flags(cache->vaddr) |
+			(unsigned long)vaddr;
+	} else {
+		struct i915_ggtt *ggtt = cache_to_ggtt(cache);
+		unsigned long offset;
+
+		offset = cache->node.start;
+		if (!drm_mm_node_allocated(&cache->node))
+			offset += cache->page << PAGE_SHIFT;
+
+		cache->vaddr = (unsigned long)
+			io_mapping_map_atomic_wc(&ggtt->iomap, offset);
+	}
+}
+
 static void reloc_cache_reset(struct reloc_cache *cache, struct i915_execbuffer *eb)
 {
 	void *vaddr;
@@ -1373,10 +1414,17 @@ eb_relocate_entry(struct i915_execbuffer *eb,
 		 * batchbuffers.
 		 */
 		if (reloc->write_domain == I915_GEM_DOMAIN_INSTRUCTION &&
-		    GRAPHICS_VER(eb->i915) == 6) {
+		    GRAPHICS_VER(eb->i915) == 6 &&
+		    !i915_vma_is_bound(target->vma, I915_VMA_GLOBAL_BIND)) {
+			struct i915_vma *vma = target->vma;
+
+			reloc_cache_unmap(&eb->reloc_cache);
+			mutex_lock(&vma->vm->mutex);
 			err = i915_vma_bind(target->vma,
 					    target->vma->obj->cache_level,
 					    PIN_GLOBAL, NULL);
+			mutex_unlock(&vma->vm->mutex);
+			reloc_cache_remap(&eb->reloc_cache, ev->vma->obj);
 			if (err)
 				return err;
 		}
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index de24e4b3b19b..be208a8f1ed0 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -393,6 +393,7 @@ int i915_vma_bind(struct i915_vma *vma,
 	u32 bind_flags;
 	u32 vma_flags;
 
+	lockdep_assert_held(&vma->vm->mutex);
 	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
 	GEM_BUG_ON(vma->size > vma->node.size);
 
-- 
2.31.1

[Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915: Asynchronous vma unbinding part1

[Patchwork]

== Series Details ==

Series: drm/i915: Asynchronous vma unbinding part1
URL   : https://patchwork.freedesktop.org/series/98278/
State : warning

== Summary ==

$ dim checkpatch origin/drm-tip
135b8658be47 drm/i915: Avoid using the i915_fence_array when collecting dependencies
1f9c56b98070 drm/i915: remove questionable fence optimization during copy
c89fcf2d1bb6 drm/i915: Break out the i915_deps utility
-:250: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating?
#250: 
new file mode 100644

total: 0 errors, 1 warnings, 0 checks, 509 lines checked
faefbdcba339 drm/i915: Require the vm mutex for i915_vma_bind()

[Intel-gfx] ✗ Fi.CI.SPARSE: warning for drm/i915: Asynchronous vma unbinding part1

[Patchwork]

== Series Details ==

Series: drm/i915: Asynchronous vma unbinding part1
URL   : https://patchwork.freedesktop.org/series/98278/
State : warning

== Summary ==

$ dim sparse --fast origin/drm-tip
Sparse version: v0.6.2
Fast mode used, each commit won't be checked separately.

[Intel-gfx] ✓ Fi.CI.BAT: success for drm/i915: Asynchronous vma unbinding part1

[Patchwork]

[-- Attachment #1: Type: text/plain, Size: 5220 bytes --]

== Series Details ==

Series: drm/i915: Asynchronous vma unbinding part1
URL   : https://patchwork.freedesktop.org/series/98278/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_11025 -> Patchwork_21887
====================================================

Summary
-------

  **SUCCESS**

  No regressions found.

  External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/index.html

Participating hosts (45 -> 37)
------------------------------

  Missing    (8): bat-dg1-6 fi-hsw-4200u bat-dg1-5 fi-bsw-cyan bat-adlp-6 bat-adlp-4 fi-ctg-p8600 fi-bdw-samus 

Known issues
------------

  Here are the changes found in Patchwork_21887 that come from known issues:

### IGT changes ###

#### Issues hit ####

  * igt@amdgpu/amd_basic@cs-multi-fence:
    - fi-blb-e6850:       NOTRUN -> [SKIP][1] ([fdo#109271]) +17 similar issues
   [1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-blb-e6850/igt@amdgpu/amd_basic@cs-multi-fence.html

  * igt@amdgpu/amd_cs_nop@sync-fork-gfx0:
    - fi-skl-6600u:       NOTRUN -> [SKIP][2] ([fdo#109271]) +18 similar issues
   [2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-skl-6600u/igt@amdgpu/amd_cs_nop@sync-fork-gfx0.html

  * igt@amdgpu/amd_prime@amd-to-i915:
    - fi-kbl-soraka:      NOTRUN -> [SKIP][3] ([fdo#109271])
   [3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-kbl-soraka/igt@amdgpu/amd_prime@amd-to-i915.html

  * igt@gem_exec_suspend@basic-s0@smem:
    - fi-tgl-1115g4:      [PASS][4] -> [FAIL][5] ([i915#1888])
   [4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s0@smem.html
   [5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s0@smem.html

  * igt@gem_exec_suspend@basic-s3@smem:
    - fi-tgl-u2:          [PASS][6] -> [FAIL][7] ([i915#1888])
   [6]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-tgl-u2/igt@gem_exec_suspend@basic-s3@smem.html
   [7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-tgl-u2/igt@gem_exec_suspend@basic-s3@smem.html

  * igt@kms_frontbuffer_tracking@basic:
    - fi-cml-u2:          [PASS][8] -> [DMESG-WARN][9] ([i915#4269])
   [8]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-cml-u2/igt@kms_frontbuffer_tracking@basic.html
   [9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-cml-u2/igt@kms_frontbuffer_tracking@basic.html

  
#### Possible fixes ####

  * igt@gem_exec_suspend@basic-s3@smem:
    - fi-tgl-1115g4:      [FAIL][10] ([i915#1888]) -> [PASS][11]
   [10]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s3@smem.html
   [11]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s3@smem.html

  * igt@i915_selftest@live@requests:
    - fi-blb-e6850:       [DMESG-FAIL][12] ([i915#4528]) -> [PASS][13]
   [12]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-blb-e6850/igt@i915_selftest@live@requests.html
   [13]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-blb-e6850/igt@i915_selftest@live@requests.html

  * igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-b:
    - fi-cfl-8109u:       [DMESG-WARN][14] ([i915#295]) -> [PASS][15] +12 similar issues
   [14]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-cfl-8109u/igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-b.html
   [15]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-cfl-8109u/igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-b.html

  * igt@kms_psr@primary_page_flip:
    - fi-skl-6600u:       [INCOMPLETE][16] ([i915#4547] / [i915#4838]) -> [PASS][17]
   [16]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/fi-skl-6600u/igt@kms_psr@primary_page_flip.html
   [17]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/fi-skl-6600u/igt@kms_psr@primary_page_flip.html

  
  [fdo#109271]: https://bugs.freedesktop.org/show_bug.cgi?id=109271
  [i915#1888]: https://gitlab.freedesktop.org/drm/intel/issues/1888
  [i915#295]: https://gitlab.freedesktop.org/drm/intel/issues/295
  [i915#4269]: https://gitlab.freedesktop.org/drm/intel/issues/4269
  [i915#4528]: https://gitlab.freedesktop.org/drm/intel/issues/4528
  [i915#4547]: https://gitlab.freedesktop.org/drm/intel/issues/4547
  [i915#4838]: https://gitlab.freedesktop.org/drm/intel/issues/4838


Build changes
-------------

  * Linux: CI_DRM_11025 -> Patchwork_21887

  CI-20190529: 20190529
  CI_DRM_11025: 22bb175d528e42a3fc0dab8fcb3071773319da6f @ git://anongit.freedesktop.org/gfx-ci/linux
  IGT_6316: 731e09c15b4fd559ce8aec30065fccde17a9b834 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
  Patchwork_21887: faefbdcba339e812caceca2171e9c98c5e2b9582 @ git://anongit.freedesktop.org/gfx-ci/linux


== Linux commits ==

faefbdcba339 drm/i915: Require the vm mutex for i915_vma_bind()
c89fcf2d1bb6 drm/i915: Break out the i915_deps utility
1f9c56b98070 drm/i915: remove questionable fence optimization during copy
135b8658be47 drm/i915: Avoid using the i915_fence_array when collecting dependencies

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/index.html

[-- Attachment #2: Type: text/html, Size: 6355 bytes --]

[Intel-gfx] ✓ Fi.CI.IGT: success for drm/i915: Asynchronous vma unbinding part1

[Patchwork]

[-- Attachment #1: Type: text/plain, Size: 30265 bytes --]

== Series Details ==

Series: drm/i915: Asynchronous vma unbinding part1
URL   : https://patchwork.freedesktop.org/series/98278/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_11025_full -> Patchwork_21887_full
====================================================

Summary
-------

  **SUCCESS**

  No regressions found.

  

Participating hosts (10 -> 10)
------------------------------

  No changes in participating hosts

Known issues
------------

  Here are the changes found in Patchwork_21887_full that come from known issues:

### CI changes ###

#### Issues hit ####

  * boot:
    - shard-glk:          ([PASS][1], [PASS][2], [PASS][3], [PASS][4], [PASS][5], [PASS][6], [PASS][7], [PASS][8], [PASS][9], [PASS][10], [PASS][11], [PASS][12], [PASS][13], [PASS][14], [PASS][15], [PASS][16], [PASS][17], [PASS][18], [PASS][19], [PASS][20], [PASS][21], [PASS][22], [PASS][23], [PASS][24], [PASS][25]) -> ([PASS][26], [PASS][27], [PASS][28], [PASS][29], [PASS][30], [PASS][31], [PASS][32], [PASS][33], [PASS][34], [PASS][35], [PASS][36], [PASS][37], [PASS][38], [PASS][39], [PASS][40], [PASS][41], [FAIL][42], [PASS][43], [PASS][44], [PASS][45], [PASS][46], [PASS][47], [PASS][48], [PASS][49], [PASS][50]) ([i915#4392])
   [1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk9/boot.html
   [2]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk9/boot.html
   [3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk9/boot.html
   [4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk8/boot.html
   [5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk8/boot.html
   [6]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk8/boot.html
   [7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk7/boot.html
   [8]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk7/boot.html
   [9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk6/boot.html
   [10]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk6/boot.html
   [11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk5/boot.html
   [12]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk5/boot.html
   [13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk5/boot.html
   [14]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk4/boot.html
   [15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk4/boot.html
   [16]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk4/boot.html
   [17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk3/boot.html
   [18]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk3/boot.html
   [19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk3/boot.html
   [20]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk2/boot.html
   [21]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk2/boot.html
   [22]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk2/boot.html
   [23]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk1/boot.html
   [24]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk1/boot.html
   [25]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk1/boot.html
   [26]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk8/boot.html
   [27]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk9/boot.html
   [28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk9/boot.html
   [29]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk9/boot.html
   [30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk8/boot.html
   [31]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk8/boot.html
   [32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk7/boot.html
   [33]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk7/boot.html
   [34]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk7/boot.html
   [35]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk6/boot.html
   [36]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk6/boot.html
   [37]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk5/boot.html
   [38]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk5/boot.html
   [39]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk4/boot.html
   [40]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk4/boot.html
   [41]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk4/boot.html
   [42]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk4/boot.html
   [43]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk3/boot.html
   [44]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk3/boot.html
   [45]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk2/boot.html
   [46]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk2/boot.html
   [47]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk2/boot.html
   [48]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk1/boot.html
   [49]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk1/boot.html
   [50]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk1/boot.html

  

### IGT changes ###

#### Issues hit ####

  * igt@feature_discovery@psr2:
    - shard-iclb:         [PASS][51] -> [SKIP][52] ([i915#658])
   [51]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-iclb2/igt@feature_discovery@psr2.html
   [52]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb4/igt@feature_discovery@psr2.html

  * igt@gem_ctx_sseu@invalid-args:
    - shard-apl:          NOTRUN -> [SKIP][53] ([fdo#109271]) +121 similar issues
   [53]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl2/igt@gem_ctx_sseu@invalid-args.html

  * igt@gem_eio@in-flight-contexts-1us:
    - shard-skl:          [PASS][54] -> [TIMEOUT][55] ([i915#3063])
   [54]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl9/igt@gem_eio@in-flight-contexts-1us.html
   [55]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl8/igt@gem_eio@in-flight-contexts-1us.html

  * igt@gem_eio@unwedge-stress:
    - shard-iclb:         [PASS][56] -> [TIMEOUT][57] ([i915#2481] / [i915#3070])
   [56]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-iclb6/igt@gem_eio@unwedge-stress.html
   [57]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb3/igt@gem_eio@unwedge-stress.html

  * igt@gem_exec_balancer@parallel-keep-in-fence:
    - shard-tglb:         NOTRUN -> [SKIP][58] ([i915#4525])
   [58]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@gem_exec_balancer@parallel-keep-in-fence.html

  * igt@gem_exec_capture@pi@bcs0:
    - shard-skl:          NOTRUN -> [INCOMPLETE][59] ([i915#4547])
   [59]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl8/igt@gem_exec_capture@pi@bcs0.html

  * igt@gem_exec_fair@basic-none@vecs0:
    - shard-apl:          [PASS][60] -> [FAIL][61] ([i915#2842])
   [60]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-apl8/igt@gem_exec_fair@basic-none@vecs0.html
   [61]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl2/igt@gem_exec_fair@basic-none@vecs0.html

  * igt@gem_exec_fair@basic-pace@rcs0:
    - shard-kbl:          [PASS][62] -> [FAIL][63] ([i915#2842]) +1 similar issue
   [62]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-kbl1/igt@gem_exec_fair@basic-pace@rcs0.html
   [63]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl3/igt@gem_exec_fair@basic-pace@rcs0.html

  * igt@gem_exec_fair@basic-pace@vcs1:
    - shard-iclb:         NOTRUN -> [FAIL][64] ([i915#2842])
   [64]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb2/igt@gem_exec_fair@basic-pace@vcs1.html

  * igt@gem_exec_fair@basic-throttle@rcs0:
    - shard-iclb:         [PASS][65] -> [FAIL][66] ([i915#2849])
   [65]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-iclb8/igt@gem_exec_fair@basic-throttle@rcs0.html
   [66]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@gem_exec_fair@basic-throttle@rcs0.html

  * igt@gem_exec_params@secure-non-root:
    - shard-iclb:         NOTRUN -> [SKIP][67] ([fdo#112283])
   [67]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@gem_exec_params@secure-non-root.html

  * igt@gem_huc_copy@huc-copy:
    - shard-tglb:         [PASS][68] -> [SKIP][69] ([i915#2190])
   [68]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-tglb8/igt@gem_huc_copy@huc-copy.html
   [69]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb7/igt@gem_huc_copy@huc-copy.html
    - shard-skl:          NOTRUN -> [SKIP][70] ([fdo#109271] / [i915#2190])
   [70]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl9/igt@gem_huc_copy@huc-copy.html

  * igt@gem_lmem_swapping@heavy-random:
    - shard-apl:          NOTRUN -> [SKIP][71] ([fdo#109271] / [i915#4613]) +1 similar issue
   [71]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl4/igt@gem_lmem_swapping@heavy-random.html

  * igt@gem_lmem_swapping@heavy-verify-random:
    - shard-skl:          NOTRUN -> [SKIP][72] ([fdo#109271] / [i915#4613])
   [72]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl10/igt@gem_lmem_swapping@heavy-verify-random.html

  * igt@gem_lmem_swapping@parallel-random-verify:
    - shard-kbl:          NOTRUN -> [SKIP][73] ([fdo#109271] / [i915#4613]) +2 similar issues
   [73]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl4/igt@gem_lmem_swapping@parallel-random-verify.html

  * igt@gem_lmem_swapping@verify:
    - shard-iclb:         NOTRUN -> [SKIP][74] ([i915#4613]) +1 similar issue
   [74]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb2/igt@gem_lmem_swapping@verify.html

  * igt@gem_lmem_swapping@verify-random:
    - shard-tglb:         NOTRUN -> [SKIP][75] ([i915#4613]) +1 similar issue
   [75]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb8/igt@gem_lmem_swapping@verify-random.html

  * igt@gem_pxp@display-protected-crc:
    - shard-iclb:         NOTRUN -> [SKIP][76] ([i915#4270])
   [76]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@gem_pxp@display-protected-crc.html

  * igt@gem_render_copy@linear-to-vebox-y-tiled:
    - shard-iclb:         NOTRUN -> [SKIP][77] ([i915#768])
   [77]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@gem_render_copy@linear-to-vebox-y-tiled.html

  * igt@gem_userptr_blits@coherency-sync:
    - shard-tglb:         NOTRUN -> [SKIP][78] ([fdo#110542])
   [78]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@gem_userptr_blits@coherency-sync.html

  * igt@gem_workarounds@suspend-resume-context:
    - shard-apl:          [PASS][79] -> [DMESG-WARN][80] ([i915#180]) +1 similar issue
   [79]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-apl1/igt@gem_workarounds@suspend-resume-context.html
   [80]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl6/igt@gem_workarounds@suspend-resume-context.html

  * igt@gen9_exec_parse@allowed-all:
    - shard-apl:          [PASS][81] -> [DMESG-WARN][82] ([i915#1436] / [i915#716])
   [81]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-apl3/igt@gen9_exec_parse@allowed-all.html
   [82]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl7/igt@gen9_exec_parse@allowed-all.html

  * igt@gen9_exec_parse@allowed-single:
    - shard-skl:          [PASS][83] -> [DMESG-WARN][84] ([i915#1436] / [i915#716])
   [83]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl8/igt@gen9_exec_parse@allowed-single.html
   [84]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl4/igt@gen9_exec_parse@allowed-single.html

  * igt@gen9_exec_parse@basic-rejected-ctx-param:
    - shard-tglb:         NOTRUN -> [SKIP][85] ([i915#2856])
   [85]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@gen9_exec_parse@basic-rejected-ctx-param.html

  * igt@i915_pm_dc@dc6-dpms:
    - shard-kbl:          NOTRUN -> [FAIL][86] ([i915#454])
   [86]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl7/igt@i915_pm_dc@dc6-dpms.html
    - shard-tglb:         NOTRUN -> [FAIL][87] ([i915#454])
   [87]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@i915_pm_dc@dc6-dpms.html

  * igt@i915_pm_rpm@dpms-mode-unset-non-lpsp:
    - shard-tglb:         NOTRUN -> [SKIP][88] ([fdo#111644] / [i915#1397] / [i915#2411]) +1 similar issue
   [88]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@i915_pm_rpm@dpms-mode-unset-non-lpsp.html

  * igt@i915_pm_rpm@gem-pread:
    - shard-skl:          [PASS][89] -> [DMESG-WARN][90] ([i915#1982]) +2 similar issues
   [89]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl7/igt@i915_pm_rpm@gem-pread.html
   [90]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl3/igt@i915_pm_rpm@gem-pread.html

  * igt@i915_pm_sseu@full-enable:
    - shard-tglb:         NOTRUN -> [SKIP][91] ([i915#4387])
   [91]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@i915_pm_sseu@full-enable.html

  * igt@kms_big_fb@linear-8bpp-rotate-90:
    - shard-tglb:         NOTRUN -> [SKIP][92] ([fdo#111614])
   [92]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_big_fb@linear-8bpp-rotate-90.html

  * igt@kms_big_fb@y-tiled-64bpp-rotate-270:
    - shard-iclb:         NOTRUN -> [SKIP][93] ([fdo#110725] / [fdo#111614])
   [93]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_big_fb@y-tiled-64bpp-rotate-270.html

  * igt@kms_big_fb@y-tiled-max-hw-stride-64bpp-rotate-180-hflip:
    - shard-skl:          NOTRUN -> [SKIP][94] ([fdo#109271] / [i915#3777]) +1 similar issue
   [94]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl9/igt@kms_big_fb@y-tiled-max-hw-stride-64bpp-rotate-180-hflip.html

  * igt@kms_big_fb@yf-tiled-8bpp-rotate-90:
    - shard-tglb:         NOTRUN -> [SKIP][95] ([fdo#111615])
   [95]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_big_fb@yf-tiled-8bpp-rotate-90.html

  * igt@kms_big_fb@yf-tiled-max-hw-stride-32bpp-rotate-0-hflip:
    - shard-apl:          NOTRUN -> [SKIP][96] ([fdo#109271] / [i915#3777]) +1 similar issue
   [96]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl2/igt@kms_big_fb@yf-tiled-max-hw-stride-32bpp-rotate-0-hflip.html

  * igt@kms_ccs@pipe-a-bad-aux-stride-yf_tiled_ccs:
    - shard-tglb:         NOTRUN -> [SKIP][97] ([fdo#111615] / [i915#3689]) +1 similar issue
   [97]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb8/igt@kms_ccs@pipe-a-bad-aux-stride-yf_tiled_ccs.html

  * igt@kms_ccs@pipe-a-crc-sprite-planes-basic-y_tiled_gen12_rc_ccs_cc:
    - shard-kbl:          NOTRUN -> [SKIP][98] ([fdo#109271] / [i915#3886]) +2 similar issues
   [98]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl4/igt@kms_ccs@pipe-a-crc-sprite-planes-basic-y_tiled_gen12_rc_ccs_cc.html

  * igt@kms_ccs@pipe-c-bad-aux-stride-y_tiled_gen12_mc_ccs:
    - shard-iclb:         NOTRUN -> [SKIP][99] ([fdo#109278] / [i915#3886])
   [99]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_ccs@pipe-c-bad-aux-stride-y_tiled_gen12_mc_ccs.html

  * igt@kms_ccs@pipe-c-bad-rotation-90-y_tiled_gen12_mc_ccs:
    - shard-apl:          NOTRUN -> [SKIP][100] ([fdo#109271] / [i915#3886]) +5 similar issues
   [100]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl7/igt@kms_ccs@pipe-c-bad-rotation-90-y_tiled_gen12_mc_ccs.html

  * igt@kms_ccs@pipe-c-ccs-on-another-bo-y_tiled_gen12_rc_ccs_cc:
    - shard-skl:          NOTRUN -> [SKIP][101] ([fdo#109271] / [i915#3886]) +2 similar issues
   [101]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl10/igt@kms_ccs@pipe-c-ccs-on-another-bo-y_tiled_gen12_rc_ccs_cc.html

  * igt@kms_ccs@pipe-d-ccs-on-another-bo-y_tiled_ccs:
    - shard-tglb:         NOTRUN -> [SKIP][102] ([i915#3689]) +2 similar issues
   [102]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_ccs@pipe-d-ccs-on-another-bo-y_tiled_ccs.html

  * igt@kms_chamelium@hdmi-crc-nonplanar-formats:
    - shard-iclb:         NOTRUN -> [SKIP][103] ([fdo#109284] / [fdo#111827])
   [103]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_chamelium@hdmi-crc-nonplanar-formats.html

  * igt@kms_chamelium@hdmi-edid-change-during-suspend:
    - shard-apl:          NOTRUN -> [SKIP][104] ([fdo#109271] / [fdo#111827]) +10 similar issues
   [104]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl1/igt@kms_chamelium@hdmi-edid-change-during-suspend.html

  * igt@kms_chamelium@vga-hpd:
    - shard-tglb:         NOTRUN -> [SKIP][105] ([fdo#109284] / [fdo#111827]) +5 similar issues
   [105]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_chamelium@vga-hpd.html

  * igt@kms_chamelium@vga-hpd-without-ddc:
    - shard-kbl:          NOTRUN -> [SKIP][106] ([fdo#109271] / [fdo#111827]) +10 similar issues
   [106]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl7/igt@kms_chamelium@vga-hpd-without-ddc.html

  * igt@kms_color@pipe-d-ctm-blue-to-red:
    - shard-iclb:         NOTRUN -> [SKIP][107] ([fdo#109278] / [i915#1149])
   [107]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_color@pipe-d-ctm-blue-to-red.html

  * igt@kms_color_chamelium@pipe-d-ctm-0-25:
    - shard-skl:          NOTRUN -> [SKIP][108] ([fdo#109271] / [fdo#111827]) +4 similar issues
   [108]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl10/igt@kms_color_chamelium@pipe-d-ctm-0-25.html

  * igt@kms_color_chamelium@pipe-d-ctm-limited-range:
    - shard-iclb:         NOTRUN -> [SKIP][109] ([fdo#109278] / [fdo#109284] / [fdo#111827])
   [109]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_color_chamelium@pipe-d-ctm-limited-range.html

  * igt@kms_content_protection@atomic:
    - shard-iclb:         NOTRUN -> [SKIP][110] ([fdo#109300] / [fdo#111066])
   [110]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_content_protection@atomic.html

  * igt@kms_content_protection@srm:
    - shard-kbl:          NOTRUN -> [TIMEOUT][111] ([i915#1319])
   [111]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl4/igt@kms_content_protection@srm.html
    - shard-apl:          NOTRUN -> [TIMEOUT][112] ([i915#1319])
   [112]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl1/igt@kms_content_protection@srm.html

  * igt@kms_cursor_crc@pipe-a-cursor-512x170-random:
    - shard-iclb:         NOTRUN -> [SKIP][113] ([fdo#109278] / [fdo#109279])
   [113]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_cursor_crc@pipe-a-cursor-512x170-random.html

  * igt@kms_cursor_crc@pipe-c-cursor-32x32-onscreen:
    - shard-tglb:         NOTRUN -> [SKIP][114] ([i915#3319])
   [114]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_cursor_crc@pipe-c-cursor-32x32-onscreen.html

  * igt@kms_cursor_crc@pipe-d-cursor-512x512-offscreen:
    - shard-tglb:         NOTRUN -> [SKIP][115] ([fdo#109279] / [i915#3359])
   [115]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_cursor_crc@pipe-d-cursor-512x512-offscreen.html

  * igt@kms_cursor_legacy@cursor-vs-flip-atomic-transitions-varying-size:
    - shard-iclb:         [PASS][116] -> [FAIL][117] ([i915#2370])
   [116]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-iclb8/igt@kms_cursor_legacy@cursor-vs-flip-atomic-transitions-varying-size.html
   [117]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb7/igt@kms_cursor_legacy@cursor-vs-flip-atomic-transitions-varying-size.html

  * igt@kms_dp_tiled_display@basic-test-pattern-with-chamelium:
    - shard-tglb:         NOTRUN -> [SKIP][118] ([i915#3528])
   [118]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_dp_tiled_display@basic-test-pattern-with-chamelium.html

  * igt@kms_fbcon_fbt@fbc-suspend:
    - shard-apl:          [PASS][119] -> [INCOMPLETE][120] ([i915#180])
   [119]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-apl3/igt@kms_fbcon_fbt@fbc-suspend.html
   [120]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl7/igt@kms_fbcon_fbt@fbc-suspend.html
    - shard-kbl:          [PASS][121] -> [INCOMPLETE][122] ([i915#180] / [i915#636])
   [121]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-kbl3/igt@kms_fbcon_fbt@fbc-suspend.html
   [122]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl1/igt@kms_fbcon_fbt@fbc-suspend.html

  * igt@kms_flip@2x-plain-flip-fb-recreate:
    - shard-iclb:         NOTRUN -> [SKIP][123] ([fdo#109274]) +3 similar issues
   [123]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_flip@2x-plain-flip-fb-recreate.html

  * igt@kms_flip@2x-plain-flip-ts-check@ab-hdmi-a1-hdmi-a2:
    - shard-glk:          [PASS][124] -> [FAIL][125] ([i915#2122])
   [124]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-glk6/igt@kms_flip@2x-plain-flip-ts-check@ab-hdmi-a1-hdmi-a2.html
   [125]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-glk5/igt@kms_flip@2x-plain-flip-ts-check@ab-hdmi-a1-hdmi-a2.html

  * igt@kms_flip@flip-vs-expired-vblank@a-edp1:
    - shard-skl:          [PASS][126] -> [FAIL][127] ([i915#79])
   [126]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl3/igt@kms_flip@flip-vs-expired-vblank@a-edp1.html
   [127]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl4/igt@kms_flip@flip-vs-expired-vblank@a-edp1.html

  * igt@kms_flip@plain-flip-fb-recreate-interruptible@c-edp1:
    - shard-skl:          [PASS][128] -> [FAIL][129] ([i915#2122])
   [128]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl9/igt@kms_flip@plain-flip-fb-recreate-interruptible@c-edp1.html
   [129]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl6/igt@kms_flip@plain-flip-fb-recreate-interruptible@c-edp1.html

  * igt@kms_flip_scaled_crc@flip-32bpp-ytile-to-32bpp-ytilegen12rcccs:
    - shard-iclb:         NOTRUN -> [SKIP][130] ([i915#2587])
   [130]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_flip_scaled_crc@flip-32bpp-ytile-to-32bpp-ytilegen12rcccs.html

  * igt@kms_frontbuffer_tracking@fbc-2p-scndscrn-pri-indfb-draw-blt:
    - shard-skl:          NOTRUN -> [SKIP][131] ([fdo#109271]) +85 similar issues
   [131]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl10/igt@kms_frontbuffer_tracking@fbc-2p-scndscrn-pri-indfb-draw-blt.html

  * igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-pri-indfb-draw-mmap-wc:
    - shard-iclb:         NOTRUN -> [SKIP][132] ([fdo#109280]) +6 similar issues
   [132]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-pri-indfb-draw-mmap-wc.html

  * igt@kms_frontbuffer_tracking@psr-2p-primscrn-cur-indfb-draw-mmap-cpu:
    - shard-tglb:         NOTRUN -> [SKIP][133] ([fdo#111825]) +13 similar issues
   [133]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_frontbuffer_tracking@psr-2p-primscrn-cur-indfb-draw-mmap-cpu.html

  * igt@kms_hdr@bpc-switch:
    - shard-skl:          [PASS][134] -> [FAIL][135] ([i915#1188])
   [134]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-skl9/igt@kms_hdr@bpc-switch.html
   [135]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl6/igt@kms_hdr@bpc-switch.html

  * igt@kms_hdr@bpc-switch-suspend:
    - shard-kbl:          [PASS][136] -> [DMESG-WARN][137] ([i915#180]) +2 similar issues
   [136]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-kbl7/igt@kms_hdr@bpc-switch-suspend.html
   [137]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl7/igt@kms_hdr@bpc-switch-suspend.html

  * igt@kms_pipe_b_c_ivb@from-pipe-c-to-b-with-3-lanes:
    - shard-tglb:         NOTRUN -> [SKIP][138] ([fdo#109289])
   [138]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_pipe_b_c_ivb@from-pipe-c-to-b-with-3-lanes.html

  * igt@kms_pipe_crc_basic@read-crc-pipe-d:
    - shard-apl:          NOTRUN -> [SKIP][139] ([fdo#109271] / [i915#533])
   [139]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl7/igt@kms_pipe_crc_basic@read-crc-pipe-d.html

  * igt@kms_pipe_crc_basic@suspend-read-crc-pipe-d:
    - shard-kbl:          NOTRUN -> [SKIP][140] ([fdo#109271] / [i915#533])
   [140]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl7/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-d.html

  * igt@kms_plane_alpha_blend@pipe-b-alpha-basic:
    - shard-apl:          NOTRUN -> [FAIL][141] ([fdo#108145] / [i915#265])
   [141]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl2/igt@kms_plane_alpha_blend@pipe-b-alpha-basic.html

  * igt@kms_plane_alpha_blend@pipe-c-alpha-opaque-fb:
    - shard-skl:          NOTRUN -> [FAIL][142] ([fdo#108145] / [i915#265])
   [142]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl9/igt@kms_plane_alpha_blend@pipe-c-alpha-opaque-fb.html

  * igt@kms_plane_cursor@pipe-d-viewport-size-128:
    - shard-iclb:         NOTRUN -> [SKIP][143] ([fdo#109278]) +8 similar issues
   [143]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_plane_cursor@pipe-d-viewport-size-128.html

  * igt@kms_plane_lowres@pipe-a-tiling-yf:
    - shard-tglb:         NOTRUN -> [SKIP][144] ([fdo#111615] / [fdo#112054])
   [144]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_plane_lowres@pipe-a-tiling-yf.html

  * igt@kms_plane_lowres@pipe-c-tiling-yf:
    - shard-iclb:         NOTRUN -> [SKIP][145] ([i915#3536])
   [145]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@kms_plane_lowres@pipe-c-tiling-yf.html

  * igt@kms_plane_scaling@scaler-with-clipping-clamping@pipe-c-scaler-with-clipping-clamping:
    - shard-kbl:          NOTRUN -> [SKIP][146] ([fdo#109271] / [i915#2733])
   [146]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl3/igt@kms_plane_scaling@scaler-with-clipping-clamping@pipe-c-scaler-with-clipping-clamping.html

  * igt@kms_psr2_sf@cursor-plane-update-sf:
    - shard-skl:          NOTRUN -> [SKIP][147] ([fdo#109271] / [i915#658]) +2 similar issues
   [147]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-skl10/igt@kms_psr2_sf@cursor-plane-update-sf.html

  * igt@kms_psr2_su@page_flip-nv12:
    - shard-tglb:         NOTRUN -> [SKIP][148] ([i915#1911])
   [148]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb5/igt@kms_psr2_su@page_flip-nv12.html
    - shard-iclb:         NOTRUN -> [SKIP][149] ([fdo#111068] / [i915#658])
   [149]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb7/igt@kms_psr2_su@page_flip-nv12.html

  * igt@kms_psr2_su@page_flip-xrgb8888:
    - shard-kbl:          NOTRUN -> [SKIP][150] ([fdo#109271] / [i915#658])
   [150]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl3/igt@kms_psr2_su@page_flip-xrgb8888.html
    - shard-apl:          NOTRUN -> [SKIP][151] ([fdo#109271] / [i915#658]) +1 similar issue
   [151]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl7/igt@kms_psr2_su@page_flip-xrgb8888.html

  * igt@kms_psr@psr2_cursor_mmap_gtt:
    - shard-tglb:         NOTRUN -> [FAIL][152] ([i915#132] / [i915#3467])
   [152]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@kms_psr@psr2_cursor_mmap_gtt.html

  * igt@kms_psr@psr2_cursor_plane_onoff:
    - shard-iclb:         [PASS][153] -> [SKIP][154] ([fdo#109441])
   [153]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_11025/shard-iclb2/igt@kms_psr@psr2_cursor_plane_onoff.html
   [154]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb4/igt@kms_psr@psr2_cursor_plane_onoff.html

  * igt@kms_universal_plane@disable-primary-vs-flip-pipe-d:
    - shard-kbl:          NOTRUN -> [SKIP][155] ([fdo#109271]) +78 similar issues
   [155]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-kbl3/igt@kms_universal_plane@disable-primary-vs-flip-pipe-d.html

  * igt@kms_writeback@writeback-pixel-formats:
    - shard-apl:          NOTRUN -> [SKIP][156] ([fdo#109271] / [i915#2437])
   [156]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl2/igt@kms_writeback@writeback-pixel-formats.html

  * igt@nouveau_crc@ctx-flip-threshold-reset-after-capture:
    - shard-iclb:         NOTRUN -> [SKIP][157] ([i915#2530]) +1 similar issue
   [157]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb7/igt@nouveau_crc@ctx-flip-threshold-reset-after-capture.html

  * igt@prime_nv_api@i915_self_import:
    - shard-tglb:         NOTRUN -> [SKIP][158] ([fdo#109291])
   [158]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-tglb1/igt@prime_nv_api@i915_self_import.html

  * igt@prime_nv_test@nv_i915_sharing:
    - shard-iclb:         NOTRUN -> [SKIP][159] ([fdo#109291]) +2 similar issues
   [159]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-iclb5/igt@prime_nv_test@nv_i915_sharing.html

  * igt@sysfs_clients@recycle:
    - shard-apl:          NOTRUN -> [SKIP][160] ([fdo#109271] / [i915#2994]) +1 similar issue
   [160]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/shard-apl1/igt@sysfs_clients@recycle.html

  * igt@sysfs_clients@sema-25:
    - shard-kbl:          NOTRUN -> [

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_21887/index.html

[-- Attachment #2: Type: text/html, Size: 33833 bytes --]