new engine discovery interface

index 16e452aa..b14ca969 100644
@@ -62,6 +62,26 @@ extern "C" {
 #define I915_ERROR_UEVENT		"ERROR"
 #define I915_RESET_UEVENT		"RESET"
 
+/*
+ * i915_user_extension: Base class for defining a chain of extensions
+ *
+ * Many interfaces need to grow over time. In most cases we can simply
+ * extend the struct and have userspace pass in more data. Another option,
+ * as demonstrated by Vulkan's approach to providing extensions for forward
+ * and backward compatibility, is to use a list of optional structs to
+ * provide those extra details.
+ *
+ * The key advantage to using an extension chain is that it allows us to
+ * redefine the interface more easily than an ever growing struct of
+ * increasing complexity, and for large parts of that interface to be
+ * entirely optional. The downside is more pointer chasing; chasing across
+ * the boundary with pointers encapsulated inside u64.
+ */
+struct i915_user_extension {
+	__u64 next_extension;
+	__u64 name;
+};
+
 /*
  * MOCS indexes used for GPU surfaces, defining the cacheability of the
  * surface data and the coherency for this data wrt. CPU vs. GPU accesses.

@@ -367,6 +387,7 @@ typedef struct _drm_i915_sarea {
 #define DRM_IOCTL_I915_GET_SPRITE_COLORKEY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GET_SPRITE_COLORKEY, struct drm_intel_sprite_colorkey)
 #define DRM_IOCTL_I915_GEM_WAIT		DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_WAIT, struct drm_i915_gem_wait)
 #define DRM_IOCTL_I915_GEM_CONTEXT_CREATE	DRM_IOWR (DRM_COMMAND_BASE + DRM_I915_GEM_CONTEXT_CREATE, struct drm_i915_gem_context_create)
+#define DRM_IOCTL_I915_GEM_CONTEXT_CREATE_v2	DRM_IOWR (DRM_COMMAND_BASE + DRM_I915_GEM_CONTEXT_CREATE, struct drm_i915_gem_context_create_v2)
 #define DRM_IOCTL_I915_GEM_CONTEXT_DESTROY	DRM_IOW (DRM_COMMAND_BASE + DRM_I915_GEM_CONTEXT_DESTROY, struct drm_i915_gem_context_destroy)
 #define DRM_IOCTL_I915_REG_READ			DRM_IOWR (DRM_COMMAND_BASE + DRM_I915_REG_READ, struct drm_i915_reg_read)
 #define DRM_IOCTL_I915_GET_RESET_STATS		DRM_IOWR (DRM_COMMAND_BASE + DRM_I915_GET_RESET_STATS, struct drm_i915_reset_stats)

@@ -412,6 +433,14 @@ typedef struct drm_i915_irq_wait {
 	int irq_seq;
 } drm_i915_irq_wait_t;
 
+/*
+ * Different modes of per-process Graphics Translation Table,
+ * see I915_PARAM_HAS_ALIASING_PPGTT
+ */
+#define I915_GEM_PPGTT_NONE	0
+#define I915_GEM_PPGTT_ALIASING	1
+#define I915_GEM_PPGTT_FULL	2
+
 /* Ioctl to query kernel params:
  */
 #define I915_PARAM_IRQ_ACTIVE            1

@@ -529,6 +558,35 @@ typedef struct drm_i915_irq_wait {
  */
 #define I915_PARAM_CS_TIMESTAMP_FREQUENCY 51
 
+/*
+ * Once upon a time we supposed that writes through the GGTT would be
+ * immediately in physical memory (once flushed out of the CPU path). However,
+ * on a few different processors and chipsets, this is not necessarily the case
+ * as the writes appear to be buffered internally. Thus a read of the backing
+ * storage (physical memory) via a different path (with different physical tags
+ * to the indirect write via the GGTT) will see stale values from before
+ * the GGTT write. Inside the kernel, we can for the most part keep track of
+ * the different read/write domains in use (e.g. set-domain), but the assumption
+ * of coherency is baked into the ABI, hence reporting its true state in this
+ * parameter.
+ *
+ * Reports true when writes via mmap_gtt are immediately visible following an
+ * lfence to flush the WCB.
+ *
+ * Reports false when writes via mmap_gtt are indeterminately delayed in an in
+ * internal buffer and are _not_ immediately visible to third parties accessing
+ * directly via mmap_cpu/mmap_wc. Use of mmap_gtt as part of an IPC
+ * communications channel when reporting false is strongly disadvised.
+ */
+#define I915_PARAM_MMAP_GTT_COHERENT	52
+
+/*
+ * Query whether DRM_I915_GEM_EXECBUFFER2 supports coordination of parallel
+ * execution through use of explicit fence support.
+ * See I915_EXEC_FENCE_OUT and I915_EXEC_FENCE_SUBMIT.
+ */
+#define I915_PARAM_HAS_EXEC_SUBMIT_FENCE 53
+
 typedef struct drm_i915_getparam {
 	__s32 param;
 	/*

@@ -942,7 +1000,7 @@ struct drm_i915_gem_execbuffer2 {
 	 * struct drm_i915_gem_exec_fence *fences.
 	 */
 	__u64 cliprects_ptr;
-#define I915_EXEC_RING_MASK              (7<<0)
+#define I915_EXEC_RING_MASK              (0x3f)
 #define I915_EXEC_DEFAULT                (0<<0)
 #define I915_EXEC_RENDER                 (1<<0)
 #define I915_EXEC_BSD                    (2<<0)

@@ -1048,7 +1106,16 @@ struct drm_i915_gem_execbuffer2 {
  */
 #define I915_EXEC_FENCE_ARRAY   (1<<19)
 
-#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_FENCE_ARRAY<<1))
+/*
+ * Setting I915_EXEC_FENCE_SUBMIT implies that lower_32_bits(rsvd2) represent
+ * a sync_file fd to wait upon (in a nonblocking manner) prior to executing
+ * the batch.
+ *
+ * Returns -EINVAL if the sync_file fd cannot be found.
+ */
+#define I915_EXEC_FENCE_SUBMIT		(1<<20)
+
+#define __I915_EXEC_UNKNOWN_FLAGS (-(I915_EXEC_FENCE_SUBMIT<<1))
 
 #define I915_EXEC_CONTEXT_ID_MASK	(0xffffffff)
 #define i915_execbuffer2_set_context_id(eb2, context) \

@@ -1387,6 +1454,16 @@ struct drm_i915_gem_context_create {
 	__u32 pad;
 };
 
+struct drm_i915_gem_context_create_v2 {
+	/*  output: id of new context*/
+	__u32 ctx_id;
+	__u32 flags;
+#define I915_GEM_CONTEXT_SHARE_GTT		0x1
+#define I915_GEM_CONTEXT_SINGLE_TIMELINE	0x2
+	__u32 share_ctx;
+	__u32 pad;
+};
+
 struct drm_i915_gem_context_destroy {
 	__u32 ctx_id;
 	__u32 pad;

@@ -1456,9 +1533,122 @@ struct drm_i915_gem_context_param {
 #define   I915_CONTEXT_MAX_USER_PRIORITY	1023 /* inclusive */
 #define   I915_CONTEXT_DEFAULT_PRIORITY		0
 #define   I915_CONTEXT_MIN_USER_PRIORITY	-1023 /* inclusive */
+
+/*
+ * I915_CONTEXT_PARAM_ENGINES:
+ *
+ * Bind this context to operate on this subset of available engines. Henceforth,
+ * the I915_EXEC_RING selector for DRM_IOCTL_I915_GEM_EXECBUFFER2 operates as
+ * an index into this array of engines; I915_EXEC_DEFAULT selecting engine[0]
+ * and upwards. The array created is offset by 1, such that by default
+ * I915_EXEC_DEFAULT is left empty, to be filled in as directed. Slots 1...N
+ * are then filled in using the specified (class, instance).
+ *
+ * Setting the number of engines bound to the context will revert back to
+ * default settings.
+ *
+ * See struct i915_context_param_engines.
+ *
+ * Extensions:
+ *   i915_context_engines_load_balance (I915_CONTEXT_ENGINES_EXT_LOAD_BALANCE)
+ *   i915_context_engines_bond (I915_CONTEXT_ENGINES_EXT_BOND)
+ */
+#define I915_CONTEXT_PARAM_ENGINES	0x7
+
+/*
+ * When using the following param, value should be a pointer to
+ * drm_i915_gem_context_param_sseu.
+ */
+#define I915_CONTEXT_PARAM_SSEU		0x8
+
 	__u64 value;
 };
 
+/*
+ * i915_context_engines_load_balance:
+ *
+ * Enable load balancing across this set of engines.
+ *
+ * Into the I915_EXEC_DEFAULT slot, a virtual engine is created that when
+ * used will proxy the execbuffer request onto one of the set of engines
+ * in such a way as to distribute the load evenly across the set.
+ *
+ * The set of engines must be compatible (e.g. the same HW class) as they
+ * will share the same logical GPU context and ring.
+ *
+ * The context must be defined to use a single timeline for all engines.
+ */
+struct i915_context_engines_load_balance {
+	struct i915_user_extension base;
+
+	__u64 flags; /* all undefined flags must be zero */
+	__u64 engines_mask;
+
+	__u64 mbz[4]; /* reserved for future use; must be zero */
+};
+
+/*
+ * i915_context_engines_bond:
+ *
+ */
+struct i915_context_engines_bond {
+	struct i915_user_extension base;
+
+	__u16 master_class;
+	__u16 master_instance;
+	__u32 flags; /* all undefined flags must be zero */
+	__u64 sibling_mask;
+};
+
+struct i915_context_param_engines {
+	__u64 extensions;
+#define I915_CONTEXT_ENGINES_EXT_LOAD_BALANCE 0
+#define I915_CONTEXT_ENGINES_EXT_BOND 1
+
+	struct {
+		__u16 class; /* see enum drm_i915_gem_engine_class */
+		__u16 instance;
+	} class_instance[0];
+};
+
+struct drm_i915_gem_context_param_sseu {
+	/*
+	 * Engine class & instance to be configured or queried.
+	 */
+	__u16 class;
+	__u16 instance;
+
+	/*
+	 * Unused for now. Must be cleared to zero.
+	 */
+	__u32 rsvd1;
+
+	/*
+	 * Mask of slices to enable for the context. Valid values are a subset
+	 * of the bitmask value returned for I915_PARAM_SLICE_MASK.
+	 */
+	__u64 slice_mask;
+
+	/*
+	 * Mask of subslices to enable for the context. Valid values are a
+	 * subset of the bitmask value return by I915_PARAM_SUBSLICE_MASK.
+	 */
+	__u64 subslice_mask;
+
+	/*
+	 * Minimum/Maximum number of EUs to enable per subslice for the
+	 * context. min_eus_per_subslice must be inferior or equal to
+	 * max_eus_per_subslice.
+	 */
+	__u16 min_eus_per_subslice;
+	__u16 max_eus_per_subslice;
+
+	/*
+	 * Unused for now. Must be cleared to zero.
+	 */
+	__u32 rsvd2;
+};
+
 enum drm_i915_oa_format {
 	I915_OA_FORMAT_A13 = 1,	    /* HSW only */
 	I915_OA_FORMAT_A29,	    /* HSW only */

@@ -1620,6 +1810,7 @@ struct drm_i915_perf_oa_config {
 struct drm_i915_query_item {
 	__u64 query_id;
 #define DRM_I915_QUERY_TOPOLOGY_INFO    1
+#define DRM_I915_QUERY_ENGINE_INFO	2
 
 	/*
 	 * When set to zero by userspace, this is filled with the size of the

@@ -1717,6 +1908,50 @@ struct drm_i915_query_topology_info {
 	__u8 data[];
 };
 
+/**
+ * struct drm_i915_engine_info
+ *
+ * Describes one engine and it's capabilities as known to the driver.
+ */
+struct drm_i915_engine_info {
+	/** Engine class as in enum drm_i915_gem_engine_class. */
+	__u16 class;
+
+	/** Engine instance number. */
+	__u16 instance;
+
+	/** Reserved field. */
+	__u32 rsvd0;
+
+	/** Engine flags. */
+	__u64 flags;
+
+	/** Capabilities of this engine. */
+	__u64 capabilities;
+#define I915_VIDEO_CLASS_CAPABILITY_HEVC		(1 << 0)
+#define I915_VIDEO_AND_ENHANCE_CLASS_CAPABILITY_SFC	(1 << 1)
+
+	/** Reserved fields. */
+	__u64 rsvd1[4];
+};
+
+/**
+ * struct drm_i915_query_engine_info
+ *
+ * Engine info query enumerates all engines known to the driver by filling in
+ * an array of struct drm_i915_engine_info structures.
+ */
+struct drm_i915_query_engine_info {
+	/** Number of struct drm_i915_engine_info structs following. */
+	__u32 num_engines;
+
+	/** MBZ */
+	__u32 rsvd[3];
+
+	/** Marker for drm_i915_engine_info structures. */
+	struct drm_i915_engine_info engines[];
+};
+
 #if defined(__cplusplus)
 }
 #endif

index a2061924..4a7b8671 100644
@@ -650,3 +650,86 @@ bool gem_ring_has_physical_engine(int fd, unsigned ring)
 
 	return gem_has_ring(fd, ring);
 }
+
+static struct drm_i915_query_engine_info *__query_engines(int fd)
+{
+	struct drm_i915_query query;
+	struct drm_i915_query_item item;
+	struct drm_i915_query_engine_info *query_engines;
+	ssize_t size = sizeof(*query_engines) + 10 * sizeof(*query_engines->engines);
+	int err;
+
+	query_engines = malloc(size);
+	if (!query_engines)
+		return NULL;
+
+	memset(&query, 0, sizeof(query));
+	memset(&item, 0, sizeof(item));
+	memset(query_engines, 0, sizeof(*query_engines));
+
+	item.query_id = DRM_I915_QUERY_ENGINE_INFO;
+	item.length = size;
+	item.flags = 0;
+	item.data_ptr = to_user_pointer(query_engines);
+
+	query.items_ptr = to_user_pointer(&item);
+	query.num_items = 1;
+
+	err = igt_ioctl(fd, DRM_IOCTL_I915_QUERY, &query);
+	if (err) {
+		free(query_engines);
+		return NULL;
+	}
+
+	return query_engines;
+}
+
+static int __set_param(int fd, unsigned ctx_id,
+		struct drm_i915_query_engine_info *query_engine)
+{
+	int i, ret;
+	struct drm_i915_gem_context_param ctx_param;
+	struct i915_context_param_engines *ctx_engine;
+	size_t size = sizeof(struct i915_context_param_engines) +
+		      query_engine->num_engines *
+		      sizeof(*ctx_engine->class_instance);
+
+	ctx_engine = malloc(size);
+	if (!ctx_engine)
+		return errno;
+
+	ctx_engine->extensions = 0;
+	for (i = 0; i < query_engine->num_engines; i++) {
+		ctx_engine->class_instance[i].class = query_engine->engines[i].class;
+		ctx_engine->class_instance[i].instance = query_engine->engines[i].instance;
+	}
+
+	ctx_param.ctx_id = ctx_id;
+	ctx_param.size = size;
+	ctx_param.param = I915_CONTEXT_PARAM_ENGINES;
+	ctx_param.value = to_user_pointer(ctx_engine);
+
+	ret = igt_ioctl(fd, DRM_IOCTL_I915_GEM_CONTEXT_SETPARAM, &ctx_param);
+	free(ctx_engine);
+
+	return ret;
+}
+
+int get_engines(int fd, uint32_t ctx_id)
+{
+	struct drm_i915_query_engine_info *query_engine;
+	int n;
+
+	query_engine = __query_engines(fd);
+	if (!query_engine)
+		return -1;
+
+	n = __set_param(fd, ctx_id, query_engine);
+	if (n < 0)
+		return n;
+
+	n = query_engine->num_engines;
+	free(query_engine);
+
+	return n;
+}

index 54e95da9..7e299c31 100644
@@ -86,8 +86,16 @@ extern const struct intel_execution_engine {
 	     e__++) \
 		for_if (gem_ring_has_physical_engine(fd__, flags__ = e__->exec_id | e__->flags))
 
+#define for_each_engine_ctx(fd, ctx, e) \
+		for (int __m, __e = 1; (__m = get_engines(fd, ctx)); ) \
+			if (__e > __m) \
+				break; \
+			else \
+				for(e = __e ; __e <= __m; e = ++__e)
+
 bool gem_ring_is_physical_engine(int fd, unsigned int ring);
 bool gem_ring_has_physical_engine(int fd, unsigned int ring);
+int get_engines(int fd, uint32_t ctx_id);
 
 bool gem_can_store_dword(int fd, unsigned int engine);
 

56 lines not on this page · raw message