* [igt-dev] [PATCH v22 1/6] include/drm-uapi: import i915_drm.h header file
2019-04-16 23:10 [igt-dev] [PATCH v22 0/6] new engine discovery interface Andi Shyti
@ 2019-04-16 23:10 ` Andi Shyti
2019-04-16 23:10 ` [igt-dev] [PATCH v22 2/6] lib/i915: add gem_engine_topology library and for_each loop definition Andi Shyti
` (7 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Andi Shyti @ 2019-04-16 23:10 UTC (permalink / raw)
To: IGT dev; +Cc: Andi Shyti
From: Andi Shyti <andi.shyti@intel.com>
This header file is imported in order to include the two new
ioctls DRM_IOCTL_I915_GEM_CONTEXT_SETPARAM,
DRM_IOCTL_I915_GEM_CONTEXT_GETPARAM and DRM_IOCTL_I915_QUERY.
Signed-off-by: Andi Shyti <andi.shyti@intel.com>
---
include/drm-uapi/i915_drm.h | 361 ++++++++++++++++++++++++++++++------
1 file changed, 304 insertions(+), 57 deletions(-)
diff --git a/include/drm-uapi/i915_drm.h b/include/drm-uapi/i915_drm.h
index 4ae1c6ff6ae6..2bbad08eb9d2 100644
--- a/include/drm-uapi/i915_drm.h
+++ b/include/drm-uapi/i915_drm.h
@@ -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.
@@ -104,6 +124,9 @@ enum drm_i915_gem_engine_class {
I915_ENGINE_CLASS_INVALID = -1
};
+#define I915_ENGINE_CLASS_INVALID_NONE -1
+#define I915_ENGINE_CLASS_INVALID_VIRTUAL 0
+
/**
* DOC: perf_events exposed by i915 through /sys/bus/event_sources/drivers/i915
*
@@ -321,6 +344,8 @@ typedef struct _drm_i915_sarea {
#define DRM_I915_PERF_ADD_CONFIG 0x37
#define DRM_I915_PERF_REMOVE_CONFIG 0x38
#define DRM_I915_QUERY 0x39
+#define DRM_I915_GEM_VM_CREATE 0x3a
+#define DRM_I915_GEM_VM_DESTROY 0x3b
/* Must be kept compact -- no holes */
#define DRM_IOCTL_I915_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_I915_INIT, drm_i915_init_t)
@@ -370,6 +395,7 @@ 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_EXT DRM_IOWR (DRM_COMMAND_BASE + DRM_I915_GEM_CONTEXT_CREATE, struct drm_i915_gem_context_create_ext)
#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)
@@ -380,6 +406,8 @@ typedef struct _drm_i915_sarea {
#define DRM_IOCTL_I915_PERF_ADD_CONFIG DRM_IOW(DRM_COMMAND_BASE + DRM_I915_PERF_ADD_CONFIG, struct drm_i915_perf_oa_config)
#define DRM_IOCTL_I915_PERF_REMOVE_CONFIG DRM_IOW(DRM_COMMAND_BASE + DRM_I915_PERF_REMOVE_CONFIG, __u64)
#define DRM_IOCTL_I915_QUERY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_QUERY, struct drm_i915_query)
+#define DRM_IOCTL_I915_GEM_VM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_VM_CREATE, struct drm_i915_gem_vm_control)
+#define DRM_IOCTL_I915_GEM_VM_DESTROY DRM_IOW (DRM_COMMAND_BASE + DRM_I915_GEM_VM_DESTROY, struct drm_i915_gem_vm_control)
/* Allow drivers to submit batchbuffers directly to hardware, relying
* on the security mechanisms provided by hardware.
@@ -563,6 +591,12 @@ typedef struct drm_i915_irq_wait {
*/
#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
/* Must be kept compact -- no holes and well documented */
typedef struct drm_i915_getparam {
@@ -1085,7 +1119,16 @@ 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) \
@@ -1421,65 +1464,18 @@ struct drm_i915_gem_wait {
};
struct drm_i915_gem_context_create {
- /* output: id of new context*/
- __u32 ctx_id;
- __u32 pad;
-};
-
-struct drm_i915_gem_context_destroy {
- __u32 ctx_id;
- __u32 pad;
-};
-
-struct drm_i915_reg_read {
- /*
- * Register offset.
- * For 64bit wide registers where the upper 32bits don't immediately
- * follow the lower 32bits, the offset of the lower 32bits must
- * be specified
- */
- __u64 offset;
-#define I915_REG_READ_8B_WA (1ul << 0)
-
- __u64 val; /* Return value */
-};
-/* Known registers:
- *
- * Render engine timestamp - 0x2358 + 64bit - gen7+
- * - Note this register returns an invalid value if using the default
- * single instruction 8byte read, in order to workaround that pass
- * flag I915_REG_READ_8B_WA in offset field.
- *
- */
-
-struct drm_i915_reset_stats {
- __u32 ctx_id;
- __u32 flags;
-
- /* All resets since boot/module reload, for all contexts */
- __u32 reset_count;
-
- /* Number of batches lost when active in GPU, for this context */
- __u32 batch_active;
-
- /* Number of batches lost pending for execution, for this context */
- __u32 batch_pending;
-
+ __u32 ctx_id; /* output: id of new context*/
__u32 pad;
};
-struct drm_i915_gem_userptr {
- __u64 user_ptr;
- __u64 user_size;
+struct drm_i915_gem_context_create_ext {
+ __u32 ctx_id; /* output: id of new context*/
__u32 flags;
-#define I915_USERPTR_READ_ONLY 0x1
-#define I915_USERPTR_UNSYNCHRONIZED 0x80000000
- /**
- * Returned handle for the object.
- *
- * Object handles are nonzero.
- */
- __u32 handle;
+#define I915_CONTEXT_CREATE_FLAGS_USE_EXTENSIONS (1u << 0)
+#define I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE (1u << 1)
+#define I915_CONTEXT_CREATE_FLAGS_UNKNOWN \
+ (-(I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE << 1))
+ __u64 extensions;
};
struct drm_i915_gem_context_param {
@@ -1520,7 +1516,43 @@ struct drm_i915_gem_context_param {
* On creation, all new contexts are marked as recoverable.
*/
#define I915_CONTEXT_PARAM_RECOVERABLE 0x8
+
+ /*
+ * The id of the associated virtual memory address space (ppGTT) of
+ * this context. Can be retrieved and passed to another context
+ * (on the same fd) for both to use the same ppGTT and so share
+ * address layouts, and avoid reloading the page tables on context
+ * switches between themselves.
+ *
+ * See DRM_I915_GEM_VM_CREATE and DRM_I915_GEM_VM_DESTROY.
+ */
+#define I915_CONTEXT_PARAM_VM 0x9
+
+/*
+ * 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. Slots 0...N are filled in using the specified (class, instance).
+ * Use
+ * engine_class: I915_ENGINE_CLASS_INVALID,
+ * engine_instance: I915_ENGINE_CLASS_INVALID_NONE
+ * to specify a gap in the array that can be filled in later, e.g. by a
+ * virtual engine used for load balancing.
+ *
+ * Setting the number of engines bound to the context to 0, by passing a zero
+ * sized argument, 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 0xa
/* Must be kept compact -- no holes and well documented */
+
__u64 value;
};
@@ -1553,9 +1585,10 @@ struct drm_i915_gem_context_param_sseu {
__u16 engine_instance;
/*
- * Unused for now. Must be cleared to zero.
+ * Unknown flags must be cleared to zero.
*/
__u32 flags;
+#define I915_CONTEXT_SSEU_FLAG_ENGINE_INDEX (1u << 0)
/*
* Mask of slices to enable for the context. Valid values are a subset
@@ -1583,6 +1616,175 @@ struct drm_i915_gem_context_param_sseu {
__u32 rsvd;
};
+/*
+ * i915_context_engines_load_balance:
+ *
+ * Enable load balancing across this set of engines.
+ *
+ * Into the I915_EXEC_DEFAULT slot [0], 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.
+ *
+ * To intermix rendering with the virtual engine and direct rendering onto
+ * the backing engines (bypassing the load balancing proxy), the context must
+ * be defined to use a single timeline for all engines.
+ */
+struct i915_context_engines_load_balance {
+ struct i915_user_extension base;
+
+ __u16 engine_index;
+ __u16 mbz16; /* reserved for future use; must be zero */
+ __u32 flags; /* all undefined flags must be zero */
+
+ __u64 engines_mask; /* selection mask of engines[] */
+
+ __u64 mbz64[4]; /* reserved for future use; must be zero */
+};
+
+/*
+ * i915_context_engines_bond:
+ *
+ */
+struct i915_context_engines_bond {
+ struct i915_user_extension base;
+
+ __u16 engine_index;
+ __u16 mbz;
+
+ __u16 master_class;
+ __u16 master_instance;
+
+ __u64 sibling_mask;
+ __u64 flags; /* all undefined flags must be zero */
+};
+
+struct i915_context_param_engines {
+ __u64 extensions; /* linked chain of extension blocks, 0 terminates */
+#define I915_CONTEXT_ENGINES_EXT_LOAD_BALANCE 0
+#define I915_CONTEXT_ENGINES_EXT_BOND 1
+
+ struct {
+ __u16 engine_class; /* see enum drm_i915_gem_engine_class */
+ __u16 engine_instance;
+ } class_instance[0];
+} __attribute__((packed));
+
+#define I915_DEFINE_CONTEXT_PARAM_ENGINES(name__, N__) struct { \
+ __u64 extensions; \
+ struct { \
+ __u16 engine_class; \
+ __u16 engine_instance; \
+ } class_instance[N__]; \
+} __attribute__((packed)) name__
+
+struct drm_i915_gem_context_create_ext_setparam {
+#define I915_CONTEXT_CREATE_EXT_SETPARAM 0
+ struct i915_user_extension base;
+ struct drm_i915_gem_context_param setparam;
+};
+
+struct drm_i915_gem_context_create_ext_clone {
+#define I915_CONTEXT_CREATE_EXT_CLONE 1
+ struct i915_user_extension base;
+ __u32 clone_id;
+ __u32 flags;
+#define I915_CONTEXT_CLONE_FLAGS (1u << 0)
+#define I915_CONTEXT_CLONE_SCHED (1u << 1)
+#define I915_CONTEXT_CLONE_SSEU (1u << 2)
+#define I915_CONTEXT_CLONE_TIMELINE (1u << 3)
+#define I915_CONTEXT_CLONE_VM (1u << 4)
+#define I915_CONTEXT_CLONE_ENGINES (1u << 5)
+#define I915_CONTEXT_CLONE_UNKNOWN -(I915_CONTEXT_CLONE_ENGINES << 1)
+ __u64 rsvd;
+};
+
+struct drm_i915_gem_context_destroy {
+ __u32 ctx_id;
+ __u32 pad;
+};
+
+/*
+ * DRM_I915_GEM_VM_CREATE -
+ *
+ * Create a new virtual memory address space (ppGTT) for use within a context
+ * on the same file. Extensions can be provided to configure exactly how the
+ * address space is setup upon creation.
+ *
+ * The id of new VM (bound to the fd) for use with I915_CONTEXT_PARAM_VM is
+ * returned in the outparam @id.
+ *
+ * No flags are defined, with all bits reserved and must be zero.
+ *
+ * An extension chain maybe provided, starting with @extensions, and terminated
+ * by the @next_extension being 0. Currently, no extensions are defined.
+ *
+ * DRM_I915_GEM_VM_DESTROY -
+ *
+ * Destroys a previously created VM id, specified in @id.
+ *
+ * No extensions or flags are allowed currently, and so must be zero.
+ */
+struct drm_i915_gem_vm_control {
+ __u64 extensions;
+ __u32 flags;
+ __u32 id;
+};
+
+struct drm_i915_reg_read {
+ /*
+ * Register offset.
+ * For 64bit wide registers where the upper 32bits don't immediately
+ * follow the lower 32bits, the offset of the lower 32bits must
+ * be specified
+ */
+ __u64 offset;
+#define I915_REG_READ_8B_WA (1ul << 0)
+
+ __u64 val; /* Return value */
+};
+
+/* Known registers:
+ *
+ * Render engine timestamp - 0x2358 + 64bit - gen7+
+ * - Note this register returns an invalid value if using the default
+ * single instruction 8byte read, in order to workaround that pass
+ * flag I915_REG_READ_8B_WA in offset field.
+ *
+ */
+
+struct drm_i915_reset_stats {
+ __u32 ctx_id;
+ __u32 flags;
+
+ /* All resets since boot/module reload, for all contexts */
+ __u32 reset_count;
+
+ /* Number of batches lost when active in GPU, for this context */
+ __u32 batch_active;
+
+ /* Number of batches lost pending for execution, for this context */
+ __u32 batch_pending;
+
+ __u32 pad;
+};
+
+struct drm_i915_gem_userptr {
+ __u64 user_ptr;
+ __u64 user_size;
+ __u32 flags;
+#define I915_USERPTR_READ_ONLY 0x1
+#define I915_USERPTR_UNSYNCHRONIZED 0x80000000
+ /**
+ * Returned handle for the object.
+ *
+ * Object handles are nonzero.
+ */
+ __u32 handle;
+};
+
enum drm_i915_oa_format {
I915_OA_FORMAT_A13 = 1, /* HSW only */
I915_OA_FORMAT_A29, /* HSW only */
@@ -1744,6 +1946,7 @@ 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
/* Must be kept compact -- no holes and well documented */
/*
@@ -1842,6 +2045,50 @@ 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 engine_class;
+
+ /** Engine instance number. */
+ __u16 engine_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
--
2.20.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 14+ messages in thread* [igt-dev] [PATCH v22 2/6] lib/i915: add gem_engine_topology library and for_each loop definition
2019-04-16 23:10 [igt-dev] [PATCH v22 0/6] new engine discovery interface Andi Shyti
2019-04-16 23:10 ` [igt-dev] [PATCH v22 1/6] include/drm-uapi: import i915_drm.h header file Andi Shyti
@ 2019-04-16 23:10 ` Andi Shyti
2019-04-16 23:10 ` [igt-dev] [PATCH v22 3/6] lib: igt_gt: add execution buffer flags to class helper Andi Shyti
` (6 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Andi Shyti @ 2019-04-16 23:10 UTC (permalink / raw)
To: IGT dev; +Cc: Andi Shyti
From: Andi Shyti <andi.shyti@intel.com>
The gem_engine_topology library is a set of functions that
interface with the query and getparam/setparam ioctls.
The library's access point is the 'intel_init_engine_list()'
function that, everytime is called, generates the list of active
engines and returns them in a 'struct intel_engine_data'. The
structure contains only the engines that are actively present in
the GPU.
The function can work in both the cases that the query and
getparam ioctls are implemented or not by the running kernel. In
case they are implemented, a query is made to the driver to fetch
the list of active engines. In case they are not implemented, the
list is taken from the 'intel_execution_engines2' array and
stored only after checking their presence.
The gem_engine_topology library provides some iteration helpers:
- intel_get_current_engine(): provides the current engine in the
iteration.
- intel_get_current_physical_engine(): provides the current
physical engine, if the current engine is a virtual engine,
it moves forward until it finds a physical engine.
- intel_next_engine() it just increments the counter so that it
points to the next engine.
Extend the 'for_each_engine_class_instance' so that it can loop
using the new 'intel_init_engine_list()' and rename it to
'for_each_context_engine'.
Move '__for_each_engine_class_instance' to gem_engine_topology.h
and rename it to '__for_each_static_engine'.
Update accordingly tests/perf_pmu.c to use correctly the new
for_each loops.
Signed-off-by: Andi Shyti <andi.shyti@intel.com>
---
lib/Makefile.sources | 2 +
lib/i915/gem_engine_topology.c | 282 +++++++++++++++++++++++++++++++++
lib/i915/gem_engine_topology.h | 79 +++++++++
lib/igt.h | 1 +
lib/igt_gt.h | 2 +
lib/meson.build | 1 +
6 files changed, 367 insertions(+)
create mode 100644 lib/i915/gem_engine_topology.c
create mode 100644 lib/i915/gem_engine_topology.h
diff --git a/lib/Makefile.sources b/lib/Makefile.sources
index a1d253511030..082049bf7c6a 100644
--- a/lib/Makefile.sources
+++ b/lib/Makefile.sources
@@ -13,6 +13,8 @@ lib_source_list = \
i915/gem_ring.c \
i915/gem_mman.c \
i915/gem_mman.h \
+ i915/gem_engine_topology.c \
+ i915/gem_engine_topology.h \
i915_3d.h \
i915_reg.h \
i915_pciids.h \
diff --git a/lib/i915/gem_engine_topology.c b/lib/i915/gem_engine_topology.c
new file mode 100644
index 000000000000..db196ca94822
--- /dev/null
+++ b/lib/i915/gem_engine_topology.c
@@ -0,0 +1,282 @@
+/*
+ * Copyright © 2019 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "drmtest.h"
+#include "ioctl_wrappers.h"
+
+#include "i915/gem_engine_topology.h"
+
+#define DEFINE_CONTEXT_PARAM(e__, p__, c__, N__) \
+ I915_DEFINE_CONTEXT_PARAM_ENGINES(e__, N__); \
+ struct drm_i915_gem_context_param p__ = { \
+ .param = I915_CONTEXT_PARAM_ENGINES, \
+ .ctx_id = c__, \
+ .size = SIZEOF_CTX_PARAM, \
+ .value = to_user_pointer(&e__), \
+ }
+
+static int __gem_query(int fd, struct drm_i915_query *q)
+{
+ int err = 0;
+
+ if (igt_ioctl(fd, DRM_IOCTL_I915_QUERY, q))
+ err = -errno;
+
+ errno = 0;
+ return err;
+}
+
+static void gem_query(int fd, struct drm_i915_query *q)
+{
+ igt_assert_eq(__gem_query(fd, q), 0);
+}
+
+static void query_engines(int fd,
+ struct drm_i915_query_engine_info *query_engines,
+ int length)
+{
+ struct drm_i915_query_item item = { };
+ struct drm_i915_query query = { };
+
+ item.query_id = DRM_I915_QUERY_ENGINE_INFO;
+ query.items_ptr = to_user_pointer(&item);
+ query.num_items = 1;
+ item.length = length;
+
+ item.data_ptr = to_user_pointer(query_engines);
+
+ gem_query(fd, &query);
+}
+
+static void ctx_map_engines(int fd, struct intel_engine_data *ed,
+ struct drm_i915_gem_context_param *param)
+{
+ struct i915_context_param_engines *engines =
+ (struct i915_context_param_engines *) param->value;
+ int i = 0;
+
+ for (typeof(engines->class_instance[0]) *p =
+ &engines->class_instance[0];
+ i < ed->nengines; i++, p++) {
+ p->engine_class = ed->engines[i].class;
+ p->engine_instance = ed->engines[i].instance;
+ }
+
+ param->size = offsetof(typeof(*engines), class_instance[i]);
+ engines->extensions = 0;
+
+ gem_context_set_param(fd, param);
+}
+
+static void init_engine(struct intel_execution_engine2 *e2,
+ int class, int instance, uint64_t flags)
+{
+ const struct intel_execution_engine2 *__e2;
+ static const char *unknown_name = "unknown",
+ *virtual_name = "virtual";
+
+ e2->class = class;
+ e2->instance = instance;
+ e2->flags = flags;
+
+ /* engine is a virtual engine */
+ if (class == I915_ENGINE_CLASS_INVALID) {
+ e2->name = virtual_name;
+ e2->is_virtual = true;
+ return;
+ }
+
+ __for_each_static_engine(__e2)
+ if (__e2->class == class && __e2->instance == instance)
+ break;
+
+ if (__e2->name) {
+ e2->name = __e2->name;
+ } else {
+ igt_warn("found unknown engine (%d, %d)", class, instance);
+ e2->name = unknown_name;
+ }
+
+ /* just to remark it */
+ e2->is_virtual = false;
+}
+
+static void query_engine_list(int fd, struct intel_engine_data *ed)
+{
+ uint8_t buff[SIZEOF_QUERY] = { };
+ struct drm_i915_query_engine_info *query_engine =
+ (struct drm_i915_query_engine_info *) buff;
+ int i;
+
+ query_engines(fd, query_engine, SIZEOF_QUERY);
+
+ for (i = 0; i < query_engine->num_engines; i++)
+ init_engine(&ed->engines[i],
+ query_engine->engines[i].engine_class,
+ query_engine->engines[i].engine_instance, i);
+
+ ed->nengines = query_engine->num_engines;
+}
+
+struct intel_execution_engine2 *
+intel_get_current_engine(struct intel_engine_data *ed)
+{
+ if (!ed->n)
+ ed->current_engine = &ed->engines[0];
+ else if (ed->n >= ed->nengines)
+ ed->current_engine = NULL;
+
+ return ed->current_engine;
+}
+
+void intel_next_engine(struct intel_engine_data *ed)
+{
+ if (ed->n + 1 < ed->nengines) {
+ ed->n++;
+ ed->current_engine = &ed->engines[ed->n];
+ } else {
+ ed->n = ed->nengines;
+ ed->current_engine = NULL;
+ }
+}
+
+struct intel_execution_engine2 *
+intel_get_current_physical_engine(struct intel_engine_data *ed)
+{
+ struct intel_execution_engine2 *e;
+
+ for (e = intel_get_current_engine(ed);
+ e && e->is_virtual;
+ intel_next_engine(ed))
+ ;
+
+ return e;
+}
+
+static int gem_topology_get_param(int fd,
+ struct drm_i915_gem_context_param *p)
+{
+ if (igt_only_list_subtests())
+ return -ENODEV;
+
+ if (__gem_context_get_param(fd, p))
+ return -1; /* using default engine map */
+
+ if (!p->size)
+ return 0;
+
+ p->size = (p->size - sizeof(struct i915_context_param_engines)) /
+ (offsetof(struct i915_context_param_engines,
+ class_instance[1]) -
+ sizeof(struct i915_context_param_engines));
+
+ igt_assert_f(p->size <= GEM_MAX_ENGINES, "unsupported engine count\n");
+
+ return 0;
+}
+
+struct intel_engine_data intel_init_engine_list(int fd, uint32_t ctx_id)
+{
+ DEFINE_CONTEXT_PARAM(engines, param, ctx_id, GEM_MAX_ENGINES);
+ struct intel_engine_data engine_data = { };
+ int i;
+
+ if (gem_topology_get_param(fd, ¶m)) {
+ /* if kernel does not support engine/context mapping */
+ const struct intel_execution_engine2 *e2;
+
+ igt_debug("using pre-allocated engine list\n");
+
+ __for_each_static_engine(e2) {
+ struct intel_execution_engine2 *__e2 =
+ &engine_data.engines[engine_data.nengines];
+
+ if (!igt_only_list_subtests()) {
+ __e2->flags = gem_class_instance_to_eb_flags(fd,
+ e2->class, e2->instance);
+
+ if (!gem_has_ring(fd, __e2->flags))
+ continue;
+ } else {
+ __e2->flags = -1; /* 0xfff... */
+ }
+
+ __e2->name = e2->name;
+ __e2->instance = e2->instance;
+ __e2->class = e2->class;
+ __e2->is_virtual = false;
+
+ engine_data.nengines++;
+ }
+ return engine_data;
+ }
+
+ if (!param.size) {
+ query_engine_list(fd, &engine_data);
+ ctx_map_engines(fd, &engine_data, ¶m);
+ } else {
+ for (i = 0; i < param.size; i++)
+ init_engine(&engine_data.engines[i],
+ engines.class_instance[i].engine_class,
+ engines.class_instance[i].engine_instance,
+ i);
+
+ engine_data.nengines = i;
+ }
+
+ return engine_data;
+}
+
+int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,
+ struct intel_execution_engine2 *e)
+{
+ DEFINE_CONTEXT_PARAM(engines, param, ctx_id, GEM_MAX_ENGINES);
+
+ if (!e || gem_topology_get_param(fd, ¶m) || !param.size)
+ return -EINVAL;
+
+ e->class = engines.class_instance[engine].engine_class;
+ e->instance = engines.class_instance[engine].engine_instance;
+
+ return 0;
+}
+
+void gem_context_set_all_engines(int fd, uint32_t ctx)
+{
+ DEFINE_CONTEXT_PARAM(engines, param, ctx, GEM_MAX_ENGINES);
+ struct intel_engine_data engine_data = { };
+
+ if (!gem_topology_get_param(fd, ¶m) && !param.size) {
+ query_engine_list(fd, &engine_data);
+ ctx_map_engines(fd, &engine_data, ¶m);
+ }
+}
+
+bool gem_has_engine_topology(int fd)
+{
+ struct drm_i915_gem_context_param param = {
+ .param = I915_CONTEXT_PARAM_ENGINES,
+ };
+
+ return !__gem_context_get_param(fd, ¶m);
+}
diff --git a/lib/i915/gem_engine_topology.h b/lib/i915/gem_engine_topology.h
new file mode 100644
index 000000000000..10ec9245aeae
--- /dev/null
+++ b/lib/i915/gem_engine_topology.h
@@ -0,0 +1,79 @@
+/*
+ * Copyright © 2019 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef GEM_ENGINE_TOPOLOGY_H
+#define GEM_ENGINE_TOPOLOGY_H
+
+#include "igt_gt.h"
+#include "i915_drm.h"
+
+/*
+ * Limit what we support for simplicity due limitation in how much we
+ * can address via execbuf2.
+ */
+#define SIZEOF_CTX_PARAM offsetof(struct i915_context_param_engines, \
+ class_instance[GEM_MAX_ENGINES])
+#define SIZEOF_QUERY offsetof(struct drm_i915_query_engine_info, \
+ engines[GEM_MAX_ENGINES])
+
+#define GEM_MAX_ENGINES I915_EXEC_RING_MASK + 1
+
+struct intel_engine_data {
+ uint32_t nengines;
+ uint32_t n;
+ struct intel_execution_engine2 *current_engine;
+ struct intel_execution_engine2 engines[GEM_MAX_ENGINES];
+};
+
+bool gem_has_engine_topology(int fd);
+struct intel_engine_data intel_init_engine_list(int fd, uint32_t ctx_id);
+
+/* iteration functions */
+struct intel_execution_engine2 *
+intel_get_current_engine(struct intel_engine_data *ed);
+
+struct intel_execution_engine2 *
+intel_get_current_physical_engine(struct intel_engine_data *ed);
+
+void intel_next_engine(struct intel_engine_data *ed);
+
+int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,
+ struct intel_execution_engine2 *e);
+
+void gem_context_set_all_engines(int fd, uint32_t ctx);
+
+#define __for_each_static_engine(e__) \
+ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
+
+#define for_each_context_engine(fd__, ctx__, e__) \
+ for (struct intel_engine_data i__ = intel_init_engine_list(fd__, ctx__); \
+ ((e__) = intel_get_current_engine(&i__)); \
+ intel_next_engine(&i__))
+
+/* needs to replace "for_each_physical_engine" when conflicts are fixed */
+#define __for_each_physical_engine(fd__, e__) \
+ for (struct intel_engine_data i__ = intel_init_engine_list(fd__, 0); \
+ ((e__) = intel_get_current_physical_engine(&i__)); \
+ intel_next_engine(&i__))
+
+#endif /* GEM_ENGINE_TOPOLOGY_H */
diff --git a/lib/igt.h b/lib/igt.h
index 6654a659c062..03f19ca2dfb6 100644
--- a/lib/igt.h
+++ b/lib/igt.h
@@ -53,5 +53,6 @@
#include "media_spin.h"
#include "rendercopy.h"
#include "i915/gem_mman.h"
+#include "i915/gem_engine_topology.h"
#endif /* IGT_H */
diff --git a/lib/igt_gt.h b/lib/igt_gt.h
index 475c0b3c3cc6..52b2f1ea95a5 100644
--- a/lib/igt_gt.h
+++ b/lib/igt_gt.h
@@ -95,6 +95,8 @@ extern const struct intel_execution_engine2 {
const char *name;
int class;
int instance;
+ uint64_t flags;
+ bool is_virtual;
} intel_execution_engines2[];
unsigned int
diff --git a/lib/meson.build b/lib/meson.build
index a846293307cb..e55e512403d9 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -5,6 +5,7 @@ lib_sources = [
'i915/gem_submission.c',
'i915/gem_ring.c',
'i915/gem_mman.c',
+ 'i915/gem_engine_topology.c',
'igt_color_encoding.c',
'igt_debugfs.c',
'igt_device.c',
--
2.20.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 14+ messages in thread* [igt-dev] [PATCH v22 6/6] test: perf_pmu: use the gem_engine_topology library
2019-04-16 23:10 [igt-dev] [PATCH v22 0/6] new engine discovery interface Andi Shyti
` (4 preceding siblings ...)
2019-04-16 23:10 ` [igt-dev] [PATCH v22 5/6] lib: igt_dummyload: use for_each_context_engine() Andi Shyti
@ 2019-04-16 23:10 ` Andi Shyti
2019-04-17 0:26 ` [igt-dev] ✓ Fi.CI.BAT: success for new engine discovery interface Patchwork
` (2 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Andi Shyti @ 2019-04-16 23:10 UTC (permalink / raw)
To: IGT dev; +Cc: Andi Shyti
From: Andi Shyti <andi.shyti@intel.com>
Replace the legacy for_each_engine* defines with the ones
implemented in the gem_engine_topology library.
Use whenever possible gem_engine_can_store_dword() that checks
class instead of flags.
Now the __for_each_engine_class_instance and
for_each_engine_class_instance are unused, remove them.
Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Andi Shyti <andi.shyti@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
lib/igt_gt.h | 7 ---
tests/perf_pmu.c | 154 ++++++++++++++++++++++++++++-------------------
2 files changed, 93 insertions(+), 68 deletions(-)
diff --git a/lib/igt_gt.h b/lib/igt_gt.h
index 0b5c7fcb4c3c..77318e2a82b8 100644
--- a/lib/igt_gt.h
+++ b/lib/igt_gt.h
@@ -119,11 +119,4 @@ void gem_require_engine(int gem_fd,
igt_require(gem_has_engine(gem_fd, class, instance));
}
-#define __for_each_engine_class_instance(e__) \
- for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
-
-#define for_each_engine_class_instance(fd__, e__) \
- for ((e__) = intel_execution_engines2; (e__)->name; (e__)++) \
- for_if (gem_has_engine((fd__), (e__)->class, (e__)->instance))
-
#endif /* IGT_GT_H */
diff --git a/tests/perf_pmu.c b/tests/perf_pmu.c
index 4f552bc2ae28..04982466bbc7 100644
--- a/tests/perf_pmu.c
+++ b/tests/perf_pmu.c
@@ -72,7 +72,7 @@ static int open_group(uint64_t config, int group)
}
static void
-init(int gem_fd, const struct intel_execution_engine2 *e, uint8_t sample)
+init(int gem_fd, struct intel_execution_engine2 *e, uint8_t sample)
{
int fd, err = 0;
bool exists;
@@ -82,7 +82,7 @@ init(int gem_fd, const struct intel_execution_engine2 *e, uint8_t sample)
if (fd < 0)
err = errno;
- exists = gem_has_engine(gem_fd, e->class, e->instance);
+ exists = gem_context_has_engine(gem_fd, 0, e->flags);
if (intel_gen(intel_get_drm_devid(gem_fd)) < 6 &&
sample == I915_SAMPLE_SEMA)
exists = false;
@@ -158,11 +158,6 @@ static unsigned int measured_usleep(unsigned int usec)
return igt_nsec_elapsed(&ts);
}
-static unsigned int e2ring(int gem_fd, const struct intel_execution_engine2 *e)
-{
- return gem_class_instance_to_eb_flags(gem_fd, e->class, e->instance);
-}
-
#define TEST_BUSY (1)
#define FLAG_SYNC (2)
#define TEST_TRAILING_IDLE (4)
@@ -170,14 +165,15 @@ static unsigned int e2ring(int gem_fd, const struct intel_execution_engine2 *e)
#define FLAG_LONG (16)
#define FLAG_HANG (32)
-static igt_spin_t * __spin_poll(int fd, uint32_t ctx, unsigned long flags)
+static igt_spin_t * __spin_poll(int fd, uint32_t ctx,
+ struct intel_execution_engine2 *e)
{
struct igt_spin_factory opts = {
.ctx = ctx,
- .engine = flags,
+ .engine = e->flags,
};
- if (gem_can_store_dword(fd, flags))
+ if (gem_class_can_store_dword(fd, e->class))
opts.flags |= IGT_SPIN_POLL_RUN;
return __igt_spin_batch_factory(fd, &opts);
@@ -209,20 +205,34 @@ static unsigned long __spin_wait(int fd, igt_spin_t *spin)
return igt_nsec_elapsed(&start);
}
-static igt_spin_t * __spin_sync(int fd, uint32_t ctx, unsigned long flags)
+static igt_spin_t * __spin_sync(int fd, uint32_t ctx,
+ struct intel_execution_engine2 *e)
{
- igt_spin_t *spin = __spin_poll(fd, ctx, flags);
+ igt_spin_t *spin = __spin_poll(fd, ctx, e);
__spin_wait(fd, spin);
return spin;
}
-static igt_spin_t * spin_sync(int fd, uint32_t ctx, unsigned long flags)
+static igt_spin_t * spin_sync(int fd, uint32_t ctx,
+ struct intel_execution_engine2 *e)
{
igt_require_gem(fd);
- return __spin_sync(fd, ctx, flags);
+ return __spin_sync(fd, ctx, e);
+}
+
+static igt_spin_t * spin_sync_flags(int fd, uint32_t ctx, unsigned int flags)
+{
+ struct intel_execution_engine2 e = { };
+
+ e.class = gem_execbuf_flags_to_engine_class(flags);
+ e.instance = (flags & (I915_EXEC_BSD_MASK | I915_EXEC_RING_MASK)) ==
+ (I915_EXEC_BSD | I915_EXEC_BSD_RING2) ? 1 : 0;
+ e.flags = flags;
+
+ return spin_sync(fd, ctx, &e);
}
static void end_spin(int fd, igt_spin_t *spin, unsigned int flags)
@@ -257,7 +267,7 @@ static void end_spin(int fd, igt_spin_t *spin, unsigned int flags)
}
static void
-single(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
+single(int gem_fd, struct intel_execution_engine2 *e, unsigned int flags)
{
unsigned long slept;
igt_spin_t *spin;
@@ -267,7 +277,7 @@ single(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
fd = open_pmu(I915_PMU_ENGINE_BUSY(e->class, e->instance));
if (flags & TEST_BUSY)
- spin = spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin = spin_sync(gem_fd, 0, e);
else
spin = NULL;
@@ -303,7 +313,7 @@ single(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
}
static void
-busy_start(int gem_fd, const struct intel_execution_engine2 *e)
+busy_start(int gem_fd, struct intel_execution_engine2 *e)
{
unsigned long slept;
uint64_t val, ts[2];
@@ -316,7 +326,7 @@ busy_start(int gem_fd, const struct intel_execution_engine2 *e)
*/
sleep(2);
- spin = __spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin = __spin_sync(gem_fd, 0, e);
fd = open_pmu(I915_PMU_ENGINE_BUSY(e->class, e->instance));
@@ -338,7 +348,7 @@ busy_start(int gem_fd, const struct intel_execution_engine2 *e)
* will depend on the CI systems running it a lot to detect issues.
*/
static void
-busy_double_start(int gem_fd, const struct intel_execution_engine2 *e)
+busy_double_start(int gem_fd, struct intel_execution_engine2 *e)
{
unsigned long slept;
uint64_t val, val2, ts[2];
@@ -347,6 +357,7 @@ busy_double_start(int gem_fd, const struct intel_execution_engine2 *e)
int fd;
ctx = gem_context_create(gem_fd);
+ gem_context_set_all_engines(gem_fd, ctx);
/*
* Defeat the busy stats delayed disable, we need to guarantee we are
@@ -359,11 +370,11 @@ busy_double_start(int gem_fd, const struct intel_execution_engine2 *e)
* re-submission in execlists mode. Make sure busyness is correctly
* reported with the engine busy, and after the engine went idle.
*/
- spin[0] = __spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin[0] = __spin_sync(gem_fd, 0, e);
usleep(500e3);
spin[1] = __igt_spin_batch_new(gem_fd,
.ctx = ctx,
- .engine = e2ring(gem_fd, e));
+ .engine = e->flags);
/*
* Open PMU as fast as possible after the second spin batch in attempt
@@ -421,10 +432,10 @@ static void log_busy(unsigned int num_engines, uint64_t *val)
}
static void
-busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
+busy_check_all(int gem_fd, struct intel_execution_engine2 *e,
const unsigned int num_engines, unsigned int flags)
{
- const struct intel_execution_engine2 *e_;
+ struct intel_execution_engine2 *e_;
uint64_t tval[2][num_engines];
unsigned int busy_idx = 0, i;
uint64_t val[num_engines];
@@ -434,8 +445,8 @@ busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
i = 0;
fd[0] = -1;
- for_each_engine_class_instance(gem_fd, e_) {
- if (e == e_)
+ __for_each_physical_engine(gem_fd, e_) {
+ if (e->class == e_->class && e->instance == e_->instance)
busy_idx = i;
fd[i++] = open_group(I915_PMU_ENGINE_BUSY(e_->class,
@@ -445,7 +456,7 @@ busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
igt_assert_eq(i, num_engines);
- spin = spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin = spin_sync(gem_fd, 0, e);
pmu_read_multi(fd[0], num_engines, tval[0]);
slept = measured_usleep(batch_duration_ns / 1000);
if (flags & TEST_TRAILING_IDLE)
@@ -472,23 +483,23 @@ busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
static void
__submit_spin_batch(int gem_fd, igt_spin_t *spin,
- const struct intel_execution_engine2 *e,
+ struct intel_execution_engine2 *e,
int offset)
{
struct drm_i915_gem_execbuffer2 eb = spin->execbuf;
eb.flags &= ~(0x3f | I915_EXEC_BSD_MASK);
- eb.flags |= e2ring(gem_fd, e) | I915_EXEC_NO_RELOC;
+ eb.flags |= e->flags | I915_EXEC_NO_RELOC;
eb.batch_start_offset += offset;
gem_execbuf(gem_fd, &eb);
}
static void
-most_busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
+most_busy_check_all(int gem_fd, struct intel_execution_engine2 *e,
const unsigned int num_engines, unsigned int flags)
{
- const struct intel_execution_engine2 *e_;
+ struct intel_execution_engine2 *e_;
uint64_t tval[2][num_engines];
uint64_t val[num_engines];
int fd[num_engines];
@@ -497,13 +508,13 @@ most_busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,
unsigned int idle_idx, i;
i = 0;
- for_each_engine_class_instance(gem_fd, e_) {
- if (e == e_)
+ __for_each_physical_engine(gem_fd, e_) {
+ if (e->class == e_->class && e->instance == e_->instance)
idle_idx = i;
else if (spin)
__submit_spin_batch(gem_fd, spin, e_, 64);
else
- spin = __spin_poll(gem_fd, 0, e2ring(gem_fd, e_));
+ spin = __spin_poll(gem_fd, 0, e_);
val[i++] = I915_PMU_ENGINE_BUSY(e_->class, e_->instance);
}
@@ -545,7 +556,7 @@ static void
all_busy_check_all(int gem_fd, const unsigned int num_engines,
unsigned int flags)
{
- const struct intel_execution_engine2 *e;
+ struct intel_execution_engine2 *e;
uint64_t tval[2][num_engines];
uint64_t val[num_engines];
int fd[num_engines];
@@ -554,11 +565,11 @@ all_busy_check_all(int gem_fd, const unsigned int num_engines,
unsigned int i;
i = 0;
- for_each_engine_class_instance(gem_fd, e) {
+ __for_each_physical_engine(gem_fd, e) {
if (spin)
__submit_spin_batch(gem_fd, spin, e, 64);
else
- spin = __spin_poll(gem_fd, 0, e2ring(gem_fd, e));
+ spin = __spin_poll(gem_fd, 0, e);
val[i++] = I915_PMU_ENGINE_BUSY(e->class, e->instance);
}
@@ -592,7 +603,7 @@ all_busy_check_all(int gem_fd, const unsigned int num_engines,
}
static void
-no_sema(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
+no_sema(int gem_fd, struct intel_execution_engine2 *e, unsigned int flags)
{
igt_spin_t *spin;
uint64_t val[2][2];
@@ -602,7 +613,7 @@ no_sema(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
open_group(I915_PMU_ENGINE_WAIT(e->class, e->instance), fd);
if (flags & TEST_BUSY)
- spin = spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin = spin_sync(gem_fd, 0, e);
else
spin = NULL;
@@ -631,7 +642,7 @@ no_sema(int gem_fd, const struct intel_execution_engine2 *e, unsigned int flags)
#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12)
static void
-sema_wait(int gem_fd, const struct intel_execution_engine2 *e,
+sema_wait(int gem_fd, struct intel_execution_engine2 *e,
unsigned int flags)
{
struct drm_i915_gem_relocation_entry reloc[2] = {};
@@ -689,7 +700,7 @@ sema_wait(int gem_fd, const struct intel_execution_engine2 *e,
eb.buffer_count = 2;
eb.buffers_ptr = to_user_pointer(obj);
- eb.flags = e2ring(gem_fd, e);
+ eb.flags = e->flags;
/**
* Start the semaphore wait PMU and after some known time let the above
@@ -792,7 +803,7 @@ static int wait_vblank(int fd, union drm_wait_vblank *vbl)
}
static void
-event_wait(int gem_fd, const struct intel_execution_engine2 *e)
+event_wait(int gem_fd, struct intel_execution_engine2 *e)
{
struct drm_i915_gem_exec_object2 obj = { };
struct drm_i915_gem_execbuffer2 eb = { };
@@ -845,7 +856,7 @@ event_wait(int gem_fd, const struct intel_execution_engine2 *e)
eb.buffer_count = 1;
eb.buffers_ptr = to_user_pointer(&obj);
- eb.flags = e2ring(gem_fd, e) | I915_EXEC_SECURE;
+ eb.flags = e->flags | I915_EXEC_SECURE;
for_each_pipe_with_valid_output(&data.display, p, output) {
struct igt_helper_process waiter = { };
@@ -917,7 +928,7 @@ event_wait(int gem_fd, const struct intel_execution_engine2 *e)
}
static void
-multi_client(int gem_fd, const struct intel_execution_engine2 *e)
+multi_client(int gem_fd, struct intel_execution_engine2 *e)
{
uint64_t config = I915_PMU_ENGINE_BUSY(e->class, e->instance);
unsigned long slept[2];
@@ -936,7 +947,7 @@ multi_client(int gem_fd, const struct intel_execution_engine2 *e)
*/
fd[1] = open_pmu(config);
- spin = spin_sync(gem_fd, 0, e2ring(gem_fd, e));
+ spin = spin_sync(gem_fd, 0, e);
val[0] = val[1] = __pmu_read_single(fd[0], &ts[0]);
slept[1] = measured_usleep(batch_duration_ns / 1000);
@@ -1039,6 +1050,7 @@ static void cpu_hotplug(int gem_fd)
igt_spin_t *spin[2];
uint64_t ts[2];
uint64_t val;
+ uint32_t ctx;
int link[2];
int fd, ret;
int cur = 0;
@@ -1046,14 +1058,18 @@ static void cpu_hotplug(int gem_fd)
igt_require(cpu0_hotplug_support());
+ ctx = gem_context_create(gem_fd);
+
fd = open_pmu(I915_PMU_ENGINE_BUSY(I915_ENGINE_CLASS_RENDER, 0));
/*
* Create two spinners so test can ensure shorter gaps in engine
* busyness as it is terminating one and re-starting the other.
*/
- spin[0] = igt_spin_batch_new(gem_fd, .engine = I915_EXEC_RENDER);
- spin[1] = __igt_spin_batch_new(gem_fd, .engine = I915_EXEC_RENDER);
+ spin[0] = igt_spin_batch_new(gem_fd,
+ .engine = I915_EXEC_RENDER, .ctx = ctx);
+ spin[1] = __igt_spin_batch_new(gem_fd,
+ .engine = I915_EXEC_RENDER, .ctx = ctx);
val = __pmu_read_single(fd, &ts[0]);
@@ -1137,6 +1153,7 @@ static void cpu_hotplug(int gem_fd)
igt_spin_batch_free(gem_fd, spin[cur]);
spin[cur] = __igt_spin_batch_new(gem_fd,
+ .ctx = ctx,
.engine = I915_EXEC_RENDER);
cur ^= 1;
}
@@ -1150,6 +1167,7 @@ static void cpu_hotplug(int gem_fd)
igt_waitchildren();
close(fd);
close(link[0]);
+ gem_context_destroy(gem_fd, ctx);
/* Skip if child signals a problem with offlining a CPU. */
igt_skip_on(buf == 's');
@@ -1165,17 +1183,21 @@ test_interrupts(int gem_fd)
igt_spin_t *spin[target];
struct pollfd pfd;
uint64_t idle, busy;
+ uint32_t ctx;
int fence_fd;
int fd;
gem_quiescent_gpu(gem_fd);
+ ctx = gem_context_create(gem_fd);
+
fd = open_pmu(I915_PMU_INTERRUPTS);
/* Queue spinning batches. */
for (int i = 0; i < target; i++) {
spin[i] = __igt_spin_batch_new(gem_fd,
.engine = I915_EXEC_RENDER,
+ .ctx = ctx,
.flags = IGT_SPIN_FENCE_OUT);
if (i == 0) {
fence_fd = spin[i]->out_fence;
@@ -1217,6 +1239,7 @@ test_interrupts(int gem_fd)
/* Check at least as many interrupts has been generated. */
busy = pmu_read_single(fd) - idle;
close(fd);
+ gem_context_destroy(gem_fd, ctx);
igt_assert_lte(target, busy);
}
@@ -1229,15 +1252,19 @@ test_interrupts_sync(int gem_fd)
igt_spin_t *spin[target];
struct pollfd pfd;
uint64_t idle, busy;
+ uint32_t ctx;
int fd;
gem_quiescent_gpu(gem_fd);
+ ctx = gem_context_create(gem_fd);
+
fd = open_pmu(I915_PMU_INTERRUPTS);
/* Queue spinning batches. */
for (int i = 0; i < target; i++)
spin[i] = __igt_spin_batch_new(gem_fd,
+ .ctx = ctx,
.flags = IGT_SPIN_FENCE_OUT);
/* Wait for idle state. */
@@ -1262,6 +1289,7 @@ test_interrupts_sync(int gem_fd)
/* Check at least as many interrupts has been generated. */
busy = pmu_read_single(fd) - idle;
close(fd);
+ gem_context_destroy(gem_fd, ctx);
igt_assert_lte(target, busy);
}
@@ -1274,6 +1302,9 @@ test_frequency(int gem_fd)
double min[2], max[2];
igt_spin_t *spin;
int fd, sysfs;
+ uint32_t ctx;
+
+ ctx = gem_context_create(gem_fd);
sysfs = igt_sysfs_open(gem_fd);
igt_require(sysfs >= 0);
@@ -1301,7 +1332,7 @@ test_frequency(int gem_fd)
igt_require(igt_sysfs_get_u32(sysfs, "gt_boost_freq_mhz") == min_freq);
gem_quiescent_gpu(gem_fd); /* Idle to be sure the change takes effect */
- spin = spin_sync(gem_fd, 0, I915_EXEC_RENDER);
+ spin = spin_sync_flags(gem_fd, ctx, I915_EXEC_RENDER);
slept = pmu_read_multi(fd, 2, start);
measured_usleep(batch_duration_ns / 1000);
@@ -1327,7 +1358,7 @@ test_frequency(int gem_fd)
igt_require(igt_sysfs_get_u32(sysfs, "gt_min_freq_mhz") == max_freq);
gem_quiescent_gpu(gem_fd);
- spin = spin_sync(gem_fd, 0, I915_EXEC_RENDER);
+ spin = spin_sync_flags(gem_fd, ctx, I915_EXEC_RENDER);
slept = pmu_read_multi(fd, 2, start);
measured_usleep(batch_duration_ns / 1000);
@@ -1348,6 +1379,8 @@ test_frequency(int gem_fd)
min_freq, igt_sysfs_get_u32(sysfs, "gt_min_freq_mhz"));
close(fd);
+ gem_context_destroy(gem_fd, ctx);
+
igt_info("Min frequency: requested %.1f, actual %.1f\n",
min[0], min[1]);
igt_info("Max frequency: requested %.1f, actual %.1f\n",
@@ -1448,7 +1481,7 @@ test_rc6(int gem_fd, unsigned int flags)
}
static void
-test_enable_race(int gem_fd, const struct intel_execution_engine2 *e)
+test_enable_race(int gem_fd, struct intel_execution_engine2 *e)
{
uint64_t config = I915_PMU_ENGINE_BUSY(e->class, e->instance);
struct igt_helper_process engine_load = { };
@@ -1458,14 +1491,14 @@ test_enable_race(int gem_fd, const struct intel_execution_engine2 *e)
int fd;
igt_require(gem_has_execlists(gem_fd));
- igt_require(gem_has_engine(gem_fd, e->class, e->instance));
+ igt_require(gem_context_has_engine(gem_fd, 0, e->flags));
obj.handle = gem_create(gem_fd, 4096);
gem_write(gem_fd, obj.handle, 0, &bbend, sizeof(bbend));
eb.buffer_count = 1;
eb.buffers_ptr = to_user_pointer(&obj);
- eb.flags = e2ring(gem_fd, e);
+ eb.flags = e->flags;
/*
* This test is probabilistic so run in a few times to increase the
@@ -1520,7 +1553,7 @@ static void __rearm_spin_batch(igt_spin_t *spin)
__assert_within(x, ref, tolerance, tolerance)
static void
-accuracy(int gem_fd, const struct intel_execution_engine2 *e,
+accuracy(int gem_fd, struct intel_execution_engine2 *e,
unsigned long target_busy_pct,
unsigned long target_iters)
{
@@ -1570,7 +1603,7 @@ accuracy(int gem_fd, const struct intel_execution_engine2 *e,
igt_spin_t *spin;
/* Allocate our spin batch and idle it. */
- spin = igt_spin_batch_new(gem_fd, .engine = e2ring(gem_fd, e));
+ spin = igt_spin_batch_new(gem_fd, .engine = e->flags);
igt_spin_batch_end(spin);
gem_sync(gem_fd, spin->handle);
@@ -1674,7 +1707,7 @@ igt_main
I915_PMU_LAST - __I915_PMU_OTHER(0) + 1;
unsigned int num_engines = 0;
int fd = -1;
- const struct intel_execution_engine2 *e;
+ struct intel_execution_engine2 *e;
unsigned int i;
igt_fixture {
@@ -1683,7 +1716,7 @@ igt_main
igt_require_gem(fd);
igt_require(i915_type_id() > 0);
- for_each_engine_class_instance(fd, e)
+ __for_each_physical_engine(fd, e)
num_engines++;
}
@@ -1693,7 +1726,7 @@ igt_main
igt_subtest("invalid-init")
invalid_init();
- __for_each_engine_class_instance(e) {
+ __for_each_physical_engine(fd, e) {
const unsigned int pct[] = { 2, 50, 98 };
/**
@@ -1711,7 +1744,7 @@ igt_main
igt_subtest_group {
igt_fixture {
- gem_require_engine(fd, e->class, e->instance);
+ gem_context_has_engine(fd, 0, e->flags);
}
/**
@@ -1897,12 +1930,11 @@ igt_main
gem_quiescent_gpu(fd);
}
- __for_each_engine_class_instance(e) {
+ __for_each_physical_engine(render_fd, e) {
igt_subtest_group {
igt_fixture {
- gem_require_engine(render_fd,
- e->class,
- e->instance);
+ gem_context_has_engine(render_fd,
+ 0, e->flags);
}
igt_subtest_f("render-node-busy-%s", e->name)
--
2.20.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
^ permalink raw reply related [flat|nested] 14+ messages in thread