Module: Mesa Branch: main Commit: dea6c8243749e63aa9b517edbfe7489dbfd103a4 URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=dea6c8243749e63aa9b517edbfe7489dbfd103a4
Author: José Roberto de Souza <[email protected]> Date: Thu Dec 14 09:41:22 2023 -0800 intel: Sync xe_drm.h final part Sync xe_drm.h with commit a8ff56e160bb ("drm/xe/uapi: Remove reset uevent for now"). This is the last xe_drm.h uAPI break. The only relevant change for ANV and Iris is that now VM bind uAPI is asynchronous only so I had to bring back the syncobj creation, wait and destruction. Is still in the Xe port TODO list to make VM binds truly asynchronous. Signed-off-by: José Roberto de Souza <[email protected]> Reviewed-by: Paulo Zanoni <[email protected]> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26699> --- include/drm-uapi/xe_drm.h | 829 ++++++++++++++++--------- src/gallium/drivers/iris/xe/iris_kmd_backend.c | 31 +- src/intel/vulkan/xe/anv_kmd_backend.c | 30 +- 3 files changed, 584 insertions(+), 306 deletions(-) diff --git a/include/drm-uapi/xe_drm.h b/include/drm-uapi/xe_drm.h index 590f7b7af4b..bacdca787c2 100644 --- a/include/drm-uapi/xe_drm.h +++ b/include/drm-uapi/xe_drm.h @@ -12,22 +12,113 @@ extern "C" { #endif -/* Please note that modifications to all structs defined here are +/* + * Please note that modifications to all structs defined here are * subject to backwards-compatibility constraints. + * Sections in this file are organized as follows: + * 1. IOCTL definition + * 2. Extension definition and helper structs + * 3. IOCTL's Query structs in the order of the Query's entries. + * 4. The rest of IOCTL structs in the order of IOCTL declaration. */ /** - * DOC: uevent generated by xe on it's pci node. + * DOC: Xe Device Block Diagram + * + * The diagram below represents a high-level simplification of a discrete + * GPU supported by the Xe driver. It shows some device components which + * are necessary to understand this API, as well as how their relations + * to each other. This diagram does not represent real hardware:: * - * DRM_XE_RESET_FAILED_UEVENT - Event is generated when attempt to reset gt - * fails. The value supplied with the event is always "NEEDS_RESET". - * Additional information supplied is tile id and gt id of the gt unit for - * which reset has failed. + * ┌──────────────────────────────────────────────────────────────────┐ + * │ ┌──────────────────────────────────────────────────┐ ┌─────────┐ │ + * │ │ ┌───────────────────────┐ ┌─────┐ │ │ ┌─────┐ │ │ + * │ │ │ VRAM0 ├───┤ ... │ │ │ │VRAM1│ │ │ + * │ │ └───────────┬───────────┘ └─GT1─┘ │ │ └──┬──┘ │ │ + * │ │ ┌──────────────────┴───────────────────────────┐ │ │ ┌──┴──┐ │ │ + * │ │ │ ┌─────────────────────┐ ┌─────────────────┐ │ │ │ │ │ │ │ + * │ │ │ │ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │ │ ┌─────┐ ┌─────┐ │ │ │ │ │ │ │ │ + * │ │ │ │ │EU│ │EU│ │EU│ │EU│ │ │ │RCS0 │ │BCS0 │ │ │ │ │ │ │ │ │ + * │ │ │ │ └──┘ └──┘ └──┘ └──┘ │ │ └─────┘ └─────┘ │ │ │ │ │ │ │ │ + * │ │ │ │ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │ │ ┌─────┐ ┌─────┐ │ │ │ │ │ │ │ │ + * │ │ │ │ │EU│ │EU│ │EU│ │EU│ │ │ │VCS0 │ │VCS1 │ │ │ │ │ │ │ │ │ + * │ │ │ │ └──┘ └──┘ └──┘ └──┘ │ │ └─────┘ └─────┘ │ │ │ │ │ │ │ │ + * │ │ │ │ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │ │ ┌─────┐ ┌─────┐ │ │ │ │ │ │ │ │ + * │ │ │ │ │EU│ │EU│ │EU│ │EU│ │ │ │VECS0│ │VECS1│ │ │ │ │ │ ... │ │ │ + * │ │ │ │ └──┘ └──┘ └──┘ └──┘ │ │ └─────┘ └─────┘ │ │ │ │ │ │ │ │ + * │ │ │ │ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │ │ ┌─────┐ ┌─────┐ │ │ │ │ │ │ │ │ + * │ │ │ │ │EU│ │EU│ │EU│ │EU│ │ │ │CCS0 │ │CCS1 │ │ │ │ │ │ │ │ │ + * │ │ │ │ └──┘ └──┘ └──┘ └──┘ │ │ └─────┘ └─────┘ │ │ │ │ │ │ │ │ + * │ │ │ └─────────DSS─────────┘ │ ┌─────┐ ┌─────┐ │ │ │ │ │ │ │ │ + * │ │ │ │ │CCS2 │ │CCS3 │ │ │ │ │ │ │ │ │ + * │ │ │ ┌─────┐ ┌─────┐ ┌─────┐ │ └─────┘ └─────┘ │ │ │ │ │ │ │ │ + * │ │ │ │ ... │ │ ... │ │ ... │ │ │ │ │ │ │ │ │ │ + * │ │ │ └─DSS─┘ └─DSS─┘ └─DSS─┘ └─────Engines─────┘ │ │ │ │ │ │ │ + * │ │ └───────────────────────────GT0────────────────┘ │ │ └─GT2─┘ │ │ + * │ └────────────────────────────Tile0─────────────────┘ └─ Tile1──┘ │ + * └─────────────────────────────Device0───────┬──────────────────────┘ + * │ + * ───────────────────────┴────────── PCI bus */ -#define DRM_XE_RESET_FAILED_UEVENT "DEVICE_STATUS" /** - * struct xe_user_extension - Base class for defining a chain of extensions + * DOC: Xe uAPI Overview + * + * This section aims to describe the Xe's IOCTL entries, its structs, and other + * Xe related uAPI such as uevents and PMU (Platform Monitoring Unit) related + * entries and usage. + * + * List of supported IOCTLs: + * - &DRM_IOCTL_XE_DEVICE_QUERY + * - &DRM_IOCTL_XE_GEM_CREATE + * - &DRM_IOCTL_XE_GEM_MMAP_OFFSET + * - &DRM_IOCTL_XE_VM_CREATE + * - &DRM_IOCTL_XE_VM_DESTROY + * - &DRM_IOCTL_XE_VM_BIND + * - &DRM_IOCTL_XE_EXEC_QUEUE_CREATE + * - &DRM_IOCTL_XE_EXEC_QUEUE_DESTROY + * - &DRM_IOCTL_XE_EXEC_QUEUE_GET_PROPERTY + * - &DRM_IOCTL_XE_EXEC + * - &DRM_IOCTL_XE_WAIT_USER_FENCE + */ + +/* + * xe specific ioctls. + * + * The device specific ioctl range is [DRM_COMMAND_BASE, DRM_COMMAND_END) ie + * [0x40, 0xa0) (a0 is excluded). The numbers below are defined as offset + * against DRM_COMMAND_BASE and should be between [0x0, 0x60). + */ +#define DRM_XE_DEVICE_QUERY 0x00 +#define DRM_XE_GEM_CREATE 0x01 +#define DRM_XE_GEM_MMAP_OFFSET 0x02 +#define DRM_XE_VM_CREATE 0x03 +#define DRM_XE_VM_DESTROY 0x04 +#define DRM_XE_VM_BIND 0x05 +#define DRM_XE_EXEC_QUEUE_CREATE 0x06 +#define DRM_XE_EXEC_QUEUE_DESTROY 0x07 +#define DRM_XE_EXEC_QUEUE_GET_PROPERTY 0x08 +#define DRM_XE_EXEC 0x09 +#define DRM_XE_WAIT_USER_FENCE 0x0a +/* Must be kept compact -- no holes */ + +#define DRM_IOCTL_XE_DEVICE_QUERY DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_DEVICE_QUERY, struct drm_xe_device_query) +#define DRM_IOCTL_XE_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_GEM_CREATE, struct drm_xe_gem_create) +#define DRM_IOCTL_XE_GEM_MMAP_OFFSET DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_GEM_MMAP_OFFSET, struct drm_xe_gem_mmap_offset) +#define DRM_IOCTL_XE_VM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_VM_CREATE, struct drm_xe_vm_create) +#define DRM_IOCTL_XE_VM_DESTROY DRM_IOW(DRM_COMMAND_BASE + DRM_XE_VM_DESTROY, struct drm_xe_vm_destroy) +#define DRM_IOCTL_XE_VM_BIND DRM_IOW(DRM_COMMAND_BASE + DRM_XE_VM_BIND, struct drm_xe_vm_bind) +#define DRM_IOCTL_XE_EXEC_QUEUE_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_CREATE, struct drm_xe_exec_queue_create) +#define DRM_IOCTL_XE_EXEC_QUEUE_DESTROY DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_DESTROY, struct drm_xe_exec_queue_destroy) +#define DRM_IOCTL_XE_EXEC_QUEUE_GET_PROPERTY DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_GET_PROPERTY, struct drm_xe_exec_queue_get_property) +#define DRM_IOCTL_XE_EXEC DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC, struct drm_xe_exec) +#define DRM_IOCTL_XE_WAIT_USER_FENCE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_WAIT_USER_FENCE, struct drm_xe_wait_user_fence) + +/** + * DOC: Xe IOCTL Extensions + * + * Before detailing the IOCTLs and its structs, it is important to highlight + * that every IOCTL in Xe is extensible. * * 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, @@ -45,29 +136,32 @@ extern "C" { * * .. code-block:: C * - * struct xe_user_extension ext3 { + * struct drm_xe_user_extension ext3 { * .next_extension = 0, // end * .name = ..., * }; - * struct xe_user_extension ext2 { + * struct drm_xe_user_extension ext2 { * .next_extension = (uintptr_t)&ext3, * .name = ..., * }; - * struct xe_user_extension ext1 { + * struct drm_xe_user_extension ext1 { * .next_extension = (uintptr_t)&ext2, * .name = ..., * }; * - * Typically the struct xe_user_extension would be embedded in some uAPI + * Typically the struct drm_xe_user_extension would be embedded in some uAPI * struct, and in this case we would feed it the head of the chain(i.e ext1), * which would then apply all of the above extensions. - * +*/ + +/** + * struct drm_xe_user_extension - Base class for defining a chain of extensions */ -struct xe_user_extension { +struct drm_xe_user_extension { /** * @next_extension: * - * Pointer to the next struct xe_user_extension, or zero if the end. + * Pointer to the next struct drm_xe_user_extension, or zero if the end. */ __u64 next_extension; @@ -78,7 +172,7 @@ struct xe_user_extension { * * Also note that the name space for this is not global for the whole * driver, but rather its scope/meaning is limited to the specific piece - * of uAPI which has embedded the struct xe_user_extension. + * of uAPI which has embedded the struct drm_xe_user_extension. */ __u32 name; @@ -90,38 +184,28 @@ struct xe_user_extension { __u32 pad; }; -/* - * xe specific ioctls. +/** + * struct drm_xe_ext_set_property - Generic set property extension * - * The device specific ioctl range is [DRM_COMMAND_BASE, DRM_COMMAND_END) ie - * [0x40, 0xa0) (a0 is excluded). The numbers below are defined as offset - * against DRM_COMMAND_BASE and should be between [0x0, 0x60). + * A generic struct that allows any of the Xe's IOCTL to be extended + * with a set_property operation. */ -#define DRM_XE_DEVICE_QUERY 0x00 -#define DRM_XE_GEM_CREATE 0x01 -#define DRM_XE_GEM_MMAP_OFFSET 0x02 -#define DRM_XE_VM_CREATE 0x03 -#define DRM_XE_VM_DESTROY 0x04 -#define DRM_XE_VM_BIND 0x05 -#define DRM_XE_EXEC_QUEUE_CREATE 0x06 -#define DRM_XE_EXEC_QUEUE_DESTROY 0x07 -#define DRM_XE_EXEC_QUEUE_GET_PROPERTY 0x08 -#define DRM_XE_EXEC 0x09 -#define DRM_XE_WAIT_USER_FENCE 0x0a -/* Must be kept compact -- no holes */ +struct drm_xe_ext_set_property { + /** @base: base user extension */ + struct drm_xe_user_extension base; -#define DRM_IOCTL_XE_DEVICE_QUERY DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_DEVICE_QUERY, struct drm_xe_device_query) -#define DRM_IOCTL_XE_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_GEM_CREATE, struct drm_xe_gem_create) -#define DRM_IOCTL_XE_GEM_MMAP_OFFSET DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_GEM_MMAP_OFFSET, struct drm_xe_gem_mmap_offset) -#define DRM_IOCTL_XE_VM_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_VM_CREATE, struct drm_xe_vm_create) -#define DRM_IOCTL_XE_VM_DESTROY DRM_IOW(DRM_COMMAND_BASE + DRM_XE_VM_DESTROY, struct drm_xe_vm_destroy) -#define DRM_IOCTL_XE_VM_BIND DRM_IOW(DRM_COMMAND_BASE + DRM_XE_VM_BIND, struct drm_xe_vm_bind) -#define DRM_IOCTL_XE_EXEC_QUEUE_CREATE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_CREATE, struct drm_xe_exec_queue_create) -#define DRM_IOCTL_XE_EXEC_QUEUE_DESTROY DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_DESTROY, struct drm_xe_exec_queue_destroy) -#define DRM_IOCTL_XE_EXEC_QUEUE_SET_PROPERTY DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_SET_PROPERTY, struct drm_xe_exec_queue_set_property) -#define DRM_IOCTL_XE_EXEC_QUEUE_GET_PROPERTY DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_EXEC_QUEUE_GET_PROPERTY, struct drm_xe_exec_queue_get_property) -#define DRM_IOCTL_XE_EXEC DRM_IOW(DRM_COMMAND_BASE + DRM_XE_EXEC, struct drm_xe_exec) -#define DRM_IOCTL_XE_WAIT_USER_FENCE DRM_IOWR(DRM_COMMAND_BASE + DRM_XE_WAIT_USER_FENCE, struct drm_xe_wait_user_fence) + /** @property: property to set */ + __u32 property; + + /** @pad: MBZ */ + __u32 pad; + + /** @value: property value */ + __u64 value; + + /** @reserved: Reserved */ + __u64 reserved[2]; +}; /** * struct drm_xe_engine_class_instance - instance of an engine class @@ -130,6 +214,15 @@ struct xe_user_extension { * the input of engine selection for both @drm_xe_exec_queue_create and * @drm_xe_query_engine_cycles * + * The @engine_class can be: + * - %DRM_XE_ENGINE_CLASS_RENDER + * - %DRM_XE_ENGINE_CLASS_COPY + * - %DRM_XE_ENGINE_CLASS_VIDEO_DECODE + * - %DRM_XE_ENGINE_CLASS_VIDEO_ENHANCE + * - %DRM_XE_ENGINE_CLASS_COMPUTE + * - %DRM_XE_ENGINE_CLASS_VM_BIND - Kernel only classes (not actual + * hardware engine class). Used for creating ordered queues of VM + * bind operations. */ struct drm_xe_engine_class_instance { #define DRM_XE_ENGINE_CLASS_RENDER 0 @@ -137,15 +230,12 @@ struct drm_xe_engine_class_instance { #define DRM_XE_ENGINE_CLASS_VIDEO_DECODE 2 #define DRM_XE_ENGINE_CLASS_VIDEO_ENHANCE 3 #define DRM_XE_ENGINE_CLASS_COMPUTE 4 - /* - * Kernel only classes (not actual hardware engine class). Used for - * creating ordered queues of VM bind operations. - */ -#define DRM_XE_ENGINE_CLASS_VM_BIND_ASYNC 5 -#define DRM_XE_ENGINE_CLASS_VM_BIND_SYNC 6 +#define DRM_XE_ENGINE_CLASS_VM_BIND 5 + /** @engine_class: engine class id */ __u16 engine_class; - + /** @engine_instance: engine instance id */ __u16 engine_instance; + /** @gt_id: Unique ID of this GT within the PCI Device */ __u16 gt_id; /** @pad: MBZ */ __u16 pad; @@ -204,10 +294,9 @@ struct drm_xe_mem_region { */ __u16 mem_class; /** - * @instance: The instance for this region. - * - * The @mem_class and @instance taken together will always give - * a unique pair. + * @instance: The unique ID for this region, which serves as the + * index in the placement bitmask used as argument for + * &DRM_IOCTL_XE_GEM_CREATE */ __u16 instance; /** @@ -265,61 +354,10 @@ struct drm_xe_mem_region { * here will always be zero). */ __u64 cpu_visible_used; - /** @reserved: MBZ */ + /** @reserved: Reserved */ __u64 reserved[6]; }; -/** - * struct drm_xe_query_engine_cycles - correlate CPU and GPU timestamps - * - * If a query is made with a struct drm_xe_device_query where .query is equal to - * DRM_XE_DEVICE_QUERY_ENGINE_CYCLES, then the reply uses struct drm_xe_query_engine_cycles - * in .data. struct drm_xe_query_engine_cycles is allocated by the user and - * .data points to this allocated structure. - * - * The query returns the engine cycles, which along with GT's @reference_clock, - * can be used to calculate the engine timestamp. In addition the - * query returns a set of cpu timestamps that indicate when the command - * streamer cycle count was captured. - */ -struct drm_xe_query_engine_cycles { - /** - * @eci: This is input by the user and is the engine for which command - * streamer cycles is queried. - */ - struct drm_xe_engine_class_instance eci; - - /** - * @clockid: This is input by the user and is the reference clock id for - * CPU timestamp. For definition, see clock_gettime(2) and - * perf_event_open(2). Supported clock ids are CLOCK_MONOTONIC, - * CLOCK_MONOTONIC_RAW, CLOCK_REALTIME, CLOCK_BOOTTIME, CLOCK_TAI. - */ - __s32 clockid; - - /** @width: Width of the engine cycle counter in bits. */ - __u32 width; - - /** - * @engine_cycles: Engine cycles as read from its register - * at 0x358 offset. - */ - __u64 engine_cycles; - - /** - * @cpu_timestamp: CPU timestamp in ns. The timestamp is captured before - * reading the engine_cycles register using the reference clockid set by the - * user. - */ - __u64 cpu_timestamp; - - /** - * @cpu_delta: Time delta in ns captured around reading the lower dword - * of the engine_cycles register. - */ - __u64 cpu_delta; -}; - /** * struct drm_xe_query_mem_regions - describe memory regions * @@ -343,6 +381,19 @@ struct drm_xe_query_mem_regions { * is equal to DRM_XE_DEVICE_QUERY_CONFIG, then the reply uses * struct drm_xe_query_config in .data. * + * The index in @info can be: + * - %DRM_XE_QUERY_CONFIG_REV_AND_DEVICE_ID - Device ID (lower 16 bits) + * and the device revision (next 8 bits) + * - %DRM_XE_QUERY_CONFIG_FLAGS - Flags describing the device + * configuration, see list below + * + * - %DRM_XE_QUERY_CONFIG_FLAG_HAS_VRAM - Flag is set if the device + * has usable VRAM + * - %DRM_XE_QUERY_CONFIG_MIN_ALIGNMENT - Minimal memory alignment + * required by this device, typically SZ_4K or SZ_64K + * - %DRM_XE_QUERY_CONFIG_VA_BITS - Maximum bits of a virtual address + * - %DRM_XE_QUERY_CONFIG_MAX_EXEC_QUEUE_PRIORITY - Value of the highest + * available exec queue priority */ struct drm_xe_query_config { /** @num_params: number of parameters returned in info */ @@ -368,6 +419,10 @@ struct drm_xe_query_config { * existing GT individual descriptions. * Graphics Technology (GT) is a subset of a GPU/tile that is responsible for * implementing graphics and/or media operations. + * + * The index in @type can be: + * - %DRM_XE_QUERY_GT_TYPE_MAIN + * - %DRM_XE_QUERY_GT_TYPE_MEDIA */ struct drm_xe_gt { #define DRM_XE_QUERY_GT_TYPE_MAIN 0 @@ -386,6 +441,10 @@ struct drm_xe_gt { * @near_mem_regions: Bit mask of instances from * drm_xe_query_mem_regions that are nearest to the current engines * of this GT. + * Each index in this mask refers directly to the struct + * drm_xe_query_mem_regions' instance, no assumptions should + * be made about order. The type of each region is described + * by struct drm_xe_query_mem_regions' mem_class. */ __u64 near_mem_regions; /** @@ -394,6 +453,10 @@ struct drm_xe_gt { * In general, they have extra indirections when compared to the * @near_mem_regions. For a discrete device this could mean system * memory and memory living in a different tile. + * Each index in this mask refers directly to the struct + * drm_xe_query_mem_regions' instance, no assumptions should + * be made about order. The type of each region is described + * by struct drm_xe_query_mem_regions' mem_class. */ __u64 far_mem_regions; /** @reserved: Reserved */ @@ -425,34 +488,30 @@ struct drm_xe_query_gt_list { * If a query is made with a struct drm_xe_device_query where .query * is equal to DRM_XE_DEVICE_QUERY_GT_TOPOLOGY, then the reply uses * struct drm_xe_query_topology_mask in .data. + * + * The @type can be: + * - %DRM_XE_TOPO_DSS_GEOMETRY - To query the mask of Dual Sub Slices + * (DSS) available for geometry operations. For example a query response + * containing the following in mask: + * ``DSS_GEOMETRY ff ff ff ff 00 00 00 00`` + * means 32 DSS are available for geometry. + * - %DRM_XE_TOPO_DSS_COMPUTE - To query the mask of Dual Sub Slices + * (DSS) available for compute operations. For example a query response + * containing the following in mask: + * ``DSS_COMPUTE ff ff ff ff 00 00 00 00`` + * means 32 DSS are available for compute. + * - %DRM_XE_TOPO_EU_PER_DSS - To query the mask of Execution Units (EU) + * available per Dual Sub Slices (DSS). For example a query response + * containing the following in mask: + * ``EU_PER_DSS ff ff 00 00 00 00 00 00`` + * means each DSS has 16 EU. */ struct drm_xe_query_topology_mask { /** @gt_id: GT ID the mask is associated with */ __u16 gt_id; - /* - * To query the mask of Dual Sub Slices (DSS) available for geometry - * operations. For example a query response containing the following - * in mask: - * DSS_GEOMETRY ff ff ff ff 00 00 00 00 - * means 32 DSS are available for geometry. - */ #define DRM_XE_TOPO_DSS_GEOMETRY (1 << 0) - /* - * To query the mask of Dual Sub Slices (DSS) available for compute - * operations. For example a query response containing the following - * in mask: - * DSS_COMPUTE ff ff ff ff 00 00 00 00 - * means 32 DSS are available for compute. - */ #define DRM_XE_TOPO_DSS_COMPUTE (1 << 1) - /* - * To query the mask of Execution Units (EU) available per Dual Sub - * Slices (DSS). For example a query response containing the following - * in mask: - * EU_PER_DSS ff ff 00 00 00 00 00 00 - * means each DSS has 16 EU. - */ #define DRM_XE_TOPO_EU_PER_DSS (1 << 2) /** @type: type of mask */ __u16 type; @@ -465,11 +524,81 @@ struct drm_xe_query_topology_mask { }; /** - * struct drm_xe_device_query - main structure to query device information + * struct drm_xe_query_engine_cycles - correlate CPU and GPU timestamps + * + * If a query is made with a struct drm_xe_device_query where .query is equal to + * DRM_XE_DEVICE_QUERY_ENGINE_CYCLES, then the reply uses struct drm_xe_query_engine_cycles + * in .data. struct drm_xe_query_engine_cycles is allocated by the user and + * .data points to this allocated structure. + * + * The query returns the engine cycles, which along with GT's @reference_clock, + * can be used to calculate the engine timestamp. In addition the + * query returns a set of cpu timestamps that indicate when the command + * streamer cycle count was captured. + */ +struct drm_xe_query_engine_cycles { + /** + * @eci: This is input by the user and is the engine for which command + * streamer cycles is queried. + */ + struct drm_xe_engine_class_instance eci; + + /** + * @clockid: This is input by the user and is the reference clock id for + * CPU timestamp. For definition, see clock_gettime(2) and + * perf_event_open(2). Supported clock ids are CLOCK_MONOTONIC, + * CLOCK_MONOTONIC_RAW, CLOCK_REALTIME, CLOCK_BOOTTIME, CLOCK_TAI. + */ + __s32 clockid; + + /** @width: Width of the engine cycle counter in bits. */ + __u32 width; + + /** + * @engine_cycles: Engine cycles as read from its register + * at 0x358 offset. + */ + __u64 engine_cycles; + + /** + * @cpu_timestamp: CPU timestamp in ns. The timestamp is captured before + * reading the engine_cycles register using the reference clockid set by the + * user. + */ + __u64 cpu_timestamp; + + /** + * @cpu_delta: Time delta in ns captured around reading the lower dword + * of the engine_cycles register. + */ + __u64 cpu_delta; +}; + +/** + * struct drm_xe_device_query - Input of &DRM_IOCTL_XE_DEVICE_QUERY - main + * structure to query device information + * + * The user selects the type of data to query among DRM_XE_DEVICE_QUERY_* + * and sets the value in the query member. This determines the type of + * the structure provided by the driver in data, among struct drm_xe_query_*. * - * If size is set to 0, the driver fills it with the required size for the - * requested type of data to query. If size is equal to the required size, - * the queried information is copied into data. + * The @query can be: + * - %DRM_XE_DEVICE_QUERY_ENGINES + * - %DRM_XE_DEVICE_QUERY_MEM_REGIONS + * - %DRM_XE_DEVICE_QUERY_CONFIG + * - %DRM_XE_DEVICE_QUERY_GT_LIST + * - %DRM_XE_DEVICE_QUERY_HWCONFIG - Query type to retrieve the hardware + * configuration of the device such as information on slices, memory, + * caches, and so on. It is provided as a table of key / value + * attributes. + * - %DRM_XE_DEVICE_QUERY_GT_TOPOLOGY + * - %DRM_XE_DEVICE_QUERY_ENGINE_CYCLES + * + * If size is set to 0, the driver fills it with the required size for + * the requested type of data to query. If size is equal to the required + * size, the queried information is copied into data. If size is set to + * a value different from 0 and different from the required size, the + * IOCTL call returns -EINVAL. * * For example the following code snippet allows retrieving and printing * information about the device engines with DRM_XE_DEVICE_QUERY_ENGINES: @@ -527,6 +656,37 @@ struct drm_xe_device_query { __u64 reserved[2]; }; +/** + * struct drm_xe_gem_create - Input of &DRM_IOCTL_XE_GEM_CREATE - A structure for + * gem creation + * + * The @flags can be: + * - %DRM_XE_GEM_CREATE_FLAG_DEFER_BACKING + * - %DRM_XE_GEM_CREATE_FLAG_SCANOUT + * - %DRM_XE_GEM_CREATE_FLAG_NEEDS_VISIBLE_VRAM - When using VRAM as a + * possible placement, ensure that the corresponding VRAM allocation + * will always use the CPU accessible part of VRAM. This is important + * for small-bar systems (on full-bar systems this gets turned into a + * noop). + * Note1: System memory can be used as an extra placement if the kernel + * should spill the allocation to system memory, if space can't be made + * available in the CPU accessible part of VRAM (giving the same + * behaviour as the i915 interface, see + * I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS). + * Note2: For clear-color CCS surfaces the kernel needs to read the + * clear-color value stored in the buffer, and on discrete platforms we + * need to use VRAM for display surfaces, therefore the kernel requires + * setting this flag for such objects, otherwise an error is thrown on + * small-bar systems. + * + * @cpu_caching supports the following values: + * - %DRM_XE_GEM_CPU_CACHING_WB - Allocate the pages with write-back + * caching. On iGPU this can't be used for scanout surfaces. Currently + * not allowed for objects placed in VRAM. + * - %DRM_XE_GEM_CPU_CACHING_WC - Allocate the pages as write-combined. This + * is uncached. Scanout surfaces should likely use this. All objects + * that can be placed in VRAM must use this. + */ struct drm_xe_gem_create { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -537,26 +697,17 @@ struct drm_xe_gem_create { */ __u64 size; - /** @placement: A mask of memory instances of where BO can be placed. */ + /** + * @placement: A mask of memory instances of where BO can be placed. + * Each index in this mask refers directly to the struct + * drm_xe_query_mem_regions' instance, no assumptions should + * be made about order. The type of each region is described + * by struct drm_xe_query_mem_regions' mem_class. + */ __u32 placement; #define DRM_XE_GEM_CREATE_FLAG_DEFER_BACKING (1 << 0) #define DRM_XE_GEM_CREATE_FLAG_SCANOUT (1 << 1) -/* - * When using VRAM as a possible placement, ensure that the corresponding VRAM - * allocation will always use the CPU accessible part of VRAM. This is important - * for small-bar systems (on full-bar systems this gets turned into a noop). - * - * Note: System memory can be used as an extra placement if the kernel should - * spill the allocation to system memory, if space can't be made available in - * the CPU accessible part of VRAM (giving the same behaviour as the i915 - * interface, see I915_GEM_CREATE_EXT_FLAG_NEEDS_CPU_ACCESS). - * - * Note: For clear-color CCS surfaces the kernel needs to read the clear-color - * value stored in the buffer, and on discrete platforms we need to use VRAM for - * display surfaces, therefore the kernel requires setting this flag for such - * objects, otherwise an error is thrown on small-bar systems. - */ #define DRM_XE_GEM_CREATE_FLAG_NEEDS_VISIBLE_VRAM (1 << 2) /** * @flags: Flags, currently a mask of memory instances of where BO can @@ -581,22 +732,12 @@ struct drm_xe_gem_create { */ __u32 handle; +#define DRM_XE_GEM_CPU_CACHING_WB 1 +#define DRM_XE_GEM_CPU_CACHING_WC 2 /** * @cpu_caching: The CPU caching mode to select for this object. If * mmaping the object the mode selected here will also be used. - * - * Supported values: - * - * DRM_XE_GEM_CPU_CACHING_WB: Allocate the pages with write-back - * caching. On iGPU this can't be used for scanout surfaces. Currently - * not allowed for objects placed in VRAM. - * - * DRM_XE_GEM_CPU_CACHING_WC: Allocate the pages as write-combined. This - * is uncached. Scanout surfaces should likely use this. All objects - * that can be placed in VRAM must use this. */ -#define DRM_XE_GEM_CPU_CACHING_WB 1 -#define DRM_XE_GEM_CPU_CACHING_WC 2 __u16 cpu_caching; /** @pad: MBZ */ __u16 pad[3]; @@ -605,6 +746,9 @@ struct drm_xe_gem_create { __u64 reserved[2]; }; +/** + * struct drm_xe_gem_mmap_offset - Input of &DRM_IOCTL_XE_GEM_MMAP_OFFSET + */ struct drm_xe_gem_mmap_offset { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -622,53 +766,36 @@ struct drm_xe_gem_mmap_offset { __u64 reserved[2]; }; -/** struct drm_xe_ext_set_property - XE set property extension */ -struct drm_xe_ext_set_property { - /** @base: base user extension */ - struct xe_user_extension base; - - /** @property: property to set */ - __u32 property; - - /** @pad: MBZ */ - __u32 pad; - - /** @value: property value */ - __u64 value; - - /** @reserved: Reserved */ - __u64 reserved[2]; -}; - +/** + * struct drm_xe_vm_create - Input of &DRM_IOCTL_XE_VM_CREATE + * + * The @flags can be: + * - %DRM_XE_VM_CREATE_FLAG_SCRATCH_PAGE + * - %DRM_XE_VM_CREATE_FLAG_LR_MODE - An LR, or Long Running VM accepts + * exec submissions to its exec_queues that don't have an upper time + * limit on the job execution time. But exec submissions to these + * don't allow any of the flags DRM_XE_SYNC_FLAG_SYNCOBJ, + * DRM_XE_SYNC_FLAG_TIMELINE_SYNCOBJ, DRM_XE_SYNC_FLAG_DMA_BUF, + * used as out-syncobjs, that is, together with DRM_XE_SYNC_FLAG_SIGNAL. + * LR VMs can be created in recoverable page-fault mode using + * DRM_XE_VM_CREATE_FLAG_FAULT_MODE, if the device supports it. + * If that flag is omitted, the UMD can not rely on the slightly + * different per-VM overcommit semantics that are enabled by + * DRM_XE_VM_CREATE_FLAG_FAULT_MODE (see below), but KMD may + * still enable recoverable pagefaults if supported by the device. + * - %DRM_XE_VM_CREATE_FLAG_FAULT_MODE - Requires also + * DRM_XE_VM_CREATE_FLAG_LR_MODE. It allows memory to be allocated on + * demand when accessed, and also allows per-VM overcommit of memory. + * The xe driver internally uses recoverable pagefaults to implement + * this. + */ struct drm_xe_vm_create { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; #define DRM_XE_VM_CREATE_FLAG_SCRATCH_PAGE (1 << 0) - /* - * An LR, or Long Running VM accepts exec submissions - * to its exec_queues that don't have an upper time limit on - * the job execution time. But exec submissions to these - * don't allow any of the flags DRM_XE_SYNC_FLAG_SYNCOBJ, - * DRM_XE_SYNC_FLAG_TIMELINE_SYNCOBJ, DRM_XE_SYNC_FLAG_DMA_BUF, - * used as out-syncobjs, that is, together with DRM_XE_SYNC_FLAG_SIGNAL. - * LR VMs can be created in recoverable page-fault mode using - * DRM_XE_VM_CREATE_FLAG_FAULT_MODE, if the device supports it. - * If that flag is omitted, the UMD can not rely on the slightly - * different per-VM overcommit semantics that are enabled by - * DRM_XE_VM_CREATE_FLAG_FAULT_MODE (see below), but KMD may - * still enable recoverable pagefaults if supported by the device. - */ #define DRM_XE_VM_CREATE_FLAG_LR_MODE (1 << 1) -#define DRM_XE_VM_CREATE_FLAG_ASYNC_DEFAULT (1 << 2) - /* - * DRM_XE_VM_CREATE_FLAG_FAULT_MODE requires also - * DRM_XE_VM_CREATE_FLAG_LR_MODE. It allows memory to be allocated - * on demand when accessed, and also allows per-VM overcommit of memory. - * The xe driver internally uses recoverable pagefaults to implement - * this. - */ -#define DRM_XE_VM_CREATE_FLAG_FAULT_MODE (1 << 3) +#define DRM_XE_VM_CREATE_FLAG_FAULT_MODE (1 << 2) /** @flags: Flags */ __u32 flags; @@ -679,6 +806,9 @@ struct drm_xe_vm_create { __u64 reserved[2]; }; +/** + * struct drm_xe_vm_destroy - Input of &DRM_IOCTL_XE_VM_DESTROY + */ struct drm_xe_vm_destroy { /** @vm_id: VM ID */ __u32 vm_id; @@ -690,6 +820,29 @@ struct drm_xe_vm_destroy { __u64 reserved[2]; }; +/** + * struct drm_xe_vm_bind_op - run bind operations + * + * The @op can be: + * - %DRM_XE_VM_BIND_OP_MAP + * - %DRM_XE_VM_BIND_OP_UNMAP + * - %DRM_XE_VM_BIND_OP_MAP_USERPTR + * - %DRM_XE_VM_BIND_OP_UNMAP_ALL + * - %DRM_XE_VM_BIND_OP_PREFETCH + * + * and the @flags can be: + * - %DRM_XE_VM_BIND_FLAG_READONLY + * - %DRM_XE_VM_BIND_FLAG_ASYNC + * - %DRM_XE_VM_BIND_FLAG_IMMEDIATE - Valid on a faulting VM only, do the + * MAP operation immediately rather than deferring the MAP to the page + * fault handler. + * - %DRM_XE_VM_BIND_FLAG_NULL - When the NULL flag is set, the page + * tables are setup with a special bit which indicates writes are + * dropped and all reads return zero. In the future, the NULL flags + * will only be valid for DRM_XE_VM_BIND_OP_MAP operations, the BO + * handle MBZ, and the BO offset MBZ. This flag is intended to + * implement VK sparse bindings. + */ struct drm_xe_vm_bind_op { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -736,6 +889,12 @@ struct drm_xe_vm_bind_op { * * Note: For userptr and externally imported dma-buf the kernel expects * either 1WAY or 2WAY for the @pat_index. + * + * For DRM_XE_VM_BIND_FLAG_NULL bindings there are no KMD restrictions + * on the @pat_index. For such mappings there is no actual memory being + * mapped (the address in the PTE is invalid), so the various PAT memory + * attributes likely do not apply. Simply leaving as zero is one + * option (still a valid pat_index). */ __u16 pat_index; @@ -770,20 +929,8 @@ struct drm_xe_vm_bind_op { __u32 op; #define DRM_XE_VM_BIND_FLAG_READONLY (1 << 0) -#define DRM_XE_VM_BIND_FLAG_ASYNC (1 << 1) - /* - * Valid on a faulting VM only, do the MAP operation immediately rather - * than deferring the MAP to the page fault handler. - */ -#define DRM_XE_VM_BIND_FLAG_IMMEDIATE (1 << 2) - /* - * When the NULL flag is set, the page tables are setup with a special - * bit which indicates writes are dropped and all reads return zero. In - * the future, the NULL flags will only be valid for DRM_XE_VM_BIND_OP_MAP - * operations, the BO handle MBZ, and the BO offset MBZ. This flag is - * intended to implement VK sparse bindings. - */ -#define DRM_XE_VM_BIND_FLAG_NULL (1 << 3) +#define DRM_XE_VM_BIND_FLAG_IMMEDIATE (1 << 1) +#define DRM_XE_VM_BIND_FLAG_NULL (1 << 2) /** @flags: Bind flags */ __u32 flags; @@ -794,13 +941,40 @@ struct drm_xe_vm_bind_op { */ __u32 prefetch_mem_region_instance; - /** @pad: MBZ */ + /** @pad2: MBZ */ __u32 pad2; /** @reserved: Reserved */ __u64 reserved[3]; }; +/** + * struct drm_xe_vm_bind - Input of &DRM_IOCTL_XE_VM_BIND + * + * Below is an example of a minimal use of @drm_xe_vm_bind to + * asynchronously bind the buffer `data` at address `BIND_ADDRESS` to + * illustrate `userptr`. It can be synchronized by using the example + * provided for @drm_xe_sync. + * + * .. code-block:: C + * + * data = aligned_alloc(ALIGNMENT, BO_SIZE); + * struct drm_xe_vm_bind bind = { + * .vm_id = vm, + * .num_binds = 1, + * .bind.obj = 0, + * .bind.obj_offset = to_user_pointer(data), + * .bind.range = BO_SIZE, + * .bind.addr = BIND_ADDRESS, + * .bind.op = DRM_XE_VM_BIND_OP_MAP_USERPTR, + * .bind.flags = 0, + * .num_syncs = 1, + * .syncs = &sync, + * .exec_queue_id = 0, + * }; + * ioctl(fd, DRM_IOCTL_XE_VM_BIND, &bind); + * + */ struct drm_xe_vm_bind { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -832,7 +1006,7 @@ struct drm_xe_vm_bind { __u64 vector_of_binds; }; - /** @pad: MBZ */ + /** @pad2: MBZ */ __u32 pad2; /** @num_syncs: amount of syncs to wait on */ @@ -845,20 +1019,28 @@ struct drm_xe_vm_bind { __u64 reserved[2]; }; -/* For use with DRM_XE_EXEC_QUEUE_SET_PROPERTY_ACC_GRANULARITY */ - -/* Monitor 128KB contiguous region with 4K sub-granularity */ -#define DRM_XE_ACC_GRANULARITY_128K 0 - -/* Monitor 2MB contiguous region with 64KB sub-granularity */ -#define DRM_XE_ACC_GRANULARITY_2M 1 - -/* Monitor 16MB contiguous region with 512KB sub-granularity */ -#define DRM_XE_ACC_GRANULARITY_16M 2 - -/* Monitor 64MB contiguous region with 2M sub-granularity */ -#define DRM_XE_ACC_GRANULARITY_64M 3 - +/** + * struct drm_xe_exec_queue_create - Input of &DRM_IOCTL_XE_EXEC_QUEUE_CREATE + * + * The example below shows how to use @drm_xe_exec_queue_create to create + * a simple exec_queue (no parallel submission) of class + * &DRM_XE_ENGINE_CLASS_RENDER. + * + * .. code-block:: C + * + * struct drm_xe_engine_class_instance instance = { + * .engine_class = DRM_XE_ENGINE_CLASS_RENDER, + * }; + * struct drm_xe_exec_queue_create exec_queue_create = { + * .extensions = 0, + * .vm_id = vm, + * .num_bb_per_exec = 1, + * .num_eng_per_bb = 1, + * .instances = to_user_pointer(&instance), + * }; + * ioctl(fd, DRM_IOCTL_XE_EXEC_QUEUE_CREATE, &exec_queue_create); + * + */ struct drm_xe_exec_queue_create { #define DRM_XE_EXEC_QUEUE_EXTENSION_SET_PROPERTY 0 #define DRM_XE_EXEC_QUEUE_SET_PROPERTY_PRIORITY 0 @@ -869,6 +1051,14 @@ struct drm_xe_exec_queue_create { #define DRM_XE_EXEC_QUEUE_SET_PROPERTY_ACC_TRIGGER 5 #define DRM_XE_EXEC_QUEUE_SET_PROPERTY_ACC_NOTIFY 6 #define DRM_XE_EXEC_QUEUE_SET_PROPERTY_ACC_GRANULARITY 7 +/* Monitor 128KB contiguous region with 4K sub-granularity */ +#define DRM_XE_ACC_GRANULARITY_128K 0 +/* Monitor 2MB contiguous region with 64KB sub-granularity */ +#define DRM_XE_ACC_GRANULARITY_2M 1 +/* Monitor 16MB contiguous region with 512KB sub-granularity */ +#define DRM_XE_ACC_GRANULARITY_16M 2 +/* Monitor 64MB contiguous region with 2M sub-granularity */ +#define DRM_XE_ACC_GRANULARITY_64M 3 /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -901,6 +1091,26 @@ struct drm_xe_exec_queue_create { __u64 reserved[2]; }; +/** + * struct drm_xe_exec_queue_destroy - Input of &DRM_IOCTL_XE_EXEC_QUEUE_DESTROY + */ +struct drm_xe_exec_queue_destroy { + /** @exec_queue_id: Exec queue ID */ + __u32 exec_queue_id; + + /** @pad: MBZ */ + __u32 pad; + + /** @reserved: Reserved */ + __u64 reserved[2]; +}; + +/** + * struct drm_xe_exec_queue_get_property - Input of &DRM_IOCTL_XE_EXEC_QUEUE_GET_PROPERTY + * + * The @property can be: + * - %DRM_XE_EXEC_QUEUE_GET_PROPERTY_BAN + */ struct drm_xe_exec_queue_get_property { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -919,17 +1129,41 @@ struct drm_xe_exec_queue_get_property { __u64 reserved[2]; }; -struct drm_xe_exec_queue_destroy { - /** @exec_queue_id: Exec queue ID */ - __u32 exec_queue_id; - - /** @pad: MBZ */ - __u32 pad; - - /** @reserved: Reserved */ - __u64 reserved[2]; -}; - +/** + * struct drm_xe_sync - sync object + * + * The @type can be: + * - %DRM_XE_SYNC_TYPE_SYNCOBJ + * - %DRM_XE_SYNC_TYPE_TIMELINE_SYNCOBJ + * - %DRM_XE_SYNC_TYPE_USER_FENCE + * + * and the @flags can be: + * - %DRM_XE_SYNC_FLAG_SIGNAL + * + * A minimal use of @drm_xe_sync looks like this: + * + * .. code-block:: C + * + * struct drm_xe_sync sync = { + * .flags = DRM_XE_SYNC_FLAG_SIGNAL, + * .type = DRM_XE_SYNC_TYPE_SYNCOBJ, + * }; + * struct drm_syncobj_create syncobj_create = { 0 }; + * ioctl(fd, DRM_IOCTL_SYNCOBJ_CREATE, &syncobj_create); + * sync.handle = syncobj_create.handle; + * ... + * use of &sync in drm_xe_exec or drm_xe_vm_bind + * ... + * struct drm_syncobj_wait wait = { + * .handles = &sync.handle, + * .timeout_nsec = INT64_MAX, + * .count_handles = 1, + * .flags = 0, + * .first_signaled = 0, + * .pad = 0, + * }; + * ioctl(fd, DRM_IOCTL_SYNCOBJ_WAIT, &wait); + */ struct drm_xe_sync { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -945,11 +1179,12 @@ struct drm_xe_sync { __u32 flags; union { + /** @handle: Handle for the object */ __u32 handle; /** - * @addr: Address of user fence. When sync passed in via exec - * IOCTL this a GPU address in the VM. When sync passed in via + * @addr: Address of user fence. When sync is passed in via exec + * IOCTL this is a GPU address in the VM. When sync passed in via * VM bind IOCTL this is a user pointer. In either case, it is * the users responsibility that this address is present and * mapped when the user fence is signalled. Must be qword @@ -958,12 +1193,36 @@ struct drm_xe_sync { __u64 addr; }; + /** + * @timeline_value: Input for the timeline sync object. Needs to be + * different than 0 when used with %DRM_XE_SYNC_FLAG_TIMELINE_SYNCOBJ. + */ __u64 timeline_value; /** @reserved: Reserved */ __u64 reserved[2]; }; +/** + * struct drm_xe_exec - Input of &DRM_IOCTL_XE_EXEC + * + * This is an example to use @drm_xe_exec for execution of the object + * at BIND_ADDRESS (see example in @drm_xe_vm_bind) by an exec_queue + * (see example in @drm_xe_exec_queue_create). It can be synchronized + * by using the example provided for @drm_xe_sync. + * + * .. code-block:: C + * + * struct drm_xe_exec exec = { + * .exec_queue_id = exec_queue, + * .syncs = &sync, + * .num_syncs = 1, + * .address = BIND_ADDRESS, + * .num_batch_buffer = 1, + * }; + * ioctl(fd, DRM_IOCTL_XE_EXEC, &exec); + * + */ struct drm_xe_exec { /** @extensions: Pointer to the first extension struct, if any */ __u64 extensions; @@ -997,7 +1256,7 @@ struct drm_xe_exec { }; /** - * struct drm_xe_wait_user_fence - wait user fence + * struct drm_xe_wait_user_fence - Input of &DRM_IOCTL_XE_WAIT_USER_FENCE * * Wait on user fence, XE will wake-up on every HW engine interrupt in the * instances list and check if user fence is complete:: @@ -1005,6 +1264,24 @@ struct drm_xe_exec { * (*addr & MASK) OP (VALUE & MASK) * * Returns to user on user fence completion or timeout. + * + * The @op can be: + * - %DRM_XE_UFENCE_WAIT_OP_EQ + * - %DRM_XE_UFENCE_WAIT_OP_NEQ + * - %DRM_XE_UFENCE_WAIT_OP_GT + * - %DRM_XE_UFENCE_WAIT_OP_GTE + * - %DRM_XE_UFENCE_WAIT_OP_LT + * - %DRM_XE_UFENCE_WAIT_OP_LTE + * + * and the @flags can be: + * - %DRM_XE_UFENCE_WAIT_FLAG_ABSTIME + * - %DRM_XE_UFENCE_WAIT_FLAG_SOFT_OP + * + * The @mask values can be for example: + * - 0xffu for u8 + * - 0xffffu for u16 + * - 0xffffffffu for u32 + * - 0xffffffffffffffffu for u64 */ struct drm_xe_wait_user_fence { /** @extensions: Pointer to the first extension struct, if any */ @@ -1024,8 +1301,7 @@ struct drm_xe_wait_user_fence { /** @op: wait operation (type of comparison) */ __u16 op; -#define DRM_XE_UFENCE_WAIT_FLAG_SOFT_OP (1 << 0) /* e.g. Wait on VM bind */ -#define DRM_XE_UFENCE_WAIT_FLAG_ABSTIME (1 << 1) +#define DRM_XE_UFENCE_WAIT_FLAG_ABSTIME (1 << 0) /** @flags: wait flags */ __u16 flags; @@ -1035,10 +1311,6 @@ struct drm_xe_wait_user_fence { /** @value: compare value */ __u64 value; -#define DRM_XE_UFENCE_WAIT_MASK_U8 0xffu -#define DRM_XE_UFENCE_WAIT_MASK_U16 0xffffu -#define DRM_XE_UFENCE_WAIT_MASK_U32 0xffffffffu -#define DRM_XE_UFENCE_WAIT_MASK_U64 0xffffffffffffffffu /** @mask: comparison mask */ __u64 mask; @@ -1058,61 +1330,16 @@ struct drm_xe_wait_user_fence { */ __s64 timeout; - /** - * @num_engines: number of engine instances to wait on, must be zero - * when DRM_XE_UFENCE_WAIT_FLAG_SOFT_OP set - */ - __u64 num_engines; + /** @exec_queue_id: exec_queue_id returned from xe_exec_queue_create_ioctl */ + __u32 exec_queue_id; - /** - * @instances: user pointer to array of drm_xe_engine_class_instance to - * wait on, must be NULL when DRM_XE_UFENCE_WAIT_FLAG_SOFT_OP set - */ - __u64 instances; + /** @pad2: MBZ */ + __u32 pad2; /** @reserved: Reserved */ __u64 reserved[2]; }; -/** - * DOC: XE PMU event config IDs - * - * Check 'man perf_event_open' to use the ID's DRM_XE_PMU_XXXX listed in xe_drm.h - * in 'struct perf_event_attr' as part of perf_event_open syscall to read a - * particular event. - * - * For example to open the DRMXE_PMU_RENDER_GROUP_BUSY(0): - * - * .. code-block:: C - * - * struct perf_event_attr attr; - * long long count; - * int cpu = 0; - * int fd; - * - * memset(&attr, 0, sizeof(struct perf_event_attr)); - * attr.type = type; // eg: /sys/bus/event_source/devices/xe_0000_56_00.0/type - * attr.read_format = PERF_FORMAT_TOTAL_TIME_ENABLED; - * attr.use_clockid = 1; - * attr.clockid = CLOCK_MONOTONIC; - * attr.config = DRM_XE_PMU_RENDER_GROUP_BUSY(0); - * - * fd = syscall(__NR_perf_event_open, &attr, -1, cpu, -1, 0); - */ - -/* - * Top bits of every counter are GT id. - */ -#define __DRM_XE_PMU_GT_SHIFT (56) - -#define ___DRM_XE_PMU_OTHER(gt, x) \ - (((__u64)(x)) | ((__u64)(gt) << __DRM_XE_PMU_GT_SHIFT)) - -#define DRM_XE_PMU_RENDER_GROUP_BUSY(gt) ___DRM_XE_PMU_OTHER(gt, 0) -#define DRM_XE_PMU_COPY_GROUP_BUSY(gt) ___DRM_XE_PMU_OTHER(gt, 1) -#define DRM_XE_PMU_MEDIA_GROUP_BUSY(gt) ___DRM_XE_PMU_OTHER(gt, 2) -#define DRM_XE_PMU_ANY_ENGINE_GROUP_BUSY(gt) ___DRM_XE_PMU_OTHER(gt, 3) - #if defined(__cplusplus) } #endif diff --git a/src/gallium/drivers/iris/xe/iris_kmd_backend.c b/src/gallium/drivers/iris/xe/iris_kmd_backend.c index 4c4d731ca6a..2f49c4d8bea 100644 --- a/src/gallium/drivers/iris/xe/iris_kmd_backend.c +++ b/src/gallium/drivers/iris/xe/iris_kmd_backend.c @@ -106,8 +106,25 @@ xe_gem_vm_bind_op(struct iris_bo *bo, uint32_t op) { const struct intel_device_info *devinfo = iris_bufmgr_get_device_info(bo->bufmgr); uint32_t handle = op == DRM_XE_VM_BIND_OP_UNMAP ? 0 : bo->gem_handle; + struct drm_xe_sync xe_sync = { + .type = DRM_XE_SYNC_TYPE_SYNCOBJ, + .flags = DRM_XE_SYNC_FLAG_SIGNAL, + }; + struct drm_syncobj_create syncobj_create = {}; + struct drm_syncobj_destroy syncobj_destroy = {}; + struct drm_syncobj_wait syncobj_wait = { + .timeout_nsec = INT64_MAX, + .count_handles = 1, + }; uint64_t range, obj_offset = 0; - int ret; + int ret, fd; + + fd = iris_bufmgr_get_fd(bo->bufmgr); + ret = intel_ioctl(fd, DRM_IOCTL_SYNCOBJ_CREATE, &syncobj_create); + if (ret) + return ret; + + xe_sync.handle = syncobj_create.handle; if (iris_bo_is_imported(bo)) range = bo->size; @@ -127,6 +144,8 @@ xe_gem_vm_bind_op(struct iris_bo *bo, uint32_t op) struct drm_xe_vm_bind args = { .vm_id = iris_bufmgr_get_global_vm_id(bo->bufmgr), + .num_syncs = 1, + .syncs = (uintptr_t)&xe_sync, .num_binds = 1, .bind.obj = handle, .bind.obj_offset = obj_offset, @@ -135,11 +154,17 @@ xe_gem_vm_bind_op(struct iris_bo *bo, uint32_t op) .bind.op = op, .bind.pat_index = pat_index, }; - ret = intel_ioctl(iris_bufmgr_get_fd(bo->bufmgr), DRM_IOCTL_XE_VM_BIND, &args); - if (ret) { + ret = intel_ioctl(fd, DRM_IOCTL_XE_VM_BIND, &args); + if (ret == 0) { + syncobj_wait.handles = (uintptr_t)&xe_sync.handle; + ret = intel_ioctl(fd, DRM_IOCTL_SYNCOBJ_WAIT, &syncobj_wait); + } else { DBG("vm_bind_op: DRM_IOCTL_XE_VM_BIND failed(%i)", ret); } + syncobj_destroy.handle = xe_sync.handle; + intel_ioctl(fd, DRM_IOCTL_SYNCOBJ_DESTROY, &syncobj_destroy); + return ret; } diff --git a/src/intel/vulkan/xe/anv_kmd_backend.c b/src/intel/vulkan/xe/anv_kmd_backend.c index db0957dd2f4..9124c04f82a 100644 --- a/src/intel/vulkan/xe/anv_kmd_backend.c +++ b/src/intel/vulkan/xe/anv_kmd_backend.c @@ -116,13 +116,24 @@ static inline int xe_vm_bind_op(struct anv_device *device, struct anv_sparse_submission *submit) { - int ret; - + struct drm_xe_sync xe_sync = { + .type = DRM_XE_SYNC_TYPE_SYNCOBJ, + .flags = DRM_XE_SYNC_FLAG_SIGNAL, + }; struct drm_xe_vm_bind args = { .vm_id = device->vm_id, .num_binds = submit->binds_len, .bind = {}, + .num_syncs = 1, + .syncs = (uintptr_t)&xe_sync, + }; + struct drm_syncobj_create syncobj_create = {}; + struct drm_syncobj_destroy syncobj_destroy = {}; + struct drm_syncobj_wait syncobj_wait = { + .timeout_nsec = INT64_MAX, + .count_handles = 1, }; + int ret; STACK_ARRAY(struct drm_xe_vm_bind_op, xe_binds_stackarray, submit->binds_len); @@ -173,7 +184,22 @@ xe_vm_bind_op(struct anv_device *device, xe_bind->userptr = (uintptr_t)bo->map; } + ret = intel_ioctl(device->fd, DRM_IOCTL_SYNCOBJ_CREATE, &syncobj_create); + if (ret) + goto out_stackarray; + + xe_sync.handle = syncobj_create.handle; ret = intel_ioctl(device->fd, DRM_IOCTL_XE_VM_BIND, &args); + if (ret) + goto out_destroy_syncobj; + + syncobj_wait.handles = (uintptr_t)&xe_sync.handle; + ret = intel_ioctl(device->fd, DRM_IOCTL_SYNCOBJ_WAIT, &syncobj_wait); + +out_destroy_syncobj: + syncobj_destroy.handle = xe_sync.handle; + intel_ioctl(device->fd, DRM_IOCTL_SYNCOBJ_DESTROY, &syncobj_destroy); +out_stackarray: STACK_ARRAY_FINISH(xe_binds_stackarray); return ret;
