/* SPDX-License-Identifier: GPL-2.0 OR MIT */ /* * Copyright 2014-2022 Advanced Micro Devices, Inc. * * 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 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 COPYRIGHT HOLDER(S) OR AUTHOR(S) 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 KFD_PRIV_H_INCLUDED #define KFD_PRIV_H_INCLUDED #include <linux/hashtable.h> #include <linux/mmu_notifier.h> #include <linux/memremap.h> #include <linux/mutex.h> #include <linux/types.h> #include <linux/atomic.h> #include <linux/workqueue.h> #include <linux/spinlock.h> #include <linux/kfd_ioctl.h> #include <linux/idr.h> #include <linux/kfifo.h> #include <linux/seq_file.h> #include <linux/kref.h> #include <linux/sysfs.h> #include <linux/device_cgroup.h> #include <drm/drm_file.h> #include <drm/drm_drv.h> #include <drm/drm_device.h> #include <drm/drm_ioctl.h> #include <kgd_kfd_interface.h> #include <linux/swap.h> #include "amd_shared.h" #include "amdgpu.h" #define KFD_MAX_RING_ENTRY_SIZE … #define KFD_SYSFS_FILE_MODE … /* GPU ID hash width in bits */ #define KFD_GPU_ID_HASH_WIDTH … /* Use upper bits of mmap offset to store KFD driver specific information. * BITS[63:62] - Encode MMAP type * BITS[61:46] - Encode gpu_id. To identify to which GPU the offset belongs to * BITS[45:0] - MMAP offset value * * NOTE: struct vm_area_struct.vm_pgoff uses offset in pages. Hence, these * defines are w.r.t to PAGE_SIZE */ #define KFD_MMAP_TYPE_SHIFT … #define KFD_MMAP_TYPE_MASK … #define KFD_MMAP_TYPE_DOORBELL … #define KFD_MMAP_TYPE_EVENTS … #define KFD_MMAP_TYPE_RESERVED_MEM … #define KFD_MMAP_TYPE_MMIO … #define KFD_MMAP_GPU_ID_SHIFT … #define KFD_MMAP_GPU_ID_MASK … #define KFD_MMAP_GPU_ID(gpu_id) … #define KFD_MMAP_GET_GPU_ID(offset) … /* * When working with cp scheduler we should assign the HIQ manually or via * the amdgpu driver to a fixed hqd slot, here are the fixed HIQ hqd slot * definitions for Kaveri. In Kaveri only the first ME queues participates * in the cp scheduling taking that in mind we set the HIQ slot in the * second ME. */ #define KFD_CIK_HIQ_PIPE … #define KFD_CIK_HIQ_QUEUE … /* Macro for allocating structures */ #define kfd_alloc_struct(ptr_to_struct) … #define KFD_MAX_NUM_OF_PROCESSES … #define KFD_MAX_NUM_OF_QUEUES_PER_PROCESS … /* * Size of the per-process TBA+TMA buffer: 2 pages * * The first chunk is the TBA used for the CWSR ISA code. The second * chunk is used as TMA for user-mode trap handler setup in daisy-chain mode. */ #define KFD_CWSR_TBA_TMA_SIZE … #define KFD_CWSR_TMA_OFFSET … #define KFD_MAX_NUM_OF_QUEUES_PER_DEVICE … #define KFD_KERNEL_QUEUE_SIZE … #define KFD_UNMAP_LATENCY_MS … #define KFD_MAX_SDMA_QUEUES … /* * 512 = 0x200 * The doorbell index distance between SDMA RLC (2*i) and (2*i+1) in the * same SDMA engine on SOC15, which has 8-byte doorbells for SDMA. * 512 8-byte doorbell distance (i.e. one page away) ensures that SDMA RLC * (2*i+1) doorbells (in terms of the lower 12 bit address) lie exactly in * the OFFSET and SIZE set in registers like BIF_SDMA0_DOORBELL_RANGE. */ #define KFD_QUEUE_DOORBELL_MIRROR_OFFSET … /** * enum kfd_ioctl_flags - KFD ioctl flags * Various flags that can be set in &amdkfd_ioctl_desc.flags to control how * userspace can use a given ioctl. */ enum kfd_ioctl_flags { … }; /* * Kernel module parameter to specify maximum number of supported queues per * device */ extern int max_num_of_queues_per_device; /* Kernel module parameter to specify the scheduling policy */ extern int sched_policy; /* * Kernel module parameter to specify the maximum process * number per HW scheduler */ extern int hws_max_conc_proc; extern int cwsr_enable; /* * Kernel module parameter to specify whether to send sigterm to HSA process on * unhandled exception */ extern int send_sigterm; /* * This kernel module is used to simulate large bar machine on non-large bar * enabled machines. */ extern int debug_largebar; /* Set sh_mem_config.retry_disable on GFX v9 */ extern int amdgpu_noretry; /* Halt if HWS hang is detected */ extern int halt_if_hws_hang; /* Whether MEC FW support GWS barriers */ extern bool hws_gws_support; /* Queue preemption timeout in ms */ extern int queue_preemption_timeout_ms; /* * Don't evict process queues on vm fault */ extern int amdgpu_no_queue_eviction_on_vm_fault; /* Enable eviction debug messages */ extern bool debug_evictions; extern struct mutex kfd_processes_mutex; enum cache_policy { … }; #define KFD_GC_VERSION(dev) … #define KFD_IS_SOC15(dev) … #define KFD_SUPPORT_XNACK_PER_PROCESS(dev) … struct kfd_node; struct kfd_event_interrupt_class { … }; struct kfd_device_info { … }; unsigned int kfd_get_num_sdma_engines(struct kfd_node *kdev); unsigned int kfd_get_num_xgmi_sdma_engines(struct kfd_node *kdev); struct kfd_mem_obj { … }; struct kfd_vmid_info { … }; #define MAX_KFD_NODES … struct kfd_dev; struct kfd_node { … }; struct kfd_dev { … }; enum kfd_mempool { … }; /* Character device interface */ int kfd_chardev_init(void); void kfd_chardev_exit(void); /** * enum kfd_unmap_queues_filter - Enum for queue filters. * * @KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES: Preempts all queues in the * running queues list. * * @KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES: Preempts all non-static queues * in the run list. * * @KFD_UNMAP_QUEUES_FILTER_BY_PASID: Preempts queues that belongs to * specific process. * */ enum kfd_unmap_queues_filter { … }; /** * enum kfd_queue_type - Enum for various queue types. * * @KFD_QUEUE_TYPE_COMPUTE: Regular user mode queue type. * * @KFD_QUEUE_TYPE_SDMA: SDMA user mode queue type. * * @KFD_QUEUE_TYPE_HIQ: HIQ queue type. * * @KFD_QUEUE_TYPE_DIQ: DIQ queue type. * * @KFD_QUEUE_TYPE_SDMA_XGMI: Special SDMA queue for XGMI interface. */ enum kfd_queue_type { … }; enum kfd_queue_format { … }; enum KFD_QUEUE_PRIORITY { … }; /** * struct queue_properties * * @type: The queue type. * * @queue_id: Queue identifier. * * @queue_address: Queue ring buffer address. * * @queue_size: Queue ring buffer size. * * @priority: Defines the queue priority relative to other queues in the * process. * This is just an indication and HW scheduling may override the priority as * necessary while keeping the relative prioritization. * the priority granularity is from 0 to f which f is the highest priority. * currently all queues are initialized with the highest priority. * * @queue_percent: This field is partially implemented and currently a zero in * this field defines that the queue is non active. * * @read_ptr: User space address which points to the number of dwords the * cp read from the ring buffer. This field updates automatically by the H/W. * * @write_ptr: Defines the number of dwords written to the ring buffer. * * @doorbell_ptr: Notifies the H/W of new packet written to the queue ring * buffer. This field should be similar to write_ptr and the user should * update this field after updating the write_ptr. * * @doorbell_off: The doorbell offset in the doorbell pci-bar. * * @is_interop: Defines if this is a interop queue. Interop queue means that * the queue can access both graphics and compute resources. * * @is_evicted: Defines if the queue is evicted. Only active queues * are evicted, rendering them inactive. * * @is_active: Defines if the queue is active or not. @is_active and * @is_evicted are protected by the DQM lock. * * @is_gws: Defines if the queue has been updated to be GWS-capable or not. * @is_gws should be protected by the DQM lock, since changing it can yield the * possibility of updating DQM state on number of GWS queues. * * @vmid: If the scheduling mode is no cp scheduling the field defines the vmid * of the queue. * * This structure represents the queue properties for each queue no matter if * it's user mode or kernel mode queue. * */ struct queue_properties { … }; #define QUEUE_IS_ACTIVE(q) … enum mqd_update_flag { … }; struct mqd_update_info { … }; /** * struct queue * * @list: Queue linked list. * * @mqd: The queue MQD (memory queue descriptor). * * @mqd_mem_obj: The MQD local gpu memory object. * * @gart_mqd_addr: The MQD gart mc address. * * @properties: The queue properties. * * @mec: Used only in no cp scheduling mode and identifies to micro engine id * that the queue should be executed on. * * @pipe: Used only in no cp scheduling mode and identifies the queue's pipe * id. * * @queue: Used only in no cp scheduliong mode and identifies the queue's slot. * * @process: The kfd process that created this queue. * * @device: The kfd device that created this queue. * * @gws: Pointing to gws kgd_mem if this is a gws control queue; NULL * otherwise. * * This structure represents user mode compute queues. * It contains all the necessary data to handle such queues. * */ struct queue { … }; enum KFD_MQD_TYPE { … }; enum KFD_PIPE_PRIORITY { … }; struct scheduling_resources { … }; struct process_queue_manager { … }; struct qcm_process_device { … }; /* KFD Memory Eviction */ /* Approx. wait time before attempting to restore evicted BOs */ #define PROCESS_RESTORE_TIME_MS … /* Approx. back off time if restore fails due to lack of memory */ #define PROCESS_BACK_OFF_TIME_MS … /* Approx. time before evicting the process again */ #define PROCESS_ACTIVE_TIME_MS … /* 8 byte handle containing GPU ID in the most significant 4 bytes and * idr_handle in the least significant 4 bytes */ #define MAKE_HANDLE(gpu_id, idr_handle) … #define GET_GPU_ID(handle) … #define GET_IDR_HANDLE(handle) … enum kfd_pdd_bound { … }; #define MAX_SYSFS_FILENAME_LEN … /* * SDMA counter runs at 100MHz frequency. * We display SDMA activity in microsecond granularity in sysfs. * As a result, the divisor is 100. */ #define SDMA_ACTIVITY_DIVISOR … /* Data that is per-process-per device. */ struct kfd_process_device { … }; #define qpd_to_pdd(x) … struct svm_range_list { … }; /* Process data */ struct kfd_process { … }; #define KFD_PROCESS_TABLE_SIZE … extern DECLARE_HASHTABLE(kfd_processes_table, KFD_PROCESS_TABLE_SIZE); extern struct srcu_struct kfd_processes_srcu; /** * typedef amdkfd_ioctl_t - typedef for ioctl function pointer. * * @filep: pointer to file structure. * @p: amdkfd process pointer. * @data: pointer to arg that was copied from user. * * Return: returns ioctl completion code. */ amdkfd_ioctl_t; struct amdkfd_ioctl_desc { … }; bool kfd_dev_is_large_bar(struct kfd_node *dev); int kfd_process_create_wq(void); void kfd_process_destroy_wq(void); void kfd_cleanup_processes(void); struct kfd_process *kfd_create_process(struct task_struct *thread); struct kfd_process *kfd_get_process(const struct task_struct *task); struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid); struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm); int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id); int kfd_process_gpuid_from_node(struct kfd_process *p, struct kfd_node *node, uint32_t *gpuid, uint32_t *gpuidx); static inline int kfd_process_gpuid_from_gpuidx(struct kfd_process *p, uint32_t gpuidx, uint32_t *gpuid) { … } static inline struct kfd_process_device *kfd_process_device_from_gpuidx( struct kfd_process *p, uint32_t gpuidx) { … } void kfd_unref_process(struct kfd_process *p); int kfd_process_evict_queues(struct kfd_process *p, uint32_t trigger); int kfd_process_restore_queues(struct kfd_process *p); void kfd_suspend_all_processes(void); int kfd_resume_all_processes(void); struct kfd_process_device *kfd_process_device_data_by_id(struct kfd_process *process, uint32_t gpu_id); int kfd_process_get_user_gpu_id(struct kfd_process *p, uint32_t actual_gpu_id); int kfd_process_device_init_vm(struct kfd_process_device *pdd, struct file *drm_file); struct kfd_process_device *kfd_bind_process_to_device(struct kfd_node *dev, struct kfd_process *p); struct kfd_process_device *kfd_get_process_device_data(struct kfd_node *dev, struct kfd_process *p); struct kfd_process_device *kfd_create_process_device_data(struct kfd_node *dev, struct kfd_process *p); bool kfd_process_xnack_mode(struct kfd_process *p, bool supported); int kfd_reserved_mem_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma); /* KFD process API for creating and translating handles */ int kfd_process_device_create_obj_handle(struct kfd_process_device *pdd, void *mem); void *kfd_process_device_translate_handle(struct kfd_process_device *p, int handle); void kfd_process_device_remove_obj_handle(struct kfd_process_device *pdd, int handle); struct kfd_process *kfd_lookup_process_by_pid(struct pid *pid); /* PASIDs */ int kfd_pasid_init(void); void kfd_pasid_exit(void); bool kfd_set_pasid_limit(unsigned int new_limit); unsigned int kfd_get_pasid_limit(void); u32 kfd_pasid_alloc(void); void kfd_pasid_free(u32 pasid); /* Doorbells */ size_t kfd_doorbell_process_slice(struct kfd_dev *kfd); int kfd_doorbell_init(struct kfd_dev *kfd); void kfd_doorbell_fini(struct kfd_dev *kfd); int kfd_doorbell_mmap(struct kfd_node *dev, struct kfd_process *process, struct vm_area_struct *vma); void __iomem *kfd_get_kernel_doorbell(struct kfd_dev *kfd, unsigned int *doorbell_off); void kfd_release_kernel_doorbell(struct kfd_dev *kfd, u32 __iomem *db_addr); u32 read_kernel_doorbell(u32 __iomem *db); void write_kernel_doorbell(void __iomem *db, u32 value); void write_kernel_doorbell64(void __iomem *db, u64 value); unsigned int kfd_get_doorbell_dw_offset_in_bar(struct kfd_dev *kfd, struct kfd_process_device *pdd, unsigned int doorbell_id); phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd); int kfd_alloc_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd); void kfd_free_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd); /* GTT Sub-Allocator */ int kfd_gtt_sa_allocate(struct kfd_node *node, unsigned int size, struct kfd_mem_obj **mem_obj); int kfd_gtt_sa_free(struct kfd_node *node, struct kfd_mem_obj *mem_obj); extern struct device *kfd_device; /* KFD's procfs */ void kfd_procfs_init(void); void kfd_procfs_shutdown(void); int kfd_procfs_add_queue(struct queue *q); void kfd_procfs_del_queue(struct queue *q); /* Topology */ int kfd_topology_init(void); void kfd_topology_shutdown(void); int kfd_topology_add_device(struct kfd_node *gpu); int kfd_topology_remove_device(struct kfd_node *gpu); struct kfd_topology_device *kfd_topology_device_by_proximity_domain( uint32_t proximity_domain); struct kfd_topology_device *kfd_topology_device_by_proximity_domain_no_lock( uint32_t proximity_domain); struct kfd_topology_device *kfd_topology_device_by_id(uint32_t gpu_id); struct kfd_node *kfd_device_by_id(uint32_t gpu_id); struct kfd_node *kfd_device_by_pci_dev(const struct pci_dev *pdev); static inline bool kfd_irq_is_from_node(struct kfd_node *node, uint32_t node_id, uint32_t vmid) { … } static inline struct kfd_node *kfd_node_by_irq_ids(struct amdgpu_device *adev, uint32_t node_id, uint32_t vmid) { … } int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_node **kdev); int kfd_numa_node_to_apic_id(int numa_node_id); /* Interrupts */ #define KFD_IRQ_FENCE_CLIENTID … #define KFD_IRQ_FENCE_SOURCEID … #define KFD_IRQ_IS_FENCE(client, source) … int kfd_interrupt_init(struct kfd_node *dev); void kfd_interrupt_exit(struct kfd_node *dev); bool enqueue_ih_ring_entry(struct kfd_node *kfd, const void *ih_ring_entry); bool interrupt_is_wanted(struct kfd_node *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *flag); int kfd_process_drain_interrupts(struct kfd_process_device *pdd); void kfd_process_close_interrupt_drain(unsigned int pasid); /* amdkfd Apertures */ int kfd_init_apertures(struct kfd_process *process); void kfd_process_set_trap_handler(struct qcm_process_device *qpd, uint64_t tba_addr, uint64_t tma_addr); void kfd_process_set_trap_debug_flag(struct qcm_process_device *qpd, bool enabled); /* CWSR initialization */ int kfd_process_init_cwsr_apu(struct kfd_process *process, struct file *filep); /* CRIU */ /* * Need to increment KFD_CRIU_PRIV_VERSION each time a change is made to any of the CRIU private * structures: * kfd_criu_process_priv_data * kfd_criu_device_priv_data * kfd_criu_bo_priv_data * kfd_criu_queue_priv_data * kfd_criu_event_priv_data * kfd_criu_svm_range_priv_data */ #define KFD_CRIU_PRIV_VERSION … struct kfd_criu_process_priv_data { … }; struct kfd_criu_device_priv_data { … }; struct kfd_criu_bo_priv_data { … }; /* * The first 4 bytes of kfd_criu_queue_priv_data, kfd_criu_event_priv_data, * kfd_criu_svm_range_priv_data is the object type */ enum kfd_criu_object_type { … }; struct kfd_criu_svm_range_priv_data { … }; struct kfd_criu_queue_priv_data { … }; struct kfd_criu_event_priv_data { … }; int kfd_process_get_queue_info(struct kfd_process *p, uint32_t *num_queues, uint64_t *priv_data_sizes); int kfd_criu_checkpoint_queues(struct kfd_process *p, uint8_t __user *user_priv_data, uint64_t *priv_data_offset); int kfd_criu_restore_queue(struct kfd_process *p, uint8_t __user *user_priv_data, uint64_t *priv_data_offset, uint64_t max_priv_data_size); int kfd_criu_checkpoint_events(struct kfd_process *p, uint8_t __user *user_priv_data, uint64_t *priv_data_offset); int kfd_criu_restore_event(struct file *devkfd, struct kfd_process *p, uint8_t __user *user_priv_data, uint64_t *priv_data_offset, uint64_t max_priv_data_size); /* CRIU - End */ /* Queue Context Management */ int init_queue(struct queue **q, const struct queue_properties *properties); void uninit_queue(struct queue *q); void print_queue_properties(struct queue_properties *q); void print_queue(struct queue *q); struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v9(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v10(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v11(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct mqd_manager *mqd_manager_init_v12(enum KFD_MQD_TYPE type, struct kfd_node *dev); struct device_queue_manager *device_queue_manager_init(struct kfd_node *dev); void device_queue_manager_uninit(struct device_queue_manager *dqm); struct kernel_queue *kernel_queue_init(struct kfd_node *dev, enum kfd_queue_type type); void kernel_queue_uninit(struct kernel_queue *kq); int kfd_dqm_evict_pasid(struct device_queue_manager *dqm, u32 pasid); /* Process Queue Manager */ struct process_queue_node { … }; void kfd_process_dequeue_from_device(struct kfd_process_device *pdd); void kfd_process_dequeue_from_all_devices(struct kfd_process *p); int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p); void pqm_uninit(struct process_queue_manager *pqm); int pqm_create_queue(struct process_queue_manager *pqm, struct kfd_node *dev, struct file *f, struct queue_properties *properties, unsigned int *qid, struct amdgpu_bo *wptr_bo, const struct kfd_criu_queue_priv_data *q_data, const void *restore_mqd, const void *restore_ctl_stack, uint32_t *p_doorbell_offset_in_process); int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid); int pqm_update_queue_properties(struct process_queue_manager *pqm, unsigned int qid, struct queue_properties *p); int pqm_update_mqd(struct process_queue_manager *pqm, unsigned int qid, struct mqd_update_info *minfo); int pqm_set_gws(struct process_queue_manager *pqm, unsigned int qid, void *gws); struct kernel_queue *pqm_get_kernel_queue(struct process_queue_manager *pqm, unsigned int qid); struct queue *pqm_get_user_queue(struct process_queue_manager *pqm, unsigned int qid); int pqm_get_wave_state(struct process_queue_manager *pqm, unsigned int qid, void __user *ctl_stack, u32 *ctl_stack_used_size, u32 *save_area_used_size); int pqm_get_queue_snapshot(struct process_queue_manager *pqm, uint64_t exception_clear_mask, void __user *buf, int *num_qss_entries, uint32_t *entry_size); int amdkfd_fence_wait_timeout(struct device_queue_manager *dqm, uint64_t fence_value, unsigned int timeout_ms); int pqm_get_queue_checkpoint_info(struct process_queue_manager *pqm, unsigned int qid, u32 *mqd_size, u32 *ctl_stack_size); /* Packet Manager */ #define KFD_FENCE_COMPLETED … #define KFD_FENCE_INIT … struct packet_manager { … }; struct packet_manager_funcs { … }; extern const struct packet_manager_funcs kfd_vi_pm_funcs; extern const struct packet_manager_funcs kfd_v9_pm_funcs; extern const struct packet_manager_funcs kfd_aldebaran_pm_funcs; int pm_init(struct packet_manager *pm, struct device_queue_manager *dqm); void pm_uninit(struct packet_manager *pm); int pm_send_set_resources(struct packet_manager *pm, struct scheduling_resources *res); int pm_send_runlist(struct packet_manager *pm, struct list_head *dqm_queues); int pm_send_query_status(struct packet_manager *pm, uint64_t fence_address, uint64_t fence_value); int pm_send_unmap_queue(struct packet_manager *pm, enum kfd_unmap_queues_filter mode, uint32_t filter_param, bool reset); void pm_release_ib(struct packet_manager *pm); int pm_update_grace_period(struct packet_manager *pm, uint32_t grace_period); /* Following PM funcs can be shared among VI and AI */ unsigned int pm_build_pm4_header(unsigned int opcode, size_t packet_size); uint64_t kfd_get_number_elems(struct kfd_dev *kfd); /* Events */ extern const struct kfd_event_interrupt_class event_interrupt_class_cik; extern const struct kfd_event_interrupt_class event_interrupt_class_v9; extern const struct kfd_event_interrupt_class event_interrupt_class_v9_4_3; extern const struct kfd_event_interrupt_class event_interrupt_class_v10; extern const struct kfd_event_interrupt_class event_interrupt_class_v11; extern const struct kfd_device_global_init_class device_global_init_class_cik; int kfd_event_init_process(struct kfd_process *p); void kfd_event_free_process(struct kfd_process *p); int kfd_event_mmap(struct kfd_process *process, struct vm_area_struct *vma); int kfd_wait_on_events(struct kfd_process *p, uint32_t num_events, void __user *data, bool all, uint32_t *user_timeout_ms, uint32_t *wait_result); void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id, uint32_t valid_id_bits); void kfd_signal_hw_exception_event(u32 pasid); int kfd_set_event(struct kfd_process *p, uint32_t event_id); int kfd_reset_event(struct kfd_process *p, uint32_t event_id); int kfd_kmap_event_page(struct kfd_process *p, uint64_t event_page_offset); int kfd_event_create(struct file *devkfd, struct kfd_process *p, uint32_t event_type, bool auto_reset, uint32_t node_id, uint32_t *event_id, uint32_t *event_trigger_data, uint64_t *event_page_offset, uint32_t *event_slot_index); int kfd_get_num_events(struct kfd_process *p); int kfd_event_destroy(struct kfd_process *p, uint32_t event_id); void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid, struct kfd_vm_fault_info *info, struct kfd_hsa_memory_exception_data *data); void kfd_signal_reset_event(struct kfd_node *dev); void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid); static inline void kfd_flush_tlb(struct kfd_process_device *pdd, enum TLB_FLUSH_TYPE type) { … } static inline bool kfd_flush_tlb_after_unmap(struct kfd_dev *dev) { … } int kfd_send_exception_to_runtime(struct kfd_process *p, unsigned int queue_id, uint64_t error_reason); bool kfd_is_locked(void); /* Compute profile */ void kfd_inc_compute_active(struct kfd_node *dev); void kfd_dec_compute_active(struct kfd_node *dev); /* Cgroup Support */ /* Check with device cgroup if @kfd device is accessible */ static inline int kfd_devcgroup_check_permission(struct kfd_node *node) { … } static inline bool kfd_is_first_node(struct kfd_node *node) { … } /* Debugfs */ #if defined(CONFIG_DEBUG_FS) void kfd_debugfs_init(void); void kfd_debugfs_fini(void); int kfd_debugfs_mqds_by_process(struct seq_file *m, void *data); int pqm_debugfs_mqds(struct seq_file *m, void *data); int kfd_debugfs_hqds_by_device(struct seq_file *m, void *data); int dqm_debugfs_hqds(struct seq_file *m, void *data); int kfd_debugfs_rls_by_device(struct seq_file *m, void *data); int pm_debugfs_runlist(struct seq_file *m, void *data); int kfd_debugfs_hang_hws(struct kfd_node *dev); int pm_debugfs_hang_hws(struct packet_manager *pm); int dqm_debugfs_hang_hws(struct device_queue_manager *dqm); #else static inline void kfd_debugfs_init(void) {} static inline void kfd_debugfs_fini(void) {} #endif #endif
Codebase Browser
linux