kfd_debug.h

  1/*
  2 * Copyright 2023 Advanced Micro Devices, Inc.
  3 *
  4 * Permission is hereby granted, free of charge, to any person obtaining a
  5 * copy of this software and associated documentation files (the "Software"),
  6 * to deal in the Software without restriction, including without limitation
  7 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  8 * and/or sell copies of the Software, and to permit persons to whom the
  9 * Software is furnished to do so, subject to the following conditions:
 10 *
 11 * The above copyright notice and this permission notice shall be included in
 12 * all copies or substantial portions of the Software.
 13 *
 14 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 15 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 16 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 17 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 18 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 19 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 20 * OTHER DEALINGS IN THE SOFTWARE.
 21 */
 22
 23#ifndef KFD_DEBUG_EVENTS_H_INCLUDED
 24#define KFD_DEBUG_EVENTS_H_INCLUDED
 25
 26#include "kfd_priv.h"
 27
 28void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind_count);
 29int kfd_dbg_trap_activate(struct kfd_process *target);
 30int kfd_dbg_ev_query_debug_event(struct kfd_process *process,
 31			unsigned int *queue_id,
 32			unsigned int *gpu_id,
 33			uint64_t exception_clear_mask,
 34			uint64_t *event_status);
 35bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev,
 36				   unsigned int pasid,
 37				   uint32_t doorbell_id,
 38				   uint64_t trap_mask,
 39				   void *exception_data,
 40				   size_t exception_data_size);
 41bool kfd_dbg_ev_raise(uint64_t event_mask,
 42			struct kfd_process *process, struct kfd_node *dev,
 43			unsigned int source_id, bool use_worker,
 44			void *exception_data,
 45			size_t exception_data_size);
 46int kfd_dbg_trap_disable(struct kfd_process *target);
 47int kfd_dbg_trap_enable(struct kfd_process *target, uint32_t fd,
 48			void __user *runtime_info,
 49			uint32_t *runtime_info_size);
 50int kfd_dbg_trap_set_wave_launch_override(struct kfd_process *target,
 51					uint32_t trap_override,
 52					uint32_t trap_mask_bits,
 53					uint32_t trap_mask_request,
 54					uint32_t *trap_mask_prev,
 55					uint32_t *trap_mask_supported);
 56int kfd_dbg_trap_set_wave_launch_mode(struct kfd_process *target,
 57					uint8_t wave_launch_mode);
 58int kfd_dbg_trap_clear_dev_address_watch(struct kfd_process_device *pdd,
 59					uint32_t watch_id);
 60int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd,
 61					uint64_t watch_address,
 62					uint32_t watch_address_mask,
 63					uint32_t *watch_id,
 64					uint32_t watch_mode);
 65int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags);
 66int kfd_dbg_trap_query_exception_info(struct kfd_process *target,
 67		uint32_t source_id,
 68		uint32_t exception_code,
 69		bool clear_exception,
 70		void __user *info,
 71		uint32_t *info_size);
 72int kfd_dbg_send_exception_to_runtime(struct kfd_process *p,
 73					unsigned int dev_id,
 74					unsigned int queue_id,
 75					uint64_t error_reason);
 76
 77static inline bool kfd_dbg_is_per_vmid_supported(struct kfd_node *dev)
 78{
 79	return (KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 2) ||
 80		KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 3) ||
 81		KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 4) ||
 82		KFD_GC_VERSION(dev) == IP_VERSION(9, 5, 0) ||
 83		KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0));
 84}
 85
 86void debug_event_write_work_handler(struct work_struct *work);
 87int kfd_dbg_trap_device_snapshot(struct kfd_process *target,
 88		uint64_t exception_clear_mask,
 89		void __user *user_info,
 90		uint32_t *number_of_device_infos,
 91		uint32_t *entry_size);
 92
 93void kfd_dbg_set_enabled_debug_exception_mask(struct kfd_process *target,
 94					uint64_t exception_set_mask);
 95/*
 96 * If GFX off is enabled, chips that do not support RLC restore for the debug
 97 * registers will disable GFX off temporarily for the entire debug session.
 98 * See disable_on_trap_action_entry and enable_on_trap_action_exit for details.
 99 */
100static inline bool kfd_dbg_is_rlc_restore_supported(struct kfd_node *dev)
101{
102	return !(KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 10) ||
103		 KFD_GC_VERSION(dev) == IP_VERSION(10, 1, 1));
104}
105
106static inline bool kfd_dbg_has_cwsr_workaround(struct kfd_node *dev)
107{
108	return KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) &&
109	       KFD_GC_VERSION(dev) <= IP_VERSION(11, 0, 3);
110}
111
112static inline bool kfd_dbg_has_gws_support(struct kfd_node *dev)
113{
114	if ((KFD_GC_VERSION(dev) == IP_VERSION(9, 0, 1)
115			&& dev->kfd->mec2_fw_version < 0x81b6) ||
116		(KFD_GC_VERSION(dev) >= IP_VERSION(9, 1, 0)
117			&& KFD_GC_VERSION(dev) <= IP_VERSION(9, 2, 2)
118			&& dev->kfd->mec2_fw_version < 0x1b6) ||
119		(KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 0)
120			&& dev->kfd->mec2_fw_version < 0x1b6) ||
121		(KFD_GC_VERSION(dev) == IP_VERSION(9, 4, 1)
122			&& dev->kfd->mec2_fw_version < 0x30) ||
123		kfd_dbg_has_cwsr_workaround(dev))
124		return false;
125
126	/* Assume debugging and cooperative launch supported otherwise. */
127	return true;
128}
129
130int kfd_dbg_set_mes_debug_mode(struct kfd_process_device *pdd, bool sq_trap_en);
131
132static inline bool kfd_dbg_has_ttmps_always_setup(struct kfd_node *dev)
133{
134	return (KFD_GC_VERSION(dev) < IP_VERSION(11, 0, 0) &&
135			KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 2)) ||
136	       (KFD_GC_VERSION(dev) >= IP_VERSION(11, 0, 0) &&
137			KFD_GC_VERSION(dev) < IP_VERSION(12, 0, 0) &&
138			(dev->adev->mes.sched_version & AMDGPU_MES_VERSION_MASK) >= 70) ||
139	       (KFD_GC_VERSION(dev) >= IP_VERSION(12, 0, 0));
140}
141#endif