| 1 | /* |
| 2 | * Copyright 2022 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 | |
| 24 | #ifndef __AMDGPU_RING_MUX__ |
| 25 | #define __AMDGPU_RING_MUX__ |
| 26 | |
| 27 | #include <linux/timer.h> |
| 28 | #include <linux/spinlock.h> |
| 29 | #include "amdgpu_ring.h" |
| 30 | |
| 31 | struct amdgpu_ring; |
| 32 | |
| 33 | /** |
| 34 | * struct amdgpu_mux_entry - the entry recording software rings copying information. |
| 35 | * @ring: the pointer to the software ring. |
| 36 | * @start_ptr_in_hw_ring: last start location copied to in the hardware ring. |
| 37 | * @end_ptr_in_hw_ring: last end location copied to in the hardware ring. |
| 38 | * @sw_cptr: the position of the copy pointer in the sw ring. |
| 39 | * @sw_rptr: the read pointer in software ring. |
| 40 | * @sw_wptr: the write pointer in software ring. |
| 41 | * @list: list head for amdgpu_mux_chunk |
| 42 | */ |
| 43 | struct amdgpu_mux_entry { |
| 44 | struct amdgpu_ring *ring; |
| 45 | u64 start_ptr_in_hw_ring; |
| 46 | u64 end_ptr_in_hw_ring; |
| 47 | u64 sw_cptr; |
| 48 | u64 sw_rptr; |
| 49 | u64 sw_wptr; |
| 50 | struct list_head list; |
| 51 | }; |
| 52 | |
| 53 | enum amdgpu_ring_mux_offset_type { |
| 54 | AMDGPU_MUX_OFFSET_TYPE_CONTROL, |
| 55 | AMDGPU_MUX_OFFSET_TYPE_DE, |
| 56 | AMDGPU_MUX_OFFSET_TYPE_CE, |
| 57 | }; |
| 58 | |
| 59 | enum ib_complete_status { |
| 60 | /* IB not started/reset value, default value. */ |
| 61 | IB_COMPLETION_STATUS_DEFAULT = 0, |
| 62 | /* IB preempted, started but not completed. */ |
| 63 | IB_COMPLETION_STATUS_PREEMPTED = 1, |
| 64 | /* IB completed. */ |
| 65 | IB_COMPLETION_STATUS_COMPLETED = 2, |
| 66 | }; |
| 67 | |
| 68 | struct amdgpu_ring_mux { |
| 69 | struct amdgpu_ring *real_ring; |
| 70 | |
| 71 | struct amdgpu_mux_entry *ring_entry; |
| 72 | unsigned int num_ring_entries; |
| 73 | unsigned int ring_entry_size; |
| 74 | /*the lock for copy data from different software rings*/ |
| 75 | spinlock_t lock; |
| 76 | bool s_resubmit; |
| 77 | uint32_t seqno_to_resubmit; |
| 78 | u64 wptr_resubmit; |
| 79 | struct timer_list resubmit_timer; |
| 80 | |
| 81 | bool pending_trailing_fence_signaled; |
| 82 | }; |
| 83 | |
| 84 | /** |
| 85 | * struct amdgpu_mux_chunk - save the location of indirect buffer's package on softare rings. |
| 86 | * @entry: the list entry. |
| 87 | * @sync_seq: the fence seqno related with the saved IB. |
| 88 | * @start:- start location on the software ring. |
| 89 | * @end:- end location on the software ring. |
| 90 | * @control_offset:- the PRE_RESUME bit position used for resubmission. |
| 91 | * @de_offset:- the anchor in write_data for de meta of resubmission. |
| 92 | * @ce_offset:- the anchor in write_data for ce meta of resubmission. |
| 93 | */ |
| 94 | struct amdgpu_mux_chunk { |
| 95 | struct list_head entry; |
| 96 | uint32_t sync_seq; |
| 97 | u64 start; |
| 98 | u64 end; |
| 99 | u64 cntl_offset; |
| 100 | u64 de_offset; |
| 101 | u64 ce_offset; |
| 102 | }; |
| 103 | |
| 104 | int amdgpu_ring_mux_init(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, |
| 105 | unsigned int entry_size); |
| 106 | void amdgpu_ring_mux_fini(struct amdgpu_ring_mux *mux); |
| 107 | int amdgpu_ring_mux_add_sw_ring(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); |
| 108 | void amdgpu_ring_mux_set_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, u64 wptr); |
| 109 | u64 amdgpu_ring_mux_get_wptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); |
| 110 | u64 amdgpu_ring_mux_get_rptr(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); |
| 111 | void amdgpu_ring_mux_start_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); |
| 112 | void amdgpu_ring_mux_end_ib(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring); |
| 113 | void amdgpu_ring_mux_ib_mark_offset(struct amdgpu_ring_mux *mux, struct amdgpu_ring *ring, |
| 114 | u64 offset, enum amdgpu_ring_mux_offset_type type); |
| 115 | bool amdgpu_mcbp_handle_trailing_fence_irq(struct amdgpu_ring_mux *mux); |
| 116 | |
| 117 | u64 amdgpu_sw_ring_get_rptr_gfx(struct amdgpu_ring *ring); |
| 118 | u64 amdgpu_sw_ring_get_wptr_gfx(struct amdgpu_ring *ring); |
| 119 | void amdgpu_sw_ring_set_wptr_gfx(struct amdgpu_ring *ring); |
| 120 | void amdgpu_sw_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); |
| 121 | void amdgpu_sw_ring_ib_begin(struct amdgpu_ring *ring); |
| 122 | void amdgpu_sw_ring_ib_end(struct amdgpu_ring *ring); |
| 123 | void amdgpu_sw_ring_ib_mark_offset(struct amdgpu_ring *ring, enum amdgpu_ring_mux_offset_type type); |
| 124 | const char *amdgpu_sw_ring_name(int idx); |
| 125 | unsigned int amdgpu_sw_ring_priority(int idx); |
| 126 | |
| 127 | #endif |
| 128 | |