xref: /linux-6.15/drivers/gpu/drm/amd/amdgpu/mca_v3_0.c (revision 20c43547)
1 /*
2  * Copyright 2021 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 #include "amdgpu_ras.h"
24 #include "amdgpu.h"
25 #include "amdgpu_mca.h"
26 
27 #define smnMCMP0_STATUST0 	0x03830408
28 #define smnMCMP1_STATUST0 	0x03b30408
29 #define smnMCMPIO_STATUST0 	0x0c930408
30 
31 
32 static void mca_v3_0_mp0_query_ras_error_count(struct amdgpu_device *adev,
33 					       void *ras_error_status)
34 {
35 	amdgpu_mca_query_ras_error_count(adev,
36 				         smnMCMP0_STATUST0,
37 				         ras_error_status);
38 }
39 
40 static void mca_v3_0_mp0_ras_fini(struct amdgpu_device *adev)
41 {
42 	amdgpu_mca_ras_fini(adev, &adev->mca.mp0);
43 }
44 
45 static int mca_v3_0_ras_block_match(struct amdgpu_ras_block_object *block_obj,
46 				enum amdgpu_ras_block block, uint32_t sub_block_index)
47 {
48 	if (!block_obj)
49 		return -EINVAL;
50 
51 	if ((block_obj->ras_comm.block == block) &&
52 		(block_obj->ras_comm.sub_block_index == sub_block_index)) {
53 		return 0;
54 	}
55 
56 	return -EINVAL;
57 }
58 
59 const struct amdgpu_ras_block_hw_ops mca_v3_0_mp0_hw_ops = {
60 	.query_ras_error_count = mca_v3_0_mp0_query_ras_error_count,
61 	.query_ras_error_address = NULL,
62 };
63 
64 struct amdgpu_mca_ras_block mca_v3_0_mp0_ras = {
65 	.ras_block = {
66 		.ras_comm = {
67 			.block = AMDGPU_RAS_BLOCK__MCA,
68 			.sub_block_index = AMDGPU_RAS_MCA_BLOCK__MP0,
69 			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
70 			.name = "mp0",
71 		},
72 		.hw_ops = &mca_v3_0_mp0_hw_ops,
73 		.ras_block_match = mca_v3_0_ras_block_match,
74 		.ras_late_init = amdgpu_ras_block_late_init,
75 		.ras_fini = mca_v3_0_mp0_ras_fini,
76 	},
77 };
78 
79 static void mca_v3_0_mp1_query_ras_error_count(struct amdgpu_device *adev,
80 					       void *ras_error_status)
81 {
82 	amdgpu_mca_query_ras_error_count(adev,
83 				         smnMCMP1_STATUST0,
84 				         ras_error_status);
85 }
86 
87 static void mca_v3_0_mp1_ras_fini(struct amdgpu_device *adev)
88 {
89 	amdgpu_mca_ras_fini(adev, &adev->mca.mp1);
90 }
91 
92 const struct amdgpu_ras_block_hw_ops mca_v3_0_mp1_hw_ops = {
93 	.query_ras_error_count = mca_v3_0_mp1_query_ras_error_count,
94 	.query_ras_error_address = NULL,
95 };
96 
97 struct amdgpu_mca_ras_block mca_v3_0_mp1_ras = {
98 	.ras_block = {
99 		.ras_comm = {
100 			.block = AMDGPU_RAS_BLOCK__MCA,
101 			.sub_block_index = AMDGPU_RAS_MCA_BLOCK__MP1,
102 			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
103 			.name = "mp1",
104 		},
105 		.hw_ops = &mca_v3_0_mp1_hw_ops,
106 		.ras_block_match = mca_v3_0_ras_block_match,
107 		.ras_late_init = amdgpu_ras_block_late_init,
108 		.ras_fini = mca_v3_0_mp1_ras_fini,
109 	},
110 };
111 
112 static void mca_v3_0_mpio_query_ras_error_count(struct amdgpu_device *adev,
113 					       void *ras_error_status)
114 {
115 	amdgpu_mca_query_ras_error_count(adev,
116 				         smnMCMPIO_STATUST0,
117 				         ras_error_status);
118 }
119 
120 static void mca_v3_0_mpio_ras_fini(struct amdgpu_device *adev)
121 {
122 	amdgpu_mca_ras_fini(adev, &adev->mca.mpio);
123 }
124 
125 const struct amdgpu_ras_block_hw_ops mca_v3_0_mpio_hw_ops = {
126 	.query_ras_error_count = mca_v3_0_mpio_query_ras_error_count,
127 	.query_ras_error_address = NULL,
128 };
129 
130 struct amdgpu_mca_ras_block mca_v3_0_mpio_ras = {
131 	.ras_block = {
132 		.ras_comm = {
133 			.block = AMDGPU_RAS_BLOCK__MCA,
134 			.sub_block_index = AMDGPU_RAS_MCA_BLOCK__MPIO,
135 			.type = AMDGPU_RAS_ERROR__MULTI_UNCORRECTABLE,
136 			.name = "mpio",
137 		},
138 		.hw_ops = &mca_v3_0_mpio_hw_ops,
139 		.ras_block_match = mca_v3_0_ras_block_match,
140 		.ras_late_init = amdgpu_ras_block_late_init,
141 		.ras_fini = mca_v3_0_mpio_ras_fini,
142 	},
143 };
144 
145 
146 static void mca_v3_0_init(struct amdgpu_device *adev)
147 {
148 	struct amdgpu_mca *mca = &adev->mca;
149 
150 	mca->mp0.ras = &mca_v3_0_mp0_ras;
151 	mca->mp1.ras = &mca_v3_0_mp1_ras;
152 	mca->mpio.ras = &mca_v3_0_mpio_ras;
153 	amdgpu_ras_register_ras_block(adev, &mca->mp0.ras->ras_block);
154 	amdgpu_ras_register_ras_block(adev, &mca->mp1.ras->ras_block);
155 	amdgpu_ras_register_ras_block(adev, &mca->mpio.ras->ras_block);
156 	mca->mp0.ras_if = &mca->mp0.ras->ras_block.ras_comm;
157 	mca->mp1.ras_if = &mca->mp1.ras->ras_block.ras_comm;
158 	mca->mpio.ras_if = &mca->mpio.ras->ras_block.ras_comm;
159 }
160 
161 const struct amdgpu_mca_funcs mca_v3_0_funcs = {
162 	.init = mca_v3_0_init,
163 };