1 /*
2 * Copyright 2013 Advanced Micro Devices, Inc.
3 * All Rights Reserved.
4 *
5 * Permission is hereby granted, free of charge, to any person obtaining a
6 * copy of this software and associated documentation files (the
7 * "Software"), to deal in the Software without restriction, including
8 * without limitation the rights to use, copy, modify, merge, publish,
9 * distribute, sub license, and/or sell copies of the Software, and to
10 * permit persons to whom the Software is furnished to do so, subject to
11 * the following conditions:
12 *
13 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
14 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
15 * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
16 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
17 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
18 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
19 * USE OR OTHER DEALINGS IN THE SOFTWARE.
20 *
21 * The above copyright notice and this permission notice (including the
22 * next paragraph) shall be included in all copies or substantial portions
23 * of the Software.
24 *
25 * Authors: Christian König <christian.koenig@amd.com>
26 */
27
28 #include <linux/firmware.h>
29
30 #include "radeon.h"
31 #include "radeon_asic.h"
32 #include "cikd.h"
33 #include "vce.h"
34
35 #define VCE_V2_0_FW_SIZE (256 * 1024)
36 #define VCE_V2_0_STACK_SIZE (64 * 1024)
37 #define VCE_V2_0_DATA_SIZE (23552 * RADEON_MAX_VCE_HANDLES)
38
vce_v2_0_set_sw_cg(struct radeon_device * rdev,bool gated)39 static void vce_v2_0_set_sw_cg(struct radeon_device *rdev, bool gated)
40 {
41 u32 tmp;
42
43 if (gated) {
44 tmp = RREG32(VCE_CLOCK_GATING_B);
45 tmp |= 0xe70000;
46 WREG32(VCE_CLOCK_GATING_B, tmp);
47
48 tmp = RREG32(VCE_UENC_CLOCK_GATING);
49 tmp |= 0xff000000;
50 WREG32(VCE_UENC_CLOCK_GATING, tmp);
51
52 tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
53 tmp &= ~0x3fc;
54 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
55
56 WREG32(VCE_CGTT_CLK_OVERRIDE, 0);
57 } else {
58 tmp = RREG32(VCE_CLOCK_GATING_B);
59 tmp |= 0xe7;
60 tmp &= ~0xe70000;
61 WREG32(VCE_CLOCK_GATING_B, tmp);
62
63 tmp = RREG32(VCE_UENC_CLOCK_GATING);
64 tmp |= 0x1fe000;
65 tmp &= ~0xff000000;
66 WREG32(VCE_UENC_CLOCK_GATING, tmp);
67
68 tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
69 tmp |= 0x3fc;
70 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
71 }
72 }
73
vce_v2_0_set_dyn_cg(struct radeon_device * rdev,bool gated)74 static void vce_v2_0_set_dyn_cg(struct radeon_device *rdev, bool gated)
75 {
76 u32 orig, tmp;
77
78 tmp = RREG32(VCE_CLOCK_GATING_B);
79 tmp &= ~0x00060006;
80 if (gated) {
81 tmp |= 0xe10000;
82 } else {
83 tmp |= 0xe1;
84 tmp &= ~0xe10000;
85 }
86 WREG32(VCE_CLOCK_GATING_B, tmp);
87
88 orig = tmp = RREG32(VCE_UENC_CLOCK_GATING);
89 tmp &= ~0x1fe000;
90 tmp &= ~0xff000000;
91 if (tmp != orig)
92 WREG32(VCE_UENC_CLOCK_GATING, tmp);
93
94 orig = tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
95 tmp &= ~0x3fc;
96 if (tmp != orig)
97 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
98
99 if (gated)
100 WREG32(VCE_CGTT_CLK_OVERRIDE, 0);
101 }
102
vce_v2_0_disable_cg(struct radeon_device * rdev)103 static void vce_v2_0_disable_cg(struct radeon_device *rdev)
104 {
105 WREG32(VCE_CGTT_CLK_OVERRIDE, 7);
106 }
107
108 /*
109 * Local variable sw_cg is used for debugging purposes, in case we
110 * ran into problems with dynamic clock gating. Don't remove it.
111 */
vce_v2_0_enable_mgcg(struct radeon_device * rdev,bool enable)112 void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable)
113 {
114 bool sw_cg = false;
115
116 if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_VCE_MGCG)) {
117 if (sw_cg)
118 vce_v2_0_set_sw_cg(rdev, true);
119 else
120 vce_v2_0_set_dyn_cg(rdev, true);
121 } else {
122 vce_v2_0_disable_cg(rdev);
123
124 if (sw_cg)
125 vce_v2_0_set_sw_cg(rdev, false);
126 else
127 vce_v2_0_set_dyn_cg(rdev, false);
128 }
129 }
130
vce_v2_0_init_cg(struct radeon_device * rdev)131 static void vce_v2_0_init_cg(struct radeon_device *rdev)
132 {
133 u32 tmp;
134
135 tmp = RREG32(VCE_CLOCK_GATING_A);
136 tmp &= ~(CGC_CLK_GATE_DLY_TIMER_MASK | CGC_CLK_GATER_OFF_DLY_TIMER_MASK);
137 tmp |= (CGC_CLK_GATE_DLY_TIMER(0) | CGC_CLK_GATER_OFF_DLY_TIMER(4));
138 tmp |= CGC_UENC_WAIT_AWAKE;
139 WREG32(VCE_CLOCK_GATING_A, tmp);
140
141 tmp = RREG32(VCE_UENC_CLOCK_GATING);
142 tmp &= ~(CLOCK_ON_DELAY_MASK | CLOCK_OFF_DELAY_MASK);
143 tmp |= (CLOCK_ON_DELAY(0) | CLOCK_OFF_DELAY(4));
144 WREG32(VCE_UENC_CLOCK_GATING, tmp);
145
146 tmp = RREG32(VCE_CLOCK_GATING_B);
147 tmp |= 0x10;
148 tmp &= ~0x100000;
149 WREG32(VCE_CLOCK_GATING_B, tmp);
150 }
151
vce_v2_0_bo_size(struct radeon_device * rdev)152 unsigned vce_v2_0_bo_size(struct radeon_device *rdev)
153 {
154 WARN_ON(rdev->vce_fw->size > VCE_V2_0_FW_SIZE);
155 return VCE_V2_0_FW_SIZE + VCE_V2_0_STACK_SIZE + VCE_V2_0_DATA_SIZE;
156 }
157
vce_v2_0_resume(struct radeon_device * rdev)158 int vce_v2_0_resume(struct radeon_device *rdev)
159 {
160 uint64_t addr = rdev->vce.gpu_addr;
161 uint32_t size;
162
163 WREG32_P(VCE_CLOCK_GATING_A, 0, ~(1 << 16));
164 WREG32_P(VCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
165 WREG32_P(VCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
166 WREG32(VCE_CLOCK_GATING_B, 0xf7);
167
168 WREG32(VCE_LMI_CTRL, 0x00398000);
169 WREG32_P(VCE_LMI_CACHE_CTRL, 0x0, ~0x1);
170 WREG32(VCE_LMI_SWAP_CNTL, 0);
171 WREG32(VCE_LMI_SWAP_CNTL1, 0);
172 WREG32(VCE_LMI_VM_CTRL, 0);
173
174 WREG32(VCE_LMI_VCPU_CACHE_40BIT_BAR, addr >> 8);
175
176 addr &= 0xff;
177 size = VCE_V2_0_FW_SIZE;
178 WREG32(VCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff);
179 WREG32(VCE_VCPU_CACHE_SIZE0, size);
180
181 addr += size;
182 size = VCE_V2_0_STACK_SIZE;
183 WREG32(VCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff);
184 WREG32(VCE_VCPU_CACHE_SIZE1, size);
185
186 addr += size;
187 size = VCE_V2_0_DATA_SIZE;
188 WREG32(VCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff);
189 WREG32(VCE_VCPU_CACHE_SIZE2, size);
190
191 WREG32_P(VCE_LMI_CTRL2, 0x0, ~0x100);
192
193 WREG32_P(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN,
194 ~VCE_SYS_INT_TRAP_INTERRUPT_EN);
195
196 vce_v2_0_init_cg(rdev);
197
198 return 0;
199 }
200