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
34 #define VCE_V2_0_FW_SIZE (256 * 1024)
35 #define VCE_V2_0_STACK_SIZE (64 * 1024)
36 #define VCE_V2_0_DATA_SIZE (23552 * RADEON_MAX_VCE_HANDLES)
37
vce_v2_0_set_sw_cg(struct radeon_device * rdev,bool gated)38 static void vce_v2_0_set_sw_cg(struct radeon_device *rdev, bool gated)
39 {
40 u32 tmp;
41
42 if (gated) {
43 tmp = RREG32(VCE_CLOCK_GATING_B);
44 tmp |= 0xe70000;
45 WREG32(VCE_CLOCK_GATING_B, tmp);
46
47 tmp = RREG32(VCE_UENC_CLOCK_GATING);
48 tmp |= 0xff000000;
49 WREG32(VCE_UENC_CLOCK_GATING, tmp);
50
51 tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
52 tmp &= ~0x3fc;
53 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
54
55 WREG32(VCE_CGTT_CLK_OVERRIDE, 0);
56 } else {
57 tmp = RREG32(VCE_CLOCK_GATING_B);
58 tmp |= 0xe7;
59 tmp &= ~0xe70000;
60 WREG32(VCE_CLOCK_GATING_B, tmp);
61
62 tmp = RREG32(VCE_UENC_CLOCK_GATING);
63 tmp |= 0x1fe000;
64 tmp &= ~0xff000000;
65 WREG32(VCE_UENC_CLOCK_GATING, tmp);
66
67 tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
68 tmp |= 0x3fc;
69 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
70 }
71 }
72
vce_v2_0_set_dyn_cg(struct radeon_device * rdev,bool gated)73 static void vce_v2_0_set_dyn_cg(struct radeon_device *rdev, bool gated)
74 {
75 u32 orig, tmp;
76
77 tmp = RREG32(VCE_CLOCK_GATING_B);
78 tmp &= ~0x00060006;
79 if (gated) {
80 tmp |= 0xe10000;
81 } else {
82 tmp |= 0xe1;
83 tmp &= ~0xe10000;
84 }
85 WREG32(VCE_CLOCK_GATING_B, tmp);
86
87 orig = tmp = RREG32(VCE_UENC_CLOCK_GATING);
88 tmp &= ~0x1fe000;
89 tmp &= ~0xff000000;
90 if (tmp != orig)
91 WREG32(VCE_UENC_CLOCK_GATING, tmp);
92
93 orig = tmp = RREG32(VCE_UENC_REG_CLOCK_GATING);
94 tmp &= ~0x3fc;
95 if (tmp != orig)
96 WREG32(VCE_UENC_REG_CLOCK_GATING, tmp);
97
98 if (gated)
99 WREG32(VCE_CGTT_CLK_OVERRIDE, 0);
100 }
101
vce_v2_0_disable_cg(struct radeon_device * rdev)102 static void vce_v2_0_disable_cg(struct radeon_device *rdev)
103 {
104 WREG32(VCE_CGTT_CLK_OVERRIDE, 7);
105 }
106
107 /*
108 * Local variable sw_cg is used for debugging purposes, in case we
109 * ran into problems with dynamic clock gating. Don't remove it.
110 */
vce_v2_0_enable_mgcg(struct radeon_device * rdev,bool enable)111 void vce_v2_0_enable_mgcg(struct radeon_device *rdev, bool enable)
112 {
113 bool sw_cg = false;
114
115 if (enable && (rdev->cg_flags & RADEON_CG_SUPPORT_VCE_MGCG)) {
116 if (sw_cg)
117 vce_v2_0_set_sw_cg(rdev, true);
118 else
119 vce_v2_0_set_dyn_cg(rdev, true);
120 } else {
121 vce_v2_0_disable_cg(rdev);
122
123 if (sw_cg)
124 vce_v2_0_set_sw_cg(rdev, false);
125 else
126 vce_v2_0_set_dyn_cg(rdev, false);
127 }
128 }
129
vce_v2_0_init_cg(struct radeon_device * rdev)130 static void vce_v2_0_init_cg(struct radeon_device *rdev)
131 {
132 u32 tmp;
133
134 tmp = RREG32(VCE_CLOCK_GATING_A);
135 tmp &= ~(CGC_CLK_GATE_DLY_TIMER_MASK | CGC_CLK_GATER_OFF_DLY_TIMER_MASK);
136 tmp |= (CGC_CLK_GATE_DLY_TIMER(0) | CGC_CLK_GATER_OFF_DLY_TIMER(4));
137 tmp |= CGC_UENC_WAIT_AWAKE;
138 WREG32(VCE_CLOCK_GATING_A, tmp);
139
140 tmp = RREG32(VCE_UENC_CLOCK_GATING);
141 tmp &= ~(CLOCK_ON_DELAY_MASK | CLOCK_OFF_DELAY_MASK);
142 tmp |= (CLOCK_ON_DELAY(0) | CLOCK_OFF_DELAY(4));
143 WREG32(VCE_UENC_CLOCK_GATING, tmp);
144
145 tmp = RREG32(VCE_CLOCK_GATING_B);
146 tmp |= 0x10;
147 tmp &= ~0x100000;
148 WREG32(VCE_CLOCK_GATING_B, tmp);
149 }
150
vce_v2_0_bo_size(struct radeon_device * rdev)151 unsigned vce_v2_0_bo_size(struct radeon_device *rdev)
152 {
153 WARN_ON(rdev->vce_fw->size > VCE_V2_0_FW_SIZE);
154 return VCE_V2_0_FW_SIZE + VCE_V2_0_STACK_SIZE + VCE_V2_0_DATA_SIZE;
155 }
156
vce_v2_0_resume(struct radeon_device * rdev)157 int vce_v2_0_resume(struct radeon_device *rdev)
158 {
159 uint64_t addr = rdev->vce.gpu_addr;
160 uint32_t size;
161
162 WREG32_P(VCE_CLOCK_GATING_A, 0, ~(1 << 16));
163 WREG32_P(VCE_UENC_CLOCK_GATING, 0x1FF000, ~0xFF9FF000);
164 WREG32_P(VCE_UENC_REG_CLOCK_GATING, 0x3F, ~0x3F);
165 WREG32(VCE_CLOCK_GATING_B, 0xf7);
166
167 WREG32(VCE_LMI_CTRL, 0x00398000);
168 WREG32_P(VCE_LMI_CACHE_CTRL, 0x0, ~0x1);
169 WREG32(VCE_LMI_SWAP_CNTL, 0);
170 WREG32(VCE_LMI_SWAP_CNTL1, 0);
171 WREG32(VCE_LMI_VM_CTRL, 0);
172
173 WREG32(VCE_LMI_VCPU_CACHE_40BIT_BAR, addr >> 8);
174
175 addr &= 0xff;
176 size = VCE_V2_0_FW_SIZE;
177 WREG32(VCE_VCPU_CACHE_OFFSET0, addr & 0x7fffffff);
178 WREG32(VCE_VCPU_CACHE_SIZE0, size);
179
180 addr += size;
181 size = VCE_V2_0_STACK_SIZE;
182 WREG32(VCE_VCPU_CACHE_OFFSET1, addr & 0x7fffffff);
183 WREG32(VCE_VCPU_CACHE_SIZE1, size);
184
185 addr += size;
186 size = VCE_V2_0_DATA_SIZE;
187 WREG32(VCE_VCPU_CACHE_OFFSET2, addr & 0x7fffffff);
188 WREG32(VCE_VCPU_CACHE_SIZE2, size);
189
190 WREG32_P(VCE_LMI_CTRL2, 0x0, ~0x100);
191
192 WREG32_P(VCE_SYS_INT_EN, VCE_SYS_INT_TRAP_INTERRUPT_EN,
193 ~VCE_SYS_INT_TRAP_INTERRUPT_EN);
194
195 vce_v2_0_init_cg(rdev);
196
197 return 0;
198 }
199