drm/amdgpu: add initial VCN2.0 support (v2)
authorLeo Liu <leo.liu@amd.com>
Mon, 15 Oct 2018 15:38:59 +0000 (11:38 -0400)
committerAlex Deucher <alexander.deucher@amd.com>
Fri, 21 Jun 2019 23:58:22 +0000 (18:58 -0500)
VCN (Video Core Next) is the video encode/decode block.

Porting over the same functions from VCN1.0

v2: squash in updates (Alex)

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/Makefile
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c [new file with mode: 0644]
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h [new file with mode: 0644]

index 354d84c6fbf8d097c20a7a8e12d7812b708aa93c..1e474598a59a02800a84b82e01de7946e12b3112 100644 (file)
@@ -136,7 +136,8 @@ amdgpu-y += \
 # add VCN block
 amdgpu-y += \
        amdgpu_vcn.o \
-       vcn_v1_0.o
+       vcn_v1_0.o \
+       vcn_v2_0.o
 
 # add ATHUB block
 amdgpu-y += \
index cc94841f2f06e6f80f16277c9b6321e254c2299a..3f6349c6f33d1b3531e70156b086dfbcf6e5966d 100644 (file)
@@ -75,6 +75,7 @@ enum engine_status_constants {
        UVD_STATUS__BUSY = 0x5,
        UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF = 0x1,
        UVD_STATUS__RBC_BUSY = 0x1,
+       UVD_PGFSM_STATUS_UVDJ_PWR_ON = 0,
 };
 
 enum internal_dpg_state {
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c
new file mode 100644 (file)
index 0000000..b5f9afc
--- /dev/null
@@ -0,0 +1,1865 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/firmware.h>
+#include <drm/drmP.h>
+#include "amdgpu.h"
+#include "amdgpu_vcn.h"
+#include "soc15.h"
+#include "soc15d.h"
+
+#include "vcn/vcn_2_0_0_offset.h"
+#include "vcn/vcn_2_0_0_sh_mask.h"
+#include "ivsrcid/vcn/irqsrcs_vcn_2_0.h"
+
+#define mmUVD_CONTEXT_ID_INTERNAL_OFFSET                       0x1fd
+#define mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET                   0x503
+#define mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET                 0x504
+#define mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET                 0x505
+#define mmUVD_NO_OP_INTERNAL_OFFSET                            0x53f
+#define mmUVD_GP_SCRATCH8_INTERNAL_OFFSET                      0x54a
+#define mmUVD_SCRATCH9_INTERNAL_OFFSET                         0xc01d
+
+#define mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET                  0x1e1
+#define mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET        0x5a6
+#define mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET         0x5a7
+#define mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET                      0x1e2
+#define mmUVD_GPCOM_SYS_CMD_INTERNAL_OFFSET                    0x1bF
+
+#define mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET                        0x1bfff
+#define mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET                           0x4029
+#define mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET                         0x402a
+#define mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET                         0x402b
+#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET         0x40ea
+#define mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET        0x40eb
+#define mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET                         0x40cf
+#define mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET                            0x40d1
+#define mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET                0x40e8
+#define mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET               0x40e9
+#define mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET                             0x4082
+#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET         0x40ec
+#define mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET        0x40ed
+#define mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET                    0x4085
+#define mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET                         0x4084
+#define mmUVD_JRBC_STATUS_INTERNAL_OFFSET                              0x4089
+#define mmUVD_JPEG_PITCH_INTERNAL_OFFSET                               0x401f
+
+#define JRBC_DEC_EXTERNAL_REG_WRITE_ADDR                               0x18000
+
+static int vcn_v2_0_stop(struct amdgpu_device *adev);
+static void vcn_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v2_0_set_jpeg_ring_funcs(struct amdgpu_device *adev);
+static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev);
+static int vcn_v2_0_set_powergating_state(void *handle,
+                               enum amd_powergating_state state);
+
+/**
+ * vcn_v2_0_early_init - set function pointers
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Set ring and irq function pointers
+ */
+static int vcn_v2_0_early_init(void *handle)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       adev->vcn.num_enc_rings = 2;
+
+       vcn_v2_0_set_dec_ring_funcs(adev);
+       vcn_v2_0_set_enc_ring_funcs(adev);
+       vcn_v2_0_set_jpeg_ring_funcs(adev);
+       vcn_v2_0_set_irq_funcs(adev);
+
+       return 0;
+}
+
+/**
+ * vcn_v2_0_sw_init - sw init for VCN block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Load firmware and sw initialization
+ */
+static int vcn_v2_0_sw_init(void *handle)
+{
+       struct amdgpu_ring *ring;
+       int i, r;
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       /* VCN DEC TRAP */
+       r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+                             VCN_2_0__SRCID__UVD_SYSTEM_MESSAGE_INTERRUPT,
+                             &adev->vcn.irq);
+       if (r)
+               return r;
+
+       /* VCN ENC TRAP */
+       for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+               r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+                                     i + VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE,
+                                     &adev->vcn.irq);
+               if (r)
+                       return r;
+       }
+
+       /* VCN JPEG TRAP */
+       r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_VCN,
+                             VCN_2_0__SRCID__JPEG_DECODE,
+                             &adev->vcn.irq);
+       if (r)
+               return r;
+
+       r = amdgpu_vcn_sw_init(adev);
+       if (r)
+               return r;
+
+       if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+               const struct common_firmware_header *hdr;
+               hdr = (const struct common_firmware_header *)adev->vcn.fw->data;
+               adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].ucode_id = AMDGPU_UCODE_ID_VCN;
+               adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].fw = adev->vcn.fw;
+               adev->firmware.fw_size +=
+                       ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
+               DRM_INFO("PSP loading VCN firmware\n");
+       }
+
+       r = amdgpu_vcn_resume(adev);
+       if (r)
+               return r;
+
+       ring = &adev->vcn.ring_dec;
+
+       ring->use_doorbell = true;
+       ring->doorbell_index = adev->doorbell_index.vcn.vcn_ring0_1 << 1;
+
+       sprintf(ring->name, "vcn_dec");
+       r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
+       if (r)
+               return r;
+
+       adev->vcn.internal.scratch9 = mmUVD_SCRATCH9_INTERNAL_OFFSET;
+       adev->vcn.external.scratch9 = SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9);
+       adev->vcn.internal.data0 = mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET;
+       adev->vcn.external.data0 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0);
+       adev->vcn.internal.data1 = mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET;
+       adev->vcn.external.data1 = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1);
+       adev->vcn.internal.cmd = mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET;
+       adev->vcn.external.cmd = SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD);
+       adev->vcn.internal.nop = mmUVD_NO_OP_INTERNAL_OFFSET;
+       adev->vcn.external.nop = SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP);
+
+       for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+               ring = &adev->vcn.ring_enc[i];
+               ring->use_doorbell = true;
+               ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 2 + i;
+               sprintf(ring->name, "vcn_enc%d", i);
+               r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
+               if (r)
+                       return r;
+       }
+
+       ring = &adev->vcn.ring_jpeg;
+       ring->use_doorbell = true;
+       ring->doorbell_index = (adev->doorbell_index.vcn.vcn_ring0_1 << 1) + 1;
+       sprintf(ring->name, "vcn_jpeg");
+       r = amdgpu_ring_init(adev, ring, 512, &adev->vcn.irq, 0);
+       if (r)
+               return r;
+
+       adev->vcn.internal.jpeg_pitch = mmUVD_JPEG_PITCH_INTERNAL_OFFSET;
+       adev->vcn.external.jpeg_pitch = SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_PITCH);
+
+       return 0;
+}
+
+/**
+ * vcn_v2_0_sw_fini - sw fini for VCN block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * VCN suspend and free up sw allocation
+ */
+static int vcn_v2_0_sw_fini(void *handle)
+{
+       int r;
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       r = amdgpu_vcn_suspend(adev);
+       if (r)
+               return r;
+
+       r = amdgpu_vcn_sw_fini(adev);
+
+       return r;
+}
+
+/**
+ * vcn_v2_0_hw_init - start and test VCN block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int vcn_v2_0_hw_init(void *handle)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+       struct amdgpu_ring *ring = &adev->vcn.ring_dec;
+       int i, r;
+
+       adev->nbio_funcs->vcn_doorbell_range(adev, ring->use_doorbell,
+               ring->doorbell_index);
+
+       ring->sched.ready = true;
+       r = amdgpu_ring_test_ring(ring);
+       if (r) {
+               ring->sched.ready = false;
+               goto done;
+       }
+
+       for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+               ring = &adev->vcn.ring_enc[i];
+               ring->sched.ready = true;
+               r = amdgpu_ring_test_ring(ring);
+               if (r) {
+                       ring->sched.ready = false;
+                       goto done;
+               }
+       }
+
+       ring = &adev->vcn.ring_jpeg;
+       ring->sched.ready = true;
+       r = amdgpu_ring_test_ring(ring);
+       if (r) {
+               ring->sched.ready = false;
+               goto done;
+       }
+
+done:
+       if (!r)
+               DRM_INFO("VCN decode and encode initialized successfully.\n");
+
+       return r;
+}
+
+/**
+ * vcn_v2_0_hw_fini - stop the hardware block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Stop the VCN block, mark ring as not ready any more
+ */
+static int vcn_v2_0_hw_fini(void *handle)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+       struct amdgpu_ring *ring = &adev->vcn.ring_dec;
+       int i;
+
+       if (RREG32_SOC15(VCN, 0, mmUVD_STATUS))
+               vcn_v2_0_set_powergating_state(adev, AMD_PG_STATE_GATE);
+
+       ring->sched.ready = false;
+
+       for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
+               ring = &adev->vcn.ring_enc[i];
+               ring->sched.ready = false;
+       }
+
+       ring = &adev->vcn.ring_jpeg;
+       ring->sched.ready = false;
+
+       return 0;
+}
+
+/**
+ * vcn_v2_0_suspend - suspend VCN block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * HW fini and suspend VCN block
+ */
+static int vcn_v2_0_suspend(void *handle)
+{
+       int r;
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       r = vcn_v2_0_hw_fini(adev);
+       if (r)
+               return r;
+
+       r = amdgpu_vcn_suspend(adev);
+
+       return r;
+}
+
+/**
+ * vcn_v2_0_resume - resume VCN block
+ *
+ * @handle: amdgpu_device pointer
+ *
+ * Resume firmware and hw init VCN block
+ */
+static int vcn_v2_0_resume(void *handle)
+{
+       int r;
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       r = amdgpu_vcn_resume(adev);
+       if (r)
+               return r;
+
+       r = vcn_v2_0_hw_init(adev);
+
+       return r;
+}
+
+/**
+ * vcn_v2_0_mc_resume - memory controller programming
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Let the VCN memory controller know it's offsets
+ */
+static void vcn_v2_0_mc_resume(struct amdgpu_device *adev)
+{
+       uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
+       uint32_t offset;
+
+       /* cache window 0: fw */
+       if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+               WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+                       (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo));
+               WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+                       (adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi));
+               WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0);
+               offset = 0;
+       } else {
+               WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
+                       lower_32_bits(adev->vcn.gpu_addr));
+               WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
+                       upper_32_bits(adev->vcn.gpu_addr));
+               offset = size;
+               /* No signed header for now from firmware
+               WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
+                       AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+               */
+               WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0);
+       }
+
+       WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size);
+
+       /* cache window 1: stack */
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
+               lower_32_bits(adev->vcn.gpu_addr + offset));
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
+               upper_32_bits(adev->vcn.gpu_addr + offset));
+       WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0);
+       WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
+
+       /* cache window 2: context */
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
+               lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
+               upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
+       WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0);
+       WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
+
+       WREG32_SOC15(UVD, 0, mmUVD_GFX10_ADDR_CONFIG, adev->gfx.config.gb_addr_config);
+       WREG32_SOC15(UVD, 0, mmJPEG_DEC_GFX10_ADDR_CONFIG, adev->gfx.config.gb_addr_config);
+}
+
+/**
+ * vcn_v2_0_disable_clock_gating - disable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @sw: enable SW clock gating
+ *
+ * Disable clock gating for VCN block
+ */
+static void vcn_v2_0_disable_clock_gating(struct amdgpu_device *adev)
+{
+       uint32_t data;
+
+       /* UVD disable CGC */
+       data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+       if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+               data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+       else
+               data &= ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK;
+       data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+       data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+       WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_CGC_GATE);
+       data &= ~(UVD_CGC_GATE__SYS_MASK
+               | UVD_CGC_GATE__UDEC_MASK
+               | UVD_CGC_GATE__MPEG2_MASK
+               | UVD_CGC_GATE__REGS_MASK
+               | UVD_CGC_GATE__RBC_MASK
+               | UVD_CGC_GATE__LMI_MC_MASK
+               | UVD_CGC_GATE__LMI_UMC_MASK
+               | UVD_CGC_GATE__IDCT_MASK
+               | UVD_CGC_GATE__MPRD_MASK
+               | UVD_CGC_GATE__MPC_MASK
+               | UVD_CGC_GATE__LBSI_MASK
+               | UVD_CGC_GATE__LRBBM_MASK
+               | UVD_CGC_GATE__UDEC_RE_MASK
+               | UVD_CGC_GATE__UDEC_CM_MASK
+               | UVD_CGC_GATE__UDEC_IT_MASK
+               | UVD_CGC_GATE__UDEC_DB_MASK
+               | UVD_CGC_GATE__UDEC_MP_MASK
+               | UVD_CGC_GATE__WCB_MASK
+               | UVD_CGC_GATE__VCPU_MASK
+               | UVD_CGC_GATE__SCPU_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_CGC_GATE, data);
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+       data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+               | UVD_CGC_CTRL__SYS_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_MODE_MASK
+               | UVD_CGC_CTRL__MPEG2_MODE_MASK
+               | UVD_CGC_CTRL__REGS_MODE_MASK
+               | UVD_CGC_CTRL__RBC_MODE_MASK
+               | UVD_CGC_CTRL__LMI_MC_MODE_MASK
+               | UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+               | UVD_CGC_CTRL__IDCT_MODE_MASK
+               | UVD_CGC_CTRL__MPRD_MODE_MASK
+               | UVD_CGC_CTRL__MPC_MODE_MASK
+               | UVD_CGC_CTRL__LBSI_MODE_MASK
+               | UVD_CGC_CTRL__LRBBM_MODE_MASK
+               | UVD_CGC_CTRL__WCB_MODE_MASK
+               | UVD_CGC_CTRL__VCPU_MODE_MASK
+               | UVD_CGC_CTRL__SCPU_MODE_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+       /* turn on */
+       data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE);
+       data |= (UVD_SUVD_CGC_GATE__SRE_MASK
+               | UVD_SUVD_CGC_GATE__SIT_MASK
+               | UVD_SUVD_CGC_GATE__SMP_MASK
+               | UVD_SUVD_CGC_GATE__SCM_MASK
+               | UVD_SUVD_CGC_GATE__SDB_MASK
+               | UVD_SUVD_CGC_GATE__SRE_H264_MASK
+               | UVD_SUVD_CGC_GATE__SRE_HEVC_MASK
+               | UVD_SUVD_CGC_GATE__SIT_H264_MASK
+               | UVD_SUVD_CGC_GATE__SIT_HEVC_MASK
+               | UVD_SUVD_CGC_GATE__SCM_H264_MASK
+               | UVD_SUVD_CGC_GATE__SCM_HEVC_MASK
+               | UVD_SUVD_CGC_GATE__SDB_H264_MASK
+               | UVD_SUVD_CGC_GATE__SDB_HEVC_MASK
+               | UVD_SUVD_CGC_GATE__SCLR_MASK
+               | UVD_SUVD_CGC_GATE__UVD_SC_MASK
+               | UVD_SUVD_CGC_GATE__ENT_MASK
+               | UVD_SUVD_CGC_GATE__SIT_HEVC_DEC_MASK
+               | UVD_SUVD_CGC_GATE__SIT_HEVC_ENC_MASK
+               | UVD_SUVD_CGC_GATE__SITE_MASK
+               | UVD_SUVD_CGC_GATE__SRE_VP9_MASK
+               | UVD_SUVD_CGC_GATE__SCM_VP9_MASK
+               | UVD_SUVD_CGC_GATE__SIT_VP9_DEC_MASK
+               | UVD_SUVD_CGC_GATE__SDB_VP9_MASK
+               | UVD_SUVD_CGC_GATE__IME_HEVC_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_GATE, data);
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
+       data &= ~(UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
+}
+
+/**
+ * jpeg_v2_0_start - start JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the JPEG block
+ */
+static int jpeg_v2_0_start(struct amdgpu_device *adev)
+{
+       struct amdgpu_ring *ring = &adev->vcn.ring_jpeg;
+       uint32_t tmp;
+       int r = 0;
+
+       /* disable power gating */
+       tmp = 1 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT;
+       WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_PGFSM_CONFIG), tmp);
+
+       SOC15_WAIT_ON_RREG(VCN, 0,
+               mmUVD_PGFSM_STATUS, UVD_PGFSM_STATUS_UVDJ_PWR_ON,
+               UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK, r);
+
+       if (r) {
+               DRM_ERROR("amdgpu: JPEG disable power gating failed\n");
+               return r;
+       }
+
+       /* Removing the anti hang mechanism to indicate the UVDJ tile is ON */
+       tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS)) & ~0x1;
+       WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS), tmp);
+
+       /* JPEG disable CGC */
+       tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
+       tmp |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+       tmp |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+       tmp |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+       WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, tmp);
+
+       tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
+       tmp &= ~(JPEG_CGC_GATE__JPEG_DEC_MASK
+               | JPEG_CGC_GATE__JPEG2_DEC_MASK
+               | JPEG_CGC_GATE__JPEG_ENC_MASK
+               | JPEG_CGC_GATE__JMCIF_MASK
+               | JPEG_CGC_GATE__JRBBM_MASK);
+       WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, tmp);
+
+       /* enable JMI channel */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JMI_CNTL), 0,
+               ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+       /* enable System Interrupt for JRBC */
+       WREG32_P(SOC15_REG_OFFSET(VCN, 0, mmJPEG_SYS_INT_EN),
+               JPEG_SYS_INT_EN__DJRBC_MASK,
+               ~JPEG_SYS_INT_EN__DJRBC_MASK);
+
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
+       WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
+               lower_32_bits(ring->gpu_addr));
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
+               upper_32_bits(ring->gpu_addr));
+       WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, 0);
+       WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, 0);
+       WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, 0x00000002L);
+       WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_SIZE, ring->ring_size / 4);
+       ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
+
+       return 0;
+}
+
+/**
+ * jpeg_v2_0_stop - stop JPEG block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the JPEG block
+ */
+static int jpeg_v2_0_stop(struct amdgpu_device *adev)
+{
+       uint32_t tmp;
+       int r = 0;
+
+       /* reset JMI */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_JMI_CNTL),
+               UVD_JMI_CNTL__SOFT_RESET_MASK,
+               ~UVD_JMI_CNTL__SOFT_RESET_MASK);
+
+       /* enable JPEG CGC */
+       tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL);
+       tmp |= 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+       tmp |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+       tmp |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+       WREG32_SOC15(VCN, 0, mmJPEG_CGC_CTRL, tmp);
+
+
+       tmp = RREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE);
+       tmp |= (JPEG_CGC_GATE__JPEG_DEC_MASK
+               |JPEG_CGC_GATE__JPEG2_DEC_MASK
+               |JPEG_CGC_GATE__JPEG_ENC_MASK
+               |JPEG_CGC_GATE__JMCIF_MASK
+               |JPEG_CGC_GATE__JRBBM_MASK);
+       WREG32_SOC15(VCN, 0, mmJPEG_CGC_GATE, tmp);
+
+       /* enable power gating */
+       tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS));
+       tmp &= ~UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_MASK;
+       tmp |=  0x1; //UVD_JPEG_POWER_STATUS__JPEG_POWER_STATUS_TILES_OFF;
+       WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_POWER_STATUS), tmp);
+
+       tmp = 2 << UVD_PGFSM_CONFIG__UVDJ_PWR_CONFIG__SHIFT;
+       WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_PGFSM_CONFIG), tmp);
+
+       SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_PGFSM_STATUS,
+               (2 << UVD_PGFSM_STATUS__UVDJ_PWR_STATUS__SHIFT),
+               UVD_PGFSM_STATUS__UVDJ_PWR_STATUS_MASK, r);
+
+       if (r) {
+               DRM_ERROR("amdgpu: JPEG enable power gating failed\n");
+               return r;
+       }
+
+       return r;
+}
+
+/**
+ * vcn_v2_0_enable_clock_gating - enable VCN clock gating
+ *
+ * @adev: amdgpu_device pointer
+ * @sw: enable SW clock gating
+ *
+ * Enable clock gating for VCN block
+ */
+static void vcn_v2_0_enable_clock_gating(struct amdgpu_device *adev)
+{
+       uint32_t data = 0;
+
+       /* enable UVD CGC */
+       data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+       if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
+               data |= 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+       else
+               data |= 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
+       data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
+       data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
+       WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL);
+       data |= (UVD_CGC_CTRL__UDEC_RE_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_CM_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_IT_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_DB_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_MP_MODE_MASK
+               | UVD_CGC_CTRL__SYS_MODE_MASK
+               | UVD_CGC_CTRL__UDEC_MODE_MASK
+               | UVD_CGC_CTRL__MPEG2_MODE_MASK
+               | UVD_CGC_CTRL__REGS_MODE_MASK
+               | UVD_CGC_CTRL__RBC_MODE_MASK
+               | UVD_CGC_CTRL__LMI_MC_MODE_MASK
+               | UVD_CGC_CTRL__LMI_UMC_MODE_MASK
+               | UVD_CGC_CTRL__IDCT_MODE_MASK
+               | UVD_CGC_CTRL__MPRD_MODE_MASK
+               | UVD_CGC_CTRL__MPC_MODE_MASK
+               | UVD_CGC_CTRL__LBSI_MODE_MASK
+               | UVD_CGC_CTRL__LRBBM_MODE_MASK
+               | UVD_CGC_CTRL__WCB_MODE_MASK
+               | UVD_CGC_CTRL__VCPU_MODE_MASK
+               | UVD_CGC_CTRL__SCPU_MODE_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_CGC_CTRL, data);
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL);
+       data |= (UVD_SUVD_CGC_CTRL__SRE_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SIT_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SMP_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SCM_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SDB_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SCLR_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__UVD_SC_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__ENT_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__IME_MODE_MASK
+               | UVD_SUVD_CGC_CTRL__SITE_MODE_MASK);
+       WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
+}
+
+static void vcn_v2_0_disable_static_power_gating(struct amdgpu_device *adev)
+{
+       uint32_t data = 0;
+       int ret;
+
+       if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+               data = (1 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDU_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDC_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDIL_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDIR_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDW_PWR_CONFIG__SHIFT);
+
+               WREG32_SOC15(VCN, 0, mmUVD_PGFSM_CONFIG, data);
+               SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_PGFSM_STATUS,
+                       UVD_PGFSM_STATUS__UVDM_UVDU_PWR_ON, 0xFFFFFF, ret);
+       } else {
+               data = (1 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDU_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDC_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDIL_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDIR_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+                       | 1 << UVD_PGFSM_CONFIG__UVDW_PWR_CONFIG__SHIFT);
+               WREG32_SOC15(VCN, 0, mmUVD_PGFSM_CONFIG, data);
+               SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_PGFSM_STATUS, 0,  0xFFFFFFFF, ret);
+       }
+
+       /* polling UVD_PGFSM_STATUS to confirm UVDM_PWR_STATUS,
+        * UVDU_PWR_STATUS are 0 (power on) */
+
+       data = RREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS);
+       data &= ~0x103;
+       if (adev->pg_flags & AMD_PG_SUPPORT_VCN)
+               data |= UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON |
+                       UVD_POWER_STATUS__UVD_PG_EN_MASK;
+
+       WREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS, data);
+}
+
+static void vcn_v2_0_enable_static_power_gating(struct amdgpu_device *adev)
+{
+       uint32_t data = 0;
+       int ret;
+
+       if (adev->pg_flags & AMD_PG_SUPPORT_VCN) {
+               /* Before power off, this indicator has to be turned on */
+               data = RREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS);
+               data &= ~UVD_POWER_STATUS__UVD_POWER_STATUS_MASK;
+               data |= UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF;
+               WREG32_SOC15(VCN, 0, mmUVD_POWER_STATUS, data);
+
+
+               data = (2 << UVD_PGFSM_CONFIG__UVDM_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDU_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDF_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDC_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDB_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDIL_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDIR_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDTD_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDTE_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDE_PWR_CONFIG__SHIFT
+                       | 2 << UVD_PGFSM_CONFIG__UVDW_PWR_CONFIG__SHIFT);
+
+               WREG32_SOC15(VCN, 0, mmUVD_PGFSM_CONFIG, data);
+
+               data = (2 << UVD_PGFSM_STATUS__UVDM_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDU_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDF_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDC_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDB_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDIL_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDIR_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDTD_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDTE_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDE_PWR_STATUS__SHIFT
+                       | 2 << UVD_PGFSM_STATUS__UVDW_PWR_STATUS__SHIFT);
+               SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_PGFSM_STATUS, data, 0xFFFFFFFF, ret);
+       }
+}
+
+static int vcn_v2_0_start(struct amdgpu_device *adev)
+{
+       struct amdgpu_ring *ring = &adev->vcn.ring_dec;
+       uint32_t rb_bufsz, tmp;
+       uint32_t lmi_swap_cntl;
+       int i, j, r;
+
+       vcn_v2_0_disable_static_power_gating(adev);
+
+       /* set uvd status busy */
+       tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
+       WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
+
+       /*SW clock gating */
+       vcn_v2_0_disable_clock_gating(adev);
+
+       /* enable VCPU clock */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
+               UVD_VCPU_CNTL__CLK_EN_MASK, ~UVD_VCPU_CNTL__CLK_EN_MASK);
+
+       /* disable master interrupt */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
+               ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+       /* setup mmUVD_LMI_CTRL */
+       tmp = RREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL);
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL, tmp |
+               UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+               UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
+               UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+               UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
+
+       /* setup mmUVD_MPC_CNTL */
+       tmp = RREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL);
+       tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
+       tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
+       WREG32_SOC15(VCN, 0, mmUVD_MPC_CNTL, tmp);
+
+       /* setup UVD_MPC_SET_MUXA0 */
+       WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0,
+               ((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
+               (0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
+               (0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
+               (0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
+
+       /* setup UVD_MPC_SET_MUXB0 */
+       WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0,
+               ((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
+               (0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
+               (0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
+               (0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
+
+       /* setup mmUVD_MPC_SET_MUX */
+       WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX,
+               ((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
+               (0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
+               (0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
+
+       vcn_v2_0_mc_resume(adev);
+
+       /* release VCPU reset to boot */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0,
+               ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+
+       /* enable LMI MC and UMC channels */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
+               ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+       tmp = RREG32_SOC15(VCN, 0, mmUVD_SOFT_RESET);
+       tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
+       tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
+       WREG32_SOC15(VCN, 0, mmUVD_SOFT_RESET, tmp);
+
+       /* disable byte swapping */
+       lmi_swap_cntl = 0;
+#ifdef __BIG_ENDIAN
+       /* swap (8 in 32) RB and IB */
+       lmi_swap_cntl = 0xa;
+#endif
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl);
+
+       for (i = 0; i < 10; ++i) {
+               uint32_t status;
+
+               for (j = 0; j < 100; ++j) {
+                       status = RREG32_SOC15(UVD, 0, mmUVD_STATUS);
+                       if (status & 2)
+                               break;
+                       mdelay(10);
+               }
+               r = 0;
+               if (status & 2)
+                       break;
+
+               DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
+               WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+                       UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK,
+                       ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+               mdelay(10);
+               WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0,
+                       ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+               mdelay(10);
+               r = -1;
+       }
+
+       if (r) {
+               DRM_ERROR("VCN decode not responding, giving up!!!\n");
+               return r;
+       }
+
+       /* enable master interrupt */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN),
+               UVD_MASTINT_EN__VCPU_EN_MASK,
+               ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+       /* clear the busy bit of VCN_STATUS */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0,
+               ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_VMID, 0);
+
+       /* force RBC into idle state */
+       rb_bufsz = order_base_2(ring->ring_size);
+       tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
+       tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
+       tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
+       tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
+       tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
+       WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp);
+
+       /* programm the RB_BASE for ring buffer */
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
+               lower_32_bits(ring->gpu_addr));
+       WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
+               upper_32_bits(ring->gpu_addr));
+
+       /* Initialize the ring buffer's read and write pointers */
+       WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0);
+
+       ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
+       WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
+                       lower_32_bits(ring->wptr));
+
+       ring = &adev->vcn.ring_enc[0];
+       WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO, ring->gpu_addr);
+       WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE, ring->ring_size / 4);
+
+       ring = &adev->vcn.ring_enc[1];
+       WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO2, ring->gpu_addr);
+       WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
+       WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
+
+       r = jpeg_v2_0_start(adev);
+
+       return r;
+}
+
+static int vcn_v2_0_stop(struct amdgpu_device *adev)
+{
+       uint32_t tmp;
+       int r;
+
+       r = jpeg_v2_0_stop(adev);
+       if (r)
+               return r;
+       /* wait for uvd idle */
+       SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, r);
+       if (r)
+               return r;
+
+       tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
+               UVD_LMI_STATUS__READ_CLEAN_MASK |
+               UVD_LMI_STATUS__WRITE_CLEAN_MASK |
+               UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
+       SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_LMI_STATUS, tmp, tmp, r);
+       if (r)
+               return r;
+
+       /* stall UMC channel */
+       tmp = RREG32_SOC15(VCN, 0, mmUVD_LMI_CTRL2);
+       tmp |= UVD_LMI_CTRL2__STALL_ARB_UMC_MASK;
+       WREG32_SOC15(VCN, 0, mmUVD_LMI_CTRL2, tmp);
+
+       tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK|
+               UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
+       SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_LMI_STATUS, tmp, tmp, r);
+       if (r)
+               return r;
+
+       /* disable VCPU clock */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0,
+               ~(UVD_VCPU_CNTL__CLK_EN_MASK));
+
+       /* reset LMI UMC */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+               UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK,
+               ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
+
+       /* reset LMI */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+               UVD_SOFT_RESET__LMI_SOFT_RESET_MASK,
+               ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK);
+
+       /* reset VCPU */
+       WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+               UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK,
+               ~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+
+       /* clear status */
+       WREG32_SOC15(VCN, 0, mmUVD_STATUS, 0);
+
+       vcn_v2_0_enable_clock_gating(adev);
+       vcn_v2_0_enable_static_power_gating(adev);
+
+       return 0;
+}
+
+static bool vcn_v2_0_is_idle(void *handle)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == UVD_STATUS__IDLE);
+}
+
+static int vcn_v2_0_wait_for_idle(void *handle)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+       int ret = 0;
+
+       SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE,
+               UVD_STATUS__IDLE, ret);
+
+       return ret;
+}
+
+static int vcn_v2_0_set_clockgating_state(void *handle,
+                                         enum amd_clockgating_state state)
+{
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+       bool enable = (state == AMD_CG_STATE_GATE) ? true : false;
+
+       if (enable) {
+               /* wait for STATUS to clear */
+               if (vcn_v2_0_is_idle(handle))
+                       return -EBUSY;
+               vcn_v2_0_enable_clock_gating(adev);
+       } else {
+               /* disable HW gating and enable Sw gating */
+               vcn_v2_0_disable_clock_gating(adev);
+       }
+       return 0;
+}
+
+/**
+ * vcn_v2_0_dec_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t vcn_v2_0_dec_ring_get_rptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
+}
+
+/**
+ * vcn_v2_0_dec_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t vcn_v2_0_dec_ring_get_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring->use_doorbell)
+               return adev->wb.wb[ring->wptr_offs];
+       else
+               return RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR);
+}
+
+/**
+ * vcn_v2_0_dec_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void vcn_v2_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring->use_doorbell) {
+               adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+               WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+       } else {
+               WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
+       }
+}
+
+/**
+ * vcn_v2_0_dec_ring_insert_start - insert a start command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a start command to the ring.
+ */
+static void vcn_v2_0_dec_ring_insert_start(struct amdgpu_ring *ring)
+{
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, 0);
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, VCN_DEC_CMD_PACKET_START << 1);
+}
+
+/**
+ * vcn_v2_0_dec_ring_insert_end - insert a end command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a end command to the ring.
+ */
+static void vcn_v2_0_dec_ring_insert_end(struct amdgpu_ring *ring)
+{
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, VCN_DEC_CMD_PACKET_END << 1);
+}
+
+/**
+ * vcn_v2_0_dec_ring_insert_nop - insert a nop command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a nop command to the ring.
+ */
+static void vcn_v2_0_dec_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+       int i;
+
+       WARN_ON(ring->wptr % 2 || count % 2);
+
+       for (i = 0; i < count / 2; i++) {
+               amdgpu_ring_write(ring, PACKET0(mmUVD_NO_OP_INTERNAL_OFFSET, 0));
+               amdgpu_ring_write(ring, 0);
+       }
+}
+
+/**
+ * vcn_v2_0_dec_ring_emit_fence - emit an fence & trap command
+ *
+ * @ring: amdgpu_ring pointer
+ * @fence: fence to emit
+ *
+ * Write a fence and a trap command to the ring.
+ */
+static void vcn_v2_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+                                    unsigned flags)
+{
+       WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_CONTEXT_ID_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, seq);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, addr & 0xffffffff);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, upper_32_bits(addr) & 0xff);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, VCN_DEC_CMD_FENCE << 1);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, 0);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, 0);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+
+       amdgpu_ring_write(ring, VCN_DEC_CMD_TRAP << 1);
+}
+
+/**
+ * vcn_v2_0_dec_ring_emit_ib - execute indirect buffer
+ *
+ * @ring: amdgpu_ring pointer
+ * @ib: indirect buffer to execute
+ *
+ * Write ring commands to execute the indirect buffer
+ */
+static void vcn_v2_0_dec_ring_emit_ib(struct amdgpu_ring *ring,
+                                     struct amdgpu_job *job,
+                                     struct amdgpu_ib *ib,
+                                     uint32_t flags)
+{
+       unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_VMID_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, vmid);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+       amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+       amdgpu_ring_write(ring, PACKET0(mmUVD_RBC_IB_SIZE_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, ib->length_dw);
+}
+
+static void vcn_v2_0_dec_ring_emit_reg_wait(struct amdgpu_ring *ring,
+                                           uint32_t reg, uint32_t val,
+                                           uint32_t mask)
+{
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, reg << 2);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, val);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GP_SCRATCH8_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, mask);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+
+       amdgpu_ring_write(ring, VCN_DEC_CMD_REG_READ_COND_WAIT << 1);
+}
+
+static void vcn_v2_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
+                                           unsigned vmid, uint64_t pd_addr)
+{
+       struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+       uint32_t data0, data1, mask;
+
+       pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+       /* wait for register write */
+       data0 = hub->ctx0_ptb_addr_lo32 + vmid * 2;
+       data1 = lower_32_bits(pd_addr);
+       mask = 0xffffffff;
+       vcn_v2_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
+}
+
+static void vcn_v2_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
+                                       uint32_t reg, uint32_t val)
+{
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, reg << 2);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1_INTERNAL_OFFSET, 0));
+       amdgpu_ring_write(ring, val);
+
+       amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD_INTERNAL_OFFSET, 0));
+
+       amdgpu_ring_write(ring, VCN_DEC_CMD_WRITE_REG << 1);
+}
+
+/**
+ * vcn_v2_0_enc_ring_get_rptr - get enc read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc read pointer
+ */
+static uint64_t vcn_v2_0_enc_ring_get_rptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring == &adev->vcn.ring_enc[0])
+               return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR);
+       else
+               return RREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2);
+}
+
+ /**
+ * vcn_v2_0_enc_ring_get_wptr - get enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc write pointer
+ */
+static uint64_t vcn_v2_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring == &adev->vcn.ring_enc[0]) {
+               if (ring->use_doorbell)
+                       return adev->wb.wb[ring->wptr_offs];
+               else
+                       return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR);
+       } else {
+               if (ring->use_doorbell)
+                       return adev->wb.wb[ring->wptr_offs];
+               else
+                       return RREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2);
+       }
+}
+
+ /**
+ * vcn_v2_0_enc_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void vcn_v2_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring == &adev->vcn.ring_enc[0]) {
+               if (ring->use_doorbell) {
+                       adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+                       WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+               } else {
+                       WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
+               }
+       } else {
+               if (ring->use_doorbell) {
+                       adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+                       WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+               } else {
+                       WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
+               }
+       }
+}
+
+/**
+ * vcn_v2_0_enc_ring_emit_fence - emit an enc fence & trap command
+ *
+ * @ring: amdgpu_ring pointer
+ * @fence: fence to emit
+ *
+ * Write enc a fence and a trap command to the ring.
+ */
+static void vcn_v2_0_enc_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+                       u64 seq, unsigned flags)
+{
+       WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+       amdgpu_ring_write(ring, VCN_ENC_CMD_FENCE);
+       amdgpu_ring_write(ring, addr);
+       amdgpu_ring_write(ring, upper_32_bits(addr));
+       amdgpu_ring_write(ring, seq);
+       amdgpu_ring_write(ring, VCN_ENC_CMD_TRAP);
+}
+
+static void vcn_v2_0_enc_ring_insert_end(struct amdgpu_ring *ring)
+{
+       amdgpu_ring_write(ring, VCN_ENC_CMD_END);
+}
+
+/**
+ * vcn_v2_0_enc_ring_emit_ib - enc execute indirect buffer
+ *
+ * @ring: amdgpu_ring pointer
+ * @ib: indirect buffer to execute
+ *
+ * Write enc ring commands to execute the indirect buffer
+ */
+static void vcn_v2_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
+                                     struct amdgpu_job *job,
+                                     struct amdgpu_ib *ib,
+                                     uint32_t flags)
+{
+       unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+
+       amdgpu_ring_write(ring, VCN_ENC_CMD_IB);
+       amdgpu_ring_write(ring, vmid);
+       amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+       amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+       amdgpu_ring_write(ring, ib->length_dw);
+}
+
+static void vcn_v2_0_enc_ring_emit_reg_wait(struct amdgpu_ring *ring,
+                                           uint32_t reg, uint32_t val,
+                                           uint32_t mask)
+{
+       amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WAIT);
+       amdgpu_ring_write(ring, reg << 2);
+       amdgpu_ring_write(ring, mask);
+       amdgpu_ring_write(ring, val);
+}
+
+static void vcn_v2_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+                                           unsigned int vmid, uint64_t pd_addr)
+{
+       struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+
+       pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+       /* wait for reg writes */
+       vcn_v2_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
+                                       lower_32_bits(pd_addr), 0xffffffff);
+}
+
+static void vcn_v2_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
+                                       uint32_t reg, uint32_t val)
+{
+       amdgpu_ring_write(ring, VCN_ENC_CMD_REG_WRITE);
+       amdgpu_ring_write(ring, reg << 2);
+       amdgpu_ring_write(ring, val);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t vcn_v2_0_jpeg_ring_get_rptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t vcn_v2_0_jpeg_ring_get_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring->use_doorbell)
+               return adev->wb.wb[ring->wptr_offs];
+       else
+               return RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void vcn_v2_0_jpeg_ring_set_wptr(struct amdgpu_ring *ring)
+{
+       struct amdgpu_device *adev = ring->adev;
+
+       if (ring->use_doorbell) {
+               adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
+               WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));
+       } else {
+               WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, lower_32_bits(ring->wptr));
+       }
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_insert_start - insert a start command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a start command to the ring.
+ */
+static void vcn_v2_0_jpeg_ring_insert_start(struct amdgpu_ring *ring)
+{
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x68e04);
+
+       amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x80010000);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_insert_end - insert a end command
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Write a end command to the ring.
+ */
+static void vcn_v2_0_jpeg_ring_insert_end(struct amdgpu_ring *ring)
+{
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x68e04);
+
+       amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x00010000);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_emit_fence - emit an fence & trap command
+ *
+ * @ring: amdgpu_ring pointer
+ * @fence: fence to emit
+ *
+ * Write a fence and a trap command to the ring.
+ */
+static void vcn_v2_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
+                                    unsigned flags)
+{
+       WARN_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_DATA0_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, seq);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_DATA1_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, seq);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_LOW_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, lower_32_bits(addr));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_WR_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, upper_32_bits(addr));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x8);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JPEG_GPCOM_CMD_INTERNAL_OFFSET,
+               0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE4));
+       amdgpu_ring_write(ring, 0);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x3fbc);
+
+       amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x1);
+
+       amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE7));
+       amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * vcn_v2_0_jpeg_ring_emit_ib - execute indirect buffer
+ *
+ * @ring: amdgpu_ring pointer
+ * @ib: indirect buffer to execute
+ *
+ * Write ring commands to execute the indirect buffer.
+ */
+static void vcn_v2_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring,
+                                      struct amdgpu_job *job,
+                                      struct amdgpu_ib *ib,
+                                      uint32_t flags)
+{
+       unsigned vmid = AMDGPU_JOB_GET_VMID(job);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_VMID_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JPEG_VMID_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, (vmid | (vmid << 4)));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_LOW_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_IB_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_IB_SIZE_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, ib->length_dw);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_LOW_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, lower_32_bits(ring->gpu_addr));
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_LMI_JRBC_RB_MEM_RD_64BIT_BAR_HIGH_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, upper_32_bits(ring->gpu_addr));
+
+       amdgpu_ring_write(ring, PACKETJ(0, 0, PACKETJ_CONDITION_CHECK0, PACKETJ_TYPE2));
+       amdgpu_ring_write(ring, 0);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x01400200);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x2);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_STATUS_INTERNAL_OFFSET,
+               0, PACKETJ_CONDITION_CHECK3, PACKETJ_TYPE3));
+       amdgpu_ring_write(ring, 0x2);
+}
+
+static void vcn_v2_0_jpeg_ring_emit_reg_wait(struct amdgpu_ring *ring,
+                                           uint32_t reg, uint32_t val,
+                                           uint32_t mask)
+{
+       uint32_t reg_offset = (reg << 2);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_COND_RD_TIMER_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, 0x01400200);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_RB_REF_DATA_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       amdgpu_ring_write(ring, val);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) {
+               amdgpu_ring_write(ring, 0);
+               amdgpu_ring_write(ring,
+                       PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE3));
+       } else {
+               amdgpu_ring_write(ring, reg_offset);
+               amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+                       0, 0, PACKETJ_TYPE3));
+       }
+       amdgpu_ring_write(ring, mask);
+}
+
+static void vcn_v2_0_jpeg_ring_emit_vm_flush(struct amdgpu_ring *ring,
+               unsigned vmid, uint64_t pd_addr)
+{
+       struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
+       uint32_t data0, data1, mask;
+
+       pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pd_addr);
+
+       /* wait for register write */
+       data0 = hub->ctx0_ptb_addr_lo32 + vmid * 2;
+       data1 = lower_32_bits(pd_addr);
+       mask = 0xffffffff;
+       vcn_v2_0_jpeg_ring_emit_reg_wait(ring, data0, data1, mask);
+}
+
+static void vcn_v2_0_jpeg_ring_emit_wreg(struct amdgpu_ring *ring,
+                                       uint32_t reg, uint32_t val)
+{
+       uint32_t reg_offset = (reg << 2);
+
+       amdgpu_ring_write(ring, PACKETJ(mmUVD_JRBC_EXTERNAL_REG_INTERNAL_OFFSET,
+               0, 0, PACKETJ_TYPE0));
+       if (reg_offset >= 0x10000 && reg_offset <= 0x105ff) {
+               amdgpu_ring_write(ring, 0);
+               amdgpu_ring_write(ring,
+                       PACKETJ((reg_offset >> 2), 0, 0, PACKETJ_TYPE0));
+       } else {
+               amdgpu_ring_write(ring, reg_offset);
+               amdgpu_ring_write(ring, PACKETJ(JRBC_DEC_EXTERNAL_REG_WRITE_ADDR,
+                       0, 0, PACKETJ_TYPE0));
+       }
+       amdgpu_ring_write(ring, val);
+}
+
+static void vcn_v2_0_jpeg_ring_nop(struct amdgpu_ring *ring, uint32_t count)
+{
+       int i;
+
+       WARN_ON(ring->wptr % 2 || count % 2);
+
+       for (i = 0; i < count / 2; i++) {
+               amdgpu_ring_write(ring, PACKETJ(0, 0, 0, PACKETJ_TYPE6));
+               amdgpu_ring_write(ring, 0);
+       }
+}
+
+static int vcn_v2_0_set_interrupt_state(struct amdgpu_device *adev,
+                                       struct amdgpu_irq_src *source,
+                                       unsigned type,
+                                       enum amdgpu_interrupt_state state)
+{
+       return 0;
+}
+
+static int vcn_v2_0_process_interrupt(struct amdgpu_device *adev,
+                                     struct amdgpu_irq_src *source,
+                                     struct amdgpu_iv_entry *entry)
+{
+       DRM_DEBUG("IH: VCN TRAP\n");
+
+       switch (entry->src_id) {
+       case VCN_2_0__SRCID__UVD_SYSTEM_MESSAGE_INTERRUPT:
+               amdgpu_fence_process(&adev->vcn.ring_dec);
+               break;
+       case VCN_2_0__SRCID__UVD_ENC_GENERAL_PURPOSE:
+               amdgpu_fence_process(&adev->vcn.ring_enc[0]);
+               break;
+       case VCN_2_0__SRCID__UVD_ENC_LOW_LATENCY:
+               amdgpu_fence_process(&adev->vcn.ring_enc[1]);
+               break;
+       case VCN_2_0__SRCID__JPEG_DECODE:
+               amdgpu_fence_process(&adev->vcn.ring_jpeg);
+               break;
+       default:
+               DRM_ERROR("Unhandled interrupt: %d %d\n",
+                         entry->src_id, entry->src_data[0]);
+               break;
+       }
+
+       return 0;
+}
+
+static int vcn_v2_0_set_powergating_state(void *handle,
+                                         enum amd_powergating_state state)
+{
+       /* This doesn't actually powergate the VCN block.
+        * That's done in the dpm code via the SMC.  This
+        * just re-inits the block as necessary.  The actual
+        * gating still happens in the dpm code.  We should
+        * revisit this when there is a cleaner line between
+        * the smc and the hw blocks
+        */
+       int ret;
+       struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+       if (state == adev->vcn.cur_state)
+               return 0;
+
+       if (state == AMD_PG_STATE_GATE)
+               ret = vcn_v2_0_stop(adev);
+       else
+               ret = vcn_v2_0_start(adev);
+
+       if (!ret)
+               adev->vcn.cur_state = state;
+       return ret;
+}
+
+static const struct amd_ip_funcs vcn_v2_0_ip_funcs = {
+       .name = "vcn_v2_0",
+       .early_init = vcn_v2_0_early_init,
+       .late_init = NULL,
+       .sw_init = vcn_v2_0_sw_init,
+       .sw_fini = vcn_v2_0_sw_fini,
+       .hw_init = vcn_v2_0_hw_init,
+       .hw_fini = vcn_v2_0_hw_fini,
+       .suspend = vcn_v2_0_suspend,
+       .resume = vcn_v2_0_resume,
+       .is_idle = vcn_v2_0_is_idle,
+       .wait_for_idle = vcn_v2_0_wait_for_idle,
+       .check_soft_reset = NULL,
+       .pre_soft_reset = NULL,
+       .soft_reset = NULL,
+       .post_soft_reset = NULL,
+       .set_clockgating_state = vcn_v2_0_set_clockgating_state,
+       .set_powergating_state = vcn_v2_0_set_powergating_state,
+};
+
+static const struct amdgpu_ring_funcs vcn_v2_0_dec_ring_vm_funcs = {
+       .type = AMDGPU_RING_TYPE_VCN_DEC,
+       .align_mask = 0xf,
+       .vmhub = AMDGPU_MMHUB,
+       .get_rptr = vcn_v2_0_dec_ring_get_rptr,
+       .get_wptr = vcn_v2_0_dec_ring_get_wptr,
+       .set_wptr = vcn_v2_0_dec_ring_set_wptr,
+       .emit_frame_size =
+               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+               8 + /* vcn_v2_0_dec_ring_emit_vm_flush */
+               14 + 14 + /* vcn_v2_0_dec_ring_emit_fence x2 vm fence */
+               6,
+       .emit_ib_size = 8, /* vcn_v2_0_dec_ring_emit_ib */
+       .emit_ib = vcn_v2_0_dec_ring_emit_ib,
+       .emit_fence = vcn_v2_0_dec_ring_emit_fence,
+       .emit_vm_flush = vcn_v2_0_dec_ring_emit_vm_flush,
+       .test_ring = amdgpu_vcn_dec_ring_test_ring,
+       .test_ib = amdgpu_vcn_dec_ring_test_ib,
+       .insert_nop = vcn_v2_0_dec_ring_insert_nop,
+       .insert_start = vcn_v2_0_dec_ring_insert_start,
+       .insert_end = vcn_v2_0_dec_ring_insert_end,
+       .pad_ib = amdgpu_ring_generic_pad_ib,
+       .begin_use = amdgpu_vcn_ring_begin_use,
+       .end_use = amdgpu_vcn_ring_end_use,
+       .emit_wreg = vcn_v2_0_dec_ring_emit_wreg,
+       .emit_reg_wait = vcn_v2_0_dec_ring_emit_reg_wait,
+       .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static const struct amdgpu_ring_funcs vcn_v2_0_enc_ring_vm_funcs = {
+       .type = AMDGPU_RING_TYPE_VCN_ENC,
+       .align_mask = 0x3f,
+       .nop = VCN_ENC_CMD_NO_OP,
+       .vmhub = AMDGPU_MMHUB,
+       .get_rptr = vcn_v2_0_enc_ring_get_rptr,
+       .get_wptr = vcn_v2_0_enc_ring_get_wptr,
+       .set_wptr = vcn_v2_0_enc_ring_set_wptr,
+       .emit_frame_size =
+               SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+               4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+               5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+               1, /* vcn_v2_0_enc_ring_insert_end */
+       .emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
+       .emit_ib = vcn_v2_0_enc_ring_emit_ib,
+       .emit_fence = vcn_v2_0_enc_ring_emit_fence,
+       .emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+       .test_ring = amdgpu_vcn_enc_ring_test_ring,
+       .test_ib = amdgpu_vcn_enc_ring_test_ib,
+       .insert_nop = amdgpu_ring_insert_nop,
+       .insert_end = vcn_v2_0_enc_ring_insert_end,
+       .pad_ib = amdgpu_ring_generic_pad_ib,
+       .begin_use = amdgpu_vcn_ring_begin_use,
+       .end_use = amdgpu_vcn_ring_end_use,
+       .emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
+       .emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+       .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static const struct amdgpu_ring_funcs vcn_v2_0_jpeg_ring_vm_funcs = {
+       .type = AMDGPU_RING_TYPE_VCN_JPEG,
+       .align_mask = 0xf,
+       .vmhub = AMDGPU_MMHUB,
+       .get_rptr = vcn_v2_0_jpeg_ring_get_rptr,
+       .get_wptr = vcn_v2_0_jpeg_ring_get_wptr,
+       .set_wptr = vcn_v2_0_jpeg_ring_set_wptr,
+       .emit_frame_size =
+               SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+               SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+               8 + /* vcn_v2_0_jpeg_ring_emit_vm_flush */
+               18 + 18 + /* vcn_v2_0_jpeg_ring_emit_fence x2 vm fence */
+               8 + 16,
+       .emit_ib_size = 22, /* vcn_v2_0_jpeg_ring_emit_ib */
+       .emit_ib = vcn_v2_0_jpeg_ring_emit_ib,
+       .emit_fence = vcn_v2_0_jpeg_ring_emit_fence,
+       .emit_vm_flush = vcn_v2_0_jpeg_ring_emit_vm_flush,
+       .test_ring = amdgpu_vcn_jpeg_ring_test_ring,
+       .test_ib = amdgpu_vcn_jpeg_ring_test_ib,
+       .insert_nop = vcn_v2_0_jpeg_ring_nop,
+       .insert_start = vcn_v2_0_jpeg_ring_insert_start,
+       .insert_end = vcn_v2_0_jpeg_ring_insert_end,
+       .pad_ib = amdgpu_ring_generic_pad_ib,
+       .begin_use = amdgpu_vcn_ring_begin_use,
+       .end_use = amdgpu_vcn_ring_end_use,
+       .emit_wreg = vcn_v2_0_jpeg_ring_emit_wreg,
+       .emit_reg_wait = vcn_v2_0_jpeg_ring_emit_reg_wait,
+       .emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
+};
+
+static void vcn_v2_0_set_dec_ring_funcs(struct amdgpu_device *adev)
+{
+       adev->vcn.ring_dec.funcs = &vcn_v2_0_dec_ring_vm_funcs;
+       DRM_INFO("VCN decode is enabled in VM mode\n");
+}
+
+static void vcn_v2_0_set_enc_ring_funcs(struct amdgpu_device *adev)
+{
+       int i;
+
+       for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+               adev->vcn.ring_enc[i].funcs = &vcn_v2_0_enc_ring_vm_funcs;
+
+       DRM_INFO("VCN encode is enabled in VM mode\n");
+}
+
+static void vcn_v2_0_set_jpeg_ring_funcs(struct amdgpu_device *adev)
+{
+       adev->vcn.ring_jpeg.funcs = &vcn_v2_0_jpeg_ring_vm_funcs;
+       DRM_INFO("VCN jpeg decode is enabled in VM mode\n");
+}
+
+static const struct amdgpu_irq_src_funcs vcn_v2_0_irq_funcs = {
+       .set = vcn_v2_0_set_interrupt_state,
+       .process = vcn_v2_0_process_interrupt,
+};
+
+static void vcn_v2_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+       adev->vcn.irq.num_types = adev->vcn.num_enc_rings + 2;
+       adev->vcn.irq.funcs = &vcn_v2_0_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version vcn_v2_0_ip_block =
+{
+               .type = AMD_IP_BLOCK_TYPE_VCN,
+               .major = 2,
+               .minor = 0,
+               .rev = 0,
+               .funcs = &vcn_v2_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h b/drivers/gpu/drm/amd/amdgpu/vcn_v2_0.h
new file mode 100644 (file)
index 0000000..a74227f
--- /dev/null
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2018 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __VCN_V2_0_H__
+#define __VCN_V2_0_H__
+
+extern const struct amdgpu_ip_block_version vcn_v2_0_ip_block;
+
+#endif /* __VCN_V2_0_H__ */