[PATCH 050/100] drm/amdgpu: add initial uvd 7.0 support for vega10
Alex Deucher
alexdeucher at gmail.com
Mon Mar 20 20:29:45 UTC 2017
From: Leo Liu <leo.liu at amd.com>
Signed-off-by: Leo Liu <leo.liu at amd.com>
Reviewed-by: Alex Deucher <alexander.deucher at amd.com>
Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
---
drivers/gpu/drm/amd/amdgpu/Makefile | 3 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c | 52 +-
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c | 1543 +++++++++++++++++++++++++++++++
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.h | 29 +
4 files changed, 1615 insertions(+), 12 deletions(-)
create mode 100644 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
create mode 100644 drivers/gpu/drm/amd/amdgpu/uvd_v7_0.h
diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index bc29569..65829fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -84,7 +84,8 @@ amdgpu-y += \
amdgpu-y += \
amdgpu_uvd.o \
uvd_v5_0.o \
- uvd_v6_0.o
+ uvd_v6_0.o \
+ uvd_v7_0.o
# add VCE block
amdgpu-y += \
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
index 02b8613..b2e1d3b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c
@@ -67,6 +67,14 @@
#define FIRMWARE_POLARIS11 "amdgpu/polaris11_uvd.bin"
#define FIRMWARE_POLARIS12 "amdgpu/polaris12_uvd.bin"
+#define FIRMWARE_VEGA10 "amdgpu/vega10_uvd.bin"
+
+#define mmUVD_GPCOM_VCPU_DATA0_VEGA10 (0x03c4 + 0x7e00)
+#define mmUVD_GPCOM_VCPU_DATA1_VEGA10 (0x03c5 + 0x7e00)
+#define mmUVD_GPCOM_VCPU_CMD_VEGA10 (0x03c3 + 0x7e00)
+#define mmUVD_NO_OP_VEGA10 (0x03ff + 0x7e00)
+#define mmUVD_ENGINE_CNTL_VEGA10 (0x03c6 + 0x7e00)
+
/**
* amdgpu_uvd_cs_ctx - Command submission parser context
*
@@ -101,6 +109,8 @@ MODULE_FIRMWARE(FIRMWARE_POLARIS10);
MODULE_FIRMWARE(FIRMWARE_POLARIS11);
MODULE_FIRMWARE(FIRMWARE_POLARIS12);
+MODULE_FIRMWARE(FIRMWARE_VEGA10);
+
static void amdgpu_uvd_idle_work_handler(struct work_struct *work);
int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
@@ -151,6 +161,9 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
case CHIP_POLARIS11:
fw_name = FIRMWARE_POLARIS11;
break;
+ case CHIP_VEGA10:
+ fw_name = FIRMWARE_VEGA10;
+ break;
case CHIP_POLARIS12:
fw_name = FIRMWARE_POLARIS12;
break;
@@ -203,9 +216,11 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
DRM_ERROR("POLARIS10/11 UVD firmware version %hu.%hu is too old.\n",
version_major, version_minor);
- bo_size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8)
- + AMDGPU_UVD_STACK_SIZE + AMDGPU_UVD_HEAP_SIZE
+ bo_size = AMDGPU_UVD_STACK_SIZE + AMDGPU_UVD_HEAP_SIZE
+ AMDGPU_UVD_SESSION_SIZE * adev->uvd.max_handles;
+ if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
+ bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
+
r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
AMDGPU_GEM_DOMAIN_VRAM, &adev->uvd.vcpu_bo,
&adev->uvd.gpu_addr, &adev->uvd.cpu_addr);
@@ -319,11 +334,13 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev)
unsigned offset;
hdr = (const struct common_firmware_header *)adev->uvd.fw->data;
- offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
- memcpy_toio(adev->uvd.cpu_addr, adev->uvd.fw->data + offset,
- le32_to_cpu(hdr->ucode_size_bytes));
- size -= le32_to_cpu(hdr->ucode_size_bytes);
- ptr += le32_to_cpu(hdr->ucode_size_bytes);
+ if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) {
+ offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
+ memcpy_toio(adev->uvd.cpu_addr, adev->uvd.fw->data + offset,
+ le32_to_cpu(hdr->ucode_size_bytes));
+ size -= le32_to_cpu(hdr->ucode_size_bytes);
+ ptr += le32_to_cpu(hdr->ucode_size_bytes);
+ }
memset_io(ptr, 0, size);
}
@@ -936,6 +953,7 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
struct fence *f = NULL;
struct amdgpu_device *adev = ring->adev;
uint64_t addr;
+ uint32_t data[4];
int i, r;
memset(&tv, 0, sizeof(tv));
@@ -961,16 +979,28 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
if (r)
goto err;
+ if (adev->asic_type >= CHIP_VEGA10) {
+ data[0] = PACKET0(mmUVD_GPCOM_VCPU_DATA0_VEGA10, 0);
+ data[1] = PACKET0(mmUVD_GPCOM_VCPU_DATA1_VEGA10, 0);
+ data[2] = PACKET0(mmUVD_GPCOM_VCPU_CMD_VEGA10, 0);
+ data[3] = PACKET0(mmUVD_NO_OP_VEGA10, 0);
+ } else {
+ data[0] = PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0);
+ data[1] = PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0);
+ data[2] = PACKET0(mmUVD_GPCOM_VCPU_CMD, 0);
+ data[3] = PACKET0(mmUVD_NO_OP, 0);
+ }
+
ib = &job->ibs[0];
addr = amdgpu_bo_gpu_offset(bo);
- ib->ptr[0] = PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0);
+ ib->ptr[0] = data[0];
ib->ptr[1] = addr;
- ib->ptr[2] = PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0);
+ ib->ptr[2] = data[1];
ib->ptr[3] = addr >> 32;
- ib->ptr[4] = PACKET0(mmUVD_GPCOM_VCPU_CMD, 0);
+ ib->ptr[4] = data[2];
ib->ptr[5] = 0;
for (i = 6; i < 16; i += 2) {
- ib->ptr[i] = PACKET0(mmUVD_NO_OP, 0);
+ ib->ptr[i] = data[3];
ib->ptr[i+1] = 0;
}
ib->length_dw = 16;
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
new file mode 100644
index 0000000..3457546
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -0,0 +1,1543 @@
+/*
+ * Copyright 2016 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_uvd.h"
+#include "soc15d.h"
+#include "soc15_common.h"
+
+#include "vega10/soc15ip.h"
+#include "vega10/UVD/uvd_7_0_offset.h"
+#include "vega10/UVD/uvd_7_0_sh_mask.h"
+#include "vega10/NBIF/nbif_6_1_offset.h"
+#include "vega10/HDP/hdp_4_0_offset.h"
+#include "vega10/MMHUB/mmhub_1_0_offset.h"
+#include "vega10/MMHUB/mmhub_1_0_sh_mask.h"
+
+static void uvd_v7_0_set_ring_funcs(struct amdgpu_device *adev);
+static void uvd_v7_0_set_enc_ring_funcs(struct amdgpu_device *adev);
+static void uvd_v7_0_set_irq_funcs(struct amdgpu_device *adev);
+static int uvd_v7_0_start(struct amdgpu_device *adev);
+static void uvd_v7_0_stop(struct amdgpu_device *adev);
+
+/**
+ * uvd_v7_0_ring_get_rptr - get read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware read pointer
+ */
+static uint64_t uvd_v7_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_RPTR));
+}
+
+/**
+ * uvd_v7_0_enc_ring_get_rptr - get enc read pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc read pointer
+ */
+static uint64_t uvd_v7_0_enc_ring_get_rptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring == &adev->uvd.ring_enc[0])
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_RPTR));
+ else
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_RPTR2));
+}
+
+/**
+ * uvd_v7_0_ring_get_wptr - get write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware write pointer
+ */
+static uint64_t uvd_v7_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_WPTR));
+}
+
+/**
+ * uvd_v7_0_enc_ring_get_wptr - get enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Returns the current hardware enc write pointer
+ */
+static uint64_t uvd_v7_0_enc_ring_get_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring == &adev->uvd.ring_enc[0])
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR));
+ else
+ return RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR2));
+}
+
+/**
+ * uvd_v7_0_ring_set_wptr - set write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the write pointer to the hardware
+ */
+static void uvd_v7_0_ring_set_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_WPTR), lower_32_bits(ring->wptr));
+}
+
+/**
+ * uvd_v7_0_enc_ring_set_wptr - set enc write pointer
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Commits the enc write pointer to the hardware
+ */
+static void uvd_v7_0_enc_ring_set_wptr(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+
+ if (ring == &adev->uvd.ring_enc[0])
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR),
+ lower_32_bits(ring->wptr));
+ else
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR2),
+ lower_32_bits(ring->wptr));
+}
+
+/**
+ * uvd_v7_0_enc_ring_test_ring - test if UVD ENC ring is working
+ *
+ * @ring: the engine to test on
+ *
+ */
+static int uvd_v7_0_enc_ring_test_ring(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+ uint32_t rptr = amdgpu_ring_get_rptr(ring);
+ unsigned i;
+ int r;
+
+ r = amdgpu_ring_alloc(ring, 16);
+ if (r) {
+ DRM_ERROR("amdgpu: uvd enc failed to lock ring %d (%d).\n",
+ ring->idx, r);
+ return r;
+ }
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_END);
+ amdgpu_ring_commit(ring);
+
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (amdgpu_ring_get_rptr(ring) != rptr)
+ break;
+ DRM_UDELAY(1);
+ }
+
+ if (i < adev->usec_timeout) {
+ DRM_INFO("ring test on %d succeeded in %d usecs\n",
+ ring->idx, i);
+ } else {
+ DRM_ERROR("amdgpu: ring %d test failed\n",
+ ring->idx);
+ r = -ETIMEDOUT;
+ }
+
+ return r;
+}
+
+/**
+ * uvd_v7_0_enc_get_create_msg - generate a UVD ENC create msg
+ *
+ * @adev: amdgpu_device pointer
+ * @ring: ring we should submit the msg to
+ * @handle: session handle to use
+ * @fence: optional fence to return
+ *
+ * Open up a stream for HW test
+ */
+static int uvd_v7_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle,
+ struct fence **fence)
+{
+ const unsigned ib_size_dw = 16;
+ struct amdgpu_job *job;
+ struct amdgpu_ib *ib;
+ struct fence *f = NULL;
+ uint64_t dummy;
+ int i, r;
+
+ r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job);
+ if (r)
+ return r;
+
+ ib = &job->ibs[0];
+ dummy = ib->gpu_addr + 1024;
+
+ ib->length_dw = 0;
+ ib->ptr[ib->length_dw++] = 0x00000018;
+ ib->ptr[ib->length_dw++] = 0x00000001; /* session info */
+ ib->ptr[ib->length_dw++] = handle;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+ ib->ptr[ib->length_dw++] = upper_32_bits(dummy);
+ ib->ptr[ib->length_dw++] = dummy;
+
+ ib->ptr[ib->length_dw++] = 0x00000014;
+ ib->ptr[ib->length_dw++] = 0x00000002; /* task info */
+ ib->ptr[ib->length_dw++] = 0x0000001c;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+
+ ib->ptr[ib->length_dw++] = 0x00000008;
+ ib->ptr[ib->length_dw++] = 0x08000001; /* op initialize */
+
+ for (i = ib->length_dw; i < ib_size_dw; ++i)
+ ib->ptr[i] = 0x0;
+
+ r = amdgpu_ib_schedule(ring, 1, ib, NULL, &f);
+ job->fence = fence_get(f);
+ if (r)
+ goto err;
+
+ amdgpu_job_free(job);
+ if (fence)
+ *fence = fence_get(f);
+ fence_put(f);
+ return 0;
+
+err:
+ amdgpu_job_free(job);
+ return r;
+}
+
+/**
+ * uvd_v7_0_enc_get_destroy_msg - generate a UVD ENC destroy msg
+ *
+ * @adev: amdgpu_device pointer
+ * @ring: ring we should submit the msg to
+ * @handle: session handle to use
+ * @fence: optional fence to return
+ *
+ * Close up a stream for HW test or if userspace failed to do so
+ */
+int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
+ bool direct, struct fence **fence)
+{
+ const unsigned ib_size_dw = 16;
+ struct amdgpu_job *job;
+ struct amdgpu_ib *ib;
+ struct fence *f = NULL;
+ uint64_t dummy;
+ int i, r;
+
+ r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job);
+ if (r)
+ return r;
+
+ ib = &job->ibs[0];
+ dummy = ib->gpu_addr + 1024;
+
+ ib->length_dw = 0;
+ ib->ptr[ib->length_dw++] = 0x00000018;
+ ib->ptr[ib->length_dw++] = 0x00000001;
+ ib->ptr[ib->length_dw++] = handle;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+ ib->ptr[ib->length_dw++] = upper_32_bits(dummy);
+ ib->ptr[ib->length_dw++] = dummy;
+
+ ib->ptr[ib->length_dw++] = 0x00000014;
+ ib->ptr[ib->length_dw++] = 0x00000002;
+ ib->ptr[ib->length_dw++] = 0x0000001c;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+ ib->ptr[ib->length_dw++] = 0x00000000;
+
+ ib->ptr[ib->length_dw++] = 0x00000008;
+ ib->ptr[ib->length_dw++] = 0x08000002; /* op close session */
+
+ for (i = ib->length_dw; i < ib_size_dw; ++i)
+ ib->ptr[i] = 0x0;
+
+ if (direct) {
+ r = amdgpu_ib_schedule(ring, 1, ib, NULL, &f);
+ job->fence = fence_get(f);
+ if (r)
+ goto err;
+
+ amdgpu_job_free(job);
+ } else {
+ r = amdgpu_job_submit(job, ring, &ring->adev->vce.entity,
+ AMDGPU_FENCE_OWNER_UNDEFINED, &f);
+ if (r)
+ goto err;
+ }
+
+ if (fence)
+ *fence = fence_get(f);
+ fence_put(f);
+ return 0;
+
+err:
+ amdgpu_job_free(job);
+ return r;
+}
+
+/**
+ * uvd_v7_0_enc_ring_test_ib - test if UVD ENC IBs are working
+ *
+ * @ring: the engine to test on
+ *
+ */
+static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+ struct fence *fence = NULL;
+ long r;
+
+ r = uvd_v7_0_enc_get_create_msg(ring, 1, NULL);
+ if (r) {
+ DRM_ERROR("amdgpu: failed to get create msg (%ld).\n", r);
+ goto error;
+ }
+
+ r = uvd_v7_0_enc_get_destroy_msg(ring, 1, true, &fence);
+ if (r) {
+ DRM_ERROR("amdgpu: failed to get destroy ib (%ld).\n", r);
+ goto error;
+ }
+
+ r = fence_wait_timeout(fence, false, timeout);
+ if (r == 0) {
+ DRM_ERROR("amdgpu: IB test timed out.\n");
+ r = -ETIMEDOUT;
+ } else if (r < 0) {
+ DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
+ } else {
+ DRM_INFO("ib test on ring %d succeeded\n", ring->idx);
+ r = 0;
+ }
+error:
+ fence_put(fence);
+ return r;
+}
+
+static int uvd_v7_0_early_init(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ adev->uvd.num_enc_rings = 2;
+ uvd_v7_0_set_ring_funcs(adev);
+ uvd_v7_0_set_enc_ring_funcs(adev);
+ uvd_v7_0_set_irq_funcs(adev);
+
+ return 0;
+}
+
+static int uvd_v7_0_sw_init(void *handle)
+{
+ struct amdgpu_ring *ring;
+ struct amd_sched_rq *rq;
+ int i, r;
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ /* UVD TRAP */
+ r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_UVD, 124, &adev->uvd.irq);
+ if (r)
+ return r;
+
+ /* UVD ENC TRAP */
+ for (i = 0; i < adev->uvd.num_enc_rings; ++i) {
+ r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_UVD, i + 119, &adev->uvd.irq);
+ if (r)
+ return r;
+ }
+
+ r = amdgpu_uvd_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->uvd.fw->data;
+ adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].ucode_id = AMDGPU_UCODE_ID_UVD;
+ adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].fw = adev->uvd.fw;
+ adev->firmware.fw_size +=
+ ALIGN(le32_to_cpu(hdr->ucode_size_bytes), PAGE_SIZE);
+ DRM_INFO("PSP loading UVD firmware\n");
+ }
+
+ ring = &adev->uvd.ring_enc[0];
+ rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL];
+ r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity_enc,
+ rq, amdgpu_sched_jobs);
+ if (r) {
+ DRM_ERROR("Failed setting up UVD ENC run queue.\n");
+ return r;
+ }
+
+ r = amdgpu_uvd_resume(adev);
+ if (r)
+ return r;
+
+ ring = &adev->uvd.ring;
+ sprintf(ring->name, "uvd");
+ r = amdgpu_ring_init(adev, ring, 512, &adev->uvd.irq, 0);
+ if (r)
+ return r;
+
+ for (i = 0; i < adev->uvd.num_enc_rings; ++i) {
+ ring = &adev->uvd.ring_enc[i];
+ sprintf(ring->name, "uvd_enc%d", i);
+ r = amdgpu_ring_init(adev, ring, 512, &adev->uvd.irq, 0);
+ if (r)
+ return r;
+ }
+
+ return r;
+}
+
+static int uvd_v7_0_sw_fini(void *handle)
+{
+ int i, r;
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ r = amdgpu_uvd_suspend(adev);
+ if (r)
+ return r;
+
+ amd_sched_entity_fini(&adev->uvd.ring_enc[0].sched, &adev->uvd.entity_enc);
+
+ for (i = 0; i < adev->uvd.num_enc_rings; ++i)
+ amdgpu_ring_fini(&adev->uvd.ring_enc[i]);
+
+ r = amdgpu_uvd_sw_fini(adev);
+ if (r)
+ return r;
+
+ return r;
+}
+
+/**
+ * uvd_v7_0_hw_init - start and test UVD block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Initialize the hardware, boot up the VCPU and do some testing
+ */
+static int uvd_v7_0_hw_init(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_ring *ring = &adev->uvd.ring;
+ uint32_t tmp;
+ int i, r;
+
+ r = uvd_v7_0_start(adev);
+ if (r)
+ goto done;
+
+ ring->ready = true;
+ r = amdgpu_ring_test_ring(ring);
+ if (r) {
+ ring->ready = false;
+ goto done;
+ }
+
+ r = amdgpu_ring_alloc(ring, 10);
+ if (r) {
+ DRM_ERROR("amdgpu: ring failed to lock UVD ring (%d).\n", r);
+ goto done;
+ }
+
+ tmp = PACKET0(SOC15_REG_OFFSET(UVD, 0,
+ mmUVD_SEMA_WAIT_FAULT_TIMEOUT_CNTL), 0);
+ amdgpu_ring_write(ring, tmp);
+ amdgpu_ring_write(ring, 0xFFFFF);
+
+ tmp = PACKET0(SOC15_REG_OFFSET(UVD, 0,
+ mmUVD_SEMA_WAIT_INCOMPLETE_TIMEOUT_CNTL), 0);
+ amdgpu_ring_write(ring, tmp);
+ amdgpu_ring_write(ring, 0xFFFFF);
+
+ tmp = PACKET0(SOC15_REG_OFFSET(UVD, 0,
+ mmUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL), 0);
+ amdgpu_ring_write(ring, tmp);
+ amdgpu_ring_write(ring, 0xFFFFF);
+
+ /* Clear timeout status bits */
+ amdgpu_ring_write(ring, PACKET0(SOC15_REG_OFFSET(UVD, 0,
+ mmUVD_SEMA_TIMEOUT_STATUS), 0));
+ amdgpu_ring_write(ring, 0x8);
+
+ amdgpu_ring_write(ring, PACKET0(SOC15_REG_OFFSET(UVD, 0,
+ mmUVD_SEMA_CNTL), 0));
+ amdgpu_ring_write(ring, 3);
+
+ amdgpu_ring_commit(ring);
+
+ for (i = 0; i < adev->uvd.num_enc_rings; ++i) {
+ ring = &adev->uvd.ring_enc[i];
+ ring->ready = true;
+ r = amdgpu_ring_test_ring(ring);
+ if (r) {
+ ring->ready = false;
+ goto done;
+ }
+ }
+
+done:
+ if (!r)
+ DRM_INFO("UVD and UVD ENC initialized successfully.\n");
+
+ return r;
+}
+
+/**
+ * uvd_v7_0_hw_fini - stop the hardware block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Stop the UVD block, mark ring as not ready any more
+ */
+static int uvd_v7_0_hw_fini(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ struct amdgpu_ring *ring = &adev->uvd.ring;
+
+ uvd_v7_0_stop(adev);
+ ring->ready = false;
+
+ return 0;
+}
+
+static int uvd_v7_0_suspend(void *handle)
+{
+ int r;
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ r = uvd_v7_0_hw_fini(adev);
+ if (r)
+ return r;
+
+ /* Skip this for APU for now */
+ if (!(adev->flags & AMD_IS_APU)) {
+ r = amdgpu_uvd_suspend(adev);
+ if (r)
+ return r;
+ }
+
+ return r;
+}
+
+static int uvd_v7_0_resume(void *handle)
+{
+ int r;
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ /* Skip this for APU for now */
+ if (!(adev->flags & AMD_IS_APU)) {
+ r = amdgpu_uvd_resume(adev);
+ if (r)
+ return r;
+ }
+ r = uvd_v7_0_hw_init(adev);
+ if (r)
+ return r;
+
+ return r;
+}
+
+/**
+ * uvd_v7_0_mc_resume - memory controller programming
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Let the UVD memory controller know it's offsets
+ */
+static void uvd_v7_0_mc_resume(struct amdgpu_device *adev)
+{
+ uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->uvd.fw->size + 4);
+ uint32_t offset;
+
+ if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+ lower_32_bits(adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].mc_addr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+ upper_32_bits(adev->firmware.ucode[AMDGPU_UCODE_ID_UVD].mc_addr));
+ offset = 0;
+ } else {
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW),
+ lower_32_bits(adev->uvd.gpu_addr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH),
+ upper_32_bits(adev->uvd.gpu_addr));
+ offset = size;
+ }
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0),
+ AMDGPU_UVD_FIRMWARE_OFFSET >> 3);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_SIZE0), size);
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW),
+ lower_32_bits(adev->uvd.gpu_addr + offset));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH),
+ upper_32_bits(adev->uvd.gpu_addr + offset));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1), (1 << 21));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_SIZE1), AMDGPU_UVD_HEAP_SIZE);
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW),
+ lower_32_bits(adev->uvd.gpu_addr + offset + AMDGPU_UVD_HEAP_SIZE));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH),
+ upper_32_bits(adev->uvd.gpu_addr + offset + AMDGPU_UVD_HEAP_SIZE));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2), (2 << 21));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CACHE_SIZE2),
+ AMDGPU_UVD_STACK_SIZE + (AMDGPU_UVD_SESSION_SIZE * 40));
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_UDEC_ADDR_CONFIG),
+ adev->gfx.config.gb_addr_config);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG),
+ adev->gfx.config.gb_addr_config);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG),
+ adev->gfx.config.gb_addr_config);
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_GP_SCRATCH4), adev->uvd.max_handles);
+}
+
+/**
+ * uvd_v7_0_start - start UVD block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup and start the UVD block
+ */
+static int uvd_v7_0_start(struct amdgpu_device *adev)
+{
+ struct amdgpu_ring *ring = &adev->uvd.ring;
+ uint32_t rb_bufsz, tmp;
+ uint32_t lmi_swap_cntl;
+ uint32_t mp_swap_cntl;
+ int i, j, r;
+
+ /* disable DPG */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0,
+ ~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
+
+ /* disable byte swapping */
+ lmi_swap_cntl = 0;
+ mp_swap_cntl = 0;
+
+ uvd_v7_0_mc_resume(adev);
+
+ /* disable clock gating */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_CTRL), 0,
+ ~UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK);
+
+ /* disable interupt */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
+ ~UVD_MASTINT_EN__VCPU_EN_MASK);
+
+ /* stall UMC and register bus before resetting VCPU */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
+ UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
+ ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+ mdelay(1);
+
+ /* put LMI, VCPU, RBC etc... into reset */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+ UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__RBC_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__CXW_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
+ UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
+ mdelay(5);
+
+ /* initialize UVD memory controller */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL),
+ (0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
+ UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
+ UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
+ UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
+ UVD_LMI_CTRL__REQ_MODE_MASK |
+ 0x00100000L);
+
+#ifdef __BIG_ENDIAN
+ /* swap (8 in 32) RB and IB */
+ lmi_swap_cntl = 0xa;
+ mp_swap_cntl = 0;
+#endif
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_SWAP_CNTL), lmi_swap_cntl);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MP_SWAP_CNTL), mp_swap_cntl);
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_MUXA0), 0x40c2040);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_MUXA1), 0x0);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_MUXB0), 0x40c2040);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_MUXB1), 0x0);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_ALU), 0);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_MPC_SET_MUX), 0x88);
+
+ /* take all subblocks out of reset, except VCPU */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+ UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+ mdelay(5);
+
+ /* enable VCPU clock */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL),
+ UVD_VCPU_CNTL__CLK_EN_MASK);
+
+ /* enable UMC */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
+ ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+
+ /* boot up the VCPU */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0);
+ mdelay(10);
+
+ for (i = 0; i < 10; ++i) {
+ uint32_t status;
+
+ for (j = 0; j < 100; ++j) {
+ status = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS));
+ if (status & 2)
+ break;
+ mdelay(10);
+ }
+ r = 0;
+ if (status & 2)
+ break;
+
+ DRM_ERROR("UVD 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("UVD 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__SYS_EN_MASK),
+ ~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK));
+
+ /* clear the bit 4 of UVD_STATUS */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0,
+ ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
+
+ /* 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_WPTR_POLL_EN, 0);
+ 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_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), tmp);
+
+ /* set the write pointer delay */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_WPTR_CNTL), 0);
+
+ /* set the wb address */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_RPTR_ADDR),
+ (upper_32_bits(ring->gpu_addr) >> 2));
+
+ /* programm the RB_BASE for ring buffer */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW),
+ lower_32_bits(ring->gpu_addr));
+ WREG32(SOC15_REG_OFFSET(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_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_RPTR), 0);
+
+ ring->wptr = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_RPTR));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_WPTR),
+ lower_32_bits(ring->wptr));
+
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0,
+ ~UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK);
+
+ ring = &adev->uvd.ring_enc[0];
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_RPTR), lower_32_bits(ring->wptr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR), lower_32_bits(ring->wptr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_BASE_LO), ring->gpu_addr);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_BASE_HI), upper_32_bits(ring->gpu_addr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_SIZE), ring->ring_size / 4);
+
+ ring = &adev->uvd.ring_enc[1];
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_RPTR2), lower_32_bits(ring->wptr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_WPTR2), lower_32_bits(ring->wptr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_BASE_LO2), ring->gpu_addr);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_BASE_HI2), upper_32_bits(ring->gpu_addr));
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RB_SIZE2), ring->ring_size / 4);
+
+ return 0;
+}
+
+/**
+ * uvd_v7_0_stop - stop UVD block
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * stop the UVD block
+ */
+static void uvd_v7_0_stop(struct amdgpu_device *adev)
+{
+ /* force RBC into idle state */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0x11010101);
+
+ /* Stall UMC and register bus before resetting VCPU */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
+ UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
+ ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+ mdelay(1);
+
+ /* put VCPU into reset */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
+ UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
+ mdelay(5);
+
+ /* disable VCPU clock */
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0x0);
+
+ /* Unstall UMC and register bus */
+ WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
+ ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
+}
+
+/**
+ * uvd_v7_0_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 uvd_v7_0_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(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0));
+ amdgpu_ring_write(ring, seq);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
+ amdgpu_ring_write(ring, addr & 0xffffffff);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
+ amdgpu_ring_write(ring, upper_32_bits(addr) & 0xff);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
+ amdgpu_ring_write(ring, 0);
+
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
+ amdgpu_ring_write(ring, 0);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
+ amdgpu_ring_write(ring, 0);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
+ amdgpu_ring_write(ring, 2);
+}
+
+/**
+ * uvd_v7_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 uvd_v7_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, HEVC_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, HEVC_ENC_CMD_TRAP);
+}
+
+/**
+ * uvd_v7_0_ring_emit_hdp_flush - emit an hdp flush
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp flush.
+ */
+static void uvd_v7_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+ amdgpu_ring_write(ring, PACKET0(SOC15_REG_OFFSET(NBIF, 0,
+ mmHDP_MEM_COHERENCY_FLUSH_CNTL), 0));
+ amdgpu_ring_write(ring, 0);
+}
+
+/**
+ * uvd_v7_0_ring_hdp_invalidate - emit an hdp invalidate
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Emits an hdp invalidate.
+ */
+static void uvd_v7_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
+{
+ amdgpu_ring_write(ring, PACKET0(SOC15_REG_OFFSET(HDP, 0, mmHDP_DEBUG0), 0));
+ amdgpu_ring_write(ring, 1);
+}
+
+/**
+ * uvd_v7_0_ring_test_ring - register write test
+ *
+ * @ring: amdgpu_ring pointer
+ *
+ * Test if we can successfully write to the context register
+ */
+static int uvd_v7_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+ struct amdgpu_device *adev = ring->adev;
+ uint32_t tmp = 0;
+ unsigned i;
+ int r;
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0xCAFEDEAD);
+ r = amdgpu_ring_alloc(ring, 3);
+ if (r) {
+ DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n",
+ ring->idx, r);
+ return r;
+ }
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0));
+ amdgpu_ring_write(ring, 0xDEADBEEF);
+ amdgpu_ring_commit(ring);
+ for (i = 0; i < adev->usec_timeout; i++) {
+ tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID));
+ if (tmp == 0xDEADBEEF)
+ break;
+ DRM_UDELAY(1);
+ }
+
+ if (i < adev->usec_timeout) {
+ DRM_INFO("ring test on %d succeeded in %d usecs\n",
+ ring->idx, i);
+ } else {
+ DRM_ERROR("amdgpu: ring %d test failed (0x%08X)\n",
+ ring->idx, tmp);
+ r = -EINVAL;
+ }
+ return r;
+}
+
+/**
+ * uvd_v7_0_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 uvd_v7_0_ring_emit_ib(struct amdgpu_ring *ring,
+ struct amdgpu_ib *ib,
+ unsigned vm_id, bool ctx_switch)
+{
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_VMID), 0));
+ amdgpu_ring_write(ring, vm_id);
+
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_LOW), 0));
+ amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr));
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH), 0));
+ amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr));
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_IB_SIZE), 0));
+ amdgpu_ring_write(ring, ib->length_dw);
+}
+
+/**
+ * uvd_v7_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 uvd_v7_0_enc_ring_emit_ib(struct amdgpu_ring *ring,
+ struct amdgpu_ib *ib, unsigned int vm_id, bool ctx_switch)
+{
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_IB_VM);
+ amdgpu_ring_write(ring, vm_id);
+ 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 uvd_v7_0_vm_reg_write(struct amdgpu_ring *ring,
+ uint32_t data0, uint32_t data1)
+{
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
+ amdgpu_ring_write(ring, data0);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
+ amdgpu_ring_write(ring, data1);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
+ amdgpu_ring_write(ring, 8);
+}
+
+static void uvd_v7_0_vm_reg_wait(struct amdgpu_ring *ring,
+ uint32_t data0, uint32_t data1, uint32_t mask)
+{
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA0), 0));
+ amdgpu_ring_write(ring, data0);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_DATA1), 0));
+ amdgpu_ring_write(ring, data1);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GP_SCRATCH8), 0));
+ amdgpu_ring_write(ring, mask);
+ amdgpu_ring_write(ring,
+ PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_GPCOM_VCPU_CMD), 0));
+ amdgpu_ring_write(ring, 12);
+}
+
+static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+ unsigned vm_id, uint64_t pd_addr)
+{
+ uint32_t data0, data1, mask;
+ unsigned eng = ring->idx;
+ unsigned i;
+
+ pd_addr = pd_addr | 0x1; /* valid bit */
+ /* now only use physical base address of PDE and valid */
+ BUG_ON(pd_addr & 0xFFFF00000000003EULL);
+
+ for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+ struct amdgpu_vmhub *hub = &ring->adev->vmhub[i];
+ uint32_t req = hub->get_invalidate_req(vm_id);
+
+ data0 = (hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2;
+ data1 = upper_32_bits(pd_addr);
+ uvd_v7_0_vm_reg_write(ring, data0, data1);
+
+ data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+ data1 = lower_32_bits(pd_addr);
+ uvd_v7_0_vm_reg_write(ring, data0, data1);
+
+ data0 = (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2;
+ data1 = lower_32_bits(pd_addr);
+ mask = 0xffffffff;
+ uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
+
+ /* flush TLB */
+ data0 = (hub->vm_inv_eng0_req + eng) << 2;
+ data1 = req;
+ uvd_v7_0_vm_reg_write(ring, data0, data1);
+
+ /* wait for flush */
+ data0 = (hub->vm_inv_eng0_ack + eng) << 2;
+ data1 = 1 << vm_id;
+ mask = 1 << vm_id;
+ uvd_v7_0_vm_reg_wait(ring, data0, data1, mask);
+ }
+}
+
+static void uvd_v7_0_enc_ring_insert_end(struct amdgpu_ring *ring)
+{
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_END);
+}
+
+static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+ unsigned int vm_id, uint64_t pd_addr)
+{
+ unsigned eng = ring->idx;
+ unsigned i;
+
+ pd_addr = pd_addr | 0x1; /* valid bit */
+ /* now only use physical base address of PDE and valid */
+ BUG_ON(pd_addr & 0xFFFF00000000003EULL);
+
+ for (i = 0; i < AMDGPU_MAX_VMHUBS; ++i) {
+ struct amdgpu_vmhub *hub = &ring->adev->vmhub[i];
+ uint32_t req = hub->get_invalidate_req(vm_id);
+
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WRITE);
+ amdgpu_ring_write(ring,
+ (hub->ctx0_ptb_addr_hi32 + vm_id * 2) << 2);
+ amdgpu_ring_write(ring, upper_32_bits(pd_addr));
+
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WRITE);
+ amdgpu_ring_write(ring,
+ (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+ amdgpu_ring_write(ring, lower_32_bits(pd_addr));
+
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
+ amdgpu_ring_write(ring,
+ (hub->ctx0_ptb_addr_lo32 + vm_id * 2) << 2);
+ amdgpu_ring_write(ring, 0xffffffff);
+ amdgpu_ring_write(ring, lower_32_bits(pd_addr));
+
+ /* flush TLB */
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WRITE);
+ amdgpu_ring_write(ring, (hub->vm_inv_eng0_req + eng) << 2);
+ amdgpu_ring_write(ring, req);
+
+ /* wait for flush */
+ amdgpu_ring_write(ring, HEVC_ENC_CMD_REG_WAIT);
+ amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
+ amdgpu_ring_write(ring, 1 << vm_id);
+ amdgpu_ring_write(ring, 1 << vm_id);
+ }
+}
+
+#if 0
+static bool uvd_v7_0_is_idle(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ return !(RREG32(mmSRBM_STATUS) & SRBM_STATUS__UVD_BUSY_MASK);
+}
+
+static int uvd_v7_0_wait_for_idle(void *handle)
+{
+ unsigned i;
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ for (i = 0; i < adev->usec_timeout; i++) {
+ if (uvd_v7_0_is_idle(handle))
+ return 0;
+ }
+ return -ETIMEDOUT;
+}
+
+#define AMDGPU_UVD_STATUS_BUSY_MASK 0xfd
+static bool uvd_v7_0_check_soft_reset(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ u32 srbm_soft_reset = 0;
+ u32 tmp = RREG32(mmSRBM_STATUS);
+
+ if (REG_GET_FIELD(tmp, SRBM_STATUS, UVD_RQ_PENDING) ||
+ REG_GET_FIELD(tmp, SRBM_STATUS, UVD_BUSY) ||
+ (RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS) &
+ AMDGPU_UVD_STATUS_BUSY_MASK)))
+ srbm_soft_reset = REG_SET_FIELD(srbm_soft_reset,
+ SRBM_SOFT_RESET, SOFT_RESET_UVD, 1);
+
+ if (srbm_soft_reset) {
+ adev->uvd.srbm_soft_reset = srbm_soft_reset;
+ return true;
+ } else {
+ adev->uvd.srbm_soft_reset = 0;
+ return false;
+ }
+}
+
+static int uvd_v7_0_pre_soft_reset(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ if (!adev->uvd.srbm_soft_reset)
+ return 0;
+
+ uvd_v7_0_stop(adev);
+ return 0;
+}
+
+static int uvd_v7_0_soft_reset(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+ u32 srbm_soft_reset;
+
+ if (!adev->uvd.srbm_soft_reset)
+ return 0;
+ srbm_soft_reset = adev->uvd.srbm_soft_reset;
+
+ if (srbm_soft_reset) {
+ u32 tmp;
+
+ tmp = RREG32(mmSRBM_SOFT_RESET);
+ tmp |= srbm_soft_reset;
+ dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+ WREG32(mmSRBM_SOFT_RESET, tmp);
+ tmp = RREG32(mmSRBM_SOFT_RESET);
+
+ udelay(50);
+
+ tmp &= ~srbm_soft_reset;
+ WREG32(mmSRBM_SOFT_RESET, tmp);
+ tmp = RREG32(mmSRBM_SOFT_RESET);
+
+ /* Wait a little for things to settle down */
+ udelay(50);
+ }
+
+ return 0;
+}
+
+static int uvd_v7_0_post_soft_reset(void *handle)
+{
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ if (!adev->uvd.srbm_soft_reset)
+ return 0;
+
+ mdelay(5);
+
+ return uvd_v7_0_start(adev);
+}
+#endif
+
+static int uvd_v7_0_set_interrupt_state(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ unsigned type,
+ enum amdgpu_interrupt_state state)
+{
+ // TODO
+ return 0;
+}
+
+static int uvd_v7_0_process_interrupt(struct amdgpu_device *adev,
+ struct amdgpu_irq_src *source,
+ struct amdgpu_iv_entry *entry)
+{
+ DRM_DEBUG("IH: UVD TRAP\n");
+ switch (entry->src_id) {
+ case 124:
+ amdgpu_fence_process(&adev->uvd.ring);
+ break;
+ case 119:
+ amdgpu_fence_process(&adev->uvd.ring_enc[0]);
+ break;
+ case 120:
+ amdgpu_fence_process(&adev->uvd.ring_enc[1]);
+ break;
+ default:
+ DRM_ERROR("Unhandled interrupt: %d %d\n",
+ entry->src_id, entry->src_data[0]);
+ break;
+ }
+
+ return 0;
+}
+
+#if 0
+static void uvd_v7_0_set_sw_clock_gating(struct amdgpu_device *adev)
+{
+ uint32_t data, data1, data2, suvd_flags;
+
+ data = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_CTRL));
+ data1 = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_GATE));
+ data2 = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_CTRL));
+
+ data &= ~(UVD_CGC_CTRL__CLK_OFF_DELAY_MASK |
+ UVD_CGC_CTRL__CLK_GATE_DLY_TIMER_MASK);
+
+ suvd_flags = 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;
+
+ data |= UVD_CGC_CTRL__DYN_CLOCK_MODE_MASK |
+ (1 << REG_FIELD_SHIFT(UVD_CGC_CTRL, CLK_GATE_DLY_TIMER)) |
+ (4 << REG_FIELD_SHIFT(UVD_CGC_CTRL, CLK_OFF_DELAY));
+
+ 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__JPEG_MODE_MASK |
+ UVD_CGC_CTRL__JPEG2_MODE_MASK |
+ UVD_CGC_CTRL__SCPU_MODE_MASK);
+ data2 &= ~(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);
+ data1 |= suvd_flags;
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_CTRL), data);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_GATE), 0);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_GATE), data1);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_CTRL), data2);
+}
+
+static void uvd_v7_0_set_hw_clock_gating(struct amdgpu_device *adev)
+{
+ uint32_t data, data1, cgc_flags, suvd_flags;
+
+ data = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_GATE));
+ data1 = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_GATE));
+
+ cgc_flags = UVD_CGC_GATE__SYS_MASK |
+ UVD_CGC_GATE__UDEC_MASK |
+ UVD_CGC_GATE__MPEG2_MASK |
+ UVD_CGC_GATE__RBC_MASK |
+ UVD_CGC_GATE__LMI_MC_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 |
+ UVD_CGC_GATE__JPEG_MASK |
+ UVD_CGC_GATE__JPEG2_MASK;
+
+ suvd_flags = 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;
+
+ data |= cgc_flags;
+ data1 |= suvd_flags;
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CGC_GATE), data);
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SUVD_CGC_GATE), data1);
+}
+
+static void uvd_v7_0_set_bypass_mode(struct amdgpu_device *adev, bool enable)
+{
+ u32 tmp = RREG32_SMC(ixGCK_DFS_BYPASS_CNTL);
+
+ if (enable)
+ tmp |= (GCK_DFS_BYPASS_CNTL__BYPASSDCLK_MASK |
+ GCK_DFS_BYPASS_CNTL__BYPASSVCLK_MASK);
+ else
+ tmp &= ~(GCK_DFS_BYPASS_CNTL__BYPASSDCLK_MASK |
+ GCK_DFS_BYPASS_CNTL__BYPASSVCLK_MASK);
+
+ WREG32_SMC(ixGCK_DFS_BYPASS_CNTL, tmp);
+}
+
+
+static int uvd_v7_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;
+
+ uvd_v7_0_set_bypass_mode(adev, enable);
+
+ if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG))
+ return 0;
+
+ if (enable) {
+ /* disable HW gating and enable Sw gating */
+ uvd_v7_0_set_sw_clock_gating(adev);
+ } else {
+ /* wait for STATUS to clear */
+ if (uvd_v7_0_wait_for_idle(handle))
+ return -EBUSY;
+
+ /* enable HW gates because UVD is idle */
+ /* uvd_v7_0_set_hw_clock_gating(adev); */
+ }
+
+ return 0;
+}
+
+static int uvd_v7_0_set_powergating_state(void *handle,
+ enum amd_powergating_state state)
+{
+ /* This doesn't actually powergate the UVD 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
+ */
+ struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+ if (!(adev->pg_flags & AMD_PG_SUPPORT_UVD))
+ return 0;
+
+ WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), UVD_POWER_STATUS__UVD_PG_EN_MASK);
+
+ if (state == AMD_PG_STATE_GATE) {
+ uvd_v7_0_stop(adev);
+ return 0;
+ } else {
+ return uvd_v7_0_start(adev);
+ }
+}
+#endif
+
+static int uvd_v7_0_set_clockgating_state(void *handle,
+ enum amd_clockgating_state state)
+{
+ /* needed for driver unload*/
+ return 0;
+}
+
+const struct amd_ip_funcs uvd_v7_0_ip_funcs = {
+ .name = "uvd_v7_0",
+ .early_init = uvd_v7_0_early_init,
+ .late_init = NULL,
+ .sw_init = uvd_v7_0_sw_init,
+ .sw_fini = uvd_v7_0_sw_fini,
+ .hw_init = uvd_v7_0_hw_init,
+ .hw_fini = uvd_v7_0_hw_fini,
+ .suspend = uvd_v7_0_suspend,
+ .resume = uvd_v7_0_resume,
+ .is_idle = NULL /* uvd_v7_0_is_idle */,
+ .wait_for_idle = NULL /* uvd_v7_0_wait_for_idle */,
+ .check_soft_reset = NULL /* uvd_v7_0_check_soft_reset */,
+ .pre_soft_reset = NULL /* uvd_v7_0_pre_soft_reset */,
+ .soft_reset = NULL /* uvd_v7_0_soft_reset */,
+ .post_soft_reset = NULL /* uvd_v7_0_post_soft_reset */,
+ .set_clockgating_state = uvd_v7_0_set_clockgating_state,
+ .set_powergating_state = NULL /* uvd_v7_0_set_powergating_state */,
+};
+
+static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
+ .type = AMDGPU_RING_TYPE_UVD,
+ .align_mask = 0xf,
+ .nop = PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_NO_OP), 0),
+ .support_64bit_ptrs = false,
+ .get_rptr = uvd_v7_0_ring_get_rptr,
+ .get_wptr = uvd_v7_0_ring_get_wptr,
+ .set_wptr = uvd_v7_0_ring_set_wptr,
+ .emit_frame_size =
+ 2 + /* uvd_v7_0_ring_emit_hdp_flush */
+ 2 + /* uvd_v7_0_ring_emit_hdp_invalidate */
+ 34 * AMDGPU_MAX_VMHUBS + /* uvd_v7_0_ring_emit_vm_flush */
+ 14 + 14, /* uvd_v7_0_ring_emit_fence x2 vm fence */
+ .emit_ib_size = 8, /* uvd_v7_0_ring_emit_ib */
+ .emit_ib = uvd_v7_0_ring_emit_ib,
+ .emit_fence = uvd_v7_0_ring_emit_fence,
+ .emit_vm_flush = uvd_v7_0_ring_emit_vm_flush,
+ .emit_hdp_flush = uvd_v7_0_ring_emit_hdp_flush,
+ .emit_hdp_invalidate = uvd_v7_0_ring_emit_hdp_invalidate,
+ .test_ring = uvd_v7_0_ring_test_ring,
+ .test_ib = amdgpu_uvd_ring_test_ib,
+ .insert_nop = amdgpu_ring_insert_nop,
+ .pad_ib = amdgpu_ring_generic_pad_ib,
+ .begin_use = amdgpu_uvd_ring_begin_use,
+ .end_use = amdgpu_uvd_ring_end_use,
+};
+
+static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
+ .type = AMDGPU_RING_TYPE_UVD_ENC,
+ .align_mask = 0x3f,
+ .nop = HEVC_ENC_CMD_NO_OP,
+ .support_64bit_ptrs = false,
+ .get_rptr = uvd_v7_0_enc_ring_get_rptr,
+ .get_wptr = uvd_v7_0_enc_ring_get_wptr,
+ .set_wptr = uvd_v7_0_enc_ring_set_wptr,
+ .emit_frame_size =
+ 17 * AMDGPU_MAX_VMHUBS + /* uvd_v7_0_enc_ring_emit_vm_flush */
+ 5 + 5 + /* uvd_v7_0_enc_ring_emit_fence x2 vm fence */
+ 1, /* uvd_v7_0_enc_ring_insert_end */
+ .emit_ib_size = 5, /* uvd_v7_0_enc_ring_emit_ib */
+ .emit_ib = uvd_v7_0_enc_ring_emit_ib,
+ .emit_fence = uvd_v7_0_enc_ring_emit_fence,
+ .emit_vm_flush = uvd_v7_0_enc_ring_emit_vm_flush,
+ .test_ring = uvd_v7_0_enc_ring_test_ring,
+ .test_ib = uvd_v7_0_enc_ring_test_ib,
+ .insert_nop = amdgpu_ring_insert_nop,
+ .insert_end = uvd_v7_0_enc_ring_insert_end,
+ .pad_ib = amdgpu_ring_generic_pad_ib,
+ .begin_use = amdgpu_uvd_ring_begin_use,
+ .end_use = amdgpu_uvd_ring_end_use,
+};
+
+static void uvd_v7_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+ adev->uvd.ring.funcs = &uvd_v7_0_ring_vm_funcs;
+ DRM_INFO("UVD is enabled in VM mode\n");
+}
+
+static void uvd_v7_0_set_enc_ring_funcs(struct amdgpu_device *adev)
+{
+ int i;
+
+ for (i = 0; i < adev->uvd.num_enc_rings; ++i)
+ adev->uvd.ring_enc[i].funcs = &uvd_v7_0_enc_ring_vm_funcs;
+
+ DRM_INFO("UVD ENC is enabled in VM mode\n");
+}
+
+static const struct amdgpu_irq_src_funcs uvd_v7_0_irq_funcs = {
+ .set = uvd_v7_0_set_interrupt_state,
+ .process = uvd_v7_0_process_interrupt,
+};
+
+static void uvd_v7_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+ adev->uvd.irq.num_types = adev->uvd.num_enc_rings + 1;
+ adev->uvd.irq.funcs = &uvd_v7_0_irq_funcs;
+}
+
+const struct amdgpu_ip_block_version uvd_v7_0_ip_block =
+{
+ .type = AMD_IP_BLOCK_TYPE_UVD,
+ .major = 7,
+ .minor = 0,
+ .rev = 0,
+ .funcs = &uvd_v7_0_ip_funcs,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.h b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.h
new file mode 100644
index 0000000..cbe82ab
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2016 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 __UVD_V7_0_H__
+#define __UVD_V7_0_H__
+
+extern const struct amdgpu_ip_block_version uvd_v7_0_ip_block;
+
+#endif
--
2.5.5
More information about the amd-gfx
mailing list