drm/amdgpu/vcn1.0: use its own idle handler and begin use funcs

Because VCN1.0 power management and DPG mode are managed together with
JPEG1.0 under both HW and FW, so separated them from general VCN code.
Also the multiple instances case got removed, since VCN1.0 HW just have
a single instance.

v2: override work func with vcn1.0's own
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c

@@ -39,9 +39,6 @@
 #include "vcn/vcn_1_0_offset.h"
 #include "vcn/vcn_1_0_sh_mask.h"
 
-/* 1 second timeout */
-#define VCN_IDLE_TIMEOUT	msecs_to_jiffies(1000)
-
 /* Firmware Names */
 #define FIRMWARE_RAVEN		"amdgpu/raven_vcn.bin"
 #define FIRMWARE_PICASSO	"amdgpu/picasso_vcn.bin"

drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h

@@ -56,6 +56,9 @@
 #define VCN_VID_IP_ADDRESS_2_0		0x0
 #define VCN_AON_IP_ADDRESS_2_0		0x30000
 
+/* 1 second timeout */
+#define VCN_IDLE_TIMEOUT	msecs_to_jiffies(1000)
+
 #define RREG32_SOC15_DPG_MODE(ip, inst, reg, mask, sram_sel) 				\
 	({	WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); 			\
 		WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, 				\

drivers/gpu/drm/amd/amdgpu/jpeg_v1_0.c

@@ -25,6 +25,7 @@
 #include "amdgpu_jpeg.h"
 #include "soc15.h"
 #include "soc15d.h"
+#include "vcn_v1_0.h"
 
 #include "vcn/vcn_1_0_offset.h"
 #include "vcn/vcn_1_0_sh_mask.h"
@@ -561,7 +562,7 @@ static const struct amdgpu_ring_funcs jpeg_v1_0_decode_ring_vm_funcs = {
 	.insert_start = jpeg_v1_0_decode_ring_insert_start,
 	.insert_end = jpeg_v1_0_decode_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
+	.begin_use = vcn_v1_0_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
 	.emit_wreg = jpeg_v1_0_decode_ring_emit_wreg,
 	.emit_reg_wait = jpeg_v1_0_decode_ring_emit_reg_wait,

drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c

@@ -25,6 +25,7 @@
 
 #include "amdgpu.h"
 #include "amdgpu_vcn.h"
+#include "amdgpu_pm.h"
 #include "soc15.h"
 #include "soc15d.h"
 #include "soc15_common.h"
@@ -51,6 +52,8 @@ static int vcn_v1_0_set_powergating_state(void *handle, enum amd_powergating_sta
 static int vcn_v1_0_pause_dpg_mode(struct amdgpu_device *adev,
 				struct dpg_pause_state *new_state);
 
+static void vcn_v1_0_idle_work_handler(struct work_struct *work);
+
 /**
  * vcn_v1_0_early_init - set function pointers
  *
@@ -105,6 +108,9 @@ static int vcn_v1_0_sw_init(void *handle)
 	if (r)
 		return r;
 
+	/* Override the work func */
+	adev->vcn.idle_work.work.func = vcn_v1_0_idle_work_handler;
+
 	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;
@@ -1758,6 +1764,86 @@ static int vcn_v1_0_set_powergating_state(void *handle,
 	return ret;
 }
 
+static void vcn_v1_0_idle_work_handler(struct work_struct *work)
+{
+	struct amdgpu_device *adev =
+		container_of(work, struct amdgpu_device, vcn.idle_work.work);
+	unsigned int fences = 0, i;
+
+	for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+		fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]);
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		struct dpg_pause_state new_state;
+
+		if (fences)
+			new_state.fw_based = VCN_DPG_STATE__PAUSE;
+		else
+			new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+
+		if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec))
+			new_state.jpeg = VCN_DPG_STATE__PAUSE;
+		else
+			new_state.jpeg = VCN_DPG_STATE__UNPAUSE;
+
+		adev->vcn.pause_dpg_mode(adev, &new_state);
+	}
+
+	fences += amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec);
+	fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_dec);
+
+	if (fences == 0) {
+		amdgpu_gfx_off_ctrl(adev, true);
+		if (adev->pm.dpm_enabled)
+			amdgpu_dpm_enable_uvd(adev, false);
+		else
+			amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
+			       AMD_PG_STATE_GATE);
+	} else {
+		schedule_delayed_work(&adev->vcn.idle_work, VCN_IDLE_TIMEOUT);
+	}
+}
+
+void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	bool set_clocks = !cancel_delayed_work_sync(&adev->vcn.idle_work);
+
+	if (set_clocks) {
+		amdgpu_gfx_off_ctrl(adev, false);
+		if (adev->pm.dpm_enabled)
+			amdgpu_dpm_enable_uvd(adev, true);
+		else
+			amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
+			       AMD_PG_STATE_UNGATE);
+	}
+
+	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
+		struct dpg_pause_state new_state;
+		unsigned int fences = 0, i;
+
+		for (i = 0; i < adev->vcn.num_enc_rings; ++i)
+			fences += amdgpu_fence_count_emitted(&adev->vcn.inst->ring_enc[i]);
+
+		if (fences)
+			new_state.fw_based = VCN_DPG_STATE__PAUSE;
+		else
+			new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
+
+		if (amdgpu_fence_count_emitted(&adev->jpeg.inst->ring_dec))
+			new_state.jpeg = VCN_DPG_STATE__PAUSE;
+		else
+			new_state.jpeg = VCN_DPG_STATE__UNPAUSE;
+
+		if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC)
+			new_state.fw_based = VCN_DPG_STATE__PAUSE;
+		else if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG)
+			new_state.jpeg = VCN_DPG_STATE__PAUSE;
+
+		adev->vcn.pause_dpg_mode(adev, &new_state);
+	}
+}
+
 static const struct amd_ip_funcs vcn_v1_0_ip_funcs = {
 	.name = "vcn_v1_0",
 	.early_init = vcn_v1_0_early_init,
@@ -1804,7 +1890,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
 	.insert_start = vcn_v1_0_dec_ring_insert_start,
 	.insert_end = vcn_v1_0_dec_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
+	.begin_use = vcn_v1_0_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
 	.emit_wreg = vcn_v1_0_dec_ring_emit_wreg,
 	.emit_reg_wait = vcn_v1_0_dec_ring_emit_reg_wait,
@@ -1836,7 +1922,7 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
 	.insert_nop = amdgpu_ring_insert_nop,
 	.insert_end = vcn_v1_0_enc_ring_insert_end,
 	.pad_ib = amdgpu_ring_generic_pad_ib,
-	.begin_use = amdgpu_vcn_ring_begin_use,
+	.begin_use = vcn_v1_0_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
 	.emit_wreg = vcn_v1_0_enc_ring_emit_wreg,
 	.emit_reg_wait = vcn_v1_0_enc_ring_emit_reg_wait,

drivers/gpu/drm/amd/amdgpu/vcn_v1_0.h

@@ -24,6 +24,8 @@
 #ifndef __VCN_V1_0_H__
 #define __VCN_V1_0_H__
 
+void vcn_v1_0_ring_begin_use(struct amdgpu_ring *ring);
+
 extern const struct amdgpu_ip_block_version vcn_v1_0_ip_block;
 
 #endif