Merge drm/drm-next into drm-misc-next

Backmerging to get a late RC of v6.10 before moving into v6.11.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>

1 files changed, 5299 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
new file mode 100644
index 000000000000..f384be0d1800
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -0,0 +1,5299 @@
+/*
+ * Copyright 2023 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/delay.h>
+#include <linux/kernel.h>
+#include <linux/firmware.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include "amdgpu.h"
+#include "amdgpu_gfx.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_smu.h"
+#include "amdgpu_atomfirmware.h"
+#include "imu_v12_0.h"
+#include "soc24.h"
+#include "nvd.h"
+
+#include "gc/gc_12_0_0_offset.h"
+#include "gc/gc_12_0_0_sh_mask.h"
+#include "soc24_enum.h"
+#include "ivsrcid/gfx/irqsrcs_gfx_11_0_0.h"
+
+#include "soc15.h"
+#include "soc15d.h"
+#include "clearstate_gfx12.h"
+#include "v12_structs.h"
+#include "gfx_v12_0.h"
+#include "nbif_v6_3_1.h"
+#include "mes_v12_0.h"
+
+#define GFX12_NUM_GFX_RINGS	1
+#define GFX12_MEC_HPD_SIZE	2048
+
+#define RLCG_UCODE_LOADING_START_ADDRESS	0x00002000L
+
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_0_toc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_pfp.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_me.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_mec.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_rlc.bin");
+MODULE_FIRMWARE("amdgpu/gc_12_0_1_toc.bin");
+
+static const struct amdgpu_hwip_reg_entry gc_reg_list_12_0[] = {
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_STALLED_STAT3),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STALLED_STAT1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_BUSY_STAT2),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPF_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HPD_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB0_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB2_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPF_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPC_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCPG_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regIA_UTCL1_STATUS_2),
+	SOC15_REG_ENTRY_STR(GC, 0, regPA_CL_CNTL_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regRMI_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQC_CACHES),
+	SOC15_REG_ENTRY_STR(GC, 0, regSQG_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regWD_UTCL1_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS_LO32),
+	SOC15_REG_ENTRY_STR(GC, 0, regGCVM_L2_PROTECTION_FAULT_STATUS_HI32),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_DEBUG),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_INSTR_PNTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_CPC_STATUS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_RS64_INSTR_PNTR1),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_RS64_INSTR_PNTR),
+
+	/* cp header registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MEC_ME1_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_PFP_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_ME_HEADER_DUMP),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_MES_HEADER_DUMP),
+	/* SE status registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE0),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE1),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE2),
+	SOC15_REG_ENTRY_STR(GC, 0, regGRBM_STATUS_SE3)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_cp_reg_list_12[] = {
+	/* compute registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PERSISTENT_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PIPE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_IB_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_EVENTS),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CNTL_STACK_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_CTX_SAVE_SIZE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_GDS_RESOURCE_STATE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_ERROR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_EOP_WPTR_MEM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_PQ_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_CNTL_STACK_DW_CNT),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_SUSPEND_WG_STATE_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_HQD_DEQUEUE_STATUS)
+};
+
+static const struct amdgpu_hwip_reg_entry gc_gfx_queue_reg_list_12[] = {
+	/* gfx queue registers */
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_ACTIVE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_VMID),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUANTUM),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_OFFSET),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CNTL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_CSMD_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_WPTR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_DEQUEUE_REQUEST),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_MAPPED),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_QUE_MGR_CONTROL),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_CONTROL0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_HQD_HQ_STATUS0),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_GFX_MQD_BASE_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_WPTR_POLL_ADDR_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_RB_RPTR),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_LO),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BASE_HI),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_CMD_BUFSZ),
+	SOC15_REG_ENTRY_STR(GC, 0, regCP_IB1_BUFSZ)
+};
+
+#define DEFAULT_SH_MEM_CONFIG \
+	((SH_MEM_ADDRESS_MODE_64 << SH_MEM_CONFIG__ADDRESS_MODE__SHIFT) | \
+	 (SH_MEM_ALIGNMENT_MODE_UNALIGNED << SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) | \
+	 (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT))
+
+static void gfx_v12_0_disable_gpa_mode(struct amdgpu_device *adev);
+static void gfx_v12_0_set_ring_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_irq_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_rlc_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_mqd_funcs(struct amdgpu_device *adev);
+static void gfx_v12_0_set_imu_funcs(struct amdgpu_device *adev);
+static int gfx_v12_0_get_cu_info(struct amdgpu_device *adev,
+				 struct amdgpu_cu_info *cu_info);
+static uint64_t gfx_v12_0_get_gpu_clock_counter(struct amdgpu_device *adev);
+static void gfx_v12_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				   u32 sh_num, u32 instance, int xcc_id);
+static u32 gfx_v12_0_get_wgp_active_bitmap_per_sh(struct amdgpu_device *adev);
+
+static void gfx_v12_0_ring_emit_frame_cntl(struct amdgpu_ring *ring, bool start, bool secure);
+static void gfx_v12_0_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				     uint32_t val);
+static int gfx_v12_0_wait_for_rlc_autoload_complete(struct amdgpu_device *adev);
+static void gfx_v12_0_ring_invalidate_tlbs(struct amdgpu_ring *ring,
+					   uint16_t pasid, uint32_t flush_type,
+					   bool all_hub, uint8_t dst_sel);
+static void gfx_v12_0_set_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v12_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id);
+static void gfx_v12_0_update_perf_clk(struct amdgpu_device *adev,
+				      bool enable);
+
+static void gfx_v12_0_kiq_set_resources(struct amdgpu_ring *kiq_ring,
+					uint64_t queue_mask)
+{
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_SET_RESOURCES, 6));
+	amdgpu_ring_write(kiq_ring, PACKET3_SET_RESOURCES_VMID_MASK(0) |
+			  PACKET3_SET_RESOURCES_QUEUE_TYPE(0));	/* vmid_mask:0 queue_type:0 (KIQ) */
+	amdgpu_ring_write(kiq_ring, lower_32_bits(queue_mask));	/* queue mask lo */
+	amdgpu_ring_write(kiq_ring, upper_32_bits(queue_mask));	/* queue mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask lo */
+	amdgpu_ring_write(kiq_ring, 0);	/* gws mask hi */
+	amdgpu_ring_write(kiq_ring, 0);	/* oac mask */
+	amdgpu_ring_write(kiq_ring, 0);
+}
+
+static void gfx_v12_0_kiq_map_queues(struct amdgpu_ring *kiq_ring,
+				     struct amdgpu_ring *ring)
+{
+	uint64_t mqd_addr = amdgpu_bo_gpu_offset(ring->mqd_obj);
+	uint64_t wptr_addr = ring->wptr_gpu_addr;
+	uint32_t me = 0, eng_sel = 0;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_COMPUTE:
+		me = 1;
+		eng_sel = 0;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+		me = 0;
+		eng_sel = 4;
+		break;
+	case AMDGPU_RING_TYPE_MES:
+		me = 2;
+		eng_sel = 5;
+		break;
+	default:
+		WARN_ON(1);
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	/* Q_sel:0, vmid:0, vidmem: 1, engine:0, num_Q:1*/
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(0) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(ring->queue) |
+			  PACKET3_MAP_QUEUES_PIPE(ring->pipe) |
+			  PACKET3_MAP_QUEUES_ME((me)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring, PACKET3_MAP_QUEUES_DOORBELL_OFFSET(ring->doorbell_index));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(mqd_addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(wptr_addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr));
+}
+
+static void gfx_v12_0_kiq_unmap_queues(struct amdgpu_ring *kiq_ring,
+				       struct amdgpu_ring *ring,
+				       enum amdgpu_unmap_queues_action action,
+				       u64 gpu_addr, u64 seq)
+{
+	struct amdgpu_device *adev = kiq_ring->adev;
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	if (adev->enable_mes && !adev->gfx.kiq[0].ring.sched.ready) {
+		amdgpu_mes_unmap_legacy_queue(adev, ring, action, gpu_addr, seq);
+		return;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_UNMAP_QUEUES, 4));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_UNMAP_QUEUES_ACTION(action) |
+			  PACKET3_UNMAP_QUEUES_QUEUE_SEL(0) |
+			  PACKET3_UNMAP_QUEUES_ENGINE_SEL(eng_sel) |
+			  PACKET3_UNMAP_QUEUES_NUM_QUEUES(1));
+	amdgpu_ring_write(kiq_ring,
+		  PACKET3_UNMAP_QUEUES_DOORBELL_OFFSET0(ring->doorbell_index));
+
+	if (action == PREEMPT_QUEUES_NO_UNMAP) {
+		amdgpu_ring_write(kiq_ring, lower_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, upper_32_bits(gpu_addr));
+		amdgpu_ring_write(kiq_ring, seq);
+	} else {
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+		amdgpu_ring_write(kiq_ring, 0);
+	}
+}
+
+static void gfx_v12_0_kiq_query_status(struct amdgpu_ring *kiq_ring,
+				       struct amdgpu_ring *ring,
+				       u64 addr, u64 seq)
+{
+	uint32_t eng_sel = ring->funcs->type == AMDGPU_RING_TYPE_GFX ? 4 : 0;
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_QUERY_STATUS, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_QUERY_STATUS_CONTEXT_ID(0) |
+			  PACKET3_QUERY_STATUS_INTERRUPT_SEL(0) |
+			  PACKET3_QUERY_STATUS_COMMAND(2));
+	amdgpu_ring_write(kiq_ring, /* Q_sel: 0, vmid: 0, engine: 0, num_Q: 1 */
+			  PACKET3_QUERY_STATUS_DOORBELL_OFFSET(ring->doorbell_index) |
+			  PACKET3_QUERY_STATUS_ENG_SEL(eng_sel));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(addr));
+	amdgpu_ring_write(kiq_ring, lower_32_bits(seq));
+	amdgpu_ring_write(kiq_ring, upper_32_bits(seq));
+}
+
+static void gfx_v12_0_kiq_invalidate_tlbs(struct amdgpu_ring *kiq_ring,
+					  uint16_t pasid,
+					  uint32_t flush_type,
+					  bool all_hub)
+{
+	gfx_v12_0_ring_invalidate_tlbs(kiq_ring, pasid, flush_type, all_hub, 1);
+}
+
+static const struct kiq_pm4_funcs gfx_v12_0_kiq_pm4_funcs = {
+	.kiq_set_resources = gfx_v12_0_kiq_set_resources,
+	.kiq_map_queues = gfx_v12_0_kiq_map_queues,
+	.kiq_unmap_queues = gfx_v12_0_kiq_unmap_queues,
+	.kiq_query_status = gfx_v12_0_kiq_query_status,
+	.kiq_invalidate_tlbs = gfx_v12_0_kiq_invalidate_tlbs,
+	.set_resources_size = 8,
+	.map_queues_size = 7,
+	.unmap_queues_size = 6,
+	.query_status_size = 7,
+	.invalidate_tlbs_size = 2,
+};
+
+static void gfx_v12_0_set_kiq_pm4_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.kiq[0].pmf = &gfx_v12_0_kiq_pm4_funcs;
+}
+
+static void gfx_v12_0_wait_reg_mem(struct amdgpu_ring *ring, int eng_sel,
+				   int mem_space, int opt, uint32_t addr0,
+				   uint32_t addr1, uint32_t ref,
+				   uint32_t mask, uint32_t inv)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring,
+			  /* memory (1) or register (0) */
+			  (WAIT_REG_MEM_MEM_SPACE(mem_space) |
+			   WAIT_REG_MEM_OPERATION(opt) | /* wait */
+			   WAIT_REG_MEM_FUNCTION(3) |  /* equal */
+			   WAIT_REG_MEM_ENGINE(eng_sel)));
+
+	if (mem_space)
+		BUG_ON(addr0 & 0x3); /* Dword align */
+	amdgpu_ring_write(ring, addr0);
+	amdgpu_ring_write(ring, addr1);
+	amdgpu_ring_write(ring, ref);
+	amdgpu_ring_write(ring, mask);
+	amdgpu_ring_write(ring, inv); /* poll interval */
+}
+
+static int gfx_v12_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t scratch = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	WREG32(scratch, 0xCAFEDEAD);
+	r = amdgpu_ring_alloc(ring, 5);
+	if (r) {
+		dev_err(adev->dev,
+			"amdgpu: cp failed to lock ring %d (%d).\n",
+			ring->idx, r);
+		return r;
+	}
+
+	if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ) {
+		gfx_v12_0_ring_emit_wreg(ring, scratch, 0xDEADBEEF);
+	} else {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+		amdgpu_ring_write(ring, scratch -
+				  PACKET3_SET_UCONFIG_REG_START);
+		amdgpu_ring_write(ring, 0xDEADBEEF);
+	}
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		if (amdgpu_emu_mode == 1)
+			msleep(1);
+		else
+			udelay(1);
+	}
+
+	if (i >= adev->usec_timeout)
+		r = -ETIMEDOUT;
+	return r;
+}
+
+static int gfx_v12_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct dma_fence *f = NULL;
+	unsigned index;
+	uint64_t gpu_addr;
+	volatile uint32_t *cpu_ptr;
+	long r;
+
+	/* MES KIQ fw hasn't indirect buffer support for now */
+	if (adev->enable_mes_kiq &&
+	    ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
+		return 0;
+
+	memset(&ib, 0, sizeof(ib));
+
+	if (ring->is_mes_queue) {
+		uint32_t padding, offset;
+
+		offset = amdgpu_mes_ctx_get_offs(ring, AMDGPU_MES_CTX_IB_OFFS);
+		padding = amdgpu_mes_ctx_get_offs(ring,
+						  AMDGPU_MES_CTX_PADDING_OFFS);
+
+		ib.gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, offset);
+		ib.ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, offset);
+
+		gpu_addr = amdgpu_mes_ctx_get_offs_gpu_addr(ring, padding);
+		cpu_ptr = amdgpu_mes_ctx_get_offs_cpu_addr(ring, padding);
+		*cpu_ptr = cpu_to_le32(0xCAFEDEAD);
+	} else {
+		r = amdgpu_device_wb_get(adev, &index);
+		if (r)
+			return r;
+
+		gpu_addr = adev->wb.gpu_addr + (index * 4);
+		adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD);
+		cpu_ptr = &adev->wb.wb[index];
+
+		r = amdgpu_ib_get(adev, NULL, 16, AMDGPU_IB_POOL_DIRECT, &ib);
+		if (r) {
+			dev_err(adev->dev, "amdgpu: failed to get ib (%ld).\n", r);
+			goto err1;
+		}
+	}
+
+	ib.ptr[0] = PACKET3(PACKET3_WRITE_DATA, 3);
+	ib.ptr[1] = WRITE_DATA_DST_SEL(5) | WR_CONFIRM;
+	ib.ptr[2] = lower_32_bits(gpu_addr);
+	ib.ptr[3] = upper_32_bits(gpu_addr);
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f);
+	if (r)
+		goto err2;
+
+	r = dma_fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		r = -ETIMEDOUT;
+		goto err2;
+	} else if (r < 0) {
+		goto err2;
+	}
+
+	if (le32_to_cpu(*cpu_ptr) == 0xDEADBEEF)
+		r = 0;
+	else
+		r = -EINVAL;
+err2:
+	if (!ring->is_mes_queue)
+		amdgpu_ib_free(adev, &ib, NULL);
+	dma_fence_put(f);
+err1:
+	if (!ring->is_mes_queue)
+		amdgpu_device_wb_free(adev, index);
+	return r;
+}
+
+static void gfx_v12_0_free_microcode(struct amdgpu_device *adev)
+{
+	amdgpu_ucode_release(&adev->gfx.pfp_fw);
+	amdgpu_ucode_release(&adev->gfx.me_fw);
+	amdgpu_ucode_release(&adev->gfx.rlc_fw);
+	amdgpu_ucode_release(&adev->gfx.mec_fw);
+
+	kfree(adev->gfx.rlc.register_list_format);
+}
+
+static int gfx_v12_0_init_toc_microcode(struct amdgpu_device *adev, const char *ucode_prefix)
+{
+	const struct psp_firmware_header_v1_0 *toc_hdr;
+	int err = 0;
+
+	err = amdgpu_ucode_request(adev, &adev->psp.toc_fw,
+				   "amdgpu/%s_toc.bin", ucode_prefix);
+	if (err)
+		goto out;
+
+	toc_hdr = (const struct psp_firmware_header_v1_0 *)adev->psp.toc_fw->data;
+	adev->psp.toc.fw_version = le32_to_cpu(toc_hdr->header.ucode_version);
+	adev->psp.toc.feature_version = le32_to_cpu(toc_hdr->sos.fw_version);
+	adev->psp.toc.size_bytes = le32_to_cpu(toc_hdr->header.ucode_size_bytes);
+	adev->psp.toc.start_addr = (uint8_t *)toc_hdr +
+			le32_to_cpu(toc_hdr->header.ucode_array_offset_bytes);
+	return 0;
+out:
+	amdgpu_ucode_release(&adev->psp.toc_fw);
+	return err;
+}
+
+static int gfx_v12_0_init_microcode(struct amdgpu_device *adev)
+{
+	char ucode_prefix[15];
+	int err;
+	const struct rlc_firmware_header_v2_0 *rlc_hdr;
+	uint16_t version_major;
+	uint16_t version_minor;
+
+	DRM_DEBUG("\n");
+
+	amdgpu_ucode_ip_version_decode(adev, GC_HWIP, ucode_prefix, sizeof(ucode_prefix));
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.pfp_fw,
+				   "amdgpu/%s_pfp.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_PFP_P0_STACK);
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.me_fw,
+				   "amdgpu/%s_me.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_ME_P0_STACK);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		err = amdgpu_ucode_request(adev, &adev->gfx.rlc_fw,
+					   "amdgpu/%s_rlc.bin", ucode_prefix);
+		if (err)
+			goto out;
+		rlc_hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
+		version_major = le16_to_cpu(rlc_hdr->header.header_version_major);
+		version_minor = le16_to_cpu(rlc_hdr->header.header_version_minor);
+		err = amdgpu_gfx_rlc_init_microcode(adev, version_major, version_minor);
+		if (err)
+			goto out;
+	}
+
+	err = amdgpu_ucode_request(adev, &adev->gfx.mec_fw,
+				   "amdgpu/%s_mec.bin", ucode_prefix);
+	if (err)
+		goto out;
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P0_STACK);
+	amdgpu_gfx_cp_init_microcode(adev, AMDGPU_UCODE_ID_CP_RS64_MEC_P1_STACK);
+
+	if (adev->firmware.load_type == AMDGPU_FW_LOAD_RLC_BACKDOOR_AUTO)
+		err = gfx_v12_0_init_toc_microcode(adev, ucode_prefix);
+
+	/* only one MEC for gfx 12 */
+	adev->gfx.mec2_fw = NULL;
+
+	if (adev->gfx.imu.funcs) {
+		if (adev->gfx.imu.funcs->init_microcode) {
+			err = adev->gfx.imu.funcs->init_microcode(adev);
+			if (err)
+				dev_err(adev->dev, "Failed to load imu firmware!\n");
+		}
+	}
+
+out:
+	if (err) {
+		amdgpu_ucode_release(&adev->gfx.pfp_fw);
+		amdgpu_ucode_release(&adev->gfx.me_fw);
+		amdgpu_ucode_release(&adev->gfx.rlc_fw);
+		amdgpu_ucode_release(&adev->gfx.mec_fw);
+	}
+
+	return err;
+}
+
+static u32 gfx_v12_0_get_csb_size(struct amdgpu_device *adev)
+{
+	u32 count = 0;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	count += 1;
+
+	for (sect = gfx12_cs_data; sect->section != NULL; ++sect) {
+		if (sect->id == SECT_CONTEXT) {
+			for (ext = sect->section; ext->extent != NULL; ++ext)
+				count += 2 + ext->reg_count;
+		} else
+			return 0;
+	}
+
+	return count;
+}
+
+static void gfx_v12_0_get_csb_buffer(struct amdgpu_device *adev,
+				     volatile u32 *buffer)
+{
+	u32 count = 0, clustercount = 0, i;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	if (adev->gfx.rlc.cs_data == NULL)
+		return;
+	if (buffer == NULL)
+		return;
+
+	count += 1;
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		if (sect->id == SECT_CONTEXT) {
+			for (ext = sect->section; ext->extent != NULL; ++ext) {
+				clustercount++;
+				buffer[count++] = ext->reg_count;
+				buffer[count++] = ext->reg_index;
+
+				for (i = 0; i < ext->reg_count; i++)
+					buffer[count++] = cpu_to_le32(ext->extent[i]);
+			}
+		} else
+			return;
+	}
+
+	buffer[0] = clustercount;
+}
+
+static void gfx_v12_0_rlc_fini(struct amdgpu_device *adev)
+{
+	/* clear state block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.clear_state_obj,
+			&adev->gfx.rlc.clear_state_gpu_addr,
+			(void **)&adev->gfx.rlc.cs_ptr);
+
+	/* jump table block */
+	amdgpu_bo_free_kernel(&adev->gfx.rlc.cp_table_obj,
+			&adev->gfx.rlc.cp_table_gpu_addr,
+			(void **)&adev->gfx.rlc.cp_table_ptr);
+}
+
+static void gfx_v12_0_init_rlcg_reg_access_ctrl(struct amdgpu_device *adev)
+{
+	struct amdgpu_rlcg_reg_access_ctrl *reg_access_ctrl;
+
+	reg_access_ctrl = &adev->gfx.rlc.reg_access_ctrl[0];
+	reg_access_ctrl->scratch_reg0 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG0);
+	reg_access_ctrl->scratch_reg1 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG1);
+	reg_access_ctrl->scratch_reg2 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG2);
+	reg_access_ctrl->scratch_reg3 = SOC15_REG_OFFSET(GC, 0, regSCRATCH_REG3);
+	reg_access_ctrl->grbm_cntl = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_CNTL);
+	reg_access_ctrl->grbm_idx = SOC15_REG_OFFSET(GC, 0, regGRBM_GFX_INDEX);
+	reg_access_ctrl->spare_int = SOC15_REG_OFFSET(GC, 0, regRLC_SPARE_INT_0);
+	adev->gfx.rlc.rlcg_reg_access_supported = true;
+}
+
+static int gfx_v12_0_rlc_init(struct amdgpu_device *adev)
+{
+	const struct cs_section_def *cs_data;
+	int r;
+
+	adev->gfx.rlc.cs_data = gfx12_cs_data;
+
+	cs_data = adev->gfx.rlc.cs_data;
+
+	if (cs_data) {
+		/* init clear state block */
+		r = amdgpu_gfx_rlc_init_csb(adev);
+		if (r)
+			return r;
+	}
+
+	/* init spm vmid with 0xf */
+	if (adev->gfx.rlc.funcs->update_spm_vmid)
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
+
+	return 0;
+}
+
+static void gfx_v12_0_mec_fini(struct amdgpu_device *adev)
+{
+	amdgpu_bo_free_kernel(&adev->gfx.mec.hpd_eop_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_obj, NULL, NULL);
+	amdgpu_bo_free_kernel(&adev->gfx.mec.mec_fw_data_obj, NULL, NULL);
+}
+
+static void gfx_v12_0_me_init(struct amdgpu_device *adev)
+{
+	bitmap_zero(adev->gfx.me.queue_bitmap, AMDGPU_MAX_GFX_QUEUES);
+
+	amdgpu_gfx_graphics_queue_acquire(adev);
+}
+
+static int gfx_v12_0_mec_init(struct amdgpu_device *adev)
+{
+	int r;
+	u32 *hpd;
+	size_t mec_hpd_size;
+
+	bitmap_zero(adev->gfx.mec_bitmap[0].queue_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
+
+	/* take ownership of the relevant compute queues */
+	amdgpu_gfx_compute_queue_acquire(adev);
+	mec_hpd_size = adev->gfx.num_compute_rings * GFX12_MEC_HPD_SIZE;
+
+	if (mec_hpd_size) {
+		r = amdgpu_bo_create_reserved(adev, mec_hpd_size, PAGE_SIZE,
+					      AMDGPU_GEM_DOMAIN_GTT,
+					      &adev->gfx.mec.hpd_eop_obj,
+					      &adev->gfx.mec.hpd_eop_gpu_addr,
+					      (void **)&hpd);
+		if (r) {
+			dev_warn(adev->dev, "(%d) create HDP EOP bo failed\n", r);
+			gfx_v12_0_mec_fini(adev);
+			return r;
+		}
+
+		memset(hpd, 0, mec_hpd_size);
+
+		amdgpu_bo_kunmap(adev->gfx.mec.hpd_eop_obj);
+		amdgpu_bo_unreserve(adev->gfx.mec.hpd_eop_obj);
+	}
+
+	return 0;
+}
+
+static uint32_t wave_read_ind(struct amdgpu_device *adev, uint32_t wave, uint32_t address)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(address << SQ_IND_INDEX__INDEX__SHIFT));
+	return RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void wave_read_regs(struct amdgpu_device *adev, uint32_t wave,
+			   uint32_t thread, uint32_t regno,
+			   uint32_t num, uint32_t *out)
+{
+	WREG32_SOC15(GC, 0, regSQ_IND_INDEX,
+		(wave << SQ_IND_INDEX__WAVE_ID__SHIFT) |
+		(regno << SQ_IND_INDEX__INDEX__SHIFT) |
+		(thread << SQ_IND_INDEX__WORKITEM_ID__SHIFT) |
+		(SQ_IND_INDEX__AUTO_INCR_MASK));
+	while (num--)
+		*(out++) = RREG32_SOC15(GC, 0, regSQ_IND_DATA);
+}
+
+static void gfx_v12_0_read_wave_data(struct amdgpu_device *adev,
+				     uint32_t xcc_id,
+				     uint32_t simd, uint32_t wave,
+				     uint32_t *dst, int *no_fields)
+{
+	/* in gfx12 the SIMD_ID is specified as part of the INSTANCE
+	 * field when performing a select_se_sh so it should be
+	 * zero here */
+	WARN_ON(simd != 0);
+
+	/* type 4 wave data */
+	dst[(*no_fields)++] = 4;
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_STATUS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_PC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXEC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_HW_ID2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_GPR_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_LDS_ALLOC);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_STS2);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_IB_DBG1);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_M0);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_MODE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_STATE_PRIV);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXCP_FLAG_PRIV);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_EXCP_FLAG_USER);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_TRAP_CTRL);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_ACTIVE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_VALID_AND_IDLE);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_DVGPR_ALLOC_LO);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_DVGPR_ALLOC_HI);
+	dst[(*no_fields)++] = wave_read_ind(adev, wave, ixSQ_WAVE_SCHED_MODE);
+}
+
+static void gfx_v12_0_read_wave_sgprs(struct amdgpu_device *adev,
+				      uint32_t xcc_id, uint32_t simd,
+				      uint32_t wave, uint32_t start,
+				      uint32_t size, uint32_t *dst)
+{
+	WARN_ON(simd != 0);
+
+	wave_read_regs(
+		adev, wave, 0, start + SQIND_WAVE_SGPRS_OFFSET, size,
+		dst);
+}
+
+static void gfx_v12_0_read_wave_vgprs(struct amdgpu_device *adev,
+				      uint32_t xcc_id, uint32_t simd,
+				      uint32_t wave, uint32_t thread,
+				      uint32_t start, uint32_t size,
+				      uint32_t *dst)
+{
+	wave_read_regs(
+		adev, wave, thread,
+		start + SQIND_WAVE_VGPRS_OFFSET, size, dst);
+}
+
+static void gfx_v12_0_select_me_pipe_q(struct amdgpu_device *adev,
+				       u32 me, u32 pipe, u32 q, u32 vm, u32 xcc_id)
+{
+	soc24_grbm_select(adev, me, pipe, q, vm);
+}
+
+static const struct amdgpu_gfx_funcs gfx_v12_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v12_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v12_0_select_se_sh,
+	.read_wave_data = &gfx_v12_0_read_wave_data,
+	.read_wave_sgprs = &gfx_v12_0_read_wave_sgprs,
+	.read_wave_vgprs = &gfx_v12_0_read_wave_vgprs,
+	.select_me_pipe_q = &gfx_v12_0_select_me_pipe_q,
+	.update_perfmon_mgcg = &gfx_v12_0_update_perf_clk,
+};
+
+static int gfx_v12_0_gpu_early_init(struct amdgpu_device *adev)
+{
+
+	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(12, 0, 0):
+	case IP_VERSION(12, 0, 1):
+		adev->gfx.config.max_hw_contexts = 8;
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x4C0;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v12_0_gfx_ring_init(struct amdgpu_device *adev, int ring_id,
+				   int me, int pipe, int queue)
+{
+	int r;
+	struct amdgpu_ring *ring;
+	unsigned int irq_type;
+
+	ring = &adev->gfx.gfx_ring[ring_id];
+
+	ring->me = me;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+
+	if (!ring_id)
+		ring->doorbell_index = adev->doorbell_index.gfx_ring0 << 1;
+	else
+		ring->doorbell_index = adev->doorbell_index.gfx_ring1 << 1;
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "gfx_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP + ring->pipe;
+	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
+	if (r)
+		return r;
+	return 0;
+}
+
+static int gfx_v12_0_compute_ring_init(struct amdgpu_device *adev, int ring_id,
+				       int mec, int pipe, int queue)
+{
+	int r;
+	unsigned irq_type;
+	struct amdgpu_ring *ring;
+	unsigned int hw_prio;
+
+	ring = &adev->gfx.compute_ring[ring_id];
+
+	/* mec0 is me1 */
+	ring->me = mec + 1;
+	ring->pipe = pipe;
+	ring->queue = queue;
+
+	ring->ring_obj = NULL;
+	ring->use_doorbell = true;
+	ring->doorbell_index = (adev->doorbell_index.mec_ring0 + ring_id) << 1;
+	ring->eop_gpu_addr = adev->gfx.mec.hpd_eop_gpu_addr
+				+ (ring_id * GFX12_MEC_HPD_SIZE);
+	ring->vm_hub = AMDGPU_GFXHUB(0);
+	sprintf(ring->name, "comp_%d.%d.%d", ring->me, ring->pipe, ring->queue);
+
+	irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP
+		+ ((ring->me - 1) * adev->gfx.mec.num_pipe_per_mec)
+		+ ring->pipe;
+	hw_prio = amdgpu_gfx_is_high_priority_compute_queue(adev, ring) ?
+			AMDGPU_GFX_PIPE_PRIO_HIGH : AMDGPU_GFX_PIPE_PRIO_NORMAL;
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	r = amdgpu_ring_init(adev, ring, 1024, &adev->gfx.eop_irq, irq_type,
+			     hw_prio, NULL);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static struct {
+	SOC24_FIRMWARE_ID	id;
+	unsigned int		offset;
+	unsigned int		size;
+	unsigned int		size_x16;
+} rlc_autoload_info[SOC24_FIRMWARE_ID_MAX];
+
+#define RLC_TOC_OFFSET_DWUNIT   8
+#define RLC_SIZE_MULTIPLE       1024
+#define RLC_TOC_UMF_SIZE_inM	23ULL
+#define RLC_TOC_FORMAT_API	165ULL
+
+static void gfx_v12_0_parse_rlc_toc(struct amdgpu_device *adev, void *rlc_toc)
+{
+	RLC_TABLE_OF_CONTENT_V2 *ucode = rlc_toc;
+
+