Merge tag 'amd-drm-next-5.15-2021-07-29' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.15-2021-07-29:

amdgpu:
- VCN/JPEG power down sequencing fixes
- Various navi pcie link handling fixes
- Clockgating fixes
- Yellow Carp fixes
- Beige Goby fixes
- Misc code cleanups
- S0ix fixes
- SMU i2c bus rework
- EEPROM handling rework
- PSP ucode handling cleanup
- SMU error handling rework
- AMD HDMI freesync fixes
- USB PD firmware update rework
- MMIO based vram access rework
- Misc display fixes
- Backlight fixes
- Add initial Cyan Skillfish support
- Overclocking fixes suspend/resume

amdkfd:
- Sysfs leak fix
- Add counters for vm faults and migration
- GPUVM TLB optimizations

radeon:
- Misc fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210730033455.3852-1-alexander.deucher@amd.com

18 files changed, 1312 insertions, 676 deletions

diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
index f6e0e7d8a007..d03e6fa2bf1a 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
@@ -450,6 +450,7 @@ struct amdgpu_pm {
 
 	/* Used for I2C access to various EEPROMs on relevant ASICs */
 	struct i2c_adapter smu_i2c;
+	struct mutex		smu_i2c_mutex;
 	struct list_head	pm_attr_list;
 };
 
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
index 3e89852e4820..c2c201b8e3cf 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
@@ -231,6 +231,7 @@ struct smu_user_dpm_profile {
 	uint32_t power_limit;
 	uint32_t fan_speed_percent;
 	uint32_t flags;
+	uint32_t user_od;
 
 	/* user clock state information */
 	uint32_t clk_mask[SMU_CLK_COUNT];
@@ -352,6 +353,7 @@ struct smu_table_context
 
 	void				*overdrive_table;
 	void                            *boot_overdrive_table;
+	void				*user_overdrive_table;
 
 	uint32_t			gpu_metrics_table_size;
 	void				*gpu_metrics_table;
@@ -624,6 +626,12 @@ struct pptable_funcs {
 				 long *input, uint32_t size);
 
 	/**
+	 * @restore_user_od_settings: Restore the user customized
+	 *                            OD settings on S3/S4/Runpm resume.
+	 */
+	int (*restore_user_od_settings)(struct smu_context *smu);
+
+	/**
 	 * @get_clock_by_type_with_latency: Get the speed and latency of a clock
 	 *                                  domain.
 	 */
diff --git a/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_cyan_skillfish.h b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_cyan_skillfish.h
new file mode 100644
index 000000000000..8a08ecc34c69
--- /dev/null
+++ b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_cyan_skillfish.h
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2021 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 __SMU11_DRIVER_IF_CYAN_SKILLFISH_H__
+#define __SMU11_DRIVER_IF_CYAN_SKILLFISH_H__
+
+// *** IMPORTANT ***
+// Always increment the interface version if
+// any structure is changed in this file
+#define MP1_DRIVER_IF_VERSION 0x8
+
+#define TABLE_BIOS_IF            0 // Called by BIOS
+#define TABLE_WATERMARKS         1 // Called by Driver; defined here, but not used, for backward compatible
+#define TABLE_PMSTATUSLOG        3 // Called by Tools for Agm logging
+#define TABLE_DPMCLOCKS          4 // Called by Driver; defined here, but not used, for backward compatible
+#define TABLE_MOMENTARY_PM       5 // Called by Tools; defined here, but not used, for backward compatible
+#define TABLE_COUNT              6
+
+#define NUM_DSPCLK_LEVELS		8
+#define NUM_SOCCLK_DPM_LEVELS	8
+#define NUM_DCEFCLK_DPM_LEVELS	4
+#define NUM_FCLK_DPM_LEVELS		4
+#define NUM_MEMCLK_DPM_LEVELS	4
+
+#define NUMBER_OF_PSTATES		8
+#define NUMBER_OF_CORES			8
+
+typedef enum {
+	S3_TYPE_ENTRY,
+	S5_TYPE_ENTRY,
+} Sleep_Type_e;
+
+typedef enum {
+	GFX_OFF = 0,
+	GFX_ON  = 1,
+} GFX_Mode_e;
+
+typedef enum {
+	CPU_P0 = 0,
+	CPU_P1,
+	CPU_P2,
+	CPU_P3,
+	CPU_P4,
+	CPU_P5,
+	CPU_P6,
+	CPU_P7
+} CPU_PState_e;
+
+typedef enum {
+	CPU_CORE0 = 0,
+	CPU_CORE1,
+	CPU_CORE2,
+	CPU_CORE3,
+	CPU_CORE4,
+	CPU_CORE5,
+	CPU_CORE6,
+	CPU_CORE7
+} CORE_ID_e;
+
+typedef enum {
+	DF_DPM0 = 0,
+	DF_DPM1,
+	DF_DPM2,
+	DF_DPM3,
+	DF_PState_Count
+} DF_PState_e;
+
+typedef enum {
+	GFX_DPM0 = 0,
+	GFX_DPM1,
+	GFX_DPM2,
+	GFX_DPM3,
+	GFX_PState_Count
+} GFX_PState_e;
+
+#endif
diff --git a/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_sienna_cichlid.h b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_sienna_cichlid.h
index 61c87c39be80..63b8701fd466 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_sienna_cichlid.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu11_driver_if_sienna_cichlid.h
@@ -131,7 +131,7 @@
 #define FEATURE_GFX_EDC_BIT             49
 #define FEATURE_GFX_PER_PART_VMIN_BIT   50
 #define FEATURE_SMART_SHIFT_BIT         51
-#define FEATURE_SPARE_52_BIT            52
+#define FEATURE_APT_BIT                 52
 #define FEATURE_SPARE_53_BIT            53
 #define FEATURE_SPARE_54_BIT            54
 #define FEATURE_SPARE_55_BIT            55
@@ -211,6 +211,7 @@ typedef enum {
 #define THROTTLER_FIT_BIT          17
 #define THROTTLER_PPM_BIT          18
 #define THROTTLER_APCC_BIT         19
+#define THROTTLER_COUNT            20
 
 // FW DState Features Control Bits
 // FW DState Features Control Bits
@@ -1406,7 +1407,67 @@ typedef struct {
 } SmuMetrics_t;
 
 typedef struct {
-  SmuMetrics_t SmuMetrics;
+  uint32_t CurrClock[PPCLK_COUNT];
+
+  uint16_t AverageGfxclkFrequencyPreDs;
+  uint16_t AverageGfxclkFrequencyPostDs;
+  uint16_t AverageFclkFrequencyPreDs;
+  uint16_t AverageFclkFrequencyPostDs;
+  uint16_t AverageUclkFrequencyPreDs  ;
+  uint16_t AverageUclkFrequencyPostDs  ;
+
+
+  uint16_t AverageGfxActivity    ;
+  uint16_t AverageUclkActivity   ;
+  uint8_t  CurrSocVoltageOffset  ;
+  uint8_t  CurrGfxVoltageOffset  ;
+  uint8_t  CurrMemVidOffset      ;
+  uint8_t  Padding8        ;
+  uint16_t AverageSocketPower    ;
+  uint16_t TemperatureEdge       ;
+  uint16_t TemperatureHotspot    ;
+  uint16_t TemperatureMem        ;
+  uint16_t TemperatureVrGfx      ;
+  uint16_t TemperatureVrMem0     ;
+  uint16_t TemperatureVrMem1     ;
+  uint16_t TemperatureVrSoc      ;
+  uint16_t TemperatureLiquid0    ;
+  uint16_t TemperatureLiquid1    ;
+  uint16_t TemperaturePlx        ;
+  uint16_t Padding16             ;
+  uint32_t AccCnt                ;
+  uint8_t  ThrottlingPercentage[THROTTLER_COUNT];
+
+
+  uint8_t  LinkDpmLevel;
+  uint8_t  CurrFanPwm;
+  uint16_t CurrFanSpeed;
+
+  //BACO metrics, PMFW-1721
+  //metrics for D3hot entry/exit and driver ARM msgs
+  uint8_t D3HotEntryCountPerMode[D3HOT_SEQUENCE_COUNT];
+  uint8_t D3HotExitCountPerMode[D3HOT_SEQUENCE_COUNT];
+  uint8_t ArmMsgReceivedCountPerMode[D3HOT_SEQUENCE_COUNT];
+
+  //PMFW-4362
+  uint32_t EnergyAccumulator;
+  uint16_t AverageVclk0Frequency  ;
+  uint16_t AverageDclk0Frequency  ;
+  uint16_t AverageVclk1Frequency  ;
+  uint16_t AverageDclk1Frequency  ;
+  uint16_t VcnActivityPercentage  ; //place holder, David N. to provide full sequence
+  uint8_t  PcieRate               ;
+  uint8_t  PcieWidth              ;
+  uint16_t AverageGfxclkFrequencyTarget;
+  uint16_t Padding16_2;
+
+} SmuMetrics_V2_t;
+
+typedef struct {
+  union {
+    SmuMetrics_t SmuMetrics;
+    SmuMetrics_V2_t SmuMetrics_V2;
+  };
   uint32_t Spare[1];
 
   // Padding - ignore
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h b/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
index f61b5c914a3d..403bc1bf8a77 100644
--- a/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v11_0.h
@@ -30,11 +30,12 @@
 #define SMU11_DRIVER_IF_VERSION_NV10 0x37
 #define SMU11_DRIVER_IF_VERSION_NV12 0x38
 #define SMU11_DRIVER_IF_VERSION_NV14 0x38
-#define SMU11_DRIVER_IF_VERSION_Sienna_Cichlid 0x3D
+#define SMU11_DRIVER_IF_VERSION_Sienna_Cichlid 0x40
 #define SMU11_DRIVER_IF_VERSION_Navy_Flounder 0xE
 #define SMU11_DRIVER_IF_VERSION_VANGOGH 0x03
 #define SMU11_DRIVER_IF_VERSION_Dimgrey_Cavefish 0xF
 #define SMU11_DRIVER_IF_VERSION_Beige_Goby 0xD
+#define SMU11_DRIVER_IF_VERSION_Cyan_Skillfish 0x8
 
 /* MP Apertures */
 #define MP0_Public			0x03800000
@@ -302,5 +303,7 @@ void smu_v11_0_interrupt_work(struct smu_context *smu);
 
 int smu_v11_0_set_light_sbr(struct smu_context *smu, bool enable);
 
+int smu_v11_0_restore_user_od_settings(struct smu_context *smu);
+
 #endif
 #endif
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v11_8_pmfw.h b/drivers/gpu/drm/amd/pm/inc/smu_v11_8_pmfw.h
new file mode 100644
index 000000000000..bd4fcb6b9610
--- /dev/null
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v11_8_pmfw.h
@@ -0,0 +1,152 @@
+/*
+ * Copyright 2021 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 __SMU_V11_8_0_PMFW_H__
+#define __SMU_V11_8_0_PMFW_H__
+
+#pragma pack(push, 1)
+
+#define ENABLE_DEBUG_FEATURES
+
+// Feature Control Defines
+#define FEATURE_CCLK_CONTROLLER_BIT       0
+#define FEATURE_GFXCLK_EFFT_FREQ_BIT      1
+#define FEATURE_DATA_CALCULATION_BIT      2
+#define FEATURE_THERMAL_BIT               3
+#define FEATURE_PLL_POWER_DOWN_BIT        4
+#define FEATURE_FCLK_DPM_BIT              5
+#define FEATURE_GFX_DPM_BIT               6
+#define FEATURE_DS_GFXCLK_BIT             7
+#define FEATURE_DS_SOCCLK_BIT             8
+#define FEATURE_DS_LCLK_BIT               9
+#define FEATURE_CORE_CSTATES_BIT          10
+#define FEATURE_G6_SSC_BIT                11 //G6 memory UCLK and UCLK_DIV SS
+#define FEATURE_RM_BIT                    12
+#define FEATURE_SOC_DPM_BIT               13
+#define FEATURE_DS_SMNCLK_BIT             14
+#define FEATURE_DS_MP1CLK_BIT             15
+#define FEATURE_DS_MP0CLK_BIT             16
+#define FEATURE_MGCG_BIT                  17
+#define FEATURE_DS_FUSE_SRAM_BIT          18
+#define FEATURE_GFX_CKS_BIT               19
+#define FEATURE_FP_THROTTLING_BIT         20
+#define FEATURE_PROCHOT_BIT               21
+#define FEATURE_CPUOFF_BIT                22
+#define FEATURE_UMC_THROTTLE_BIT          23
+#define FEATURE_DF_THROTTLE_BIT           24
+#define FEATURE_DS_MP3CLK_BIT             25
+#define FEATURE_DS_SHUBCLK_BIT            26
+#define FEATURE_TDC_BIT                   27 //Legacy APM_BIT
+#define FEATURE_UMC_CAL_SHARING_BIT       28
+#define FEATURE_DFLL_BTC_CALIBRATION_BIT  29
+#define FEATURE_EDC_BIT                   30
+#define FEATURE_DLDO_BIT                  31
+#define FEATURE_MEAS_DRAM_BLACKOUT_BIT    32
+#define FEATURE_CC1_BIT                   33
+#define FEATURE_PPT_BIT                   34
+#define FEATURE_STAPM_BIT                 35
+#define FEATURE_CSTATE_BOOST_BIT          36
+#define FEATURE_SPARE_37_BIT              37
+#define FEATURE_SPARE_38_BIT              38
+#define FEATURE_SPARE_39_BIT              39
+#define FEATURE_SPARE_40_BIT              40
+#define FEATURE_SPARE_41_BIT              41
+#define FEATURE_SPARE_42_BIT              42
+#define FEATURE_SPARE_43_BIT              43
+#define FEATURE_SPARE_44_BIT              44
+#define FEATURE_SPARE_45_BIT              45
+#define FEATURE_SPARE_46_BIT              46
+#define FEATURE_SPARE_47_BIT              47
+#define FEATURE_SPARE_48_BIT              48
+#define FEATURE_SPARE_49_BIT              49
+#define FEATURE_SPARE_50_BIT              50
+#define FEATURE_SPARE_51_BIT              51
+#define FEATURE_SPARE_52_BIT              52
+#define FEATURE_SPARE_53_BIT              53
+#define FEATURE_SPARE_54_BIT              54
+#define FEATURE_SPARE_55_BIT              55
+#define FEATURE_SPARE_56_BIT              56
+#define FEATURE_SPARE_57_BIT              57
+#define FEATURE_SPARE_58_BIT              58
+#define FEATURE_SPARE_59_BIT              59
+#define FEATURE_SPARE_60_BIT              60
+#define FEATURE_SPARE_61_BIT              61
+#define FEATURE_SPARE_62_BIT              62
+#define FEATURE_SPARE_63_BIT              63
+
+#define NUM_FEATURES                      64
+
+#define FEATURE_CCLK_CONTROLLER_MASK  (1 << FEATURE_CCLK_CONTROLLER_BIT)
+#define FEATURE_DATA_CALCULATION_MASK (1 << FEATURE_DATA_CALCULATION_BIT)
+#define FEATURE_THERMAL_MASK          (1 << FEATURE_THERMAL_BIT)
+#define FEATURE_PLL_POWER_DOWN_MASK   (1 << FEATURE_PLL_POWER_DOWN_BIT)
+#define FEATURE_FCLK_DPM_MASK         (1 << FEATURE_FCLK_DPM_BIT)
+#define FEATURE_GFX_DPM_MASK          (1 << FEATURE_GFX_DPM_BIT)
+#define FEATURE_DS_GFXCLK_MASK        (1 << FEATURE_DS_GFXCLK_BIT)
+#define FEATURE_DS_SOCCLK_MASK        (1 << FEATURE_DS_SOCCLK_BIT)
+#define FEATURE_DS_LCLK_MASK          (1 << FEATURE_DS_LCLK_BIT)
+#define FEATURE_RM_MASK               (1 << FEATURE_RM_BIT)
+#define FEATURE_DS_SMNCLK_MASK        (1 << FEATURE_DS_SMNCLK_BIT)
+#define FEATURE_DS_MP1CLK_MASK        (1 << FEATURE_DS_MP1CLK_BIT)
+#define FEATURE_DS_MP0CLK_MASK        (1 << FEATURE_DS_MP0CLK_BIT)
+#define FEATURE_MGCG_MASK             (1 << FEATURE_MGCG_BIT)
+#define FEATURE_DS_FUSE_SRAM_MASK     (1 << FEATURE_DS_FUSE_SRAM_BIT)
+#define FEATURE_PROCHOT_MASK          (1 << FEATURE_PROCHOT_BIT)
+#define FEATURE_CPUOFF_MASK           (1 << FEATURE_CPUOFF_BIT)
+#define FEATURE_GFX_CKS_MASK          (1 << FEATURE_GFX_CKS_BIT)
+#define FEATURE_UMC_THROTTLE_MASK     (1 << FEATURE_UMC_THROTTLE_BIT)
+#define FEATURE_DF_THROTTLE_MASK      (1 << FEATURE_DF_THROTTLE_BIT)
+#define FEATURE_SOC_DPM_MASK          (1 << FEATURE_SOC_DPM_BIT)
+
+typedef struct {
+	// MP1_EXT_SCRATCH0
+	uint32_t SPARE1            : 4;
+	uint32_t SPARE2            : 4;
+	uint32_t SPARE3            : 4;
+	uint32_t CurrLevel_LCLK    : 4;
+	uint32_t CurrLevel_MP0CLK  : 4;
+	uint32_t CurrLevel_FCLK    : 4;
+	uint32_t CurrLevel_SOCCLK  : 4;
+	uint32_t CurrLevel_DCEFCLK : 4;
+	// MP1_EXT_SCRATCH1
+	uint32_t SPARE4            : 4;
+	uint32_t SPARE5            : 4;
+	uint32_t SPARE6            : 4;
+	uint32_t TargLevel_LCLK    : 4;
+	uint32_t TargLevel_MP0CLK  : 4;
+	uint32_t TargLevel_FCLK    : 4;
+	uint32_t TargLevel_SOCCLK  : 4;
+	uint32_t TargLevel_DCEFCLK : 4;
+	// MP1_EXT_SCRATCH2
+	uint32_t CurrLevel_SHUBCLK  : 4;
+	uint32_t TargLevel_SHUBCLK  : 4;
+	uint32_t Reserved          : 24;
+	// MP1_EXT_SCRATCH3-4
+	uint32_t Reserved2[2];
+	// MP1_EXT_SCRATCH5
+	uint32_t FeatureStatus[NUM_FEATURES / 32];
+} FwStatus_t;
+
+#pragma pack(pop)
+
+#endif
diff --git a/drivers/gpu/drm/amd/pm/inc/smu_v11_8_ppsmc.h b/drivers/gpu/drm/amd/pm/inc/smu_v11_8_ppsmc.h
new file mode 100644
index 000000000000..6e6088760b18
--- /dev/null
+++ b/drivers/gpu/drm/amd/pm/inc/smu_v11_8_ppsmc.h
@@ -0,0 +1,70 @@
+/*
+ * Copyright 2021 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 SMU_11_8_0_PPSMC_H
+#define SMU_11_8_0_PPSMC_H
+
+// SMU Response Codes:
+#define PPSMC_Result_OK                    0x1
+#define PPSMC_Result_Failed                0xFF
+#define PPSMC_Result_UnknownCmd            0xFE
+#define PPSMC_Result_CmdRejectedPrereq     0xFD
+#define PPSMC_Result_CmdRejectedBusy       0xFC
+
+// Message Definitions:
+#define PPSMC_MSG_TestMessage                           0x1
+#define PPSMC_MSG_GetSmuVersion                         0x2
+#define PPSMC_MSG_GetDriverIfVersion                    0x3
+#define PPSMC_MSG_SetDriverTableDramAddrHigh            0x4
+#define PPSMC_MSG_SetDriverTableDramAddrLow             0x5
+#define PPSMC_MSG_TransferTableSmu2Dram                 0x6
+#define PPSMC_MSG_TransferTableDram2Smu                 0x7
+#define PPSMC_MSG_Rsvd1                                 0xA
+#define PPSMC_MSG_RequestCorePstate                     0xB
+#define PPSMC_MSG_QueryCorePstate                       0xC
+#define PPSMC_MSG_Rsvd2                                 0xD
+#define PPSMC_MSG_RequestGfxclk                         0xE
+#define PPSMC_MSG_QueryGfxclk                           0xF
+#define PPSMC_MSG_QueryVddcrSocClock                    0x11
+#define PPSMC_MSG_QueryDfPstate                         0x13
+#define PPSMC_MSG_Rsvd3                                 0x14
+#define PPSMC_MSG_ConfigureS3PwrOffRegisterAddressHigh  0x16
+#define PPSMC_MSG_ConfigureS3PwrOffRegisterAddressLow   0x17
+#define PPSMC_MSG_RequestActiveWgp                      0x18
+#define PPSMC_MSG_SetMinDeepSleepGfxclkFreq             0x19
+#define PPSMC_MSG_SetMaxDeepSleepDfllGfxDiv             0x1A
+#define PPSMC_MSG_StartTelemetryReporting               0x1B
+#define PPSMC_MSG_StopTelemetryReporting                0x1C
+#define PPSMC_MSG_ClearTelemetryMax                     0x1D
+#define PPSMC_MSG_QueryActiveWgp                        0x1E
+#define PPSMC_MSG_SetCoreEnableMask                     0x2C
+#define PPSMC_MSG_InitiateGcRsmuSoftReset               0x2E
+#define PPSMC_MSG_GfxCacWeightOperation                 0x2F
+#define PPSMC_MSG_L3CacWeightOperation                  0x30
+#define PPSMC_MSG_PackCoreCacWeight                     0x31
+#define PPSMC_MSG_SetDriverTableVMID                    0x34
+#define PPSMC_MSG_SetSoftMinCclk                        0x35
+#define PPSMC_MSG_SetSoftMaxCclk                        0x36
+#define PPSMC_Message_Count                             0x37
+
+#endif
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index ebe672142808..71afc2d20b12 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -36,6 +36,7 @@
 #include "vangogh_ppt.h"
 #include "aldebaran_ppt.h"
 #include "yellow_carp_ppt.h"
+#include "cyan_skillfish_ppt.h"
 #include "amd_pcie.h"
 
 /*
@@ -416,6 +417,15 @@ static void smu_restore_dpm_user_profile(struct smu_context *smu)
 		}
 	}
 
+	/* Restore user customized OD settings */
+	if (smu->user_dpm_profile.user_od) {
+		if (smu->ppt_funcs->restore_user_od_settings) {
+			ret = smu->ppt_funcs->restore_user_od_settings(smu);
+			if (ret)
+				dev_err(smu->adev->dev, "Failed to upload customized OD settings\n");
+		}
+	}
+
 	/* Disable restore flag */
 	smu->user_dpm_profile.flags &= ~SMU_DPM_USER_PROFILE_RESTORE;
 }
@@ -588,6 +598,9 @@ static int smu_set_funcs(struct amdgpu_device *adev)
 	case CHIP_YELLOW_CARP:
 		yellow_carp_set_ppt_funcs(smu);
 		break;
+	case CHIP_CYAN_SKILLFISH:
+		cyan_skillfish_set_ppt_funcs(smu);
+		break;
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/Makefile b/drivers/gpu/drm/amd/pm/swsmu/smu11/Makefile
index 0138c982dfd3..f9b2e16f6431 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/Makefile
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/Makefile
@@ -27,6 +27,7 @@ SMU11_MGR = arcturus_ppt.o \
 	    navi10_ppt.o \
 	    sienna_cichlid_ppt.o \
 	    vangogh_ppt.o \
+	    cyan_skillfish_ppt.o \
 	    smu_v11_0.o
 
 AMD_SWSMU_SMU11MGR = $(addprefix $(AMD_SWSMU_PATH)/smu11/,$(SMU11_MGR))
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
index 094df6f87cfc..6ec8492f71f5 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
@@ -1936,197 +1936,77 @@ static int arcturus_dpm_set_vcn_enable(struct smu_context *smu, bool enable)
 	return ret;
 }
 
-static void arcturus_fill_i2c_req(SwI2cRequest_t  *req, bool write,
-				  uint8_t address, uint32_t numbytes,
-				  uint8_t *data)
-{
-	int i;
-
-	req->I2CcontrollerPort = 0;
-	req->I2CSpeed = 2;
-	req->SlaveAddress = address;
-	req->NumCmds = numbytes;
-
-	for (i = 0; i < numbytes; i++) {
-		SwI2cCmd_t *cmd =  &req->SwI2cCmds[i];
-
-		/* First 2 bytes are always write for lower 2b EEPROM address */
-		if (i < 2)
-			cmd->Cmd = 1;
-		else
-			cmd->Cmd = write;
-
-
-		/* Add RESTART for read  after address filled */
-		cmd->CmdConfig |= (i == 2 && !write) ? CMDCONFIG_RESTART_MASK : 0;
-
-		/* Add STOP in the end */
-		cmd->CmdConfig |= (i == (numbytes - 1)) ? CMDCONFIG_STOP_MASK : 0;
-
-		/* Fill with data regardless if read or write to simplify code */
-		cmd->RegisterAddr = data[i];
-	}
-}
-
-static int arcturus_i2c_read_data(struct i2c_adapter *control,
-					       uint8_t address,
-					       uint8_t *data,
-					       uint32_t numbytes)
+static int arcturus_i2c_xfer(struct i2c_adapter *i2c_adap,
+			     struct i2c_msg *msg, int num_msgs)
 {
-	uint32_t  i, ret = 0;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_device *adev = to_amdgpu_device(i2c_adap);
 	struct smu_table_context *smu_table = &adev->smu.smu_table;
 	struct smu_table *table = &smu_table->driver_table;
+	SwI2cRequest_t *req, *res = (SwI2cRequest_t *)table->cpu_addr;
+	int i, j, r, c;
+	u16 dir;
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
-	}
-
-	memset(&req, 0, sizeof(req));
-	arcturus_fill_i2c_req(&req, false, address, numbytes, data);
-
-	mutex_lock(&adev->smu.mutex);
-	/* Now read data starting with that address */
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req,
-					true);
-	mutex_unlock(&adev->smu.mutex);
-
-	if (!ret) {
-		SwI2cRequest_t *res = (SwI2cRequest_t *)table->cpu_addr;
-
-		/* Assume SMU  fills res.SwI2cCmds[i].Data with read bytes */
-		for (i = 0; i < numbytes; i++)
-			data[i] = res->SwI2cCmds[i].Data;
-
-		dev_dbg(adev->dev, "arcturus_i2c_read_data, address = %x, bytes = %d, data :",
-				  (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-	} else
-		dev_err(adev->dev, "arcturus_i2c_read_data - error occurred :%x", ret);
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (!req)
+		return -ENOMEM;
 
-	return ret;
-}
+	req->I2CcontrollerPort = 0;
+	req->I2CSpeed = I2C_SPEED_FAST_400K;
+	req->SlaveAddress = msg[0].addr << 1; /* wants an 8-bit address */
+	dir = msg[0].flags & I2C_M_RD;
+
+	for (c = i = 0; i < num_msgs; i++) {
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &req->SwI2cCmds[c];
+
+			if (!(msg[i].flags & I2C_M_RD)) {
+				/* write */
+				cmd->Cmd = I2C_CMD_WRITE;
+				cmd->RegisterAddr = msg[i].buf[j];
+			}
 
-static int arcturus_i2c_write_data(struct i2c_adapter *control,
-						uint8_t address,
-						uint8_t *data,
-						uint32_t numbytes)
-{
-	uint32_t ret;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+			if ((dir ^ msg[i].flags) & I2C_M_RD) {
+				/* The direction changes.
+				 */
+				dir = msg[i].flags & I2C_M_RD;
+				cmd->CmdConfig |= CMDCONFIG_RESTART_MASK;
+			}
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
+			req->NumCmds++;
+
+			/*
+			 * Insert STOP if we are at the last byte of either last
+			 * message for the transaction or the client explicitly
+			 * requires a STOP at this particular message.
+			 */
+			if ((j == msg[i].len - 1) &&
+			    ((i == num_msgs - 1) || (msg[i].flags & I2C_M_STOP))) {
+				cmd->CmdConfig &= ~CMDCONFIG_RESTART_MASK;
+				cmd->CmdConfig |= CMDCONFIG_STOP_MASK;
+			}
+		}
 	}
-
-	memset(&req, 0, sizeof(req));
-	arcturus_fill_i2c_req(&req, true, address, numbytes, data);
-
 	mutex_lock(&adev->smu.mutex);
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req, true);
+	r = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, req, true);
 	mutex_unlock(&adev->smu.mutex);
+	if (r)
+		goto fail;
 
-	if (!ret) {
-		dev_dbg(adev->dev, "arcturus_i2c_write(), address = %x, bytes = %d , data: ",
-					 (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-		/*
-		 * According to EEPROM spec there is a MAX of 10 ms required for
-		 * EEPROM to flush internal RX buffer after STOP was issued at the
-		 * end of write transaction. During this time the EEPROM will not be
-		 * responsive to any more commands - so wait a bit more.
-		 */
-		msleep(10);
-
-	} else
-		dev_err(adev->dev, "arcturus_i2c_write- error occurred :%x", ret);
-
-	return ret;
-}
-
-static int arcturus_i2c_xfer(struct i2c_adapter *i2c_adap,
-			      struct i2c_msg *msgs, int num)
-{
-	uint32_t  i, j, ret, data_size, data_chunk_size, next_eeprom_addr = 0;
-	uint8_t *data_ptr, data_chunk[MAX_SW_I2C_COMMANDS] = { 0 };
-
-	for (i = 0; i < num; i++) {
-		/*
-		 * SMU interface allows at most MAX_SW_I2C_COMMANDS bytes of data at
-		 * once and hence the data needs to be spliced into chunks and sent each
-		 * chunk separately
-		 */
-		data_size = msgs[i].len - 2;
-		data_chunk_size = MAX_SW_I2C_COMMANDS - 2;
-		next_eeprom_addr = (msgs[i].buf[0] << 8 & 0xff00) | (msgs[i].buf[1] & 0xff);
-		data_ptr = msgs[i].buf + 2;
-
-		for (j = 0; j < data_size / data_chunk_size; j++) {
-			/* Insert the EEPROM dest addess, bits 0-15 */
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = arcturus_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, MAX_SW_I2C_COMMANDS);
-
-				memcpy(data_ptr, data_chunk + 2, data_chunk_size);
-			} else {
-
-				memcpy(data_chunk + 2, data_ptr, data_chunk_size);
-
-				ret = arcturus_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, MAX_SW_I2C_COMMANDS);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
-
-			next_eeprom_addr += data_chunk_size;
-			data_ptr += data_chunk_size;
+	for (c = i = 0; i < num_msgs; i++) {
+		if (!(msg[i].flags & I2C_M_RD)) {
+			c += msg[i].len;
+			continue;
 		}
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &res->SwI2cCmds[c];
 
-		if (data_size % data_chunk_size) {
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = arcturus_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, (data_size % data_chunk_size) + 2);
-
-				memcpy(data_ptr, data_chunk + 2, data_size % data_chunk_size);
-			} else {
-				memcpy(data_chunk + 2, data_ptr, data_size % data_chunk_size);
-
-				ret = arcturus_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, (data_size % data_chunk_size) + 2);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
+			msg[i].buf[j] = cmd->Data;
 		}
 	}
-
+	r = num_msgs;
 fail:
-	return num;
+	kfree(req);
+	return r;
 }
 
 static u32 arcturus_i2c_func(struct i2c_adapter *adap)
@@ -2140,15 +2020,25 @@ static const struct i2c_algorithm arcturus_i2c_algo = {
 	.functionality = arcturus_i2c_func,
 };
 
+
+static const struct i2c_adapter_quirks arcturus_i2c_control_quirks = {
+	.flags = I2C_AQ_COMB | I2C_AQ_COMB_SAME_ADDR | I2C_AQ_NO_ZERO_LEN,
+	.max_read_len  = MAX_SW_I2C_COMMANDS,
+	.max_write_len = MAX_SW_I2C_COMMANDS,
+	.max_comb_1st_msg_len = 2,
+	.max_comb_2nd_msg_len = MAX_SW_I2C_COMMANDS - 2,
+};
+
 static int arcturus_i2c_control_init(struct smu_context *smu, struct i2c_adapter *control)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
 	int res;
 
 	control->owner = THIS_MODULE;
-	control->class = I2C_CLASS_SPD;
+	control->class = I2C_CLASS_HWMON;
 	control->dev.parent = &adev->pdev->dev;
 	control->algo = &arcturus_i2c_algo;
+	control->quirks = &arcturus_i2c_control_quirks;
 	snprintf(control->name, sizeof(control->name), "AMDGPU SMU");
 
 	res = i2c_add_adapter(control);
@@ -2288,7 +2178,9 @@ static void arcturus_log_thermal_throttling_event(struct smu_context *smu)
 
 	dev_warn(adev->dev, "WARN: GPU thermal throttling temperature reached, expect performance decrease. %s.\n",
 			log_buf);
-	kgd2kfd_smi_event_throttle(smu->adev->kfd.dev, throttler_status);
+	kgd2kfd_smi_event_throttle(smu->adev->kfd.dev,
+		smu_cmn_get_indep_throttler_status(throttler_status,
+						   arcturus_throttler_map));
 }
 
 static uint16_t arcturus_get_current_pcie_link_speed(struct smu_context *smu)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.c
new file mode 100644
index 000000000000..b05f9541accc
--- /dev/null
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.c
@@ -0,0 +1,76 @@
+/*
+ * Copyright 2021 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.
+ *
+ */
+
+#define SWSMU_CODE_LAYER_L2
+
+#include "amdgpu.h"
+#include "amdgpu_smu.h"
+#include "smu_v11_0.h"
+#include "smu11_driver_if_cyan_skillfish.h"
+#include "cyan_skillfish_ppt.h"
+#include "smu_v11_8_ppsmc.h"
+#include "smu_v11_8_pmfw.h"
+#include "smu_cmn.h"
+#include "soc15_common.h"
+
+/*
+ * DO NOT use these for err/warn/info/debug messages.
+ * Use dev_err, dev_warn, dev_info and dev_dbg instead.
+ * They are more MGPU friendly.
+ */
+
+#undef pr_err
+#undef pr_warn
+#undef pr_info
+#undef pr_debug
+
+static struct cmn2asic_msg_mapping cyan_skillfish_message_map[SMU_MSG_MAX_COUNT] = {
+	MSG_MAP(TestMessage,                    PPSMC_MSG_TestMessage,			0),
+	MSG_MAP(GetSmuVersion,                  PPSMC_MSG_GetSmuVersion,		0),
+	MSG_MAP(GetDriverIfVersion,             PPSMC_MSG_GetDriverIfVersion,		0),
+	MSG_MAP(SetDriverDramAddrHigh,          PPSMC_MSG_SetDriverTableDramAddrHigh,	0),
+	MSG_MAP(SetDriverDramAddrLow,           PPSMC_MSG_SetDriverTableDramAddrLow,	0),
+	MSG_MAP(TransferTableSmu2Dram,          PPSMC_MSG_TransferTableSmu2Dram,	0),
+	MSG_MAP(TransferTableDram2Smu,          PPSMC_MSG_TransferTableDram2Smu,	0),
+};
+
+static const struct pptable_funcs cyan_skillfish_ppt_funcs = {
+
+	.check_fw_status = smu_v11_0_check_fw_status,
+	.check_fw_version = smu_v11_0_check_fw_version,
+	.init_power = smu_v11_0_init_power,
+	.fini_power = smu_v11_0_fini_power,
+	.register_irq_handler = smu_v11_0_register_irq_handler,
+	.notify_memory_pool_location = smu_v11_0_notify_memory_pool_location,
+	.send_smc_msg_with_param = smu_cmn_send_smc_msg_with_param,
+	.send_smc_msg = smu_cmn_send_smc_msg,
+	.set_driver_table_location = smu_v11_0_set_driver_table_location,
+	.interrupt_work = smu_v11_0_interrupt_work,
+};
+
+void cyan_skillfish_set_ppt_funcs(struct smu_context *smu)
+{
+	smu->ppt_funcs = &cyan_skillfish_ppt_funcs;
+	smu->message_map = cyan_skillfish_message_map;
+	smu->is_apu = true;
+}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.h b/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.h
new file mode 100644
index 000000000000..76cd7229e383
--- /dev/null
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/cyan_skillfish_ppt.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright 2021 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 __CYAN_SKILLFISH_PPT_H__
+#define __CYAN_SKILLFISH_PPT_H__
+
+extern void cyan_skillfish_set_ppt_funcs(struct smu_context *smu);
+
+#endif
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
index 1ba42b69ce74..d7722c229ddd 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
@@ -2294,41 +2294,52 @@ static int navi10_set_default_od_settings(struct smu_context *smu)
 		(OverDriveTable_t *)smu->smu_table.overdrive_table;
 	OverDriveTable_t *boot_od_table =
 		(OverDriveTable_t *)smu->smu_table.boot_overdrive_table;
+	OverDriveTable_t *user_od_table =
+		(OverDriveTable_t *)smu->smu_table.user_overdrive_table;
 	int ret = 0;
 
-	ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)od_table, false);
+	/*
+	 * For S3/S4/Runpm resume, no need to setup those overdrive tables again as
+	 *   - either they already have the default OD settings got during cold bootup
+	 *   - or they have some user customized OD settings which cannot be overwritten
+	 */
+	if (smu->adev->in_suspend)
+		return 0;
+
+	ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)boot_od_table, false);
 	if (ret) {
 		dev_err(smu->adev->dev, "Failed to get overdrive table!\n");
 		return ret;
 	}
 
-	if (!od_table->GfxclkVolt1) {
+	if (!boot_od_table->GfxclkVolt1) {
 		ret = navi10_overdrive_get_gfx_clk_base_voltage(smu,
-								&od_table->GfxclkVolt1,
-								od_table->GfxclkFreq1);
+								&boot_od_table->GfxclkVolt1,
+								boot_od_table->GfxclkFreq1);
 		if (ret)
 			return ret;
 	}
 
-	if (!od_table->GfxclkVolt2) {
+	if (!boot_od_table->GfxclkVolt2) {
 		ret = navi10_overdrive_get_gfx_clk_base_voltage(smu,
-								&od_table->GfxclkVolt2,
-								od_table->GfxclkFreq2);
+								&boot_od_table->GfxclkVolt2,
+								boot_od_table->GfxclkFreq2);
 		if (ret)
 			return ret;
 	}
 
-	if (!od_table->GfxclkVolt3) {
+	if (!boot_od_table->GfxclkVolt3) {
 		ret = navi10_overdrive_get_gfx_clk_base_voltage(smu,
-								&od_table->GfxclkVolt3,
-								od_table->GfxclkFreq3);
+								&boot_od_table->GfxclkVolt3,
+								boot_od_table->GfxclkFreq3);
 		if (ret)
 			return ret;
 	}
 
-	memcpy(boot_od_table, od_table, sizeof(OverDriveTable_t));
+	navi10_dump_od_table(smu, boot_od_table);
 
-	navi10_dump_od_table(smu, od_table);
+	memcpy(od_table, boot_od_table, sizeof(OverDriveTable_t));
+	memcpy(user_od_table, boot_od_table, sizeof(OverDriveTable_t));
 
 	return 0;
 }
@@ -2429,11 +2440,20 @@ static int navi10_od_edit_dpm_table(struct smu_context *smu, enum PP_OD_DPM_TABL
 		memcpy(table_context->overdrive_table, table_context->boot_overdrive_table, sizeof(OverDriveTable_t));
 		break;
 	case PP_OD_COMMIT_DPM_TABLE:
-		navi10_dump_od_table(smu, od_table);
-		ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)od_table, true);
-		if (ret) {
-			dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
-			return ret;
+		if (memcmp(od_table, table_context->user_overdrive_table, sizeof(OverDriveTable_t))) {
+			navi10_dump_od_table(smu, od_table);
+			ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)od_table, true);
+			if (ret) {
+				dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
+				return ret;
+			}
+			memcpy(table_context->user_overdrive_table, od_table, sizeof(OverDriveTable_t));
+			smu->user_dpm_profile.user_od = true;
+
+			if (!memcmp(table_context->user_overdrive_table,
+				    table_context->boot_overdrive_table,
+				    sizeof(OverDriveTable_t)))
+				smu->user_dpm_profile.user_od = false;
 		}
 		break;
 	case PP_OD_EDIT_VDDC_CURVE:
@@ -2735,6 +2755,122 @@ static ssize_t navi10_get_legacy_gpu_metrics(struct smu_context *smu,
 	return sizeof(struct gpu_metrics_v1_3);
 }
 
+static int navi10_i2c_xfer(struct i2c_adapter *i2c_adap,
+			   struct i2c_msg *msg, int num_msgs)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(i2c_adap);
+	struct smu_table_context *smu_table = &adev->smu.smu_table;
+	struct smu_table *table = &smu_table->driver_table;
+	SwI2cRequest_t *req, *res = (SwI2cRequest_t *)table->cpu_addr;
+	int i, j, r, c;
+	u16 dir;
+
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (!req)
+		return -ENOMEM;
+
+	req->I2CcontrollerPort = 0;
+	req->I2CSpeed = I2C_SPEED_FAST_400K;
+	req->SlaveAddress = msg[0].addr << 1; /* wants an 8-bit address */
+	dir = msg[0].flags & I2C_M_RD;
+
+	for (c = i = 0; i < num_msgs; i++) {
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &req->SwI2cCmds[c];
+
+			if (!(msg[i].flags & I2C_M_RD)) {
+				/* write */
+				cmd->Cmd = I2C_CMD_WRITE;
+				cmd->RegisterAddr = msg[i].buf[j];
+			}
+
+			if ((dir ^ msg[i].flags) & I2C_M_RD) {
+				/* The direction changes.
+				 */
+				dir = msg[i].flags & I2C_M_RD;
+				cmd->CmdConfig |= CMDCONFIG_RESTART_MASK;
+			}
+
+			req->NumCmds++;
+
+			/*
+			 * Insert STOP if we are at the last byte of either last
+			 * message for the transaction or the client explicitly
+			 * requires a STOP at this particular message.
+			 */
+			if ((j == msg[i].len - 1) &&
+			    ((i == num_msgs - 1) || (msg[i].flags & I2C_M_STOP))) {
+				cmd->CmdConfig &= ~CMDCONFIG_RESTART_MASK;
+				cmd->CmdConfig |= CMDCONFIG_STOP_MASK;
+			}
+		}
+	}
+	mutex_lock(&adev->smu.mutex);
+	r = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, req, true);
+	mutex_unlock(&adev->smu.mutex);
+	if (r)
+		goto fail;
+
+	for (c = i = 0; i < num_msgs; i++) {
+		if (!(msg[i].flags & I2C_M_RD)) {
+			c += msg[i].len;
+			continue;
+		}
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &res->SwI2cCmds[c];
+
+			msg[i].buf[j] = cmd->Data;
+		}
+	}
+	r = num_msgs;
+fail:
+	kfree(req);
+	return r;
+}
+
+static u32 navi10_i2c_func(struct i2c_adapter *adap)
+{
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
+}
+
+
+static const struct i2c_algorithm navi10_i2c_algo = {
+	.master_xfer = navi10_i2c_xfer,
+	.functionality = navi10_i2c_func,
+};
+
+static const struct i2c_adapter_quirks navi10_i2c_control_quirks = {
+	.flags = I2C_AQ_COMB | I2C_AQ_COMB_SAME_ADDR | I2C_AQ_NO_ZERO_LEN,
+	.max_read_len  = MAX_SW_I2C_COMMANDS,
+	.max_write_len = MAX_SW_I2C_COMMANDS,
+	.max_comb_1st_msg_len = 2,
+	.max_comb_2nd_msg_len = MAX_SW_I2C_COMMANDS - 2,
+};
+
+static int navi10_i2c_control_init(struct smu_context *smu, struct i2c_adapter *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+	int res;
+
+	control->owner = THIS_MODULE;
+	control->class = I2C_CLASS_HWMON;
+	control->dev.parent = &adev->pdev->dev;
+	control->algo = &navi10_i2c_algo;
+	snprintf(control->name, sizeof(control->name), "AMDGPU SMU");
+	control->quirks = &navi10_i2c_control_quirks;
+
+	res = i2c_add_adapter(control);
+	if (res)
+		DRM_ERROR("Failed to register hw i2c, err: %d\n", res);
+
+	return res;
+}
+
+static void navi10_i2c_control_fini(struct smu_context *smu, struct i2c_adapter *control)
+{
+	i2c_del_adapter(control);
+}
+
 static ssize_t navi10_get_gpu_metrics(struct smu_context *smu,
 				      void **table)
 {
@@ -3078,6 +3214,8 @@ static const struct pptable_funcs navi10_ppt_funcs = {
 	.set_default_dpm_table = navi10_set_default_dpm_table,
 	.dpm_set_vcn_enable = navi10_dpm_set_vcn_enable,
 	.dpm_set_jpeg_enable = navi10_dpm_set_jpeg_enable,
+	.i2c_init = navi10_i2c_control_init,
+	.i2c_fini = navi10_i2c_control_fini,
 	.print_clk_levels = navi10_print_clk_levels,
 	.force_clk_levels = navi10_force_clk_levels,
 	.populate_umd_state_clk = navi10_populate_umd_state_clk,
@@ -3144,6 +3282,7 @@ static const struct pptable_funcs navi10_ppt_funcs = {
 	.set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range,
 	.set_default_od_settings = navi10_set_default_od_settings,
 	.od_edit_dpm_table = navi10_od_edit_dpm_table,
+	.restore_user_od_settings = smu_v11_0_restore_user_od_settings,
 	.run_btc = navi10_run_btc,
 	.set_power_source = smu_v11_0_set_power_source,
 	.get_pp_feature_mask = smu_cmn_get_pp_feature_mask,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
index c751f717a0da..90e40aacb8f6 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
@@ -489,6 +489,26 @@ err0_out:
 	return -ENOMEM;
 }
 
+static uint32_t sienna_cichlid_get_throttler_status_locked(struct smu_context *smu)
+{
+	struct smu_table_context *smu_table= &smu->smu_table;
+	SmuMetricsExternal_t *metrics_ext =
+		(SmuMetricsExternal_t *)(smu_table->metrics_table);
+	uint32_t throttler_status = 0;
+	int i;
+
+	if ((smu->adev->asic_type == CHIP_SIENNA_CICHLID) &&
+	     (smu->smc_fw_version >= 0x3A4300)) {
+		for (i = 0; i < THROTTLER_COUNT; i++)
+			throttler_status |=
+				(metrics_ext->SmuMetrics_V2.ThrottlingPercentage[i] ? 1U << i : 0);
+	} else {
+		throttler_status = metrics_ext->SmuMetrics.ThrottlerStatus;
+	}
+
+	return throttler_status;
+}
+
 static int sienna_cichlid_get_smu_metrics_data(struct smu_context *smu,
 					       MetricsMember_t member,
 					       uint32_t *value)
@@ -496,6 +516,11 @@ static int sienna_cichlid_get_smu_metrics_data(struct smu_context *smu,
 	struct smu_table_context *smu_table= &smu->smu_table;
 	SmuMetrics_t *metrics =
 		&(((SmuMetricsExternal_t *)(smu_table->metrics_table))->SmuMetrics);
+	SmuMetrics_V2_t *metrics_v2 =
+		&(((SmuMetricsExternal_t *)(smu_table->metrics_table))->SmuMetrics_V2);
+	bool use_metrics_v2 = ((smu->adev->asic_type == CHIP_SIENNA_CICHLID) &&
+		(smu->smc_fw_version >= 0x3A4300)) ? true : false;
+	uint16_t average_gfx_activity;
 	int ret = 0;
 
 	mutex_lock(&smu->metrics_lock);
@@ -510,78 +535,96 @@ static int sienna_cichlid_get_smu_metrics_data(struct smu_context *smu,
 
 	switch (member) {
 	case METRICS_CURR_GFXCLK:
-		*value = metrics->CurrClock[PPCLK_GFXCLK];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_GFXCLK] :
+			metrics->CurrClock[PPCLK_GFXCLK];
 		break;
 	case METRICS_CURR_SOCCLK:
-		*value = metrics->CurrClock[PPCLK_SOCCLK];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_SOCCLK] :
+			metrics->CurrClock[PPCLK_SOCCLK];
 		break;
 	case METRICS_CURR_UCLK:
-		*value = metrics->CurrClock[PPCLK_UCLK];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_UCLK] :
+			metrics->CurrClock[PPCLK_UCLK];
 		break;
 	case METRICS_CURR_VCLK:
-		*value = metrics->CurrClock[PPCLK_VCLK_0];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_VCLK_0] :
+			metrics->CurrClock[PPCLK_VCLK_0];
 		break;
 	case METRICS_CURR_VCLK1:
-		*value = metrics->CurrClock[PPCLK_VCLK_1];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_VCLK_1] :
+			metrics->CurrClock[PPCLK_VCLK_1];
 		break;
 	case METRICS_CURR_DCLK:
-		*value = metrics->CurrClock[PPCLK_DCLK_0];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCLK_0] :
+			metrics->CurrClock[PPCLK_DCLK_0];
 		break;
 	case METRICS_CURR_DCLK1:
-		*value = metrics->CurrClock[PPCLK_DCLK_1];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCLK_1] :
+			metrics->CurrClock[PPCLK_DCLK_1];
 		break;
 	case METRICS_CURR_DCEFCLK:
-		*value = metrics->CurrClock[PPCLK_DCEFCLK];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCEFCLK] :
+			metrics->CurrClock[PPCLK_DCEFCLK];
 		break;
 	case METRICS_CURR_FCLK:
-		*value = metrics->CurrClock[PPCLK_FCLK];
+		*value = use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_FCLK] :
+			metrics->CurrClock[PPCLK_FCLK];
 		break;
 	case METRICS_AVERAGE_GFXCLK:
-		if (metrics->AverageGfxActivity <= SMU_11_0_7_GFX_BUSY_THRESHOLD)
-			*value = metrics->AverageGfxclkFrequencyPostDs;
+		average_gfx_activity = use_metrics_v2 ? metrics_v2->AverageGfxActivity :
+			metrics->AverageGfxActivity;
+		if (average_gfx_activity <= SMU_11_0_7_GFX_BUSY_THRESHOLD)
+			*value = use_metrics_v2 ? metrics_v2->AverageGfxclkFrequencyPostDs :
+				metrics->AverageGfxclkFrequencyPostDs;
 		else
-			*value = metrics->AverageGfxclkFrequencyPreDs;
+			*value = use_metrics_v2 ? metrics_v2->AverageGfxclkFrequencyPreDs :
+				metrics->AverageGfxclkFrequencyPreDs;
 		break;
 	case METRICS_AVERAGE_FCLK:
-		*value = metrics->AverageFclkFrequencyPostDs;
+		*value = use_metrics_v2 ? metrics_v2->AverageFclkFrequencyPostDs :
+			metrics->AverageFclkFrequencyPostDs;
 		break;
 	case METRICS_AVERAGE_UCLK:
-		*value = metrics->AverageUclkFrequencyPostDs;
+		*value = use_metrics_v2 ? metrics_v2->AverageUclkFrequencyPostDs :
+			metrics->AverageUclkFrequencyPostDs;
 		break;
 	case METRICS_AVERAGE_GFXACTIVITY:
-		*value = metrics->AverageGfxActivity;
+		*value = use_metrics_v2 ? metrics_v2->AverageGfxActivity :
+			metrics->AverageGfxActivity;
 		break;
 	case METRICS_AVERAGE_MEMACTIVITY:
-		*value = metrics->AverageUclkActivity;
+		*value = use_metrics_v2 ? metrics_v2->AverageUclkActivity :
+			metrics->AverageUclkActivity;
 		break;
 	case METRICS_AVERAGE_SOCKETPOWER:
-		*value = metrics->AverageSocketPower << 8;
+		*value = use_metrics_v2 ? metrics_v2->AverageSocketPower << 8 :
+			metrics->AverageSocketPower << 8;
 		break;
 	case METRICS_TEMPERATURE_EDGE:
-		*value = metrics->TemperatureEdge *
+		*value = (use_metrics_v2 ? metrics_v2->TemperatureEdge : metrics->TemperatureEdge) *
 			SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_HOTSPOT:
-		*value = metrics->TemperatureHotspot *
+		*value = (use_metrics_v2 ? metrics_v2->TemperatureHotspot : metrics->TemperatureHotspot) *
 			SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_MEM:
-		*value = metrics->TemperatureMem *
+		*value = (use_metrics_v2 ? metrics_v2->TemperatureMem : metrics->TemperatureMem) *
 			SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_VRGFX:
-		*value = metrics->TemperatureVrGfx *
+		*value = (use_metrics_v2 ? metrics_v2->TemperatureVrGfx : metrics->TemperatureVrGfx) *
 			SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_VRSOC:
-		*value = metrics->TemperatureVrSoc *
+		*value = (use_metrics_v2 ? metrics_v2->TemperatureVrSoc : metrics->TemperatureVrSoc) *
 			SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_THROTTLER_STATUS:
-		*value = metrics->ThrottlerStatus;
+		*value = sienna_cichlid_get_throttler_status_locked(smu);
 		break;
 	case METRICS_CURR_FANSPEED:
-		*value = metrics->CurrFanSpeed;
+		*value = use_metrics_v2 ? metrics_v2->CurrFanSpeed : metrics->CurrFanSpeed;
 		break;
 	default:
 		*value = UINT_MAX;
@@ -1911,18 +1954,29 @@ static int sienna_cichlid_set_default_od_settings(struct smu_context *smu)
 		(OverDriveTable_t *)smu->smu_table.overdrive_table;
 	OverDriveTable_t *boot_od_table =
 		(OverDriveTable_t *)smu->smu_table.boot_overdrive_table;
+	OverDriveTable_t *user_od_table =
+		(OverDriveTable_t *)smu->smu_table.user_overdrive_table;
 	int ret = 0;
 
+	/*
+	 * For S3/S4/Runpm resume, no need to setup those overdrive tables again as
+	 *   - either they already have the default OD settings got during cold bootup
+	 *   - or they have some user customized OD settings which cannot be overwritten
+	 */
+	if (smu->adev->in_suspend)
+		return 0;
+
 	ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE,
-				   0, (void *)od_table, false);
+				   0, (void *)boot_od_table, false);
 	if (ret) {
 		dev_err(smu->adev->dev, "Failed to get overdrive table!\n");
 		return ret;
 	}
 
-	memcpy(boot_od_table, od_table, sizeof(OverDriveTable_t));
+	sienna_cichlid_dump_od_table(smu, boot_od_table);
 
-	sienna_cichlid_dump_od_table(smu, od_table);
+	memcpy(od_table, boot_od_table, sizeof(OverDriveTable_t));
+	memcpy(user_od_table, boot_od_table, sizeof(OverDriveTable_t));
 
 	return 0;
 }
@@ -2085,13 +2139,20 @@ static int sienna_cichlid_od_edit_dpm_table(struct smu_context *smu,
 		fallthrough;
 
 	case PP_OD_COMMIT_DPM_TABLE:
-		sienna_cichlid_dump_od_table(smu, od_table);
+		if (memcmp(od_table, table_context->user_overdrive_table, sizeof(OverDriveTable_t))) {
+			sienna_cichlid_dump_od_table(smu, od_table);
+			ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)od_table, true);
+			if (ret) {
+				dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
+				return ret;
+			}
+			memcpy(table_context->user_overdrive_table, od_table, sizeof(OverDriveTable_t));
+			smu->user_dpm_profile.user_od = true;
 
-		ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE,
-					   0, (void *)od_table, true);
-		if (ret) {
-			dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
-			return ret;
+			if (!memcmp(table_context->user_overdrive_table,
+				    table_context->boot_overdrive_table,
+				    sizeof(OverDriveTable_t)))
+				smu->user_dpm_profile.user_od = false;
 		}
 		break;
 
@@ -3442,197 +3503,77 @@ static void sienna_cichlid_dump_pptable(struct smu_context *smu)
 	dev_info(smu->adev->dev, "MmHubPadding[7] = 0x%x\n", pptable->MmHubPadding[7]);
 }
 
-static void sienna_cichlid_fill_i2c_req(SwI2cRequest_t  *req, bool write,
-				  uint8_t address, uint32_t numbytes,
-				  uint8_t *data)
-{
-	int i;
-
-	req->I2CcontrollerPort = 1;
-	req->I2CSpeed = 2;
-	req->SlaveAddress = address;
-	req->NumCmds = numbytes;
-
-	for (i = 0; i < numbytes; i++) {
-		SwI2cCmd_t *cmd =  &req->SwI2cCmds[i];
-
-		/* First 2 bytes are always write for lower 2b EEPROM address */
-		if (i < 2)
-			cmd->CmdConfig = CMDCONFIG_READWRITE_MASK;
-		else
-			cmd->CmdConfig = write ? CMDCONFIG_READWRITE_MASK : 0;
-
-
-		/* Add RESTART for read  after address filled */
-		cmd->CmdConfig |= (i == 2 && !write) ? CMDCONFIG_RESTART_MASK : 0;
-
-		/* Add STOP in the end */
-		cmd->CmdConfig |= (i == (numbytes - 1)) ? CMDCONFIG_STOP_MASK : 0;
-
-		/* Fill with data regardless if read or write to simplify code */
-		cmd->ReadWriteData = data[i];
-	}
-}
-
-static int sienna_cichlid_i2c_read_data(struct i2c_adapter *control,
-					       uint8_t address,
-					       uint8_t *data,
-					       uint32_t numbytes)
+static int sienna_cichlid_i2c_xfer(struct i2c_adapter *i2c_adap,
+				   struct i2c_msg *msg, int num_msgs)
 {
-	uint32_t  i, ret = 0;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_device *adev = to_amdgpu_device(i2c_adap);
 	struct smu_table_context *smu_table = &adev->smu.smu_table;
 	struct smu_table *table = &smu_table->driver_table;
+	SwI2cRequest_t *req, *res = (SwI2cRequest_t *)table->cpu_addr;
+	int i, j, r, c;
+	u16 dir;
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
-	}
-
-	memset(&req, 0, sizeof(req));
-	sienna_cichlid_fill_i2c_req(&req, false, address, numbytes, data);
-
-	mutex_lock(&adev->smu.mutex);
-	/* Now read data starting with that address */
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req,
-					true);
-	mutex_unlock(&adev->smu.mutex);
-
-	if (!ret) {
-		SwI2cRequest_t *res = (SwI2cRequest_t *)table->cpu_addr;
-
-		/* Assume SMU  fills res.SwI2cCmds[i].Data with read bytes */
-		for (i = 0; i < numbytes; i++)
-			data[i] = res->SwI2cCmds[i].ReadWriteData;
-
-		dev_dbg(adev->dev, "sienna_cichlid_i2c_read_data, address = %x, bytes = %d, data :",
-				  (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-	} else
-		dev_err(adev->dev, "sienna_cichlid_i2c_read_data - error occurred :%x", ret);
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (!req)
+		return -ENOMEM;
 
-	return ret;
-}
+	req->I2CcontrollerPort = 1;
+	req->I2CSpeed = I2C_SPEED_FAST_400K;
+	req->SlaveAddress = msg[0].addr << 1; /* wants an 8-bit address */
+	dir = msg[0].flags & I2C_M_RD;
+
+	for (c = i = 0; i < num_msgs; i++) {
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &req->SwI2cCmds[c];
+
+			if (!(msg[i].flags & I2C_M_RD)) {
+				/* write */
+				cmd->CmdConfig |= CMDCONFIG_READWRITE_MASK;
+				cmd->ReadWriteData = msg[i].buf[j];
+			}
 
-static int sienna_cichlid_i2c_write_data(struct i2c_adapter *control,
-						uint8_t address,
-						uint8_t *data,
-						uint32_t numbytes)
-{
-	uint32_t ret;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+			if ((dir ^ msg[i].flags) & I2C_M_RD) {
+				/* The direction changes.
+				 */
+				dir = msg[i].flags & I2C_M_RD;
+				cmd->CmdConfig |= CMDCONFIG_RESTART_MASK;
+			}
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
+			req->NumCmds++;
+
+			/*
+			 * Insert STOP if we are at the last byte of either last
+			 * message for the transaction or the client explicitly
+			 * requires a STOP at this particular message.
+			 */
+			if ((j == msg[i].len - 1) &&
+			    ((i == num_msgs - 1) || (msg[i].flags & I2C_M_STOP))) {
+				cmd->CmdConfig &= ~CMDCONFIG_RESTART_MASK;
+				cmd->CmdConfig |= CMDCONFIG_STOP_MASK;
+			}
+		}
 	}
-
-	memset(&req, 0, sizeof(req));
-	sienna_cichlid_fill_i2c_req(&req, true, address, numbytes, data);
-
 	mutex_lock(&adev->smu.mutex);
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req, true);
+	r = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, req, true);
 	mutex_unlock(&adev->smu.mutex);
+	if (r)
+		goto fail;
 
-	if (!ret) {
-		dev_dbg(adev->dev, "sienna_cichlid_i2c_write(), address = %x, bytes = %d , data: ",
-					 (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-		/*
-		 * According to EEPROM spec there is a MAX of 10 ms required for
-		 * EEPROM to flush internal RX buffer after STOP was issued at the
-		 * end of write transaction. During this time the EEPROM will not be
-		 * responsive to any more commands - so wait a bit more.
-		 */
-		msleep(10);
-
-	} else
-		dev_err(adev->dev, "sienna_cichlid_i2c_write- error occurred :%x", ret);
-
-	return ret;
-}
-
-static int sienna_cichlid_i2c_xfer(struct i2c_adapter *i2c_adap,
-			      struct i2c_msg *msgs, int num)
-{
-	uint32_t  i, j, ret, data_size, data_chunk_size, next_eeprom_addr = 0;
-	uint8_t *data_ptr, data_chunk[MAX_SW_I2C_COMMANDS] = { 0 };
-
-	for (i = 0; i < num; i++) {
-		/*
-		 * SMU interface allows at most MAX_SW_I2C_COMMANDS bytes of data at
-		 * once and hence the data needs to be spliced into chunks and sent each
-		 * chunk separately
-		 */
-		data_size = msgs[i].len - 2;
-		data_chunk_size = MAX_SW_I2C_COMMANDS - 2;
-		next_eeprom_addr = (msgs[i].buf[0] << 8 & 0xff00) | (msgs[i].buf[1] & 0xff);
-		data_ptr = msgs[i].buf + 2;
-
-		for (j = 0; j < data_size / data_chunk_size; j++) {
-			/* Insert the EEPROM dest addess, bits 0-15 */
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = sienna_cichlid_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, MAX_SW_I2C_COMMANDS);
-
-				memcpy(data_ptr, data_chunk + 2, data_chunk_size);
-			} else {
-
-				memcpy(data_chunk + 2, data_ptr, data_chunk_size);
-
-				ret = sienna_cichlid_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, MAX_SW_I2C_COMMANDS);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
-
-			next_eeprom_addr += data_chunk_size;
-			data_ptr += data_chunk_size;
+	for (c = i = 0; i < num_msgs; i++) {
+		if (!(msg[i].flags & I2C_M_RD)) {
+			c += msg[i].len;
+			continue;
 		}
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &res->SwI2cCmds[c];
 
-		if (data_size % data_chunk_size) {
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = sienna_cichlid_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, (data_size % data_chunk_size) + 2);
-
-				memcpy(data_ptr, data_chunk + 2, data_size % data_chunk_size);
-			} else {
-				memcpy(data_chunk + 2, data_ptr, data_size % data_chunk_size);
-
-				ret = sienna_cichlid_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, (data_size % data_chunk_size) + 2);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
+			msg[i].buf[j] = cmd->ReadWriteData;
 		}
 	}
-
+	r = num_msgs;
 fail:
-	return num;
+	kfree(req);
+	return r;
 }
 
 static u32 sienna_cichlid_i2c_func(struct i2c_adapter *adap)
@@ -3646,16 +3587,25 @@ static const struct i2c_algorithm sienna_cichlid_i2c_algo = {
 	.functionality = sienna_cichlid_i2c_func,
 };
 
+static const struct i2c_adapter_quirks sienna_cichlid_i2c_control_quirks = {
+	.flags = I2C_AQ_COMB | I2C_AQ_COMB_SAME_ADDR | I2C_AQ_NO_ZERO_LEN,
+	.max_read_len  = MAX_SW_I2C_COMMANDS,
+	.max_write_len = MAX_SW_I2C_COMMANDS,
+	.max_comb_1st_msg_len = 2,
+	.max_comb_2nd_msg_len = MAX_SW_I2C_COMMANDS - 2,
+};
+
 static int sienna_cichlid_i2c_control_init(struct smu_context *smu, struct i2c_adapter *control)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
 	int res;
 
 	control->owner = THIS_MODULE;
-	control->class = I2C_CLASS_SPD;
+	control->class = I2C_CLASS_HWMON;
 	control->dev.parent = &adev->pdev->dev;
 	control->algo = &sienna_cichlid_i2c_algo;
 	snprintf(control->name, sizeof(control->name), "AMDGPU SMU");
+	control->quirks = &sienna_cichlid_i2c_control_quirks;
 
 	res = i2c_add_adapter(control);
 	if (res)
@@ -3678,65 +3628,94 @@ static ssize_t sienna_cichlid_get_gpu_metrics(struct smu_context *smu,
 	SmuMetricsExternal_t metrics_external;
 	SmuMetrics_t *metrics =
 		&(metrics_external.SmuMetrics);
+	SmuMetrics_V2_t *metrics_v2 =
+		&(metrics_external.SmuMetrics_V2);
 	struct amdgpu_device *adev = smu->adev;
-	uint32_t smu_version;
+	bool use_metrics_v2 = ((adev->asic_type == CHIP_SIENNA_CICHLID) &&
+		(smu->smc_fw_version >= 0x3A4300)) ? true : false;
+	uint16_t average_gfx_activity;
 	int ret = 0;
 
-	ret = smu_cmn_get_metrics_table(smu,
-					&metrics_external,
-					true);
-	if (ret)
+	mutex_lock(&smu->metrics_lock);
+	ret = smu_cmn_get_metrics_table_locked(smu,
+					       &metrics_external,
+					       true);
+	if (ret) {
+		mutex_unlock(&smu->metrics_lock);
 		return ret;
+	}
 
 	smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 3);
 
-	gpu_metrics->temperature_edge = metrics->TemperatureEdge;
-	gpu_metrics->temperature_hotspot = metrics->TemperatureHotspot;
-	gpu_metrics->temperature_mem = metrics->TemperatureMem;
-	gpu_metrics->temperature_vrgfx = metrics->TemperatureVrGfx;
-	gpu_metrics->temperature_vrsoc = metrics->TemperatureVrSoc;
-	gpu_metrics->temperature_vrmem = metrics->TemperatureVrMem0;
-
-	gpu_metrics->average_gfx_activity = metrics->AverageGfxActivity;
-	gpu_metrics->average_umc_activity = metrics->AverageUclkActivity;
-	gpu_metrics->average_mm_activity = metrics->VcnActivityPercentage;
-
-	gpu_metrics->average_socket_power = metrics->AverageSocketPower;
-	gpu_metrics->energy_accumulator = metrics->EnergyAccumulator;
-
-	if (metrics->AverageGfxActivity <= SMU_11_0_7_GFX_BUSY_THRESHOLD)
-		gpu_metrics->average_gfxclk_frequency = metrics->AverageGfxclkFrequencyPostDs;
+	gpu_metrics->temperature_edge =
+		use_metrics_v2 ? metrics_v2->TemperatureEdge : metrics->TemperatureEdge;
+	gpu_metrics->temperature_hotspot =
+		use_metrics_v2 ? metrics_v2->TemperatureHotspot : metrics->TemperatureHotspot;
+	gpu_metrics->temperature_mem =
+		use_metrics_v2 ? metrics_v2->TemperatureMem : metrics->TemperatureMem;
+	gpu_metrics->temperature_vrgfx =
+		use_metrics_v2 ? metrics_v2->TemperatureVrGfx : metrics->TemperatureVrGfx;
+	gpu_metrics->temperature_vrsoc =
+		use_metrics_v2 ? metrics_v2->TemperatureVrSoc : metrics->TemperatureVrSoc;
+	gpu_metrics->temperature_vrmem =
+		use_metrics_v2 ? metrics_v2->TemperatureVrMem0 : metrics->TemperatureVrMem0;
+
+	gpu_metrics->average_gfx_activity =
+		use_metrics_v2 ? metrics_v2->AverageGfxActivity : metrics->AverageGfxActivity;
+	gpu_metrics->average_umc_activity =
+		use_metrics_v2 ? metrics_v2->AverageUclkActivity : metrics->AverageUclkActivity;
+	gpu_metrics->average_mm_activity =
+		use_metrics_v2 ? metrics_v2->VcnActivityPercentage : metrics->VcnActivityPercentage;
+
+	gpu_metrics->average_socket_power =
+		use_metrics_v2 ? metrics_v2->AverageSocketPower : metrics->AverageSocketPower;
+	gpu_metrics->energy_accumulator =
+		use_metrics_v2 ? metrics_v2->EnergyAccumulator : metrics->EnergyAccumulator;
+
+	average_gfx_activity = use_metrics_v2 ? metrics_v2->AverageGfxActivity : metrics->AverageGfxActivity;
+	if (average_gfx_activity <= SMU_11_0_7_GFX_BUSY_THRESHOLD)
+		gpu_metrics->average_gfxclk_frequency =
+			use_metrics_v2 ? metrics_v2->AverageGfxclkFrequencyPostDs : metrics->AverageGfxclkFrequencyPostDs;
 	else
-		gpu_metrics->average_gfxclk_frequency = metrics->AverageGfxclkFrequencyPreDs;
-	gpu_metrics->average_uclk_frequency = metrics->AverageUclkFrequencyPostDs;
-	gpu_metrics->average_vclk0_frequency = metrics->AverageVclk0Frequency;
-	gpu_metrics->average_dclk0_frequency = metrics->AverageDclk0Frequency;
-	gpu_metrics->average_vclk1_frequency = metrics->AverageVclk1Frequency;
-	gpu_metrics->average_dclk1_frequency = metrics->AverageDclk1Frequency;
-
-	gpu_metrics->current_gfxclk = metrics->CurrClock[PPCLK_GFXCLK];
-	gpu_metrics->current_socclk = metrics->CurrClock[PPCLK_SOCCLK];
-	gpu_metrics->current_uclk = metrics->CurrClock[PPCLK_UCLK];
-	gpu_metrics->current_vclk0 = metrics->CurrClock[PPCLK_VCLK_0];
-	gpu_metrics->current_dclk0 = metrics->CurrClock[PPCLK_DCLK_0];
-	gpu_metrics->current_vclk1 = metrics->CurrClock[PPCLK_VCLK_1];
-	gpu_metrics->current_dclk1 = metrics->CurrClock[PPCLK_DCLK_1];
-
-	gpu_metrics->throttle_status = metrics->ThrottlerStatus;
+		gpu_metrics->average_gfxclk_frequency =
+			use_metrics_v2 ? metrics_v2->AverageGfxclkFrequencyPreDs : metrics->AverageGfxclkFrequencyPreDs;
+	gpu_metrics->average_uclk_frequency =
+		use_metrics_v2 ? metrics_v2->AverageUclkFrequencyPostDs : metrics->AverageUclkFrequencyPostDs;
+	gpu_metrics->average_vclk0_frequency =
+		use_metrics_v2 ? metrics_v2->AverageVclk0Frequency : metrics->AverageVclk0Frequency;
+	gpu_metrics->average_dclk0_frequency =
+		use_metrics_v2 ? metrics_v2->AverageDclk0Frequency : metrics->AverageDclk0Frequency;
+	gpu_metrics->average_vclk1_frequency =
+		use_metrics_v2 ? metrics_v2->AverageVclk1Frequency : metrics->AverageVclk1Frequency;
+	gpu_metrics->average_dclk1_frequency =
+		use_metrics_v2 ? metrics_v2->AverageDclk1Frequency : metrics->AverageDclk1Frequency;
+
+	gpu_metrics->current_gfxclk =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_GFXCLK] : metrics->CurrClock[PPCLK_GFXCLK];
+	gpu_metrics->current_socclk =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_SOCCLK] : metrics->CurrClock[PPCLK_SOCCLK];
+	gpu_metrics->current_uclk =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_UCLK] : metrics->CurrClock[PPCLK_UCLK];
+	gpu_metrics->current_vclk0 =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_VCLK_0] : metrics->CurrClock[PPCLK_VCLK_0];
+	gpu_metrics->current_dclk0 =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCLK_0] : metrics->CurrClock[PPCLK_DCLK_0];
+	gpu_metrics->current_vclk1 =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_VCLK_1] : metrics->CurrClock[PPCLK_VCLK_1];
+	gpu_metrics->current_dclk1 =
+		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCLK_1] : metrics->CurrClock[PPCLK_DCLK_1];
+
+	gpu_metrics->throttle_status = sienna_cichlid_get_throttler_status_locked(smu);
 	gpu_metrics->indep_throttle_status =
-			smu_cmn_get_indep_throttler_status(metrics->ThrottlerStatus,
+			smu_cmn_get_indep_throttler_status(gpu_metrics->throttle_status,
 							   sienna_cichlid_throttler_map);
 
-	gpu_metrics->current_fan_speed = metrics->CurrFanSpeed;
+	gpu_metrics->current_fan_speed = use_metrics_v2 ? metrics_v2->CurrFanSpeed : metrics->CurrFanSpeed;
 
-	ret = smu_cmn_get_smc_version(smu, NULL, &smu_version);
-	if (ret)
-		return ret;
-
-	if (((adev->asic_type == CHIP_SIENNA_CICHLID) && smu_version > 0x003A1E00) ||
-	      ((adev->asic_type == CHIP_NAVY_FLOUNDER) && smu_version > 0x00410400)) {
-		gpu_metrics->pcie_link_width = metrics->PcieWidth;
-		gpu_metrics->pcie_link_speed = link_speed[metrics->PcieRate];
+	if (((adev->asic_type == CHIP_SIENNA_CICHLID) && smu->smc_fw_version > 0x003A1E00) ||
+	      ((adev->asic_type == CHIP_NAVY_FLOUNDER) && smu->smc_fw_version > 0x00410400)) {
+		gpu_metrics->pcie_link_width = use_metrics_v2 ? metrics_v2->PcieWidth : metrics->PcieWidth;
+		gpu_metrics->pcie_link_speed = link_speed[use_metrics_v2 ? metrics_v2->PcieRate : metrics->PcieRate];
 	} else {
 		gpu_metrics->pcie_link_width =
 				smu_v11_0_get_current_pcie_link_width(smu);
@@ -3744,6 +3723,8 @@ static ssize_t sienna_cichlid_get_gpu_metrics(struct smu_context *smu,
 				smu_v11_0_get_current_pcie_link_speed(smu);
 	}
 
+	mutex_unlock(&smu->metrics_lock);
+
 	gpu_metrics->system_clock_counter = ktime_get_boottime_ns();
 
 	*table = (void *)gpu_metrics;
@@ -3939,6 +3920,7 @@ static const struct pptable_funcs sienna_cichlid_ppt_funcs = {
 	.set_soft_freq_limited_range = smu_v11_0_set_soft_freq_limited_range,
 	.set_default_od_settings = sienna_cichlid_set_default_od_settings,
 	.od_edit_dpm_table = sienna_cichlid_od_edit_dpm_table,
+	.restore_user_od_settings = smu_v11_0_restore_user_od_settings,
 	.run_btc = sienna_cichlid_run_btc,
 	.set_power_source = smu_v11_0_set_power_source,
 	.get_pp_feature_mask = smu_cmn_get_pp_feature_mask,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
index 0a5d46ac9ccd..b5419e8eba89 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c
@@ -266,6 +266,9 @@ int smu_v11_0_check_fw_version(struct smu_context *smu)
 	case CHIP_BEIGE_GOBY:
 		smu->smc_driver_if_version = SMU11_DRIVER_IF_VERSION_Beige_Goby;
 		break;
+	case CHIP_CYAN_SKILLFISH:
+		smu->smc_driver_if_version = SMU11_DRIVER_IF_VERSION_Cyan_Skillfish;
+		break;
 	default:
 		dev_err(smu->adev->dev, "smu unsupported asic type:%d.\n", smu->adev->asic_type);
 		smu->smc_driver_if_version = SMU11_DRIVER_IF_VERSION_INV;
@@ -422,10 +425,20 @@ int smu_v11_0_init_smc_tables(struct smu_context *smu)
 			ret = -ENOMEM;
 			goto err3_out;
 		}
+
+		smu_table->user_overdrive_table =
+			kzalloc(tables[SMU_TABLE_OVERDRIVE].size, GFP_KERNEL);
+		if (!smu_table->user_overdrive_table) {
+			ret = -ENOMEM;
+			goto err4_out;
+		}
+
 	}
 
 	return 0;
 
+err4_out:
+	kfree(smu_table->boot_overdrive_table);
 err3_out:
 	kfree(smu_table->overdrive_table);
 err2_out:
@@ -442,12 +455,14 @@ int smu_v11_0_fini_smc_tables(struct smu_context *smu)
 	struct smu_dpm_context *smu_dpm = &smu->smu_dpm;
 
 	kfree(smu_table->gpu_metrics_table);
+	kfree(smu_table->user_overdrive_table);
 	kfree(smu_table->boot_overdrive_table);
 	kfree(smu_table->overdrive_table);
 	kfree(smu_table->max_sustainable_clocks);
 	kfree(smu_table->driver_pptable);
 	kfree(smu_table->clocks_table);
 	smu_table->gpu_metrics_table = NULL;
+	smu_table->user_overdrive_table = NULL;
 	smu_table->boot_overdrive_table = NULL;
 	smu_table->overdrive_table = NULL;
 	smu_table->max_sustainable_clocks = NULL;
@@ -2101,3 +2116,16 @@ int smu_v11_0_deep_sleep_control(struct smu_context *smu,
 
 	return ret;
 }
+
+int smu_v11_0_restore_user_od_settings(struct smu_context *smu)
+{
+	struct smu_table_context *table_context = &smu->smu_table;
+	void *user_od_table = table_context->user_overdrive_table;
+	int ret = 0;
+
+	ret = smu_cmn_update_table(smu, SMU_TABLE_OVERDRIVE, 0, (void *)user_od_table, true);
+	if (ret)
+		dev_err(smu->adev->dev, "Failed to import overdrive table!\n");
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
index cb5485cf243f..856eeaf293b8 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
@@ -1451,197 +1451,77 @@ static bool aldebaran_is_dpm_running(struct smu_context *smu)
 	return !!(feature_enabled & SMC_DPM_FEATURE);
 }
 
-static void aldebaran_fill_i2c_req(SwI2cRequest_t  *req, bool write,
-				  uint8_t address, uint32_t numbytes,
-				  uint8_t *data)
-{
-	int i;
-
-	req->I2CcontrollerPort = 0;
-	req->I2CSpeed = 2;
-	req->SlaveAddress = address;
-	req->NumCmds = numbytes;
-
-	for (i = 0; i < numbytes; i++) {
-		SwI2cCmd_t *cmd =  &req->SwI2cCmds[i];
-
-		/* First 2 bytes are always write for lower 2b EEPROM address */
-		if (i < 2)
-			cmd->CmdConfig = CMDCONFIG_READWRITE_MASK;
-		else
-			cmd->CmdConfig = write ? CMDCONFIG_READWRITE_MASK : 0;
-
-
-		/* Add RESTART for read  after address filled */
-		cmd->CmdConfig |= (i == 2 && !write) ? CMDCONFIG_RESTART_MASK : 0;
-
-		/* Add STOP in the end */
-		cmd->CmdConfig |= (i == (numbytes - 1)) ? CMDCONFIG_STOP_MASK : 0;
-
-		/* Fill with data regardless if read or write to simplify code */
-		cmd->ReadWriteData = data[i];
-	}
-}
-
-static int aldebaran_i2c_read_data(struct i2c_adapter *control,
-					       uint8_t address,
-					       uint8_t *data,
-					       uint32_t numbytes)
+static int aldebaran_i2c_xfer(struct i2c_adapter *i2c_adap,
+			      struct i2c_msg *msg, int num_msgs)
 {
-	uint32_t  i, ret = 0;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_device *adev = to_amdgpu_device(i2c_adap);
 	struct smu_table_context *smu_table = &adev->smu.smu_table;
 	struct smu_table *table = &smu_table->driver_table;
+	SwI2cRequest_t *req, *res = (SwI2cRequest_t *)table->cpu_addr;
+	int i, j, r, c;
+	u16 dir;
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
-	}
-
-	memset(&req, 0, sizeof(req));
-	aldebaran_fill_i2c_req(&req, false, address, numbytes, data);
-
-	mutex_lock(&adev->smu.mutex);
-	/* Now read data starting with that address */
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req,
-					true);
-	mutex_unlock(&adev->smu.mutex);
-
-	if (!ret) {
-		SwI2cRequest_t *res = (SwI2cRequest_t *)table->cpu_addr;
-
-		/* Assume SMU  fills res.SwI2cCmds[i].Data with read bytes */
-		for (i = 0; i < numbytes; i++)
-			data[i] = res->SwI2cCmds[i].ReadWriteData;
-
-		dev_dbg(adev->dev, "aldebaran_i2c_read_data, address = %x, bytes = %d, data :",
-				  (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-	} else
-		dev_err(adev->dev, "aldebaran_i2c_read_data - error occurred :%x", ret);
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (!req)
+		return -ENOMEM;
 
-	return ret;
-}
+	req->I2CcontrollerPort = 0;
+	req->I2CSpeed = I2C_SPEED_FAST_400K;
+	req->SlaveAddress = msg[0].addr << 1; /* wants an 8-bit address */
+	dir = msg[0].flags & I2C_M_RD;
+
+	for (c = i = 0; i < num_msgs; i++) {
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &req->SwI2cCmds[c];
+
+			if (!(msg[i].flags & I2C_M_RD)) {
+				/* write */
+				cmd->CmdConfig |= CMDCONFIG_READWRITE_MASK;
+				cmd->ReadWriteData = msg[i].buf[j];
+			}
 
-static int aldebaran_i2c_write_data(struct i2c_adapter *control,
-						uint8_t address,
-						uint8_t *data,
-						uint32_t numbytes)
-{
-	uint32_t ret;
-	SwI2cRequest_t req;
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+			if ((dir ^ msg[i].flags) & I2C_M_RD) {
+				/* The direction changes.
+				 */
+				dir = msg[i].flags & I2C_M_RD;
+				cmd->CmdConfig |= CMDCONFIG_RESTART_MASK;
+			}
 
-	if (numbytes > MAX_SW_I2C_COMMANDS) {
-		dev_err(adev->dev, "numbytes requested %d is over max allowed %d\n",
-			numbytes, MAX_SW_I2C_COMMANDS);
-		return -EINVAL;
+			req->NumCmds++;
+
+			/*
+			 * Insert STOP if we are at the last byte of either last
+			 * message for the transaction or the client explicitly
+			 * requires a STOP at this particular message.
+			 */
+			if ((j == msg[i].len - 1) &&
+			    ((i == num_msgs - 1) || (msg[i].flags & I2C_M_STOP))) {
+				cmd->CmdConfig &= ~CMDCONFIG_RESTART_MASK;
+				cmd->CmdConfig |= CMDCONFIG_STOP_MASK;
+			}
+		}
 	}
-
-	memset(&req, 0, sizeof(req));
-	aldebaran_fill_i2c_req(&req, true, address, numbytes, data);
-
 	mutex_lock(&adev->smu.mutex);
-	ret = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, &req, true);
+	r = smu_cmn_update_table(&adev->smu, SMU_TABLE_I2C_COMMANDS, 0, req, true);
 	mutex_unlock(&adev->smu.mutex);
+	if (r)
+		goto fail;
 
-	if (!ret) {
-		dev_dbg(adev->dev, "aldebaran_i2c_write(), address = %x, bytes = %d , data: ",
-					 (uint16_t)address, numbytes);
-
-		print_hex_dump(KERN_DEBUG, "data: ", DUMP_PREFIX_NONE,
-			       8, 1, data, numbytes, false);
-		/*
-		 * According to EEPROM spec there is a MAX of 10 ms required for
-		 * EEPROM to flush internal RX buffer after STOP was issued at the
-		 * end of write transaction. During this time the EEPROM will not be
-		 * responsive to any more commands - so wait a bit more.
-		 */
-		msleep(10);
-
-	} else
-		dev_err(adev->dev, "aldebaran_i2c_write- error occurred :%x", ret);
-
-	return ret;
-}
-
-static int aldebaran_i2c_xfer(struct i2c_adapter *i2c_adap,
-			      struct i2c_msg *msgs, int num)
-{
-	uint32_t  i, j, ret, data_size, data_chunk_size, next_eeprom_addr = 0;
-	uint8_t *data_ptr, data_chunk[MAX_SW_I2C_COMMANDS] = { 0 };
-
-	for (i = 0; i < num; i++) {
-		/*
-		 * SMU interface allows at most MAX_SW_I2C_COMMANDS bytes of data at
-		 * once and hence the data needs to be spliced into chunks and sent each
-		 * chunk separately
-		 */
-		data_size = msgs[i].len - 2;
-		data_chunk_size = MAX_SW_I2C_COMMANDS - 2;
-		next_eeprom_addr = (msgs[i].buf[0] << 8 & 0xff00) | (msgs[i].buf[1] & 0xff);
-		data_ptr = msgs[i].buf + 2;
-
-		for (j = 0; j < data_size / data_chunk_size; j++) {
-			/* Insert the EEPROM dest addess, bits 0-15 */
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = aldebaran_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, MAX_SW_I2C_COMMANDS);
-
-				memcpy(data_ptr, data_chunk + 2, data_chunk_size);
-			} else {
-
-				memcpy(data_chunk + 2, data_ptr, data_chunk_size);
-
-				ret = aldebaran_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, MAX_SW_I2C_COMMANDS);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
-
-			next_eeprom_addr += data_chunk_size;
-			data_ptr += data_chunk_size;
+	for (c = i = 0; i < num_msgs; i++) {
+		if (!(msg[i].flags & I2C_M_RD)) {
+			c += msg[i].len;
+			continue;
 		}
+		for (j = 0; j < msg[i].len; j++, c++) {
+			SwI2cCmd_t *cmd = &res->SwI2cCmds[c];
 
-		if (data_size % data_chunk_size) {
-			data_chunk[0] = ((next_eeprom_addr >> 8) & 0xff);
-			data_chunk[1] = (next_eeprom_addr & 0xff);
-
-			if (msgs[i].flags & I2C_M_RD) {
-				ret = aldebaran_i2c_read_data(i2c_adap,
-							     (uint8_t)msgs[i].addr,
-							     data_chunk, (data_size % data_chunk_size) + 2);
-
-				memcpy(data_ptr, data_chunk + 2, data_size % data_chunk_size);
-			} else {
-				memcpy(data_chunk + 2, data_ptr, data_size % data_chunk_size);
-
-				ret = aldebaran_i2c_write_data(i2c_adap,
-							      (uint8_t)msgs[i].addr,
-							      data_chunk, (data_size % data_chunk_size) + 2);
-			}
-
-			if (ret) {
-				num = -EIO;
-				goto fail;
-			}
+			msg[i].buf[j] = cmd->ReadWriteData;
 		}
 	}
-
+	r = num_msgs;
 fail:
-	return num;
+	kfree(req);
+	return r;
 }
 
 static u32 aldebaran_i2c_func(struct i2c_adapter *adap)
@@ -1655,6 +1535,14 @@ static const struct i2c_algorithm aldebaran_i2c_algo = {
 	.functionality = aldebaran_i2c_func,
 };
 
+static const struct i2c_adapter_quirks aldebaran_i2c_control_quirks = {
+	.flags = I2C_AQ_COMB | I2C_AQ_COMB_SAME_ADDR | I2C_AQ_NO_ZERO_LEN,
+	.max_read_len  = MAX_SW_I2C_COMMANDS,
+	.max_write_len = MAX_SW_I2C_COMMANDS,
+	.max_comb_1st_msg_len = 2,
+	.max_comb_2nd_msg_len = MAX_SW_I2C_COMMANDS - 2,
+};
+
 static int aldebaran_i2c_control_init(struct smu_context *smu, struct i2c_adapter *control)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
@@ -1665,6 +1553,7 @@ static int aldebaran_i2c_control_init(struct smu_context *smu, struct i2c_adapte
 	control->dev.parent = &adev->pdev->dev;
 	control->algo = &aldebaran_i2c_algo;
 	snprintf(control->name, sizeof(control->name), "AMDGPU SMU");
+	control->quirks = &aldebaran_i2c_control_quirks;
 
 	res = i2c_add_adapter(control);
 	if (res)
@@ -1764,7 +1653,9 @@ static void aldebaran_log_thermal_throttling_event(struct smu_context *smu)
 
 	dev_warn(adev->dev, "WARN: GPU thermal throttling temperature reached, expect performance decrease. %s.\n",
 		 log_buf);
-	kgd2kfd_smi_event_throttle(smu->adev->kfd.dev, throttler_status);
+	kgd2kfd_smi_event_throttle(smu->adev->kfd.dev,
+		smu_cmn_get_indep_throttler_status(throttler_status,
+						   aldebaran_throttler_map));
 }
 
 static int aldebaran_get_current_pcie_link_speed(struct smu_context *smu)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
index e802f9a95f08..415be74df28c 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
@@ -55,7 +55,7 @@
 
 #undef __SMU_DUMMY_MAP
 #define __SMU_DUMMY_MAP(type)	#type
-static const char* __smu_message_names[] = {
+static const char * const __smu_message_names[] = {
 	SMU_MESSAGE_TYPES
 };
 
@@ -76,55 +76,256 @@ static void smu_cmn_read_arg(struct smu_context *smu,
 	*arg = RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_82);
 }
 
-int smu_cmn_wait_for_response(struct smu_context *smu)
+/* Redefine the SMU error codes here.
+ *
+ * Note that these definitions are redundant and should be removed
+ * when the SMU has exported a unified header file containing these
+ * macros, which header file we can just include and use the SMU's
+ * macros. At the moment, these error codes are defined by the SMU
+ * per-ASIC unfortunately, yet we're a one driver for all ASICs.
+ */
+#define SMU_RESP_NONE           0
+#define SMU_RESP_OK             1
+#define SMU_RESP_CMD_FAIL       0xFF
+#define SMU_RESP_CMD_UNKNOWN    0xFE
+#define SMU_RESP_CMD_BAD_PREREQ 0xFD
+#define SMU_RESP_BUSY_OTHER     0xFC
+#define SMU_RESP_DEBUG_END      0xFB
+
+/**
+ * __smu_cmn_poll_stat -- poll for a status from the SMU
+ * smu: a pointer to SMU context
+ *
+ * Returns the status of the SMU, which could be,
+ *    0, the SMU is busy with your previous command;
+ *    1, execution status: success, execution result: success;
+ * 0xFF, execution status: success, execution result: failure;
+ * 0xFE, unknown command;
+ * 0xFD, valid command, but bad (command) prerequisites;
+ * 0xFC, the command was rejected as the SMU is busy;
+ * 0xFB, "SMC_Result_DebugDataDumpEnd".
+ *
+ * The values here are not defined by macros, because I'd rather we
+ * include a single header file which defines them, which is
+ * maintained by the SMU FW team, so that we're impervious to firmware
+ * changes. At the moment those values are defined in various header
+ * files, one for each ASIC, yet here we're a single ASIC-agnostic
+ * interface. Such a change can be followed-up by a subsequent patch.
+ */
+static u32 __smu_cmn_poll_stat(struct smu_context *smu)
 {
 	struct amdgpu_device *adev = smu->adev;
-	uint32_t cur_value, i, timeout = adev->usec_timeout * 20;
+	int timeout = adev->usec_timeout * 20;
+	u32 reg;
 
-	for (i = 0; i < timeout; i++) {
-		cur_value = RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_90);
-		if ((cur_value & MP1_C2PMSG_90__CONTENT_MASK) != 0)
-			return cur_value;
+	for ( ; timeout > 0; timeout--) {
+		reg = RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_90);
+		if ((reg & MP1_C2PMSG_90__CONTENT_MASK) != 0)
+			break;
 
 		udelay(1);
 	}
 
-	/* timeout means wrong logic */
-	if (i == timeout)
-		return -ETIME;
-
-	return RREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_90);
+	return reg;
 }
 
-int smu_cmn_send_msg_without_waiting(struct smu_context *smu,
-				     uint16_t msg, uint32_t param)
+static void __smu_cmn_reg_print_error(struct smu_context *smu,
+				      u32 reg_c2pmsg_90,
+				      int msg_index,
+				      u32 param,
+				      enum smu_message_type msg)
 {
 	struct amdgpu_device *adev = smu->adev;
-	int ret;
+	const char *message = smu_get_message_name(smu, msg);
 
-	ret = smu_cmn_wait_for_response(smu);
-	if (ret != 0x1) {
-		dev_err(adev->dev, "Msg issuing pre-check failed(0x%x) and "
-		       "SMU may be not in the right state!\n", ret);
-		if (ret != -ETIME)
-			ret = -EIO;
-		return ret;
+	switch (reg_c2pmsg_90) {
+	case SMU_RESP_NONE:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: I'm not done with your previous command!");
+		break;
+	case SMU_RESP_OK:
+		/* The SMU executed the command. It completed with a
+		 * successful result.
+		 */
+		break;
+	case SMU_RESP_CMD_FAIL:
+		/* The SMU executed the command. It completed with an
+		 * unsuccessful result.
+		 */
+		break;
+	case SMU_RESP_CMD_UNKNOWN:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: unknown command: index:%d param:0x%08X message:%s",
+				    msg_index, param, message);
+		break;
+	case SMU_RESP_CMD_BAD_PREREQ:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: valid command, bad prerequisites: index:%d param:0x%08X message:%s",
+				    msg_index, param, message);
+		break;
+	case SMU_RESP_BUSY_OTHER:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: I'm very busy for your command: index:%d param:0x%08X message:%s",
+				    msg_index, param, message);
+		break;
+	case SMU_RESP_DEBUG_END:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: I'm debugging!");
+		break;
+	default:
+		dev_err_ratelimited(adev->dev,
+				    "SMU: response:0x%08X for index:%d param:0x%08X message:%s?",
+				    reg_c2pmsg_90, msg_index, param, message);
+		break;
 	}
+}
+
+static int __smu_cmn_reg2errno(struct smu_context *smu, u32 reg_c2pmsg_90)
+{
+	int res;
+
+	switch (reg_c2pmsg_90) {
+	case SMU_RESP_NONE:
+		/* The SMU is busy--still executing your command.
+		 */
+		res = -ETIME;
+		break;
+	case SMU_RESP_OK:
+		res = 0;
+		break;
+	case SMU_RESP_CMD_FAIL:
+		/* Command completed successfully, but the command
+		 * status was failure.
+		 */
+		res = -EIO;
+		break;
+	case SMU_RESP_CMD_UNKNOWN:
+		/* Unknown command--ignored by the SMU.
+		 */
+		res = -EOPNOTSUPP;
+		break;
+	case SMU_RESP_CMD_BAD_PREREQ:
+		/* Valid command--bad prerequisites.
+		 */
+		res = -EINVAL;
+		break;
+	case SMU_RESP_BUSY_OTHER:
+		/* The SMU is busy with other commands. The client
+		 * should retry in 10 us.
+		 */
+		res = -EBUSY;
+		break;
+	default:
+		/* Unknown or debug response from the SMU.
+		 */
+		res = -EREMOTEIO;
+		break;
+	}
+
+	return res;
+}
+
+static void __smu_cmn_send_msg(struct smu_context *smu,
+			       u16 msg,
+			       u32 param)
+{
+	struct amdgpu_device *adev = smu->adev;
 
 	WREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_90, 0);
 	WREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_82, param);
 	WREG32_SOC15(MP1, 0, mmMP1_SMN_C2PMSG_66, msg);
+}
 
-	return 0;
+/**
+ * smu_cmn_send_msg_without_waiting -- send the message; don't wait for status
+ * @smu: pointer to an SMU context
+ * @msg_index: message index
+ * @param: message parameter to send to the SMU
+ *
+ * Send a message to the SMU with the parameter passed. Do not wait
+ * for status/result of the message, thus the "without_waiting".
+ *
+ * Return 0 on success, -errno on error if we weren't able to _send_
+ * the message for some reason. See __smu_cmn_reg2errno() for details
+ * of the -errno.
+ */
+int smu_cmn_send_msg_without_waiting(struct smu_context *smu,
+				     uint16_t msg_index,
+				     uint32_t param)
+{
+	u32 reg;
+	int res;
+
+	if (smu->adev->no_hw_access)
+		return 0;
+
+	reg = __smu_cmn_poll_stat(smu);
+	res = __smu_cmn_reg2errno(smu, reg);
+	if (reg == SMU_RESP_NONE ||
+	    reg == SMU_RESP_BUSY_OTHER ||
+	    res == -EREMOTEIO)
+		goto Out;
+	__smu_cmn_send_msg(smu, msg_index, param);
+	res = 0;
+Out:
+	return res;
+}
+
+/**
+ * smu_cmn_wait_for_response -- wait for response from the SMU
+ * @smu: pointer to an SMU context
+ *
+ * Wait for status from the SMU.
+ *
+ * Return 0 on success, -errno on error, indicating the execution
+ * status and result of the message being waited for. See
+ * __smu_cmn_reg2errno() for details of the -errno.
+ */
+int smu_cmn_wait_for_response(struct smu_context *smu)
+{
+	u32 reg;
+
+	reg = __smu_cmn_poll_stat(smu);
+	return __smu_cmn_reg2errno(smu, reg);
 }
 
+/**
+ * smu_cmn_send_smc_msg_with_param -- send a message with parameter
+ * @smu: pointer to an SMU context
+ * @msg: message to send
+ * @param: parameter to send to the SMU
+ * @read_arg: pointer to u32 to return a value from the SMU back
+ *            to the caller
+ *
+ * Send the message @msg with parameter @param to the SMU, wait for
+ * completion of the command, and return back a value from the SMU in
+ * @read_arg pointer.
+ *
+ * Return 0 on success, -errno on error, if we weren't able to send
+ * the message or if the message completed with some kind of
+ * error. See __smu_cmn_reg2errno() for details of the -errno.
+ *
+ * If we weren't able to send the message to the SMU, we also print
+ * the error to the standard log.
+ *
+ * Command completion status is printed only if the -errno is
+ * -EREMOTEIO, indicating that the SMU returned back an
+ * undefined/unknown/unspecified result. All other cases are
+ * well-defined, not printed, but instead given back to the client to
+ * decide what further to do.
+ *
+ * The return value, @read_arg is read back regardless, to give back
+ * more information to the client, which on error would most likely be
+ * @param, but we can't assume that. This also eliminates more
+ * conditionals.
+ */
 int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
 				    enum smu_message_type msg,
 				    uint32_t param,
 				    uint32_t *read_arg)
 {
-	struct amdgpu_device *adev = smu->adev;
-	int ret = 0, index = 0;
+	int res, index;
+	u32 reg;
 
 	if (smu->adev->no_hw_access)
 		return 0;
@@ -136,31 +337,24 @@ int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
 		return index == -EACCES ? 0 : index;
 
 	mutex_lock(&smu->message_lock);
-	ret = smu_cmn_send_msg_without_waiting(smu, (uint16_t)index, param);
-	if (ret)
-		goto out;
-
-	ret = smu_cmn_wait_for_response(smu);
-	if (ret != 0x1) {
-		if (ret == -ETIME) {
-			dev_err(adev->dev, "message: %15s (%d) \tparam: 0x%08x is timeout (no response)\n",
-				smu_get_message_name(smu, msg), index, param);
-		} else {
-			dev_err(adev->dev, "failed send message: %15s (%d) \tparam: 0x%08x response %#x\n",
-				smu_get_message_name(smu, msg), index, param,
-				ret);
-			ret = -EIO;
-		}
-		goto out;
+	reg = __smu_cmn_poll_stat(smu);
+	res = __smu_cmn_reg2errno(smu, reg);
+	if (reg == SMU_RESP_NONE ||
+	    reg == SMU_RESP_BUSY_OTHER ||
+	    res == -EREMOTEIO) {
+		__smu_cmn_reg_print_error(smu, reg, index, param, msg);
+		goto Out;
 	}
-
+	__smu_cmn_send_msg(smu, (uint16_t) index, param);
+	reg = __smu_cmn_poll_stat(smu);
+	res = __smu_cmn_reg2errno(smu, reg);
+	if (res == -EREMOTEIO)
+		__smu_cmn_reg_print_error(smu, reg, index, param, msg);
 	if (read_arg)
 		smu_cmn_read_arg(smu, read_arg);
-
-	ret = 0; /* 0 as driver return value */
-out:
+Out:
 	mutex_unlock(&smu->message_lock);
-	return ret;
+	return res;
 }
 
 int smu_cmn_send_smc_msg(struct smu_context *smu,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
index 9add5f16ff56..16993daa2ae0 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
@@ -27,7 +27,8 @@
 
 #if defined(SWSMU_CODE_LAYER_L2) || defined(SWSMU_CODE_LAYER_L3) || defined(SWSMU_CODE_LAYER_L4)
 int smu_cmn_send_msg_without_waiting(struct smu_context *smu,
-				     uint16_t msg, uint32_t param);
+				     uint16_t msg_index,
+				     uint32_t param);
 int smu_cmn_send_smc_msg_with_param(struct smu_context *smu,
 				    enum smu_message_type msg,
 				    uint32_t param,