Merge tag 'amd-drm-next-5.10-2020-09-03' of git://people.freedesktop.org/~agd5f/linux into drm-next

amd-drm-next-5.10-2020-09-03:

amdgpu:
- RAS fixes
- Sienna Cichlid updates
- Navy Flounder updates
- DCE6 (SI) support in DC
- Enable plane rotation
- Rework pre-OS vram reservation handling during driver init
- Add standard interface to dump GPU metrics table from SMU
- Rework tiling and tmz state handling in atomic commits
- Pstate fixes
- Add voltage and power hwmon interfaces for renoir
- SW CTF fixes
- S/G display fix for Raven
- Print client strings for vmfaults for vega and newer
- Manual fan control fixes
- Display updates
- Reorg power management directory structure
- Misc bug fixes
- Misc code cleanups

amdkfd:
- Topology fixes
- Add SMI events for thermal throttling and GPU resets

radeon:
- switch from pci_* to dma_* for dma allocations
- PLL fix

Scheduler:
- Clean up priority levels

UAPI:
- amdgpu INFO IOCTL query update for TMZ state
  https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6049
- amdkfd SMI event interface updates
  https://github.com/RadeonOpenCompute/rocm_smi_lib/tree/therm_thrott

From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200903222921.4152-1-alexander.deucher@amd.com

15 files changed, 280 insertions, 169 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
index 577d901fdb63..affbca7c0050 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h
@@ -911,7 +911,7 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0x705d0000, 0x807c817c,
 	0x8070ff70, 0x00000080,
 	0xbf0a7b7c, 0xbf85fff8,
-	0xbf82014f, 0xbef4037e,
+	0xbf820151, 0xbef4037e,
 	0x8775ff7f, 0x0000ffff,
 	0x8875ff75, 0x00040000,
 	0xbef60380, 0xbef703ff,
@@ -1024,61 +1024,62 @@ static const uint32_t cwsr_trap_nv1x_hex[] = {
 	0xbe883108, 0xbe8a310a,
 	0xbe8c310c, 0xbe8e310e,
 	0xbf06807c, 0xbf84fff0,
-	0xb9782a05, 0x80788178,
-	0xbf0d9972, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb96e1e06,
-	0x8f6e8a6e, 0x80786e78,
-	0x8078ff78, 0x00000200,
-	0xbef603ff, 0x01000000,
-	0xf4211bfa, 0xf0000000,
-	0x80788478, 0xf4211b3a,
+	0xba80f801, 0x00000000,
+	0xbf8a0000, 0xb9782a05,
+	0x80788178, 0xbf0d9972,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb96e1e06, 0x8f6e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0xf4211bfa,
 	0xf0000000, 0x80788478,
-	0xf4211b7a, 0xf0000000,
-	0x80788478, 0xf4211c3a,
+	0xf4211b3a, 0xf0000000,
+	0x80788478, 0xf4211b7a,
 	0xf0000000, 0x80788478,
-	0xf4211c7a, 0xf0000000,
-	0x80788478, 0xf4211eba,
+	0xf4211c3a, 0xf0000000,
+	0x80788478, 0xf4211c7a,
 	0xf0000000, 0x80788478,
-	0xf4211efa, 0xf0000000,
-	0x80788478, 0xf4211e7a,
+	0xf4211eba, 0xf0000000,
+	0x80788478, 0xf4211efa,
 	0xf0000000, 0x80788478,
-	0xf4211cfa, 0xf0000000,
-	0x80788478, 0xf4211bba,
+	0xf4211e7a, 0xf0000000,
+	0x80788478, 0xf4211cfa,
 	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef815, 0xbefc036f,
-	0xbefe0370, 0xbeff0371,
-	0x876f7bff, 0x000003ff,
-	0xb9ef4803, 0xb9f9f816,
-	0x876f7bff, 0xfffff800,
-	0x906f8b6f, 0xb9efa2c3,
-	0xb9f3f801, 0xb96e2a05,
-	0x806e816e, 0xbf0d9972,
-	0xbf850002, 0x8f6e896e,
-	0xbf820001, 0x8f6e8a6e,
-	0x806eff6e, 0x00000200,
-	0x806e746e, 0x826f8075,
-	0x876fff6f, 0x0000ffff,
-	0xf4091c37, 0xfa000050,
-	0xf4091d37, 0xfa000060,
-	0xf4011e77, 0xfa000074,
-	0xbf8cc07f, 0x876fff6d,
-	0xfc000000, 0x906f9a6f,
-	0x8f6f906f, 0xbeee0380,
+	0xb9eef814, 0xf4211bba,
+	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef815,
+	0xbefc036f, 0xbefe0370,
+	0xbeff0371, 0x876f7bff,
+	0x000003ff, 0xb9ef4803,
+	0xb9f9f816, 0x876f7bff,
+	0xfffff800, 0x906f8b6f,
+	0xb9efa2c3, 0xb9f3f801,
+	0xb96e2a05, 0x806e816e,
+	0xbf0d9972, 0xbf850002,
+	0x8f6e896e, 0xbf820001,
+	0x8f6e8a6e, 0x806eff6e,
+	0x00000200, 0x806e746e,
+	0x826f8075, 0x876fff6f,
+	0x0000ffff, 0xf4091c37,
+	0xfa000050, 0xf4091d37,
+	0xfa000060, 0xf4011e77,
+	0xfa000074, 0xbf8cc07f,
+	0x876fff6d, 0xfc000000,
+	0x906f9a6f, 0x8f6f906f,
+	0xbeee0380, 0x886e6f6e,
+	0x876fff6d, 0x02000000,
+	0x906f996f, 0x8f6f8f6f,
 	0x886e6f6e, 0x876fff6d,
-	0x02000000, 0x906f996f,
-	0x8f6f8f6f, 0x886e6f6e,
-	0x876fff6d, 0x01000000,
-	0x906f986f, 0x8f6f996f,
-	0x886e6f6e, 0x876fff7a,
-	0x00800000, 0x906f976f,
-	0xb9eef807, 0x876dff6d,
-	0x0000ffff, 0x87fe7e7e,
-	0x87ea6a6a, 0xb9faf802,
-	0xbf8a0000, 0xbe80226c,
+	0x01000000, 0x906f986f,
+	0x8f6f996f, 0x886e6f6e,
+	0x876fff7a, 0x00800000,
+	0x906f976f, 0xb9eef807,
+	0x876dff6d, 0x0000ffff,
+	0x87fe7e7e, 0x87ea6a6a,
+	0xb9faf802, 0xbe80226c,
 	0xbf810000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
@@ -1807,7 +1808,7 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0x705d0000, 0x807c817c,
 	0x8070ff70, 0x00000080,
 	0xbf0a7b7c, 0xbf85fff8,
-	0xbf82013a, 0xbef4037e,
+	0xbf82013c, 0xbef4037e,
 	0x8775ff7f, 0x0000ffff,
 	0x8875ff75, 0x00040000,
 	0xbef60380, 0xbef703ff,
@@ -1920,50 +1921,51 @@ static const uint32_t cwsr_trap_gfx10_hex[] = {
 	0xbe883108, 0xbe8a310a,
 	0xbe8c310c, 0xbe8e310e,
 	0xbf06807c, 0xbf84fff0,
-	0xb9782a05, 0x80788178,
-	0xbf0d9972, 0xbf850002,
-	0x8f788978, 0xbf820001,
-	0x8f788a78, 0xb96e1e06,
-	0x8f6e8a6e, 0x80786e78,
-	0x8078ff78, 0x00000200,
-	0xbef603ff, 0x01000000,
-	0xf4211bfa, 0xf0000000,
-	0x80788478, 0xf4211b3a,
+	0xba80f801, 0x00000000,
+	0xbf8a0000, 0xb9782a05,
+	0x80788178, 0xbf0d9972,
+	0xbf850002, 0x8f788978,
+	0xbf820001, 0x8f788a78,
+	0xb96e1e06, 0x8f6e8a6e,
+	0x80786e78, 0x8078ff78,
+	0x00000200, 0xbef603ff,
+	0x01000000, 0xf4211bfa,
 	0xf0000000, 0x80788478,
-	0xf4211b7a, 0xf0000000,
-	0x80788478, 0xf4211c3a,
+	0xf4211b3a, 0xf0000000,
+	0x80788478, 0xf4211b7a,
 	0xf0000000, 0x80788478,
-	0xf4211c7a, 0xf0000000,
-	0x80788478, 0xf4211eba,
+	0xf4211c3a, 0xf0000000,
+	0x80788478, 0xf4211c7a,
 	0xf0000000, 0x80788478,
-	0xf4211efa, 0xf0000000,
-	0x80788478, 0xf4211e7a,
+	0xf4211eba, 0xf0000000,
+	0x80788478, 0xf4211efa,
 	0xf0000000, 0x80788478,
-	0xf4211cfa, 0xf0000000,
-	0x80788478, 0xf4211bba,
+	0xf4211e7a, 0xf0000000,
+	0x80788478, 0xf4211cfa,
 	0xf0000000, 0x80788478,
-	0xbf8cc07f, 0xb9eef814,
 	0xf4211bba, 0xf0000000,
 	0x80788478, 0xbf8cc07f,
-	0xb9eef815, 0xbefc036f,
-	0xbefe0370, 0xbeff0371,
-	0x876f7bff, 0x000003ff,
-	0xb9ef4803, 0x876f7bff,
-	0xfffff800, 0x906f8b6f,
-	0xb9efa2c3, 0xb9f3f801,
-	0xb96e2a05, 0x806e816e,
-	0xbf0d9972, 0xbf850002,
-	0x8f6e896e, 0xbf820001,
-	0x8f6e8a6e, 0x806eff6e,
-	0x00000200, 0x806e746e,
-	0x826f8075, 0x876fff6f,
-	0x0000ffff, 0xf4091c37,
-	0xfa000050, 0xf4091d37,
-	0xfa000060, 0xf4011e77,
-	0xfa000074, 0xbf8cc07f,
-	0x876dff6d, 0x0000ffff,
-	0x87fe7e7e, 0x87ea6a6a,
-	0xb9faf802, 0xbf8a0000,
+	0xb9eef814, 0xf4211bba,
+	0xf0000000, 0x80788478,
+	0xbf8cc07f, 0xb9eef815,
+	0xbefc036f, 0xbefe0370,
+	0xbeff0371, 0x876f7bff,
+	0x000003ff, 0xb9ef4803,
+	0x876f7bff, 0xfffff800,
+	0x906f8b6f, 0xb9efa2c3,
+	0xb9f3f801, 0xb96e2a05,
+	0x806e816e, 0xbf0d9972,
+	0xbf850002, 0x8f6e896e,
+	0xbf820001, 0x8f6e8a6e,
+	0x806eff6e, 0x00000200,
+	0x806e746e, 0x826f8075,
+	0x876fff6f, 0x0000ffff,
+	0xf4091c37, 0xfa000050,
+	0xf4091d37, 0xfa000060,
+	0xf4011e77, 0xfa000074,
+	0xbf8cc07f, 0x876dff6d,
+	0x0000ffff, 0x87fe7e7e,
+	0x87ea6a6a, 0xb9faf802,
 	0xbe80226c, 0xbf810000,
 	0xbf9f0000, 0xbf9f0000,
 	0xbf9f0000, 0xbf9f0000,
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
index 5b220f2a7501..5081f91190b8 100644
--- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
+++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm
@@ -894,6 +894,11 @@ L_RESTORE_SGPR:
 	s_cmp_eq_u32	m0, 0							//scc = (m0 < s_sgpr_save_num) ? 1 : 0
 	s_cbranch_scc0	L_RESTORE_SGPR_LOOP
 
+	// s_barrier with MODE.DEBUG_EN=1, STATUS.PRIV=1 incorrectly asserts debug exception.
+	// Clear DEBUG_EN before and restore MODE after the barrier.
+	s_setreg_imm32_b32	hwreg(HW_REG_MODE), 0
+	s_barrier								//barrier to ensure the readiness of LDS before access attemps from any other wave in the same TG
+
 	/* restore HW registers */
 L_RESTORE_HWREG:
 	// HWREG SR memory offset : size(VGPR)+size(SVGPR)+size(SGPR)
@@ -976,8 +981,6 @@ L_RESTORE_HWREG:
 	s_and_b64	vcc, vcc, vcc						// Restore STATUS.VCCZ, not writable by s_setreg_b32
 	s_setreg_b32	hwreg(HW_REG_STATUS), s_restore_status			// SCC is included, which is changed by previous salu
 
-	s_barrier								//barrier to ensure the readiness of LDS before access attemps from any other wave in the same TG
-
 	s_rfe_b64	s_restore_pc_lo						//Return to the main shader program and resume execution
 
 L_END_PGM:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index e9b96ad3d9a5..b7b16adb0615 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1254,7 +1254,7 @@ bool kfd_dev_is_large_bar(struct kfd_dev *dev)
 		return true;
 	}
 
-	if (dev->device_info->needs_iommu_device)
+	if (dev->use_iommu_v2)
 		return false;
 
 	amdgpu_amdkfd_get_local_mem_info(dev->kgd, &mem_info);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index 6a250f8fcfb8..3fac06b281ce 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -742,6 +742,22 @@ static int kfd_fill_gpu_cache_info(struct kfd_dev *kdev,
 	return 0;
 }
 
+static bool kfd_ignore_crat(void)
+{
+	bool ret;
+
+	if (ignore_crat)
+		return true;
+
+#ifndef KFD_SUPPORT_IOMMU_V2
+	ret = true;
+#else
+	ret = false;
+#endif
+
+	return ret;
+}
+
 /*
  * kfd_create_crat_image_acpi - Allocates memory for CRAT image and
  * copies CRAT from ACPI (if available).
@@ -776,7 +792,7 @@ int kfd_create_crat_image_acpi(void **crat_image, size_t *size)
 		return -EINVAL;
 	}
 
-	if (ignore_crat) {
+	if (kfd_ignore_crat()) {
 		pr_info("CRAT table disabled by module option\n");
 		return -ENODATA;
 	}
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 4bfedaab183f..0e71a0543f98 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -29,6 +29,7 @@
 #include "cwsr_trap_handler.h"
 #include "kfd_iommu.h"
 #include "amdgpu_amdkfd.h"
+#include "kfd_smi_events.h"
 
 #define MQD_SIZE_ALIGNED 768
 
@@ -115,6 +116,7 @@ static const struct kfd_device_info carrizo_device_info = {
 	.num_xgmi_sdma_engines = 0,
 	.num_sdma_queues_per_engine = 2,
 };
+#endif
 
 static const struct kfd_device_info raven_device_info = {
 	.asic_family = CHIP_RAVEN,
@@ -133,7 +135,6 @@ static const struct kfd_device_info raven_device_info = {
 	.num_xgmi_sdma_engines = 0,
 	.num_sdma_queues_per_engine = 2,
 };
-#endif
 
 static const struct kfd_device_info hawaii_device_info = {
 	.asic_family = CHIP_HAWAII,
@@ -711,11 +712,9 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 		goto kfd_doorbell_error;
 	}
 
-	if (kfd->kfd2kgd->get_hive_id)
-		kfd->hive_id = kfd->kfd2kgd->get_hive_id(kfd->kgd);
+	kfd->hive_id = amdgpu_amdkfd_get_hive_id(kfd->kgd);
 
-	if (kfd->kfd2kgd->get_unique_id)
-		kfd->unique_id = kfd->kfd2kgd->get_unique_id(kfd->kgd);
+	kfd->unique_id = amdgpu_amdkfd_get_unique_id(kfd->kgd);
 
 	if (kfd_interrupt_init(kfd)) {
 		dev_err(kfd_device, "Error initializing interrupts\n");
@@ -737,6 +736,9 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 		goto gws_error;
 	}
 
+	/* If CRAT is broken, won't set iommu enabled */
+	kfd_double_confirm_iommu_support(kfd);
+
 	if (kfd_iommu_device_init(kfd)) {
 		dev_err(kfd_device, "Error initializing iommuv2\n");
 		goto device_iommu_error;
@@ -810,6 +812,8 @@ int kgd2kfd_pre_reset(struct kfd_dev *kfd)
 	if (!kfd->init_complete)
 		return 0;
 
+	kfd_smi_event_update_gpu_reset(kfd, false);
+
 	kfd->dqm->ops.pre_reset(kfd->dqm);
 
 	kgd2kfd_suspend(kfd, false);
@@ -838,6 +842,8 @@ int kgd2kfd_post_reset(struct kfd_dev *kfd)
 
 	atomic_set(&kfd->sram_ecc_flag, 0);
 
+	kfd_smi_event_update_gpu_reset(kfd, true);
+
 	return 0;
 }
 
@@ -1245,6 +1251,12 @@ void kfd_dec_compute_active(struct kfd_dev *kfd)
 	WARN_ONCE(count < 0, "Compute profile ref. count error");
 }
 
+void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint32_t throttle_bitmask)
+{
+	if (kfd)
+		kfd_smi_event_update_thermal_throttling(kfd, throttle_bitmask);
+}
+
 #if defined(CONFIG_DEBUG_FS)
 
 /* This function will send a package to HIQ to hang the HWS
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index e0e60b0d0669..560adc57a050 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -153,30 +153,6 @@ static void decrement_queue_count(struct device_queue_manager *dqm,
 		dqm->active_cp_queue_count--;
 }
 
-int read_sdma_queue_counter(uint64_t q_rptr, uint64_t *val)
-{
-	int ret;
-	uint64_t tmp = 0;
-
-	if (!val)
-		return -EINVAL;
-	/*
-	 * SDMA activity counter is stored at queue's RPTR + 0x8 location.
-	 */
-	if (!access_ok((const void __user *)(q_rptr +
-					sizeof(uint64_t)), sizeof(uint64_t))) {
-		pr_err("Can't access sdma queue activity counter\n");
-		return -EFAULT;
-	}
-
-	ret = get_user(tmp, (uint64_t *)(q_rptr + sizeof(uint64_t)));
-	if (!ret) {
-		*val = tmp;
-	}
-
-	return ret;
-}
-
 static int allocate_doorbell(struct qcm_process_device *qpd, struct queue *q)
 {
 	struct kfd_dev *dev = qpd->dqm->dev;
@@ -552,7 +528,7 @@ static int destroy_queue_nocpsch(struct device_queue_manager *dqm,
 	/* Get the SDMA queue stats */
 	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
 	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
-		retval = read_sdma_queue_counter((uint64_t)q->properties.read_ptr,
+		retval = read_sdma_queue_counter((uint64_t __user *)q->properties.read_ptr,
 							&sdma_val);
 		if (retval)
 			pr_err("Failed to read SDMA queue counter for queue: %d\n",
@@ -1473,7 +1449,7 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 	/* Get the SDMA queue stats */
 	if ((q->properties.type == KFD_QUEUE_TYPE_SDMA) ||
 	    (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)) {
-		retval = read_sdma_queue_counter((uint64_t)q->properties.read_ptr,
+		retval = read_sdma_queue_counter((uint64_t __user *)q->properties.read_ptr,
 							&sdma_val);
 		if (retval)
 			pr_err("Failed to read SDMA queue counter for queue: %d\n",
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index 49d8e324c636..16262e5d93f5 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -251,5 +251,11 @@ static inline void dqm_unlock(struct device_queue_manager *dqm)
 	mutex_unlock(&dqm->lock_hidden);
 }
 
-int read_sdma_queue_counter(uint64_t q_rptr, uint64_t *val);
+static inline int read_sdma_queue_counter(uint64_t __user *q_rptr, uint64_t *val)
+{
+        /*
+         * SDMA activity counter is stored at queue's RPTR + 0x8 location.
+         */
+	return get_user(*val, q_rptr + 1);
+}
 #endif /* KFD_DEVICE_QUEUE_MANAGER_H_ */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
index 95a82ac455f2..309f63a0b34a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
@@ -62,7 +62,7 @@ static int update_qpd_v9(struct device_queue_manager *dqm,
 				SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
 					SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
 		if (amdgpu_noretry &&
-		    !dqm->dev->device_info->needs_iommu_device)
+		    !dqm->dev->use_iommu_v2)
 			qpd->sh_mem_config |=
 				1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
index c1166c40ac15..3c22909470f2 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_flat_memory.c
@@ -321,7 +321,7 @@ static void kfd_init_apertures_vi(struct kfd_process_device *pdd, uint8_t id)
 	pdd->lds_base = MAKE_LDS_APP_BASE_VI();
 	pdd->lds_limit = MAKE_LDS_APP_LIMIT(pdd->lds_base);
 
-	if (!pdd->dev->device_info->needs_iommu_device) {
+	if (!pdd->dev->use_iommu_v2) {
 		/* dGPUs: SVM aperture starting at 0
 		 * with small reserved space for kernel.
 		 * Set them to CANONICAL addresses.
@@ -425,7 +425,7 @@ int kfd_init_apertures(struct kfd_process *process)
 				return -EINVAL;
 			}
 
-			if (!dev->device_info->needs_iommu_device) {
+			if (!dev->use_iommu_v2) {
 				/* dGPUs: the reserved space for kernel
 				 * before SVM
 				 */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c
index 7c8786b9eb0a..5a64915abaf7 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_iommu.c
@@ -41,7 +41,7 @@ int kfd_iommu_check_device(struct kfd_dev *kfd)
 	struct amd_iommu_device_info iommu_info;
 	int err;
 
-	if (!kfd->device_info->needs_iommu_device)
+	if (!kfd->use_iommu_v2)
 		return -ENODEV;
 
 	iommu_info.flags = 0;
@@ -63,7 +63,7 @@ int kfd_iommu_device_init(struct kfd_dev *kfd)
 	unsigned int pasid_limit;
 	int err;
 
-	if (!kfd->device_info->needs_iommu_device)
+	if (!kfd->use_iommu_v2)
 		return 0;
 
 	iommu_info.flags = 0;
@@ -109,7 +109,7 @@ int kfd_iommu_bind_process_to_device(struct kfd_process_device *pdd)
 	struct kfd_process *p = pdd->process;
 	int err;
 
-	if (!dev->device_info->needs_iommu_device || pdd->bound == PDD_BOUND)
+	if (!dev->use_iommu_v2 || pdd->bound == PDD_BOUND)
 		return 0;
 
 	if (unlikely(pdd->bound == PDD_BOUND_SUSPENDED)) {
@@ -284,7 +284,7 @@ static void kfd_unbind_processes_from_device(struct kfd_dev *kfd)
  */
 void kfd_iommu_suspend(struct kfd_dev *kfd)
 {
-	if (!kfd->device_info->needs_iommu_device)
+	if (!kfd->use_iommu_v2)
 		return;
 
 	kfd_unbind_processes_from_device(kfd);
@@ -304,7 +304,7 @@ int kfd_iommu_resume(struct kfd_dev *kfd)
 	unsigned int pasid_limit;
 	int err;
 
-	if (!kfd->device_info->needs_iommu_device)
+	if (!kfd->use_iommu_v2)
 		return 0;
 
 	pasid_limit = kfd_get_pasid_limit();
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index 6727e9de5b8b..023629f28495 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -297,6 +297,9 @@ struct kfd_dev {
 
 	bool pci_atomic_requested;
 
+	/* Use IOMMU v2 flag */
+	bool use_iommu_v2;
+
 	/* SRAM ECC flag */
 	atomic_t sram_ecc_flag;
 
@@ -309,6 +312,8 @@ struct kfd_dev {
 	/* Clients watching SMI events */
 	struct list_head smi_clients;
 	spinlock_t smi_lock;
+
+	uint32_t reset_seq_num;
 };
 
 enum kfd_mempool {
@@ -892,6 +897,7 @@ struct kfd_dev *kfd_device_by_pci_dev(const struct pci_dev *pdev);
 struct kfd_dev *kfd_device_by_kgd(const struct kgd_dev *kgd);
 int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev);
 int kfd_numa_node_to_apic_id(int numa_node_id);
+void kfd_double_confirm_iommu_support(struct kfd_dev *gpu);
 
 /* Interrupts */
 int kfd_interrupt_init(struct kfd_dev *dev);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index 40695d52e9a8..a0e12a79ab7d 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -87,7 +87,7 @@ struct kfd_sdma_activity_handler_workarea {
 };
 
 struct temp_sdma_queue_list {
-	uint64_t rptr;
+	uint64_t __user *rptr;
 	uint64_t sdma_val;
 	unsigned int queue_id;
 	struct list_head list;
@@ -159,7 +159,7 @@ static void kfd_sdma_activity_worker(struct work_struct *work)
 		}
 
 		INIT_LIST_HEAD(&sdma_q->list);
-		sdma_q->rptr = (uint64_t)q->properties.read_ptr;
+		sdma_q->rptr = (uint64_t __user *)q->properties.read_ptr;
 		sdma_q->queue_id = q->properties.queue_id;
 		list_add_tail(&sdma_q->list, &sdma_q_list.list);
 	}
@@ -218,7 +218,7 @@ static void kfd_sdma_activity_worker(struct work_struct *work)
 			continue;
 
 		list_for_each_entry_safe(sdma_q, next, &sdma_q_list.list, list) {
-			if (((uint64_t)q->properties.read_ptr == sdma_q->rptr) &&
+			if (((uint64_t __user *)q->properties.read_ptr == sdma_q->rptr) &&
 			     (sdma_q->queue_id == q->properties.queue_id)) {
 				list_del(&sdma_q->list);
 				kfree(sdma_q);
@@ -270,6 +270,7 @@ static ssize_t kfd_procfs_show(struct kobject *kobj, struct attribute *attr,
 					kfd_sdma_activity_worker);
 
 		sdma_activity_work_handler.pdd = pdd;
+		sdma_activity_work_handler.sdma_activity_counter = 0;
 
 		schedule_work(&sdma_activity_work_handler.sdma_activity_work);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
index 7b348bf9df21..17d1736367ea 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.c
@@ -24,6 +24,7 @@
 #include <linux/wait.h>
 #include <linux/anon_inodes.h>
 #include <uapi/linux/kfd_ioctl.h>
+#include "amdgpu.h"
 #include "amdgpu_vm.h"
 #include "kfd_priv.h"
 #include "kfd_smi_events.h"
@@ -148,15 +149,94 @@ static int kfd_smi_ev_release(struct inode *inode, struct file *filep)
 	return 0;
 }
 
+static void add_event_to_kfifo(struct kfd_dev *dev, unsigned int smi_event,
+			      char *event_msg, int len)
+{
+	struct kfd_smi_client *client;
+
+	rcu_read_lock();
+
+	list_for_each_entry_rcu(client, &dev->smi_clients, list) {
+		if (!(READ_ONCE(client->events) &
+				KFD_SMI_EVENT_MASK_FROM_INDEX(smi_event)))
+			continue;
+		spin_lock(&client->lock);
+		if (kfifo_avail(&client->fifo) >= len) {
+			kfifo_in(&client->fifo, event_msg, len);
+			wake_up_all(&client->wait_queue);
+		} else {
+			pr_debug("smi_event(EventID: %u): no space left\n",
+					smi_event);
+		}
+		spin_unlock(&client->lock);
+	}
+
+	rcu_read_unlock();
+}
+
+void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset)
+{
+	/*
+	 * GpuReset msg = Reset seq number (incremented for
+	 * every reset message sent before GPU reset).
+	 * 1 byte event + 1 byte space + 8 bytes seq num +
+	 * 1 byte \n + 1 byte \0 = 12
+	 */
+	char fifo_in[12];
+	int len;
+	unsigned int event;
+
+	if (list_empty(&dev->smi_clients))
+		return;
+
+	memset(fifo_in, 0x0, sizeof(fifo_in));
+
+	if (post_reset) {
+		event = KFD_SMI_EVENT_GPU_POST_RESET;
+	} else {
+		event = KFD_SMI_EVENT_GPU_PRE_RESET;
+		++(dev->reset_seq_num);
+	}
+
+	len = snprintf(fifo_in, sizeof(fifo_in), "%x %x\n", event,
+						dev->reset_seq_num);
+
+	add_event_to_kfifo(dev, event, fifo_in, len);
+}
+
+void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev,
+					     uint32_t throttle_bitmask)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)dev->kgd;
+	/*
+	 * ThermalThrottle msg = throttle_bitmask(8):
+	 * 			 thermal_interrupt_count(16):
+	 * 1 byte event + 1 byte space + 8 byte throttle_bitmask +
+	 * 1 byte : + 16 byte thermal_interupt_counter + 1 byte \n +
+	 * 1 byte \0 = 29
+	 */
+	char fifo_in[29];
+	int len;
+
+	if (list_empty(&dev->smi_clients))
+		return;
+
+	len = snprintf(fifo_in, sizeof(fifo_in), "%x %x:%llx\n",
+		       KFD_SMI_EVENT_THERMAL_THROTTLE, throttle_bitmask,
+		       atomic64_read(&adev->smu.throttle_int_counter));
+
+	add_event_to_kfifo(dev, KFD_SMI_EVENT_THERMAL_THROTTLE,	fifo_in, len);
+}
+
 void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)dev->kgd;
 	struct amdgpu_task_info task_info;
 	/* VmFault msg = (hex)uint32_pid(8) + :(1) + task name(16) = 25 */
-	/* 16 bytes event + 1 byte space + 25 bytes msg + 1 byte \n = 43
+	/* 1 byte event + 1 byte space + 25 bytes msg + 1 byte \n +
+	 * 1 byte \0 = 29
 	 */
-	char fifo_in[43];
-	struct kfd_smi_client *client;
+	char fifo_in[29];
 	int len;
 
 	if (list_empty(&dev->smi_clients))
@@ -168,25 +248,10 @@ void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid)
 	if (!task_info.pid)
 		return;
 
-	len = snprintf(fifo_in, 43, "%x %x:%s\n", KFD_SMI_EVENT_VMFAULT,
+	len = snprintf(fifo_in, sizeof(fifo_in), "%x %x:%s\n", KFD_SMI_EVENT_VMFAULT,
 		task_info.pid, task_info.task_name);
 
-	rcu_read_lock();
-
-	list_for_each_entry_rcu(client, &dev->smi_clients, list) {
-		if (!(READ_ONCE(client->events) & KFD_SMI_EVENT_VMFAULT))
-			continue;
-		spin_lock(&client->lock);
-		if (kfifo_avail(&client->fifo) >= len) {
-			kfifo_in(&client->fifo, fifo_in, len);
-			wake_up_all(&client->wait_queue);
-		}
-		else
-			pr_debug("smi_event(vmfault): no space left\n");
-		spin_unlock(&client->lock);
-	}
-
-	rcu_read_unlock();
+	add_event_to_kfifo(dev, KFD_SMI_EVENT_VMFAULT, fifo_in, len);
 }
 
 int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
index a9cb218fef96..b9b0438202e2 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_smi_events.h
@@ -25,5 +25,8 @@
 
 int kfd_smi_event_open(struct kfd_dev *dev, uint32_t *fd);
 void kfd_smi_event_update_vmfault(struct kfd_dev *dev, uint16_t pasid);
+void kfd_smi_event_update_thermal_throttling(struct kfd_dev *dev,
+					     uint32_t throttle_bitmask);
+void kfd_smi_event_update_gpu_reset(struct kfd_dev *dev, bool post_reset);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index f185f6cbc05c..2b31c3066aaa 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -446,7 +446,7 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr,
 	sysfs_show_32bit_prop(buffer, offs, "cpu_cores_count",
 			      dev->node_props.cpu_cores_count);
 	sysfs_show_32bit_prop(buffer, offs, "simd_count",
-			      dev->node_props.simd_count);
+			      dev->gpu ? dev->node_props.simd_count : 0);
 	sysfs_show_32bit_prop(buffer, offs, "mem_banks_count",
 			      dev->node_props.mem_banks_count);
 	sysfs_show_32bit_prop(buffer, offs, "caches_count",
@@ -1139,7 +1139,7 @@ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu)
 		/* Discrete GPUs need their own topology device list
 		 * entries. Don't assign them to CPU/APU nodes.
 		 */
-		if (!gpu->device_info->needs_iommu_device &&
+		if (!gpu->use_iommu_v2 &&
 		    dev->node_props.cpu_cores_count)
 			continue;
 
@@ -1239,7 +1239,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	void *crat_image = NULL;
 	size_t image_size = 0;
 	int proximity_domain;
-	struct amdgpu_ras *ctx;
+	struct amdgpu_device *adev;
 
 	INIT_LIST_HEAD(&temp_topology_device_list);
 
@@ -1388,7 +1388,7 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 	* Overwrite ATS capability according to needs_iommu_device to fix
 	* potential missing corresponding bit in CRAT of BIOS.
 	*/
-	if (dev->gpu->device_info->needs_iommu_device)
+	if (dev->gpu->use_iommu_v2)
 		dev->node_props.capability |= HSA_CAP_ATS_PRESENT;
 	else
 		dev->node_props.capability &= ~HSA_CAP_ATS_PRESENT;
@@ -1404,19 +1404,17 @@ int kfd_topology_add_device(struct kfd_dev *gpu)
 		dev->node_props.max_waves_per_simd = 10;
 	}
 
-	ctx = amdgpu_ras_get_context((struct amdgpu_device *)(dev->gpu->kgd));
-	if (ctx) {
-		/* kfd only concerns sram ecc on GFX/SDMA and HBM ecc on UMC */
-		dev->node_props.capability |=
-			(((ctx->features & BIT(AMDGPU_RAS_BLOCK__SDMA)) != 0) ||
-			 ((ctx->features & BIT(AMDGPU_RAS_BLOCK__GFX)) != 0)) ?
-			HSA_CAP_SRAM_EDCSUPPORTED : 0;
-		dev->node_props.capability |= ((ctx->features & BIT(AMDGPU_RAS_BLOCK__UMC)) != 0) ?
-			HSA_CAP_MEM_EDCSUPPORTED : 0;
-
-		dev->node_props.capability |= (ctx->features != 0) ?
+	adev = (struct amdgpu_device *)(dev->gpu->kgd);
+	/* kfd only concerns sram ecc on GFX and HBM ecc on UMC */
+	dev->node_props.capability |=
+		((adev->ras_features & BIT(AMDGPU_RAS_BLOCK__GFX)) != 0) ?
+		HSA_CAP_SRAM_EDCSUPPORTED : 0;
+	dev->node_props.capability |= ((adev->ras_features & BIT(AMDGPU_RAS_BLOCK__UMC)) != 0) ?
+		HSA_CAP_MEM_EDCSUPPORTED : 0;
+
+	if (adev->asic_type != CHIP_VEGA10)
+		dev->node_props.capability |= (adev->ras_features != 0) ?
 			HSA_CAP_RASEVENTNOTIFY : 0;
-	}
 
 	kfd_debug_print_topology();
 
@@ -1515,6 +1513,29 @@ int kfd_numa_node_to_apic_id(int numa_node_id)
 	return kfd_cpumask_to_apic_id(cpumask_of_node(numa_node_id));
 }
 
+void kfd_double_confirm_iommu_support(struct kfd_dev *gpu)
+{
+	struct kfd_topology_device *dev;
+
+	gpu->use_iommu_v2 = false;
+
+	if (!gpu->device_info->needs_iommu_device)
+		return;
+
+	down_read(&topology_lock);
+
+	/* Only use IOMMUv2 if there is an APU topology node with no GPU
+	 * assigned yet. This GPU will be assigned to it.
+	 */
+	list_for_each_entry(dev, &topology_device_list, list)
+		if (dev->node_props.cpu_cores_count &&
+		    dev->node_props.simd_count &&
+		    !dev->gpu)
+			gpu->use_iommu_v2 = true;
+
+	up_read(&topology_lock);
+}
+
 #if defined(CONFIG_DEBUG_FS)
 
 int kfd_debugfs_hqds_by_device(struct seq_file *m, void *data)